1 2017-02-13 Jakub Jelinek <jakub@redhat.com>
3 * config/i386/i386.h (REG_CLASS_NAMES): Add , in between
4 "MOD4_SSE_REGS" and "ALL_REGS".
6 * spellcheck.c (test_data): Add , in between "foo" and "food".
8 2017-02-13 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
11 * config/rs6000/rs6000.c (expand_block_compare): Make sure runtime
12 boundary crossing check and subsequent code generation agree.
14 2017-02-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16 * config/aarch64/aarch64.c (has_memory_op): Delete.
17 (aarch64_madd_needs_nop): Use contains_mem_rtx_p instead of
20 2017-02-13 Jakub Jelinek <jakub@redhat.com>
22 PR rtl-optimization/79388
23 PR rtl-optimization/79450
24 * combine.c (distribute_notes): When removing TEM_INSN for which
25 corresponding dest has last value recorded, invalidate that last
28 2017-02-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
30 * config/arm/arm.c (arm_print_tune_info): Use ASM_COMMENT_START instead
31 of explicit '@'. Add missing assembly comment marker on branch costs
34 2017-02-13 Nathan Sidwell <nathan@acm.org>
36 * gengtype-lex.l (<in_struct>): Add '/'.
38 2017-02-13 Martin Liska <mliska@suse.cz>
41 * calls.c (expand_call): Replace XALLOCAVEC with XCNEWVEC.
43 2017-02-13 Richard Biener <rguenther@suse.de>
45 * configure.ac (HAVE_ISL_OPTIONS_SET_SCHEDULE_SERIALIZE_SCCS):
47 * configure: Re-generate.
48 * config.in: Likewise.
49 * graphite-dependences.c: Simplify as if
50 HAVE_ISL_OPTIONS_SET_SCHEDULE_SERIALIZE_SCCS was defined.
51 * graphite-isl-ast-to-gimple.c: Likewise.
52 * graphite-optimize-isl.c: Likewise.
53 * graphite-poly.c: Likewise.
54 * graphite-sese-to-poly.c: Likewise.
55 * graphite.h: Likewise.
56 * toplev.c: Include isl/version.h and use isl_version () for
57 printing the ISL version.
58 * doc/install.texi: Update ISL requirement.
60 2017-02-12 Gerald Pfeifer <gerald@pfeifer.com>
62 * doc/standards.texi (Standards): Update reference to
65 2017-02-12 Gerald Pfeifer <gerald@pfeifer.com>
67 * doc/extend.texi (Named Address Spaces): sourceware.org now
69 * doc/install.texi (Binaries): Ditto.
72 2017-02-11 Sandra Loosemore <sandra@codesourcery.com>
74 * doc/cpp.texi: Replace "stringify"/"stringification" with C
75 standard terminology "stringize"/"stringizing" throughout.
76 * doc/cppinternals.texi: Likewise.
78 2017-02-11 Sandra Loosemore <sandra@codesourcery.com>
80 * doc/extend.texi: Fix some spelling mistakes and typos.
81 * doc/invoke.texi: Likewise.
83 2017-02-11 Jan Hubicka <hubicka@ucw.cz>
86 * params.def (inline-min-speedup) Change from 10 to 8.
88 2017-02-11 Jakub Jelinek <jakub@redhat.com>
90 * doc/invoke.texi (fopenmp): Bump OpenMP version from 4.0 to
93 2017-02-11 Jan Hubicka <hubicka@ucw.cz>
96 * ipa-inline-analysis.c (get_minimal_bb): New function.
97 (record_modified): Use it.
98 (remap_edge_change_prob): Handle also ancestor functions.
100 2017-02-11 Gerald Pfeifer <gerald@pfeifer.com>
102 * doc/contrib.texi (Contributors): Remove broken link into
103 the Mauve CVS repository.
105 2017-02-11 Jakub Jelinek <jakub@redhat.com>
108 * internal-fn.c (expand_vector_ubsan_overflow): Use piece-wise
109 result computation whenever lhs doesn't have vector mode, not
110 just when it has BLKmode.
112 2017-02-10 Gerald Pfeifer <gerald@pfeifer.com>
114 * doc/makefile.texi (profiledbootstrap): Refer to the
115 installation instructions only in textual form.
117 2017-02-10 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
120 * config/rs6000/altivec.md (bcd<bcd_add_sub>): Fix constraints.
122 2017-02-10 Gerald Pfeifer <gerald@pfeifer.com>
124 * doc/install.texi (Specific): Use https for blackfin.uclinux.org.
125 (Specific): Update mingw-w64 reference.
127 (Specific): Remove broken link to Renesas RX processor.
129 2017-02-10 Richard Biener <rguenther@suse.de>
131 * toplev.c (process_options): Do not mention obsolete graphite
132 options when printing sorry message about missing graphite support.
133 Mention -floop-nest-optimize.
135 2017-02-10 Christophe Lyon <christophe.lyon@linaro.org>
137 * config/aarch64/arm_neon.h (vtst_p8): Rewrite without asm.
138 (vtst_p16): Likewise.
139 (vtstq_p8): Likewise.
140 (vtstq_p16): Likewise.
142 (vtstq_p64): Likewise.
143 * config/arm/arm_neon.h (vgetq_lane_p64): New.
144 (vset_lane_p64): New.
145 (vsetq_lane_p64): New.
147 2017-02-10 Jakub Jelinek <jakub@redhat.com>
149 PR tree-optimization/79411
150 * tree-ssa-reassoc.c (is_reassociable_op): Return false if
151 stmt operands are SSA_NAMEs used in abnormal phis.
152 (can_reassociate_p): Return false if op is SSA_NAME used in abnormal
155 2017-02-09 Jan Hubicka <hubicka@ucw.cz>
158 * cgraphunit.c (cgraph_node::add_new_function): Set externally_visible
161 2017-02-09 Jan Hubicka <hubicka@ucw.cz>
163 * tree-ssa-loop-unswitch.c (hoist_guard): Update profile.
165 2017-02-09 Jakub Jelinek <jakub@redhat.com>
167 * omp-offload.c (oacc_loop_auto_partitions): Use || instead of |
171 * gimplify.h (is_gimple_sizepos): Only test for INTEGER_CST constants,
172 not arbitrary TREE_CONSTANT.
175 * gimplify.c (gimplify_adjust_omp_clauses): Ignore
176 "omp declare target link" attribute unless is_global_var.
177 * omp-offload.c (find_link_var_op): Likewise.
179 2017-02-09 Nathan Sidwell <nathan@codesourcery.com>
180 Chung-Lin Tang <cltang@codesourcery.com>
182 * gimplify.c (gimplify_scan_omp_clauses): No special handling for
184 (gimplify_adjust_omp_clauses): Don't delete TILE.
185 (gimplify_omp_for): Deal with TILE.
186 * internal-fn.c (expand_GOACC_TILE): New function.
187 * internal-fn.def (GOACC_DIM_POS): Comment may be overly conservative.
189 * omp-expand.c (struct oacc_collapse): Add tile and outer fields.
190 (expand_oacc_collapse_init): Add LOC paramter. Initialize tile
192 (expand_oacc_collapse_vars): Add INNER parm, adjust for tiling,
193 avoid DIV for outermost collapse var.
194 (expand_oacc_for): Insert tile element loop as needed. Adjust.
195 Remove out of date comments, fix whitespace.
196 * omp-general.c (omp_extract_for_data): Deal with tiling.
197 * omp-general.h (enum oacc_loop_flags): Add OLF_TILE flag,
198 adjust OLF_DIM_BASE value.
199 (struct omp_for_data): Add tiling field.
200 * omp-low.c (scan_sharing_clauses): Allow OMP_CLAUSE_TILE.
201 (lower_oacc_head_mark): Add OLF_TILE as appropriate. Ensure 2 levels
202 for auto loops. Remove default auto determining, moved to
203 oacc_loop_fixed_partitions.
204 * omp-offload.c (struct oacc_loop): Change 'ifns' to vector of call
205 stmts, add e_mask field.
206 (oacc_dim_call): New function, abstracted out from oacc_thread_numbers.
207 (oacc_thread_numbers): Use oacc_dim_call.
208 (oacc_xform_tile): New.
209 (new_oacc_loop_raw): Initialize e_mask, adjust for ifns vector.
210 (finish_oacc_loop): Adjust for ifns vector.
211 (oacc_loop_discover_walk): Append loop abstraction sites to list,
212 add case for GOACC_TILE fns.
213 (oacc_loop_xform_loop): Delete.
214 (oacc_loop_process): Iterate over call list directly, and add
215 handling for GOACC_TILE fns.
216 (oacc_loop_fixed_partitions): Determine default auto, deal with TILE,
218 (oacc_loop_auto_partitions): Add outer_assign parm. Assign all but
219 vector partitioning to outer loops. Assign 2 partitions to loops
220 when available. Add TILE handling.
221 (oacc_loop_partition): Adjust oacc_loop_auto_partitions call.
222 (execite_oacc_device_lower): Process GOACC_TILE fns, ignore unknown specs.
223 * tree-nested.c (convert_nonlocal_omp_clauses): Allow OMP_CLAUSE_TILE.
224 * tree.c (omp_clause_num_ops): Adjust TILE ops.
225 * tree.h (OMP_CLAUSE_TILE_ITERVAR, OMP_CLAUSE_TILE_COUNT): New.
227 2017-02-09 Gerald Pfeifer <gerald@pfeifer.com>
229 * configure.ac (ACX_BUGURL): Update.
230 * configure: Regenerate.
232 2017-02-09 Richard Biener <rguenther@suse.de>
234 PR tree-optimization/69823
235 * graphite-scop-detection.c (scop_detection::harmful_loop_in_region):
236 Properly enumerate all BBs in the region. Use auto_vec/auto_bitmap.
238 2017-02-09 Andrew Burgess <andrew.burgess@embecosm.com>
240 * config/arc/arc-c.def: Add __NPS400__ definition.
241 * config/arc/arc.h (CPP_SPEC): Don't define __NPS400__ here.
242 (TARGET_NPS400): Define.
244 2017-02-09 Andrew Burgess <andrew.burgess@embecosm.com>
246 * config/arc/arc-arch.h (arc_arch_t): Move unchanged to earlier in
248 (arc_cpu_t): Change base_architecture field, arch, to a arc_arc_t
250 (arc_cpu_types): Fill the arch_info field with a pointer into the
251 arc_arch_types table.
252 (arc_selected_cpu): Declare.
253 * config/arc/arc.c (arc_selected_cpu): Make global.
254 (arc_selected_arch): Delete.
255 (arc_base_cpu): Delete.
256 (arc_override_options): Remove references to deleted variables,
257 update access to arch information.
258 (ARC_OPT): Update access to arch information.
259 (ARC_OPTX): Likewise.
260 * config/arc/arc.h (arc_base_cpu): Remove declaration.
261 (TARGET_ARC600): Update access to arch information.
262 (TARGET_ARC601): Likewise.
263 (TARGET_ARC700): Likewise.
264 (TARGET_EM): Likewise.
265 (TARGET_HS): Likewise.
266 * config/arc/driver-arc.c (arc_cpu_to_as): Update access to arch
269 2017-02-08 Pat Haugen <pthaugen@us.ibm.com>
272 * config/rs6000/rs6000.c (rs6000_emit_vector_cond_expr): Invert
273 condition/operands for integer GE/LE/GEU/LEU operations.
275 2017-02-08 Segher Boessenkool <segher@kernel.crashing.org>
278 * config/rs6000/rs6000.opt (maltivec=le, maltivec=be): Fix spelling
281 2017-02-08 Martin Jambor <mjambor@suse.cz>
284 * ipa-prop.c (ipa_alloc_node_params): Make static, return bool
285 whether allocation happened.
286 (ipa_initialize_node_params): Do not call ipa_alloc_node_params if
287 nothing was allocated.
289 2017-02-08 Jakub Jelinek <jakub@redhat.com>
291 PR tree-optimization/79408
292 * tree-vrp.c (simplify_div_or_mod_using_ranges): If op1 is not
293 constant, but SSA_NAME with a known integer range, use the minimum
294 of that range instead of op1 to determine if modulo can be replaced
295 with its first operand.
297 2016-02-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
299 * config/riscv/riscv.c (riscv_build_integer_1): Avoid use of INT16_MAX.
301 2017-02-08 Richard Biener <rguenther@suse.de>
303 PR tree-optimization/71824
304 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
305 Check all loops contained in the merged region.
307 2017-02-07 Andrew Pinski <apinski@cavium.com>
309 * config/aarch64/aarch64.md (popcount<mode>2): New pattern.
311 2017-02-07 Andrew Pinski <apinski@cavium.com>
313 * config/aarch64/aarch64-cores.def (thunderx): Disable LSE.
314 (thunderxt88): Likewise.
315 (thunderxt81): Disable LSE and change v8.1 to v8.
316 (thunderxt83): Likewise.
318 2017-02-07 Jakub Jelinek <jakub@redhat.com>
319 Richard Biener <rguenther@suse.de>
322 * ira-int.h (struct target_ira_int): Change x_max_struct_costs_size
323 type from int to size_t.
324 * ira-costs.c (struct_costs_size): Change type from int to size_t.
326 2017-02-07 Jakub Jelinek <jakub@redhat.com>
328 PR rtl-optimization/79386
329 * cprop.c (bypass_conditional_jumps): Initialize
330 bypass_last_basic_block already before splitting bbs after
331 unconditional traps...
332 (bypass_conditional_jumps): ... rather than here.
335 * config/i386/sse.md (xtg_mode, gatherq_mode): New mode attrs.
336 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2,
337 *avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2): Use them,
338 fix -masm=intel patterns.
340 2017-02-07 Richard Biener <rguenther@suse.de>
342 PR tree-optimization/79256
344 * builtins.c (get_object_alignment_2): Use min_align_of_type
345 to extract alignment for MEM_REFs to honor BIGGEST_FIELD_ALIGNMENT
346 and ADJUST_FIELD_ALIGN.
348 * doc/tm.texi.in (ADJUST_FIELD_ALIGN): Adjust to take additional
350 * doc/tm.texi: Regenerate.
351 * stor-layout.c (layout_decl): Adjust.
352 (update_alignment_for_field): Likewise.
353 (place_field): Likewise.
354 (min_align_of_type): Likewise.
355 * config/arc/arc.h (ADJUST_FIELD_ALIGN): Adjust.
356 * config/epiphany/epiphany.h (ADJUST_FIELD_ALIGN): Likewise.
357 * config/epiphany/epiphany.c (epiphany_adjust_field_align): Likewise.
358 * config/frv/frv.h (ADJUST_FIELD_ALIGN): Likewise.
359 * config/frv/frv.c (frv_adjust_field_align): Likewise.
360 * config/i386/i386.h (ADJUST_FIELD_ALIGN): Likewise.
361 * config/i386/i386.c (x86_field_alignment): Likewise.
362 * config/rs6000/aix.h (ADJUST_FIELD_ALIGN): Likewise.
363 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Likewise.
364 * config/rs6000/freebsd64.h (ADJUST_FIELD_ALIGN): Likewise.
365 * config/rs6000/linux64.h (ADJUST_FIELD_ALIGN): Likewise.
366 * config/rs6000/sysv4.h (ADJUST_FIELD_ALIGN): Likewise.
367 * config/rs6000/rs6000.c (rs6000_special_adjust_field_align_p):
371 2017-01-30 Richard Biener <rguenther@suse.de>
373 PR tree-optimization/79256
374 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
375 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
378 2017-02-07 Toma Tabacu <toma.tabacu@imgtec.com>
380 * config/mips/mips.c (mips_expand_builtin_insn): Convert the QImode
381 argument of the pshufh, psllh, psllw, psrah, psraw, psrlh, psrlw
382 builtins to SImode and emit a zero-extend, if necessary.
384 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
386 * docs/invoke.texi (RISC-V Options): Alphabetize.
388 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
390 * doc/invoke.texi (RISC-V Options): Use two spaces to separate
393 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
395 * config/riscv/riscv.c: New file.
396 * gcc/common/config/riscv/riscv-common.c: Likewise.
397 * config.gcc: Likewise.
398 * config/riscv/constraints.md: Likewise.
399 * config/riscv/elf.h: Likewise.
400 * config/riscv/generic.md: Likewise.
401 * config/riscv/linux.h: Likewise.
402 * config/riscv/multilib-generator: Likewise.
403 * config/riscv/peephole.md: Likewise.
404 * config/riscv/pic.md: Likewise.
405 * config/riscv/predicates.md: Likewise.
406 * config/riscv/riscv-builtins.c: Likewise.
407 * config/riscv/riscv-c.c: Likewise.
408 * config/riscv/riscv-ftypes.def: Likewise.
409 * config/riscv/riscv-modes.def: Likewise.
410 * config/riscv/riscv-opts.h: Likewise.
411 * config/riscv/riscv-protos.h: Likewise.
412 * config/riscv/riscv.h: Likewise.
413 * config/riscv/riscv.md: Likewise.
414 * config/riscv/riscv.opt: Likewise.
415 * config/riscv/sync.md: Likewise.
416 * config/riscv/t-elf-multilib: Likewise.
417 * config/riscv/t-linux: Likewise.
418 * config/riscv/t-linux-multilib: Likewise.
419 * config/riscv/t-riscv: Likewise.
420 * configure.ac: Likewise.
421 * doc/contrib.texi: Add Kito Cheng, Palmer Dabbelt, and Andrew
422 Waterman as RISC-V maintainers.
423 * doc/install.texi: Add RISC-V entries.
424 * doc/invoke.texi: Add RISC-V options section.
425 * doc/md.texi: Add RISC-V constraints section.
426 * configure: Regenerated.
428 2017-02-06 Michael Meissner <meissner@linux.vnet.ibm.com>
431 * config/rs6000/vector.md (vcond<mode><mode>): Allow the true and
432 false values to be constant vectors with all 0 or all 1 bits set.
433 (vcondu<mode><mode>): Likewise.
434 * config/rs6000/predicates.md (vector_int_reg_or_same_bit): New
436 (fpmask_comparison_operator): Update comment.
437 (vecint_comparison_operator): New predicate.
438 * config/rs6000/rs6000.c (rs6000_emit_vector_cond_expr): Optimize
439 vector conditionals when the true and false values are constant
440 vectors with all 0 bits or all 1 bits set.
442 2017-02-06 Martin Sebor <msebor@redhat.com>
444 PR tree-optimization/79376
445 * gimple-fold.c (get_range_strlen): Set the minimum length to zero.
447 2017-02-06 Uros Bizjak <ubizjak@gmail.com>
449 * config/i386/sse.md (vector modes -> vec_extract* splitter): Use
450 explicit subreg RTX with operand 1. Use VECTOR_MODE_P predicate
451 to simplify split condition.
453 2017-02-06 Jakub Jelinek <jakub@redhat.com>
455 * omp-expand.c (oxpand_omp_atomic_fetch_op,
456 expand_omp_atomic_pipeline): Return false if can_atomic_load_p is
459 2017-02-06 Segher Boessenkool <segher@kernel.crashing.org>
461 PR rtl-optimization/68664
462 * target.def (can_speculate_insn): New hook.
463 * doc/tm.texi.in (TARGET_SCHED_CAN_SPECULATE_INSN): New hook.
464 * doc/tm.texi: Regenerate.
465 * sched-rgn.c (can_schedule_ready_p): Use the new hook.
466 * config/rs6000/rs6000.c (TARGET_SCHED_CAN_SPECULATE_INSN): New macro.
467 (rs6000_sched_can_speculate_insn): New function.
469 2017-02-06 Jakub Jelinek <jakub@redhat.com>
471 PR tree-optimization/79284
472 * tree-vectorizer.h (VECT_SCALAR_BOOLEAN_TYPE_P): Define.
473 * tree-vect-stmts.c (vect_get_vec_def_for_operand,
474 vectorizable_mask_load_store, vectorizable_operation,
475 vect_is_simple_cond, get_same_sized_vectype): Use it instead
476 of comparing TREE_CODE of a type against BOOLEAN_TYPE.
477 * tree-vect-patterns.c (check_bool_pattern, search_type_for_mask_1,
478 vect_recog_bool_pattern, vect_recog_mask_conversion_pattern): Likewise.
479 * tree-vect-slp.c (vect_get_constant_vectors): Likewise.
480 * tree-vect-loop.c (vect_determine_vectorization_factor): Likewise.
481 Remove redundant gimple_code (stmt) == GIMPLE_ASSIGN test after
482 is_gimple_assign (stmt). Replace another such test with
483 is_gimple_assign (stmt).
485 2017-02-06 Georg-Johann Lay <avr@gjlay.de>
488 * config/avr/avr.c (rtl-iter.h): Include it.
489 (TARGET_LEGITIMATE_COMBINED_INSN): New hook define...
490 (avr_legitimate_combined_insn): ...and implementation.
492 2017-02-06 Dominik Vogt <vogt@linux.vnet.ibm.com>
494 * config/s390/predicates.md ("larl_operand"): Use macros from hwint.h.
495 * config/s390/s390.c (s390_const_operand_ok)
496 (s390_canonicalize_comparison, s390_extract_part)
497 (s390_single_part, s390_contiguous_bitmask_nowrap_p)
498 (s390_contiguous_bitmask_p, s390_rtx_costs)
499 (legitimize_pic_address): Likewise.
500 * config/s390/s390.md ("clzdi2", "clztidi2"): Likewise.
501 * config/s390/vx-builtins.md ("vec_genbytemaskv16qi")
502 ("vec_permi<mode>", "vfae<mode>", "*vfaes<mode>", "vstrc<mode>")
503 ("*vstrcs<mode>"): Use UINTVAL() to set unsigned HOST_WIDE_INT.
504 * config/s390/vector.md ("vec_vfenes<mode>"): Likewise.
506 2017-02-06 Georg-Johann Lay <avr@gjlay.de>
508 * config/avr/avr.md (*addhi3_zero_extend): Add alternative where
509 REGNO($0) == REGNO($1).
511 2017-02-06 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
513 * config/s390/linux.h(SIZE_TYPE): Add comment.
515 2017-02-06 Julian Brown <julian@codesourcery.com>
516 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
517 Virendra Pathak <virendra.pathak@broadcom.com>
519 * config/aarch64/aarch64-cores.def: Change the scheduler
521 * config/aarch64/aarch64.md: Include thunderx2t99.md.
522 * config/aarch64/thunderx2t99.md: New file.
524 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
526 * doc/standards.texi (Go Language): Update link to language
529 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
531 * tree-eh.c (lower_resx): Sanitize profile.
532 (cleanup_empty_eh_move_lp): Likewise.
534 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
537 * cfgloopmanip.c (lv_adjust_loop_entry_edge, loop_version): Add
539 * cfgloopmanip.h (loop_version): Update prototype.
540 * modulo-sched.c (sms_schedule): Update call of loop_version.
541 * tree-if-conv.c(version_loop_for_if_conversion): Likewise.
542 * tree-parloops.c (gen_parallel_loop): Likewise.
543 * tree-ssa-loop-manip.c (tree_transform_and_unroll_loop): Likewise.
544 * tree-ssa-loop-split.c (split_loop): Likewise.
545 * tree-ssa-loop-unswitch.c (tree_unswitch_loop): Likewise.
546 * tree-vect-loop-manip.c (vect_loop_versioning): Likewise.
548 2017-02-05 Martin Liska <mliska@suse.cz>
551 * config/s390/s390.c (s390_gimplify_va_arg): Initialize local
553 (print_operand_address): Initialize a struct to zero.
555 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
557 * doc/contrib.texi (Contributors): Refer to Hans Boehm's
558 garbage collector only in textual form.
560 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
562 * doc/extend.texi (x86 specific memory model extensions for
563 transactional memory): Simplify a phrase.
565 2017-02-05 Eric Botcazou <ebotcazou@adacore.com>
568 * config/sparc/sync.md (atomic_loaddi_1): Replace 'U' constraint with
569 'r', 'm' constraint with 'T' and !TARGET_ARCH64 with TARGET_ARCH32.
570 (atomic_storedi_1): Likewise.
572 2017-02-04 Jakub Jelinek <jakub@redhat.com>
574 PR tree-optimization/79338
575 * tree-parloops.c (gather_scalar_reductions): Don't call
576 vect_analyze_loop_form for loop->inner before destroying loop's
579 2017-02-03 Martin Sebor <msebor@redhat.com>
581 PR tree-optimization/79327
582 * gimple-ssa-sprintf.c (tree_digits): Avoid adding the base prefix
583 when precision has resulted in leading zeros.
584 (format_integer): Adjust the likely counter to assume an unknown
585 argument that may be zero is non-zero.
587 2017-02-03 Jason Merrill <jason@redhat.com>
590 * tree-inline.c (copy_tree_body_r) [COND_EXPR]: Revert change to
591 avoid copying non-taken branch.
593 2017-02-03 Jakub Jelinek <jakub@redhat.com>
595 PR tree-optimization/79340
596 * tree-vect-loop.c (vectorizable_reduction): Release
597 vec_defs elements after safe_splicing them into other vectors.
600 PR tree-optimization/79327
601 * gimple-ssa-sprintf.c (adjust_range_for_overflow): If returning
602 true, always set *argmin and *argmax to TYPE_{MIN,MAX}_VALUE of
604 (format_integer): Use wide_int_to_tree instead of build_int_cst
605 + to_?hwi. If argmin is NULL, just set argmin and argmax to
606 TYPE_{MIN,MAX}_VALUE of argtype. Simplify and fix computation
607 of shortest and longest sequence.
609 2017-02-03 Uros Bizjak <ubizjak@gmail.com>
611 * config/i386/i386.c (dimode_scalar_chain::convert_reg):
612 Use pextrd for TARGET_SSE4_1 when creating scalar copy.
614 2017-02-03 Walter Lee <walt@tilera.com>
617 * config/tilegx/tilegx.md (tilegx_expand_prologue): Add blockage
618 after initial stackframe link reg save.
619 * config/tilepro/tilepro.md (tilepro_expand_prologue): Likewise.
621 2017-02-03 Jakub Jelinek <jakub@redhat.com>
624 * config/rs6000/rs6000.md (movsi_from_sf): Use wb constraint instead of
625 wu for stxssp alternative.
627 2017-02-03 Martin Sebor <msebor@redhat.com>
629 PR tree-optimization/79352
630 * gimple-fold.c (get_range_strlen): Add argument.
631 (get_range_strlen): Change return type to bool.
632 (get_maxval_strlen): Pass in a dummy argument.
633 * gimple-fold.h (get_range_strlen): Change return type to bool.
634 * gimple-ssa-sprintf.c (get_string_length): Set unlikely counter.
635 * tree.h (array_at_struct_end_p): Add argument.
636 * tree.c (array_at_struct_end_p): Handle it.
638 2017-02-03 Martin Liska <mliska@suse.cz>
641 * multiple_target.c (create_dispatcher_calls): Redirect edge
642 from a caller of a dispatcher.
643 (expand_target_clones): Make the clones local.
644 (ipa_target_clone): Do both target clones and resolvers.
645 (ipa_dispatcher_calls): Remove the pass.
646 (pass_dispatcher_calls::gate): Likewise.
647 (make_pass_dispatcher_calls): Likewise.
648 * passes.def (pass_target_clone): Put as very first IPA early
651 2017-02-03 Martin Liska <mliska@suse.cz>
653 * symtab.c (symtab_node::binds_to_current_def_p): Bail out
654 in case of a function with ifunc attribute.
656 2017-02-03 Martin Liska <mliska@suse.cz>
658 * cgraph.c (cgraph_node::dump): Dump function version info.
659 * symtab.c (symtab_node::dump_base): Add missing new line.
661 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
663 * tree-ssa-ifcombine.c (update_profile_after_ifcombine): New function.
664 (ifcombine_ifandif): Use it.
666 2017-02-03 Martin Liska <mliska@suse.cz>
668 * doc/invoke.texi: Document default value for
669 use-after-scope-direct-emission-threshold.
671 2017-02-03 Martin Liska <mliska@suse.cz>
673 PR tree-optimization/79339
674 * gimple-ssa-sprintf.c (format_floating_max): Call mpfr_clear.
675 (format_floating): Likewise.
677 2017-02-03 Martin Liska <mliska@suse.cz>
680 * ipa-prop.c (ipa_node_params_t::insert): Remove current
682 (ipa_node_params_t::remove): Likewise.
683 * ipa-prop.h (ipa_node_params::ipa_node_params): Make default
684 initialization from removed ipa_node_params_t::insert.
685 (ipa_node_params::~ipa_node_params): Move from removed
686 ipa_node_params_t::release.
687 * symbol-summary.h (symbol_summary::m_released): New member.
688 Do not release a summary twice. Do not allow to call finalizer
689 for types of a summary that live in GGC memory.
691 2017-02-02 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
693 * config/aarch64/aarch64.c (thunderx2t99_tunings): Enable AES and
696 2017-02-02 Martin Sebor <msebor@redhat.com>
699 * gimple-ssa-sprintf.c (get_string_length): Set lower bound to zero.
700 (format_string): Tighten up the range of output for non-constant
701 strings and correct the expected range for wide non-constant strings.
703 2017-02-02 Martin Sebor <msebor@redhat.com>
705 * doc/invoke.texi (-maccumulate-args): Fix bad grammar.
708 * doc/invoke.texi (-fdump-rtl-): Remove pass-specific options from
710 (-fdump-tree-@var): Add to index and document how to come up
711 with pass-specific option and dump file names.
712 (-fdump-passes): Clarify where to look for output.
714 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
717 * tree-ssa-threadbackward.c (profitable_jump_thread_path): Dump
718 statistics of the analyzed path; allow threading for speed when
719 any of BBs along the path are optimized for speed.
721 2017-02-02 Eric Botcazou <ebotcazou@adacore.com>
724 * emit-rtl.c (init_emit): Add ??? comment for problematic alignment
725 settings of the virtual registers.
728 2016-08-23 Dominik Vogt <vogt@linux.vnet.ibm.com>
730 * explow.c (get_dynamic_stack_size): Take known alignment of stack
731 pointer + STACK_DYNAMIC_OFFSET into account when calculating the size
734 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
736 * config/s390/vx-builtins.md ("vec_ceil", "vec_floor")
737 ("vec_trunc", "vec_roundc", "vec_round"): Remove expanders.
739 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
741 * config/s390/s390.md: Add missing comments with the expanded
743 * config/s390/vector.md: Likewise.
744 * config/s390/vx-builtins.md: Likewise.
746 2017-02-02 Jakub Jelinek <jakub@redhat.com>
749 * config/rs6000/rs6000.md (*fixuns_trunc<mode>di2_fctiduz): Rename to ...
750 (fixuns_trunc<mode>di2): ... this, remove previous expander. Put all
751 conditions on a single line.
753 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
755 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
756 __S390_VX__ to __VX__.
758 2017-02-01 Andrew Pinski <apinski@cavium.com>
760 * tree-vect-loop.c (vect_compute_single_scalar_iteration_cost): Pass
761 stmt_info to record_stmt_cost.
762 (vect_get_known_peeling_cost): Pass stmt_info if known to
764 * config/aarch64/aarch64-protos.h (cpu_vector_cost): Split
765 cpu_vector_cost field into
766 scalar_int_stmt_cost and scalar_fp_stmt_cost. Split vec_stmt_cost
767 field into vec_int_stmt_cost and vec_fp_stmt_cost.
768 * config/aarch64/aarch64.c (generic_vector_cost): Update for the
769 splitting of scalar_stmt_cost and vec_stmt_cost.
770 (thunderx_vector_cost): Likewise.
771 (cortexa57_vector_cost): LIkewise.
772 (exynosm1_vector_cost): Likewise.
773 (xgene1_vector_cost): Likewise.
774 (thunderx2t99_vector_cost): Improve after the splitting of the two
776 (aarch64_builtin_vectorization_cost): Update for the splitting of
777 scalar_stmt_cost and vec_stmt_cost.
779 2017-02-01 Torvald Riegel <triegel@redhat.com>
780 Richard Henderson <rth@redhat.com>
782 * builtins.c (fold_builtin_atomic_always_lock_free): Make "lock-free"
783 conditional on existance of a fast atomic load.
784 * optabs-query.c (can_atomic_load_p): New function.
785 * optabs-query.h (can_atomic_load_p): Declare it.
786 * optabs.c (expand_atomic_exchange): Always delegate to libatomic if
787 no fast atomic load is available for the particular size of access.
788 (expand_atomic_compare_and_swap): Likewise.
789 (expand_atomic_load): Likewise.
790 (expand_atomic_store): Likewise.
791 (expand_atomic_fetch_op): Likewise.
792 * testsuite/lib/target-supports.exp
793 (check_effective_target_sync_int_128): Remove x86 because it provides
795 (check_effective_target_sync_int_128_runtime): Likewise.
797 2017-02-01 Richard Biener <rguenther@suse.de>
799 * graphite.c: Include tree-vectorizer.h for find_loop_location.
800 (graphite_transform_loops): Provide opt-info for optimized nests.
801 * tree-parloop.c (parallelize_loops): Provide opt-info for
804 2017-02-01 Richard Biener <rguenther@suse.de>
807 * tree-cfg.c (move_stmt_op): Never set TREE_BLOCK when it
810 2017-02-01 Richard Biener <rguenther@suse.de>
812 PR tree-optimization/71824
813 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
814 Verify the loops are valid in the merged SESE region.
815 (scop_detection::can_represent_loop_1): Check analyzing the
816 evolution of the number of iterations in the region succeeds.
818 2017-01-31 Ian Lance Taylor <iant@golang.org>
820 * config/i386/i386.c (ix86_expand_split_stack_prologue): Add
821 REG_ARGS_SIZE note to 32-bit push insns and call insn.
823 2017-01-31 David Malcolm <dmalcolm@redhat.com>
825 PR preprocessor/79210
826 * input.c (get_substring_ranges_for_loc): Replace line_width
827 assertion with error-handling.
829 2017-01-31 Richard Biener <rguenther@suse.de>
831 PR tree-optimization/77318
832 * graphite-sese-to-poly.c (extract_affine): Fix assert.
833 (create_pw_aff_from_tree): Take loop parameter.
834 (add_condition_to_pbb): Pass loop of the condition to
835 create_pw_aff_from_tree.
837 2017-01-31 Jakub Jelinek <jakub@redhat.com>
839 * config/s390/s390.c (s390_asan_shadow_offset): New function.
840 (TARGET_ASAN_SHADOW_OFFSET): Redefine.
842 2017-01-31 Michael Meissner <meissner@linux.vnet.ibm.com>
846 * config/rs6000/rs6000-protos.h (convert_float128_to_int): Delete,
848 (convert_int_to_float128): Likewise.
849 * config/rs6000/rs6000.c (convert_float128_to_int): Likewise.
850 (convert_int_to_float128): Likewise.
851 * config/rs6000/rs6000.md (UNSPEC_IEEE128_MOVE): Likewise.
852 (UNSPEC_IEEE128_CONVERT): Likewise.
853 (floatsi<mode>2, FLOAT128 iterator): Bypass calling
854 rs6000_expand_float128_convert if we have IEEE 128-bit hardware.
855 Use local variables for IBM extended format.
856 (fix_trunc<mode>si2, FLOAT128 iterator): Likewise.
857 (fix_trunc<mode>si2_fprs): Likewise.
858 (fixuns_trunc<IEEE128:mode><SDI:mode>2): Likewise.
859 (floatuns<IEEE128:mode>2, IEEE128 iterator): Likewise.
860 (fix<uns>_<mode>si2_hw): Rework the IEEE 128-bt hardware support
861 to know that we can now have integers of all sizes in vector
863 (fix<uns>_<mode>di2_hw): Likewise.
864 (float<uns>_<mode>si2_hw): Likewise.
865 (fix_<mode>si2_hw): Likewise.
866 (fixuns_<mode>si2_hw): Likewise.
867 (float<uns>_<mode>di2_hw): Likewise.
868 (float_<mode>di2_hw): Likewise.
869 (float_<mode>si2_hw): Likewise.
870 (floatuns_<mode>di2_hw): Likewise.
871 (floatuns_<mode>si2_hw): Likewise.
872 (xscvqp<su>wz_<mode>): Delete, no longer used.
873 (xscvqp<su>dz_<mode>): Likewise.
874 (xscv<su>dqp_<mode>): Likewise.
875 (ieee128_mfvsrd_64bit): Likewise.
876 (ieee128_mfvsrd_32bit): Likewise.
877 (ieee128_mfvsrwz): Likewise.
878 (ieee128_mtvsrw): Likewise.
879 (ieee128_mtvsrd_64bit): Likewise.
880 (ieee128_mtvsrd_32bit): Likewise.
882 2017-01-31 Martin Liska <mliska@suse.cz>
885 * ipa-prop.c (ipa_free_all_node_params): Call release method
886 instead of ~sumbol_summary to not to trigger double times
889 2017-01-31 Aldy Hernandez <aldyh@redhat.com>
891 PR tree-optimization/71691
892 * bitmap.h (class auto_bitmap): New.
893 * tree-ssa-loop-unswitch.c (tree_may_unswitch_on): Call
894 is_maybe_undefined instead of ssa_undefined_value_p.
896 2017-01-31 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
898 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
899 __S390_ARCH_LEVEL__ to __ARCH__.
901 2017-01-31 Jakub Jelinek <jakub@redhat.com>
903 PR tree-optimization/79267
904 * value-prof.c (gimple_ic): Only drop lhs for noreturn calls
905 if should_remove_lhs_p is true.
907 2017-01-30 Alexandre Oliva <aoliva@redhat.com>
910 * dwarf2out.c (clone_as_declaration): Drop DW_AT_alignment.
911 (add_alignment_attribute): New.
912 (base_type_die): Add alignment attribute.
913 (subrange_type_die): Likewise.
914 (modified_type_die): Likewise.
915 (gen_array_type_die): Likewise.
916 (gen_descr_array_type_die: Likewise.
917 (gen_enumeration_type_die): Likewise.
918 (gen_subprogram_die): Likewise.
919 (gen_variable_die): Likewise.
920 (gen_field_die): Likewise.
921 (gen_ptr_to_mbr_type_die): Likewise.
922 (gen_struct_or_union_type_die): Likewise.
923 (gen_subroutine_type_die): Likewise.
924 (gen_typedef_die): Likewise.
925 (base_type_cmp): Compare alignment attribute.
927 2017-01-30 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
930 * config/rs6000/altivec.md (*setb_internal): Rename to setb_signed.
931 (setb_unsigned) New pattern for setb with CCUNS.
932 * config/rs6000/rs6000.c (expand_block_compare): Use a different
933 subfc./subfe sequence to avoid overflow problems. Generate a
934 shorter sequence with cmpld/setb for power9.
935 * config/rs6000/rs6000.md (subf<mode>3_carry_dot2): Add a new pattern
936 for generating subfc. instruction.
937 (cmpstrsi): Add TARGET_POPCNTD predicate as the generate sequence
938 now uses this instruction.
940 2017-01-30 Ian Lance Taylor <iant@google.com>
943 * dwarf2out.c (gen_type_die_with_usage): When picking a variant
944 for FUNCTION_TYPE/METHOD_TYPE, use the first matching one.
946 2017-01-30 Martin Sebor <msebor@redhat.com>
948 * gimple-ssa-sprintf.c (fmtresult::adjust_for_width_or_precision):
949 Move constant to the right of a relational operator.
950 (get_mpfr_format_length, format_character, format_string): Ditto.
951 (should_warn_p, maybe_warn): Same.
953 * doc/invoke.texi (-Wformat-truncation=1): Fix typo.
955 2017-01-30 Maxim Ostapenko <m.ostapenko@samsung.com>
958 * asan.c (get_translation_unit_decl): Remove function.
959 (asan_add_global): Force has_dynamic_init to zero in LTO mode.
961 2017-01-30 Martin Liska <mliska@suse.cz>
963 PR gcov-profile/79259
964 * opts.c (common_handle_option): Enable flag_ipa_bit_cp w/
967 2017-01-30 Martin Liska <mliska@suse.cz>
970 * config/aarch64/cortex-a57-fma-steering.c (func_fma_steering::analyze):
971 Initialize variables with NULL value.
973 2017-01-30 Richard Earnshaw <rearnsh@arm.com>
976 * config.gcc (arm*-*-*): Add arm/arm-flags.h and arm/arm-isa.h to
978 * arm/arm-protos.h: Don't directly include arm-flags.h and arm-isa.h.
980 2017-01-30 Richard Biener <rguenther@suse.de>
982 PR tree-optimization/79276
983 * tree-vrp.c (process_assert_insertions): Properly adjust common
984 when removing a duplicate.
986 * gcc.dg/torture/pr79276.c: New testcase.
988 2017-01-30 Richard Biener <rguenther@suse.de>
990 PR tree-optimization/79256
991 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
992 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
994 * tree.c (build_aligned_type): Set TYPE_USER_ALIGN.
996 2017-01-30 Dominik Vogt <vogt@linux.vnet.ibm.com>
999 * config/s390/s390.md ("*r<noxa>sbg_<mode>_srl_bitmask")
1000 ("*r<noxa>sbg_<mode>_sll_bitmask")
1001 ("*extzv_<mode>_srl<clobbercc_or_nocc>")
1002 ("*extzv_<mode>_sll<clobbercc_or_nocc>"):
1003 Use contiguous_bitmask_nowrap_operand.
1005 2017-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1008 * config/rs6000/altivec.h (vec_xl): Revise #define.
1009 (vec_xst): Likewise.
1011 2017-01-27 Uros Bizjak <ubizjak@gmail.com>
1013 * config/i386/i386.c (print_reg): Use REGNO instead of true_regnum.
1015 2017-01-27 Bernd Schmidt <bschmidt@redhat.com>
1017 PR rtl-optimization/79194
1018 * cprop.c (one_cprop_pass): Move deletion of code after unconditional
1019 traps before call to bypass_conditional_jumps.
1021 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
1023 PR tree-optimization/71374
1024 * lra-constraints.c (check_conflict_input_operands): New.
1025 (match_reload): Use it.
1027 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
1030 * lra-assigns.c (find_hard_regno_for_1): Take endianess for into
1031 account to calculate conflict_set.
1033 2017-01-27 Bin Cheng <bin.cheng@arm.com>
1035 PR rtl-optimization/78559
1036 * combine.c (try_combine): Discard REG_EQUAL and REG_EQUIV for
1037 other_insn in combine.
1039 2017-01-27 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
1041 * builtin-types.def: Use unsigned_char_type_node for BT_UINT8. Use
1042 uint16_type_node for BT_UINT16.
1044 2017-01-27 David Malcolm <dmalcolm@redhat.com>
1046 * doc/sourcebuild.texi (Testsuites): Add "GIMPLE Tests" and
1047 "RTL Tests" to menu.
1048 (GIMPLE Tests): New node.
1049 (RTL Tests): New node.
1051 2017-01-27 Richard Biener <rguenther@suse.de>
1053 PR tree-optimization/79245
1054 * tree-loop-distribution.c (distribute_loop): Apply cost
1055 modeling also to detected patterns.
1057 2017-01-27 Richard Biener <rguenther@suse.de>
1059 PR tree-optimization/71433
1060 * tree-vrp.c (register_new_assert_for): Revert earlier changes.
1061 (compare_assert_loc): New function.
1062 (process_assert_insertions): Sort and optimize assert locations
1063 to remove duplicates and push down identical assertions on
1064 edges to their destination block.
1066 2017-01-27 Richard Biener <rguenther@suse.de>
1068 PR tree-optimization/79244
1069 * tree-vrp.c (remove_range_assertions): Forcefully propagate
1070 out SSA names even if abnormal.
1072 2017-01-27 Jakub Jelinek <jakub@redhat.com>
1074 * realmpfr.h: Poison MPFR_RND{N,Z,U,D}.
1075 * gimple-ssa-sprintf.c (format_floating_max): Use GMP_RNDN
1076 instead of MPFR_RNDN.
1078 2017-01-27 Richard Earnshaw <rearnsha@arm.com>
1081 * arm.c (arm_option_override): Don't call build_target_option_node
1082 until after doing all option overrides.
1083 (arm_valid_target_attribute_tree): Likewise.
1085 2017-01-27 Martin Liska <mliska@suse.cz>
1087 * doc/invoke.texi (-fprofile-arcs): Document profiling support
1088 for {cd}tors and C++ {cd}tors.
1090 2017-01-27 Dominik Vogt <vogt@linux.vnet.ibm.com>
1092 * config/s390/s390.md ("*setmem_long_and")
1093 ("*setmem_long_and_31z"): Use zero_extend instead of and.
1095 2017-01-26 Martin Sebor <msebor@redhat.com>
1097 * gimple-ssa-sprintf.c (format_floating): Simplify the computation
1100 2017-01-26 Martin Sebor <msebor@redhat.com>
1102 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
1103 HAVE_DFmode before using XFmode or DFmode.
1104 (parse_directive): Avoid using the z length modifier to avoid
1105 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
1108 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
1109 to accept adjustment as an array.
1110 (get_int_range): New function.
1111 (struct directive): Make width and prec arrays.
1112 (directive::set_width, directive::set_precision): Call get_int_range.
1113 (format_integer, format_floating): Handle width and precision ranges.
1114 (format_string, parse_directive): Same.
1116 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1119 * dwarf2out.c (generate_skeleton_bottom_up): For children with
1120 comdat_type_p set, just clone them, but keep the children in the
1124 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
1125 which have direct callers with -fvar-tracking-assignments enabled
1127 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
1128 inside of type units.
1130 2017-01-26 Martin Sebor <msebor@redhat.com>
1133 * gimple-ssa-sprintf.c (struct result_range): Add likely and
1135 (struct format_result): Replace number_chars, number_chars_min,
1136 and number_chars_max with a single member of struct result_range.
1138 (format_result::operator+=): Adjust.
1139 (struct fmtresult): Remove bounded. Handle likely and unlikely
1141 (fmtresult::adjust_for_width_or_precision): New function.
1142 (fmtresult:type_max_digits): New function.
1143 (bytes_remaining): Handle likely and unlikely counters.
1144 (min_bytes_remaining): Remove.
1145 (format_percent): Simplify.
1146 (format_integer, format_floating): Set likely and unlikely counters.
1147 (get_string_length, format_character, format_string): Same.
1148 (format_plain, should_warn_p): New function.
1149 (maybe_warn): Call should_warn_p. Update diagnostic messages
1150 and handle those for all directives, including plain strings.
1151 (format_directive): Handle likely and unlikely counters.
1152 Remove unnecessary quoting from diagnostics. Add an informational
1154 (add_bytes): Remove.
1155 (pass_sprintf_length::compute_format_length): Simplify.
1156 (try_substitute_return_value): Handle likely and unlikely counters.
1158 2017-01-26 Carl Love <cel@us.ibm.com>
1160 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
1161 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
1163 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
1166 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
1167 endianess for subregs into account.
1168 * lra-constraints.c (lra_constraints): Do risky transformations
1169 always on the first iteration.
1170 * lra-lives.c (check_pseudos_live_through_calls): Add arg
1171 last_call_used_reg_set.
1172 (process_bb_lives): Define and use last_call_used_reg_set.
1173 * lra.c (lra): Always continue after lra_constraints on the first
1176 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
1178 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
1180 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
1182 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1184 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
1185 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
1186 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
1187 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
1188 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
1189 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
1190 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
1191 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
1192 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
1194 2017-01-26 Marek Polacek <polacek@redhat.com>
1197 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
1198 for the third operand.
1200 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1203 * omp-low.c (struct omp_context): Add simt_stmt field.
1204 (scan_omp_for): Return omp_context *.
1205 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
1206 context to the _simt_ SIMD stmt.
1207 (lower_omp_for): For combined SIMD with sibling _simt_
1208 SIMD, make sure to use the same decls in _looptemp_
1209 clauses as in the sibling.
1211 2017-01-26 David Sherwood <david.sherwood@arm.com>
1214 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
1217 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1220 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
1221 emit fld b; fld a; if possible.
1223 * brig-builtins.def: Update copyright years.
1224 * config/arm/arm_acle_builtins.def: Update copyright years.
1226 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
1229 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
1230 constraint instead of o for the stxsd instruction.
1232 2017-01-25 Carl Love <cel@us.ibm.com>
1234 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
1235 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
1237 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
1239 * doc/invoke.texi (C++ Dialect Options): Fix typo.
1241 2017-01-25 Richard Biener <rguenther@suse.de>
1243 PR tree-optimization/69264
1244 * target.def (vector_alignment_reachable): Improve documentation.
1245 * doc/tm.texi: Regenerate.
1246 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
1248 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
1249 earlier changes with respect to TYPE_USER_ALIGN.
1250 (vector_alignment_reachable_p): Likewise. Improve dumping.
1252 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1255 * config/arm/arm.md (xordi3): Force constant operand into a register
1258 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1260 * doc/invoke.texi (-fstore-merging): Correct default optimization
1261 levels at which it is enabled.
1262 (-O): Move -fstore-merging from list to...
1265 2017-01-25 Richard Biener <rguenther@suse.de>
1268 * omp-expand.c: Include debug.h.
1269 (expand_omp_taskreg): Make sure to generate early debug before
1270 outlining anything from a function.
1271 (expand_omp_target): Likewise.
1272 (grid_expand_target_grid_body): Likewise.
1274 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
1277 * asan.c (get_translation_unit_decl): New function.
1278 (asan_add_global): Extract modules file name from globals
1279 TRANSLATION_UNIT_DECL name.
1281 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
1284 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
1285 for long calls with APCS frame and VFP.
1287 2017-01-24 David Malcolm <dmalcolm@redhat.com>
1289 * cfg.c (original_copy_tables_initialized_p): New function.
1290 * cfg.h (original_copy_tables_initialized_p): New decl.
1291 * cfgrtl.c (relink_block_chain): Guard the call to
1292 free_original_copy_tables with a call to
1293 original_copy_tables_initialized_p.
1294 * cgraph.h (symtab_node::native_rtl_p): New decl.
1295 * cgraphunit.c (symtab_node::native_rtl_p): New function.
1296 (symtab_node::needed_p): Don't assert for early assembly output
1297 for __RTL functions.
1298 (cgraph_node::finalize_function): Set "force_output" for __RTL
1300 (cgraph_node::analyze): Bail out early for __RTL functions.
1301 (analyze_functions): Update assertion to support __RTL functions.
1302 (cgraph_node::expand): Bail out early for __RTL functions.
1303 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
1305 * function.h (struct function): Update comment for field
1307 * gimple-expr.c: Include "tree-pass.h".
1308 (gimple_has_body_p): Return false for __RTL functions.
1309 * Makefile.in (OBJS): Add run-rtl-passes.o.
1310 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
1312 (gcc::pass_manager::get_clean_slate): New accessor.
1313 * passes.c: Include "insn-addr.h".
1314 (should_skip_pass_p): Add logging. Update logic for running
1315 "expand" to be compatible with both __GIMPLE and __RTL. Guard
1316 property-provider override so it is only done for gimple passes.
1318 (skip_pass): New function.
1319 (execute_one_pass): Call skip_pass when skipping passes.
1320 * read-md.c (md_reader::read_char): Support filtering
1321 the input to a subset of line numbers.
1322 (md_reader::md_reader): Initialize fields
1323 m_first_line and m_last_line.
1324 (md_reader::read_file_fragment): New function.
1325 * read-md.h (md_reader::read_file_fragment): New decl.
1326 (md_reader::m_first_line): New field.
1327 (md_reader::m_last_line): New field.
1328 * read-rtl-function.c (function_reader::create_function): Only
1329 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
1330 curr_properties. Set DECL_INITIAL to a dummy block.
1331 (read_rtl_function_body_from_file_range): New function.
1332 * read-rtl-function.h (read_rtl_function_body_from_file_range):
1334 * run-rtl-passes.c: New file.
1335 * run-rtl-passes.h: New file.
1337 2017-01-24 Jeff Law <law@redhat.com>
1339 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
1342 2017-01-24 Bin Cheng <bin.cheng@arm.com>
1344 PR tree-optimization/79159
1345 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
1346 (record_nonwrapping_iv): Improve boundary using above function if no
1347 value range information.
1349 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
1350 Martin Jambor <mjambor@suse.cz>
1352 * brig-builtins.def: New file.
1353 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
1354 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
1355 (DEF_HSAIL_SAT_BUILTIN): Likewise.
1356 (DEF_HSAIL_INTR_BUILTIN): Likewise.
1357 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
1358 * builtin-types.def (BT_INT8): New.
1359 (BT_INT16): Likewise.
1360 (BT_UINT8): Likewise.
1361 (BT_UINT16): Likewise.
1362 (BT_FN_ULONG): Likewise.
1363 (BT_FN_UINT_INT): Likewise.
1364 (BT_FN_UINT_ULONG): Likewise.
1365 (BT_FN_UINT_LONG): Likewise.
1366 (BT_FN_UINT_PTR): Likewise.
1367 (BT_FN_ULONG_PTR): Likewise.
1368 (BT_FN_INT8_FLOAT): Likewise.
1369 (BT_FN_INT16_FLOAT): Likewise.
1370 (BT_FN_UINT32_FLOAT): Likewise.
1371 (BT_FN_UINT16_FLOAT): Likewise.
1372 (BT_FN_UINT8_FLOAT): Likewise.
1373 (BT_FN_UINT64_FLOAT): Likewise.
1374 (BT_FN_UINT16_UINT32): Likewise.
1375 (BT_FN_UINT32_UINT16): Likewise.
1376 (BT_FN_UINT16_UINT16_UINT16): Likewise.
1377 (BT_FN_INT_PTR_INT): Likewise.
1378 (BT_FN_UINT_PTR_UINT): Likewise.
1379 (BT_FN_LONG_PTR_LONG): Likewise.
1380 (BT_FN_ULONG_PTR_ULONG): Likewise.
1381 (BT_FN_VOID_UINT64_UINT64): Likewise.
1382 (BT_FN_UINT8_UINT8_UINT8): Likewise.
1383 (BT_FN_INT8_INT8_INT8): Likewise.
1384 (BT_FN_INT16_INT16_INT16): Likewise.
1385 (BT_FN_INT_INT_INT): Likewise.
1386 (BT_FN_UINT_FLOAT_UINT): Likewise.
1387 (BT_FN_FLOAT_UINT_UINT): Likewise.
1388 (BT_FN_ULONG_UINT_UINT): Likewise.
1389 (BT_FN_ULONG_UINT_PTR): Likewise.
1390 (BT_FN_ULONG_ULONG_ULONG): Likewise.
1391 (BT_FN_UINT_UINT_UINT): Likewise.
1392 (BT_FN_VOID_UINT_PTR): Likewise.
1393 (BT_FN_UINT_UINT_PTR: Likewise.
1394 (BT_FN_UINT32_UINT64_PTR): Likewise.
1395 (BT_FN_INT_INT_UINT_UINT): Likewise.
1396 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
1397 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
1398 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
1399 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
1400 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
1401 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
1402 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
1403 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
1404 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
1405 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
1406 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
1407 * doc/frontends.texi: List BRIG FE.
1408 * doc/install.texi (Testing): Add BRIG tesring requirements.
1409 * doc/invoke.texi (Overall Options): Mention BRIG.
1410 * doc/standards.texi (Standards): Doucment BRIG HSA version.
1412 2017-01-24 Richard Biener <rguenther@suse.de>
1414 PR translation/79208
1415 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
1417 2017-01-24 Martin Jambor <mjambor@suse.cz>
1420 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
1421 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
1424 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
1427 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
1428 casts from signed to unsigned really don't have a range.
1430 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
1432 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
1433 GMP_RNDx for compatiblity.
1435 2017-01-24 Martin Liska <mliska@suse.cz>
1438 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
1439 that would prevent us to call alloca with -1 as argument.
1441 2017-01-24 Jakub Jelinek <jakub@redhat.com>
1443 * dwarf2out.c (output_compilation_unit_header, output_file_names):
1444 Avoid -Wformat-security warning.
1446 2017-01-23 Andrew Pinski <apinski@cavium.com>
1448 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
1451 2017-01-23 Martin Sebor <msebor@redhat.com>
1454 * gimple-ssa-sprintf.c (warn_level): New global.
1455 (format_integer): Use it here and throughout the rest of the file.
1456 Use the same switch to compute sign as base.
1457 (maybe_warn): New function.
1458 (format_directive): Factor out warnings into maybe_warn.
1459 Add debugging output. Use warn_level.
1460 (add_bytes): Use warn_level.
1461 (pass_sprintf_length::compute_format_length): Add debugging output.
1462 (try_substitute_return_value): Same.
1463 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
1466 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
1467 (struct fmtresult, format_integer, format_floating): Adjust.
1468 (fmtresult::fmtresult): Set max correctly in two argument ctor.
1469 (get_string_length, format_string,format_directive): Same.
1470 (pass_sprintf_length::compute_format_length): Same.
1471 (try_substitute_return_value): Simplify slightly.
1474 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
1475 (fmtresult::operator+=): Outlined.
1476 (struct fmtresult): Add ctors.
1477 (struct conversion_spec): Rename...
1478 (struct directive): ...to this. Add and remove data members.
1479 (directive::set_width, directive::set_precision): New functions.
1480 (format_percent): Use fmtresult ctor.
1481 (get_width_and_precision): Remove.
1482 (format_integer): Make naming changes. Avoid computing width and
1484 (format_floating): Same. Adjust indentation.
1485 (format_character, format_none): New functions.
1486 (format_string): Moved character handling to format_character.
1487 (format_directive): Remove arguments, change return type.
1488 (parse_directive): New function.
1489 (pass_sprintf_length::compute_format_length): Move directive
1490 parsing to parse_directive.
1492 2017-01-23 Jakub Jelinek <jakub@redhat.com>
1494 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
1495 (assign_assembler_name_if_needed): ... this.
1496 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
1497 (assign_assembler_name_if_needed): ... this.
1498 (free_lang_data_in_cgraph): Adjust callers.
1499 * cgraphunit.c (cgraph_node::analyze): Likewise.
1500 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
1503 2017-01-23 Richard Biener <rguenther@suse.de>
1505 PR tree-optimization/79088
1506 PR tree-optimization/79188
1507 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
1508 resetting loop bounds after last path deletion. Reset loop
1509 bounds of the target loop, make code match the comments.
1510 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
1511 Make sure loops need no fixups.
1513 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
1515 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
1516 exponent support with double type for first argument.
1517 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
1518 type returned by __builtin_vec_extract_sig,
1519 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
1520 functions from "vector int" to "vector unsigned int" or from
1521 "vector long long int" to "vector unsigned long long int".
1522 Changed type returned by __builtin_vec_extract_exp,
1523 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
1524 functions from "vector int" to "vector unsigned int" or from
1525 "vector long long int" to "vector unsigned long long int".
1526 Changed return type of __builtin_vec_test_data_class,
1527 __builtin_vec_test_data_class_sp, and
1528 __builtin_vec_test_data_class_dp from "vector int" to
1529 "vector bool int" or from "vector long long int" to "vector bool
1530 long long int" and changed second argument type from "unsigned
1531 int" to "int". Added new overloaded function forms "vector float
1532 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
1533 "vector float __builtin_vec_insert_exp_sp (vector float, vector
1534 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
1535 double, vector unsigned long long int)" and "vector double
1536 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
1537 long int)". Changed return type of
1538 __builtin_scalar_test_data_class and
1539 __builtin_scalar_test_data_class_sp and
1540 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
1541 int" and changed second argument from "unsigned int" to "int".
1542 Changed type returned by __builtin_scalar_test_neg,
1543 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
1544 from "int" to "bool int". Added new overloaded function form
1545 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
1546 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
1547 exponent double-precision with floating point first argument.
1548 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
1549 documentation of scalar_test_data_class, scalar_test_neg,
1550 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
1551 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
1552 vec_test_data_class built-in functions to reflect refinements in
1553 their type signatures.
1555 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
1557 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
1559 (aarch64_elf_asm_destructor): Likewise.
1561 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
1563 PR rtl-optimization/78634
1564 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
1565 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
1566 * ifcvt.c (noce_try_cmove): Add missing cost check.
1568 PR rtl-optimization/71724
1569 * combine.c (if_then_else_cond): Look for situations where it is
1570 beneficial to undo the work of one of the recursive calls.
1572 2017-01-23 Bin Cheng <bin.cheng@arm.com>
1574 PR tree-optimization/70754
1575 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
1576 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
1577 combined stmt before it if not NULL.
1578 (combine_chains): Process refs reversely and compute dominance point
1581 2017-01-23 Martin Liska <mliska@suse.cz>
1583 PR tree-optimization/79196
1584 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
1585 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
1587 (strlen_optimize_stmt): Call the renamed function.
1589 2017-01-23 Michael Matz <matz@suse.de>
1591 PR tree-optimization/78384
1592 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
1594 2017-01-23 Richard Biener <rguenther@suse.de>
1596 PR tree-optimization/79186
1597 * tree-vrp.c (register_new_assert_for): Make sure we've seen
1598 both incoming edges before moving an assert.
1600 2017-01-23 Martin Jambor <mjambor@suse.cz>
1602 * ipa-prop.c (load_from_param_1): Removed.
1603 (load_from_unmodified_param): Bits from load_from_param_1 put back
1605 (load_from_param): Removed.
1606 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
1607 with stmt. Reverted back to use of load_from_unmodified_param.
1609 2017-01-23 Martin Jambor <mjambor@suse.cz>
1612 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
1613 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
1614 field a pointer to garbage collected vector, mark lattices and
1615 ipcp_orig_node with GTY((skip)).
1616 (ipa_get_param_count): Adjust to descriptors being a pointer.
1617 (ipa_get_param): Likewise.
1618 (ipa_get_type): Likewise.
1619 (ipa_get_param_move_cost): Likewise.
1620 (ipa_set_param_used): Likewise.
1621 (ipa_get_controlled_uses): Likewise.
1622 (ipa_set_controlled_uses): Likewise.
1623 (ipa_is_param_used): Likewise.
1624 (ipa_node_params_t): Move into garbage collector. New methods insert
1626 (ipa_node_params_sum): Annotate wth GTY(()).
1627 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
1629 (ipa_load_from_parm_agg): Adjust declaration.
1630 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
1631 * ipa-profile.c (ipa_profile): Likewise.
1632 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
1633 (ipa_populate_param_decls): Make descriptors parameter garbage
1635 (ipa_dump_param): Adjust to descriptors being a pointer.
1636 (ipa_alloc_node_params): Likewise.
1637 (ipa_initialize_node_params): Likewise.
1638 (load_from_param_1): Make descriptors parameter garbage collected.
1639 (load_from_unmodified_param): Likewise.
1640 (load_from_param): Likewise.
1641 (ipa_load_from_parm_agg): Likewise.
1642 (ipa_node_params::~ipa_node_params): Removed.
1643 (ipa_free_all_node_params): Remove call to delete operator.
1644 (ipa_node_params_t::insert): New.
1645 (ipa_node_params_t::remove): Likewise.
1646 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
1647 copy known_csts and known_contexts vectors.
1648 (ipa_read_node_info): Adjust to descriptors being a pointer.
1649 (ipcp_modif_dom_walker): Make m_descriptors field garbage
1651 (ipcp_transform_function): Make descriptors variable garbage
1654 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
1656 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
1657 * config/i386/avx512dqintrin.h: Ditto.
1658 * config/i386/avx512fintrin.h: Ditto.
1659 * gcc/config/i386/i386.c: Handle new builtins.
1660 * config/i386/i386-builtin.def: Add new builtins.
1661 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
1662 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
1664 2017-01-23 Jakub Jelinek <jakub@redhat.com>
1665 Martin Liska <mliska@suse.cz>
1667 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
1668 * asan.c (asan_expand_poison_ifn): Support stores and use
1669 appropriate ASAN report function.
1670 * internal-fn.c (expand_ASAN_POISON_USE): New function.
1671 * internal-fn.def (ASAN_POISON_USE): Declare.
1672 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
1673 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
1674 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
1675 ASAN_POISON calls w/o LHS.
1676 * tree-ssa.c (execute_update_addresses_taken): Create clobber
1677 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
1678 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
1679 * gimplify.c (asan_poison_variables): Add attribute
1680 use_after_scope_memory to variables that really needs to live
1682 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
1683 having the attribute.
1685 2017-01-23 Martin Liska <mliska@suse.cz>
1687 * asan.c (create_asan_shadow_var): New function.
1688 (asan_expand_poison_ifn): Likewise.
1689 * asan.h (asan_expand_poison_ifn): New declaration.
1690 * internal-fn.c (expand_ASAN_POISON): Likewise.
1691 * internal-fn.def (ASAN_POISON): New builtin.
1692 * sanopt.c (pass_sanopt::execute): Expand
1693 asan_expand_poison_ifn.
1694 * tree-inline.c (copy_decl_for_dup_finish): Make function
1696 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
1697 * tree-ssa.c (is_asan_mark_p): New function.
1698 (execute_update_addresses_taken): Rewrite local variables
1699 (identified just by use-after-scope as addressable) into SSA.
1701 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
1703 * doc/install.texi (Specific): opensource.apple.com uses https
1704 now. Remove trailing slash.
1706 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
1708 * README.Portability: Remove note on an Irix compatibility issue.
1710 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
1712 * gcov.c (INCLUDE_ALGORITHM): Define.
1713 (INCLUDE_VECTOR): Define.
1714 No longer include <vector> and <algorithm> directly.
1716 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
1718 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
1720 * doc/invoke.texi (Code Gen Options): Ditto.
1722 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
1725 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
1727 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
1729 rtl-optimization/79125
1730 * cprop.c (local_cprop_pass): Handle cases where we make an
1733 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
1737 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
1738 read from, for big endian.
1740 2017-01-20 Jiong Wang <jiong.wang@arm.com>
1742 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
1743 register pauth builtins for LP64 only.
1745 2017-01-20 Marek Polacek <polacek@redhat.com>
1748 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
1751 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1753 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
1755 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
1756 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
1757 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
1759 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1762 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
1763 in big-endian mode when they are not a single duplicated value.
1765 2017-01-20 Richard Biener <rguenther@suse.de>
1767 * BASE-VER: Bump to 7.0.1.
1769 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1771 * omp-low.c (omplow_simd_context): New struct. Use it...
1772 (lower_rec_simd_input_clauses): ...here and...
1773 (lower_rec_input_clauses): ...here to hold common data. Adjust all
1774 references to idx, lane, max_vf, is_simt.
1776 2017-01-20 Graham Markall <graham.markall@embecosm.com>
1778 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
1781 2017-01-20 Martin Jambor <mjambor@suse.cz>
1783 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
1784 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
1786 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
1787 (GTFILES): Rename hsa.c to hsa-common.c.
1788 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
1789 * hsa-dump.c: Likewise.
1790 * hsa-gen.c: Likewise.
1791 * hsa-regalloc.c: Likewise.
1792 * ipa-hsa.c: Likewise.
1793 * omp-expand.c: Likewise.
1794 * omp-low.c: Likewise.
1795 * toplev.c: Likewise.
1797 2017-01-20 Marek Polacek <polacek@redhat.com>
1800 * doc/invoke.texi: Document -Wduplicated-branches.
1801 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
1802 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
1803 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
1804 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
1805 return 0 only when not OEP_LEXICOGRAPHIC.
1806 (fold_build_cleanup_point_expr): Use the expression
1807 location when building CLEANUP_POINT_EXPR.
1808 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
1809 * tree.c (add_expr): Handle error_mark_node.
1811 2017-01-20 Martin Liska <mliska@suse.cz>
1814 * tree-profile.c (init_ic_make_global_vars): Do not call
1816 (gimple_init_gcov_profiler): Likewise.
1818 2017-01-20 Martin Liska <mliska@suse.cz>
1821 * cgraph.h (maybe_create_reference): Remove argument and
1823 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
1825 * ipa-cp.c (create_specialized_node): Likewise.
1826 * symtab.c (symtab_node::maybe_create_reference): Handle
1827 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
1829 2017-01-20 Martin Liska <mliska@suse.cz>
1831 * read-rtl-function.c (function_reader::create_function): Use
1832 build_decl instread of build_decl_stat.
1834 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
1836 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
1837 * config/i386/avx512dqintrin.h: Ditto.
1838 * config/i386/avx512fintrin.h: Ditto.
1839 * config/i386/i386-builtin-types.def: Add new types.
1840 * gcc/config/i386/i386.c: Handle new types.
1841 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
1842 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
1843 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
1844 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
1845 (__builtin_ia32_kshiftridi): New.
1846 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
1848 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
1852 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
1853 define to rs6000_init_stack_protect_guard.
1854 (rs6000_init_stack_protect_guard): New function.
1856 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
1857 Yunqiang Su <yunqiang.su@imgtec.com>
1859 * config.gcc (supported_defaults): Add madd4.
1860 (with_madd4): Add validation.
1861 (all_defaults): Add madd4.
1862 * config/mips/mips.opt (mmadd4): New option.
1863 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
1865 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
1867 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
1868 (ISA_HAS_FUSED_MADD4): Likewise.
1869 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
1870 * gcc/doc/install.texi (--with-madd4): Document the new option.
1872 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1874 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
1875 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
1876 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
1877 (aarch64_init_pauth_hint_builtins): New.
1878 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
1879 (aarch64_expand_builtin): Expand new builtins.
1881 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1883 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
1884 * combine-stack-adj.c (no_unhandled_cfa): Handle
1885 REG_CFA_TOGGLE_RA_MANGLE.
1886 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
1887 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
1888 info for return address signing.
1889 (aarch64_expand_epilogue): Likewise.
1891 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1893 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
1894 * config/aarch64/aarch64-protos.h
1895 (aarch64_return_address_signing_enabled): New declaration.
1896 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
1898 (aarch64_expand_prologue): Sign return address before it's pushed onto
1900 (aarch64_expand_epilogue): Authenticate return address fetched from
1902 (aarch64_override_options): Sanity check for ILP32 and ISA level.
1903 (aarch64_attributes): New function attributes for "sign-return-address".
1904 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
1905 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
1906 ("*do_return"): Generate combined instructions according to key index.
1907 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
1908 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
1910 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
1911 * config/aarch64/aarch64.opt (msign-return-address=): New.
1912 * doc/extend.texi (AArch64 Function Attributes): Documents
1913 "sign-return-address=".
1914 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
1916 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
1918 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
1919 overall option summary.
1921 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1923 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
1924 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
1925 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
1926 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
1928 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
1930 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
1931 -mpower9-minmax by default for -mcpu=power9.
1932 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
1933 128-bit floating point.
1935 2017-01-20 Alan Modra <amodra@gmail.com>
1937 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
1938 optimizing for size.
1940 2017-01-20 Alan Modra <amodra@gmail.com>
1943 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
1944 for strcmp and strncmp from corresponding builtin decl.
1946 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
1948 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
1949 instead of i386/rtems-64.h.
1950 * config/i386/rtems-64.h: Remove.
1952 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
1956 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
1958 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
1960 2017-01-19 Tamar Christina <tamar.christina@arm.com>
1962 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
1963 Change int to HOST_WIDE_INT.
1964 * config/aarch64/aarch64-protos.h
1965 (aarch64_simd_gen_const_vector_dup): Likewise.
1966 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
1968 2017-01-19 David Malcolm <dmalcolm@redhat.com>
1970 * langhooks-def.h (lhd_type_for_size): New decl.
1971 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
1972 * langhooks.c (lhd_type_for_size): New function, taken from
1975 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
1977 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
1978 define_bypass for CR latency.
1979 (power9-cracked-alu): Update bypass latency and remove power9-branch.
1980 (power9-alu2): Add define_bypass for CR latency.
1982 (power9-mul): Update insn latency.
1983 (power9-mul-compare): Update insn latency, bypass latency and remove
1986 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1988 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
1990 * config/aarch64/aarch64.md
1991 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
1992 aarch64_nopcrelative_literal_loads.
1993 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
1995 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
1997 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
1999 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
2001 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
2004 * config.gcc (supported_defaults): Add lxc1-sxc1.
2005 (with_lxc1_sxc1): Add validation.
2006 (all_defaults): Add lxc1-sxc1.
2007 * config/mips/mips.opt (mlxc1-sxc1): New option.
2008 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
2010 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
2011 __mips_no_lxc1_sxc1.
2012 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
2013 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
2014 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
2016 2017-01-19 Richard Biener <rguenther@suse.de>
2018 PR tree-optimization/72488
2019 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
2020 sure to restore SSA info.
2021 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
2023 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
2025 PR rtl-optimization/79121
2026 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
2027 of the inner type when shifting an extended value.
2029 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
2032 * symtab.c (symtab_node::equal_address_to): Fix comparing of
2033 interposable aliases.
2035 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
2038 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
2039 Use the evmergelohi instruction.
2040 (mov_si<mode>_e500_subreg4_2_le): Likewise.
2041 (mov_sitf_e500_subreg8_2_be): Likewise.
2042 (mov_sitf_e500_subreg12_2_le): Likewise.
2043 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
2044 (mov_si<mode>_e500_subreg4_2_be): Likewise.
2045 (mov_sitf_e500_subreg8_2_le): Likewise.
2046 (mov_sitf_e500_subreg12_2_be): Likewise.
2048 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2050 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
2051 attribute from vecsimple to vecperm.
2052 (altivec_vbpermq2): Likewise.
2054 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2057 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
2059 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
2060 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
2061 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
2062 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
2063 case where N arg is SIZE_MAX.
2064 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
2065 (cmpstrsi): Add pattern.
2067 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
2069 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2070 __builtin_vec_revb builtins.
2071 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
2072 built-in functions to support generation of the ISA 3.0 XXBR<x>
2073 vector byte reverse instructions.
2074 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
2075 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
2076 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
2077 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
2078 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
2079 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
2080 (P9V_BUILTIN_VEC_REVB): Likewise.
2081 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
2082 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
2083 (p9_xxbrq_v16qi): Likewise.
2084 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
2085 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
2086 (p9_xxbrh_v8hi): Likewise.
2087 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
2088 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
2089 vec_revb built-in functions.
2091 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
2093 PR rtl-optimization/78952
2094 * config/i386/i386.md (any_extract): New code iterator.
2095 (*insvqi_2): Use any_extract for source operand.
2096 (*insvqi_3): Use any_shiftrt for source operand.
2098 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
2100 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
2102 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
2104 2017-01-18 Matthias Klose <doko@ubuntu.com>
2106 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
2108 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2110 * config/rs6000/altivec.h (vec_bperm): Change #define.
2111 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
2112 (altivec_vbpermq2): New define_insn.
2113 (altivec_vbpermd): Likewise.
2114 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
2116 (VBPERMD): Likewise.
2117 (VBPERM): New polymorphic function interface.
2118 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
2119 Add entries for P9V_BUILTIN_VEC_VBPERM.
2120 * doc/extend.texi: Add interfaces for vec_bperm.
2122 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2124 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
2125 first letter of error messages.
2126 (s390_resolve_overloaded_builtin): Likewise.
2127 * config/s390/s390.c (s390_expand_builtin): Likewise.
2128 (s390_invalid_arg_for_unprototyped_fn): Likewise.
2129 (s390_valid_target_attribute_inner_p): Likewise.
2130 * config/s390/s390.md ("tabort"): Likewise.
2132 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
2134 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
2135 (ISA_AVOID_DIV_HILO): New macro.
2136 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
2137 (ISA_HAS_DDIV): Likewise.
2139 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2141 * doc/invoke.texi (fabi-version): Correct number of occurrences.
2143 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2145 * doc/invoke.texi (fabi-version): Spelling fix.
2147 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2150 * doc/invoke.texi (fabi-version): Mention mangling fix for
2153 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2156 * doc/invoke.texi (fabi-version): Document discriminator mangling.
2158 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
2161 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
2162 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
2164 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
2166 (stack_protect_test): Ditto.
2167 * config/rs6000/rs6000.opt (mstack-protector-guard=,
2168 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
2170 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
2171 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
2172 -mstack-protector-guard-offset=.
2173 (RS/6000 and PowerPC Options): Ditto.
2175 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
2177 * config/i386/i386.h (MASK_CLASS_P): New define.
2178 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
2179 there are no registers from different register sets also when
2180 mask registers are used. Update function comment.
2181 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
2182 to (*k/*r) and (*k/*km) alternatives.
2184 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
2186 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
2187 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
2188 (EH_RETURN_HANDLER_RTX): New define.
2189 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
2190 Force frame pointer in EH return functions.
2191 (aarch64_expand_epilogue): Add barrier for eh_return.
2192 (aarch64_final_eh_return_addr): Remove.
2193 (aarch64_eh_return_handler_rtx): New function.
2194 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
2196 (aarch64_eh_return_handler_rtx): New prototype.
2198 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2200 * config/rs6000/altivec.h (vec_rlmi): New #define.
2201 (vec_vrlnm): Likewise.
2202 (vec_rlnm): Likewise.
2203 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
2204 (UNSPEC_VRLNM): Likewise.
2205 (VIlong): New mode iterator.
2206 (altivec_vrl<VI_char>mi): New define_insn.
2207 (altivec_vrl<VI_char>nm): Likewise.
2208 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
2211 (RLNM): New polymorphic function entry.
2212 (VRLWMI): New monomorphic function entry.
2214 (RLMI): New polymorphic function entry.
2215 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
2216 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
2217 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
2220 2017-01-17 Jakub Jelinek <jakub@redhat.com>
2223 * dwarf2out.c (field_byte_offset): Restore the
2224 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
2225 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
2226 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
2229 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
2232 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
2234 2017-01-17 Jakub Jelinek <jakub@redhat.com>
2237 * dwarf2out.c (add_data_member_location_attribute): For constant
2238 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
2239 instead of DW_AT_data_member_location, DW_AT_bit_offset and
2240 DW_AT_byte_size attributes.
2242 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
2244 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
2245 after forcing to constant memory when the code model is medium.
2247 2017-01-17 Julia Koval <julia.koval@intel.com>
2250 * config/i386/avx512fintrin.h
2251 (_mm512_i32gather_ps): Change __addr type to void const*.
2252 (_mm512_mask_i32gather_ps): Ditto.
2253 (_mm512_i32gather_pd): Ditto.
2254 (_mm512_mask_i32gather_pd): Ditto.
2255 (_mm512_i64gather_ps): Ditto.
2256 (_mm512_mask_i64gather_ps): Ditto.
2257 (_mm512_i64gather_pd): Ditto.
2258 (_mm512_mask_i64gather_pd): Ditto.
2259 (_mm512_i32gather_epi32): Ditto.
2260 (_mm512_mask_i32gather_epi32): Ditto.
2261 (_mm512_i32gather_epi64): Ditto.
2262 (_mm512_mask_i32gather_epi64): Ditto.
2263 (_mm512_i64gather_epi32): Ditto.
2264 (_mm512_mask_i64gather_epi32): Ditto.
2265 (_mm512_i64gather_epi64): Ditto.
2266 (_mm512_mask_i64gather_epi64): Ditto.
2267 (_mm512_i32scatter_ps): Change __addr type to void*.
2268 (_mm512_mask_i32scatter_ps): Ditto.
2269 (_mm512_i32scatter_pd): Ditto.
2270 (_mm512_mask_i32scatter_pd): Ditto.
2271 (_mm512_i64scatter_ps): Ditto.
2272 (_mm512_mask_i64scatter_ps): Ditto.
2273 (_mm512_i64scatter_pd): Ditto.
2274 (_mm512_mask_i64scatter_pd): Ditto.
2275 (_mm512_i32scatter_epi32): Ditto.
2276 (_mm512_mask_i32scatter_epi32): Ditto.
2277 (_mm512_i32scatter_epi64): Ditto.
2278 (_mm512_mask_i32scatter_epi64): Ditto.
2279 (_mm512_i64scatter_epi32): Ditto.
2280 (_mm512_mask_i64scatter_epi32): Ditto.
2281 (_mm512_i64scatter_epi64): Ditto.
2282 (_mm512_mask_i64scatter_epi64): Ditto.
2283 * config/i386/avx512pfintrin.h
2284 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
2285 (_mm512_mask_prefetch_i32gather_ps): Ditto.
2286 (_mm512_mask_prefetch_i64gather_pd): Ditto.
2287 (_mm512_mask_prefetch_i64gather_ps): Ditto.
2288 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
2289 (_mm512_prefetch_i32scatter_ps): Ditto.
2290 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
2291 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
2292 (_mm512_prefetch_i64scatter_pd): Ditto.
2293 (_mm512_prefetch_i64scatter_ps): Ditto.
2294 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
2295 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
2296 * config/i386/avx512vlintrin.h
2297 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
2298 (_mm_mmask_i32gather_ps): Ditto.
2299 (_mm256_mmask_i32gather_pd): Ditto.
2300 (_mm_mmask_i32gather_pd): Ditto.
2301 (_mm256_mmask_i64gather_ps): Ditto.
2302 (_mm_mmask_i64gather_ps): Ditto.
2303 (_mm256_mmask_i64gather_pd): Ditto.
2304 (_mm_mmask_i64gather_pd): Ditto.
2305 (_mm256_mmask_i32gather_epi32): Ditto.
2306 (_mm_mmask_i32gather_epi32): Ditto.
2307 (_mm256_mmask_i32gather_epi64): Ditto.
2308 (_mm_mmask_i32gather_epi64): Ditto.
2309 (_mm256_mmask_i64gather_epi32): Ditto.
2310 (_mm_mmask_i64gather_epi32): Ditto.
2311 (_mm256_mmask_i64gather_epi64): Ditto.
2312 (_mm_mmask_i64gather_epi64): Ditto.
2313 (_mm256_i32scatter_ps): Change __addr type to void*.
2314 (_mm256_mask_i32scatter_ps): Ditto.
2315 (_mm_i32scatter_ps): Ditto.
2316 (_mm_mask_i32scatter_ps): Ditto.
2317 (_mm256_i32scatter_pd): Ditto.
2318 (_mm256_mask_i32scatter_pd): Ditto.
2319 (_mm_i32scatter_pd): Ditto.
2320 (_mm_mask_i32scatter_pd): Ditto.
2321 (_mm256_i64scatter_ps): Ditto.
2322 (_mm256_mask_i64scatter_ps): Ditto.
2323 (_mm_i64scatter_ps): Ditto.
2324 (_mm_mask_i64scatter_ps): Ditto.
2325 (_mm256_i64scatter_pd): Ditto.
2326 (_mm256_mask_i64scatter_pd): Ditto.
2327 (_mm_i64scatter_pd): Ditto.
2328 (_mm_mask_i64scatter_pd): Ditto.
2329 (_mm256_i32scatter_epi32): Ditto.
2330 (_mm256_mask_i32scatter_epi32): Ditto.
2331 (_mm_i32scatter_epi32): Ditto.
2332 (_mm_mask_i32scatter_epi32): Ditto.
2333 (_mm256_i32scatter_epi64): Ditto.
2334 (_mm256_mask_i32scatter_epi64): Ditto.
2335 (_mm_i32scatter_epi64): Ditto.
2336 (_mm_mask_i32scatter_epi64): Ditto.
2337 (_mm256_i64scatter_epi32): Ditto.
2338 (_mm256_mask_i64scatter_epi32): Ditto.
2339 (_mm_i64scatter_epi32): Ditto.
2340 (_mm_mask_i64scatter_epi32): Ditto.
2341 (_mm256_i64scatter_epi64): Ditto.
2342 (_mm256_mask_i64scatter_epi64): Ditto.
2343 (_mm_i64scatter_epi64): Ditto.
2344 (_mm_mask_i64scatter_epi64): Ditto.
2345 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
2346 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
2347 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
2348 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
2349 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
2350 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
2351 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
2352 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
2353 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
2354 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
2355 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
2356 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
2357 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
2358 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
2359 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
2360 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
2361 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
2362 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
2363 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
2364 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
2365 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
2366 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
2367 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
2368 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
2369 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
2370 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
2371 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
2372 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
2373 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
2374 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
2375 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
2376 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
2377 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
2378 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
2379 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
2380 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
2381 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
2382 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
2383 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
2384 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
2385 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
2386 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
2387 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
2388 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
2389 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
2390 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
2391 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
2392 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
2393 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
2394 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
2395 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
2396 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
2397 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
2398 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
2399 definitions accordingly.
2401 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
2402 Kuan-Lin Chen <kuanlinchentw@gmail.com>
2405 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
2408 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
2411 * ira-conflicts.c (ira_build_conflicts): Update total conflict
2412 hard regs for inner regno.
2414 2017-01-17 Martin Liska <mliska@suse.cz>
2417 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
2418 assumption and add comment.
2420 2017-01-17 Nathan Sidwell <nathan@acm.org>
2422 * ipa-visibility.c (localize_node): New function, broken out of ...
2423 (function_and_variable_visibility): ... here. Call it.
2425 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
2428 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
2429 correctly set frequency of oudgoing edge.
2430 (duplicate_thread_path): Fix profile updating.
2432 2017-01-17 Jakub Jelinek <jakub@redhat.com>
2435 * configure.ac: Add GCC_BASE_VER.
2436 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
2437 version from BASE-VER file.
2438 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
2439 (gcc.o): Depend on $(BASEVER).
2440 * common.opt (dumpfullversion): New option.
2441 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
2442 * doc/invoke.texi: Document -dumpfullversion.
2443 * doc/install.texi: Document --with-gcc-major-version-only.
2444 * configure: Regenerated.
2446 2017-01-17 Richard Biener <rguenther@suse.de>
2448 PR tree-optimization/71433
2449 * tree-vrp.c (register_new_assert_for): Merge same asserts
2450 on all incoming edges.
2451 (process_assert_insertions_for): Handle insertions at the
2454 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
2456 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
2457 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
2459 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
2462 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
2465 2017-01-17 Alan Modra <amodra@gmail.com>
2468 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
2469 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
2470 symbolic stack limit when pic.
2472 2017-01-16 Martin Sebor <msebor@redhat.com>
2474 PR tree-optimization/78608
2475 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
2477 2017-01-16 Jeff Law <law@redhat.com>
2480 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
2481 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
2482 for several include directories that may be relative to sysroot.
2483 * config/i386/x-mingw32 (gplus_includedir): Define.
2484 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
2485 (native_system_includedir): Likewise.
2486 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
2487 override if TARGET_SYSTEM_ROOT is defined.
2488 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
2490 PR tree-optimization/79090
2491 PR tree-optimization/33562
2492 PR tree-optimization/61912
2493 PR tree-optimization/77485
2494 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
2495 and computed trims into the dump file.
2497 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
2499 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
2501 2017-01-16 Jakub Jelinek <jakub@redhat.com>
2504 * gimplify.c (gimplify_init_constructor): If want_value and
2505 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
2509 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
2510 sequence. Formatting fixes.
2511 (doloop_optimize): Formatting fixes.
2514 * gcc.c (debug_level_greater_than_spec_func): New function.
2515 (static_spec_functions): Add debug-level-gt spec function.
2516 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
2518 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
2519 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
2520 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
2521 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
2522 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
2523 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
2525 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
2527 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
2528 QImode fixups to general and mask registers only.
2530 2017-01-16 Carl Love <cel@us.ibm.com>
2532 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
2533 for built-in functions
2534 vector signed char vec_nabs (vector signed char)
2535 vector signed short vec_nabs (vector signed short)
2536 vector signed int vec_nabs (vector signed int)
2537 vector signed long long vec_nabs (vector signed long long)
2538 vector float vec_nabs (vector float)
2539 vector double vec_nabs (vector double)
2540 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
2542 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
2543 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
2544 * doc/extend.texi: Update the documentation file for the new built-in
2547 2017-01-16 Martin Sebor <msebor@redhat.com>
2549 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
2552 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2554 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
2555 UNSPEC_VSX__XXSPLTD to require special splat handling.
2557 2017-01-16 David Malcolm <dmalcolm@redhat.com>
2560 * system.h: Poison strndup.
2562 2017-01-16 Alan Modra <amodra@gmail.com>
2565 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
2568 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
2570 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
2572 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
2574 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
2575 call recog here. Assert that INSN_CODE (insn) is non-negative.
2577 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
2580 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
2582 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
2583 in the currently scheduled RTL fragment.
2585 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
2587 PR rtl-optimization/78751
2588 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
2591 2017-01-14 Jeff Law <law@redhat.com>
2593 PR tree-optimization/79090
2594 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
2595 variable length stores.
2596 (compute_trims): Delete dead assignment to *trim_tail.
2597 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
2600 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
2602 PR rtl-optimization/78626
2603 PR rtl-optimization/78727
2604 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
2605 of a block, and split such blocks after everything else is finished.
2607 2017-01-14 Alan Modra <amodra@gmail.com>
2610 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
2611 target legitimate_combined_insn.
2612 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
2613 (rs6000_legitimate_combined_insn): New function.
2614 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
2616 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
2617 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
2618 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
2620 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
2622 * doc/frontends.texi (G++ and GCC): Remove references to Java.
2624 2017-01-13 Jeff Law <law@redhat.com>
2626 PR tree-optimization/33562
2627 PR tree-optimization/61912
2628 PR tree-optimization/77485
2629 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
2631 (delete_dead_assignment): Likewise.
2632 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
2633 statement to delete_dead_call and delete_dead_assignment.
2635 2017-01-13 David Malcolm <dmalcolm@redhat.com>
2638 * substring-locations.c (format_warning_va): Strengthen case 1 so
2639 that both endpoints of the substring must be within the format
2640 range for just the substring to be printed.
2642 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
2644 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
2645 * config/i386/i386.c (ix86_target_string): Add missing options
2646 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
2647 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
2648 flags_other and ix86_target_other to flags2_other. Display unknown
2650 (ix86_valid_target_attribute_inner_p): Add missing options and
2651 reorder options by implied ISAs, as in ix86_target_string.
2653 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
2655 * hash-table.h (hash_table::too_empty_p): New function.
2656 (hash_table::expand): Use it.
2657 (hash_table::traverse): Likewise.
2658 (hash_table::empty_slot): Use sizeof (value_type) instead of
2659 sizeof (PTR) to convert bytes to elements. Shrink the table
2660 if the current size is excessive for the current number of
2663 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
2665 * ira-costs.c (record_reg_classes): Break from the inner loop
2666 early once alt_fail is known to be true. Update outer loop
2667 handling accordingly.
2669 2017-01-13 Jeff Law <law@redhat.com>
2671 * tree-ssa-dse.c (decrement_count): New function.
2672 (increment_start_addr, maybe_trim_memstar_call): Likewise.
2673 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
2674 when we know the partially dead statement is a mem* function.
2676 PR tree-optimization/61912
2677 PR tree-optimization/77485
2678 * tree-ssa-dse.c: Include expr.h.
2679 (maybe_trim_constructor_store): New function.
2680 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
2682 PR tree-optimization/33562
2683 PR tree-optimization/61912
2684 PR tree-optimization/77485
2685 * doc/invoke.texi: Document new dse-max-object-size param.
2686 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
2687 * tree-ssa-dse.c: Include params.h.
2688 (dse_store_status): New enum.
2689 (initialize_ao_ref_for_dse): New, partially extracted from
2691 (valid_ao_ref_for_dse, normalize_ref): New.
2692 (setup_live_bytes_from_ref, compute_trims): Likewise.
2693 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
2694 (maybe_trim_partially_dead_store): Likewise.
2695 (maybe_trim_complex_store): Likewise.
2696 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
2697 Track what bytes live from the original store. Return tri-state
2698 for dead, partially dead or live.
2699 (dse_dom_walker): Add constructor, destructor and new private members.
2700 (delete_dead_call, delete_dead_assignment): New extracted from
2702 (dse_optimize_stmt): Make a member of dse_dom_walker.
2703 Use initialize_ao_ref_for_dse.
2705 PR tree-optimization/33562
2706 PR tree-optimization/61912
2707 PR tree-optimization/77485
2708 * sbitmap.h (bitmap_count_bits): Prototype.
2709 (bitmap_clear_range, bitmap_set_range): Likewise.
2710 * sbitmap.c (bitmap_clear_range): New function.
2711 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
2713 2017-01-13 Martin Liska <mliska@suse.cz>
2716 * function.c (set_cfun): Add new argument force.
2717 * function.h (set_cfun): Likewise.
2718 * ipa-inline-transform.c (inline_call): Use the function when
2719 strict alising from is dropped for function we inline to.
2721 2017-01-13 Richard Biener <rguenther@suse.de>
2723 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
2724 for dumping GIMPLE INTEGER_CSTs.
2726 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2728 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
2729 to 201112L since C++17.
2731 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
2734 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
2735 if -fsanitize=kernel-address is present.
2737 2017-01-13 Richard Biener <rguenther@suse.de>
2739 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
2740 as _Literal ( type ) number in case usual suffixes do not
2741 preserve all information.
2743 2017-01-13 Richard Biener <rguenther@suse.de>
2745 PR tree-optimization/77283
2746 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
2747 and ssa-iterators.h.
2748 (is_feasible_trace): Implement a cost model based on joiner
2751 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
2754 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
2755 char or short to __float128/_Float128 directly.
2757 2017-01-12 Martin Sebor <msebor@redhat.com>
2759 to -Wformat-overflow.
2760 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
2761 (min_bytes_remaining): Same.
2762 (get_string_length): Same.
2763 (format_string): Same.
2764 (format_directive): Same.
2766 (pass_sprintf_length::handle_gimple_call): Same.
2768 2017-01-12 Jakub Jelinek <jakub@redhat.com>
2770 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
2771 info.nowrite calls with no lhs that can't throw. Return bool
2772 whether gsi_remove has been called or not.
2773 (pass_sprintf_length::handle_gimple_call): Return bool whether
2774 try_substitute_return_value called gsi_remove. Formatting fix.
2775 (pass_sprintf_length::execute): Don't use gsi_remove if
2776 handle_gimple_call returned true.
2779 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
2780 be removed due to side-effects, don't remove following barrier nor
2781 turn the successor edge into fallthru edge.
2783 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2786 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
2787 element-reversing loads and stores as not swappable.
2789 2017-01-12 Nathan Sidwell <nathan@acm.org>
2790 Nicolai Stange <nicstange@gmail.com>
2792 * combine.c (try_combine): Don't ignore result of overlap checking
2793 loop. Combine overlap & asm check into single loop.
2795 2017-01-12 Richard Biener <rguenther@suse.de>
2797 * tree-pretty-print.c (dump_generic_node): Provide -gimple
2798 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
2800 2017-01-12 Richard Biener <rguenther@suse.de>
2802 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
2803 and TS_TARGET_OPTION directly derive from TS_BASE.
2804 * tree-core.h (tree_optimization_option): Derive from tree_base.
2805 (tree_target_option): Likewise.
2807 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
2809 * config/i386/i386.c (memory_address_length): Increase len
2810 only when rip_relative_addr_p returns false.
2812 2017-01-11 Julia Koval <julia.koval@intel.com>
2814 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
2815 (OPTION_MASK_ISA_SGX_SET): New.
2816 (ix86_handle_option): Handle OPT_msgx.
2817 * config.gcc: Added sgxintrin.h.
2818 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
2819 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
2820 * config/i386/i386.c (ix86_target_string): Add -msgx.
2822 (ix86_option_override_internal): Handle new options.
2823 (ix86_valid_target_attribute_inner_p): Add sgx.
2824 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
2825 * config/i386/i386.opt: Add msgx.
2826 * config/i386/sgxintrin.h: New file.
2827 * config/i386/x86intrin.h: Add sgxintrin.h.
2829 2017-01-11 Jakub Jelinek <jakub@redhat.com>
2832 * fold-const.c (maybe_nonzero_address): Return 1 for function
2834 (tree_single_nonzero_warnv_p): Don't handle function local objects
2838 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
2841 2017-01-11 David Malcolm <dmalcolm@redhat.com>
2844 * opts.c: Include "spellcheck.h"
2845 (struct string_fragment): New struct.
2846 (struct edit_distance_traits<const string_fragment &>): New
2848 (get_closest_sanitizer_option): New function.
2849 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
2851 2017-01-11 Jakub Jelinek <jakub@redhat.com>
2853 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
2855 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
2856 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
2857 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
2858 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
2859 for initial die_offset if dwarf_split_debug_info.
2860 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
2861 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
2863 (output_skeleton_debug_sections): Formatting fix. Use
2864 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
2865 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
2867 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
2869 * config/arm/cortex-a53.md: Add bypasses for
2871 (cortex_a53_r2f): Only use for transfers.
2872 (cortex_a53_f2r): Likewise.
2873 (cortex_a53_r2f_cvt): Add reservation for conversions.
2874 (cortex_a53_f2r_cvt): Likewise.
2876 2017-01-11 Tamar Christina <tamar.christina@arm.com>
2878 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
2879 to all inlined functions, change static to extern.
2881 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
2884 * config/arm/arm.c (legitimize_pic_address): Handle reference to
2886 (arm_assemble_integer): Likewise.
2888 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
2890 * config.gcc: Use new awk script to check CPU, FPU and architecture
2891 parameters for --with-... options.
2892 * config/arm/parsecpu.awk: New file
2893 * config/arm/arm-cpus.in: New file.
2894 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
2896 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
2898 * config/arm/t-arm: Update dependency rules.
2899 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
2900 of processing .def files.
2901 * config/arm/genopt.sh: Deleted.
2902 * config/arm/gentune.sh: Deleted.
2903 * config/arm/arm-cores.def: Deleted.
2904 * config/arm/arm-arches.def: Deleted.
2905 * config/arm/arm-fpus.def: Deleted.
2906 * config/arm/arm-tune.md: Regenerated.
2907 * config/arm/arm-tables.opt: Regenerated.
2908 * config/arm/arm-cpu.h: New generated file.
2909 * config/arm/arm-cpu-data.h: New generated file.
2910 * config/arm/arm-cpu-cdata.h: New generated file.
2912 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
2915 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
2917 (input_varpool_node): Unpack dynamically_initialized bit.
2919 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
2921 PR rtl-optimization/79032
2922 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
2923 the alignment of the adjusted memory reference against that of MODE,
2924 instead of the alignment of the original memory reference.
2926 2017-01-11 Martin Jambor <mjambor@suse.cz>
2928 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
2930 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
2931 decorated functions.
2933 2017-01-11 Richard Biener <rguenther@suse.de>
2935 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
2936 set range/nonnull info for PHI results. Do not set it on
2937 stmts marked for removal.
2939 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
2941 * expr.c (store_field): In the bitfield case, fetch the return value
2942 from the registers before applying a single big-endian adjustment.
2943 Always do a final load for a BLKmode value not larger than a word.
2945 2017-01-10 David Malcolm <dmalcolm@redhat.com>
2948 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2949 that we correctly handle column numbers greater than
2950 LINE_MAP_MAX_COLUMN_NUMBER.
2952 2017-01-10 Martin Sebor <msebor@redhat.com>
2955 * gimple-ssa-sprintf.c (get_destination_size): Call
2956 {init,fini}object_sizes.
2957 * tree-object-size.c (addr_object_size): Adjust.
2958 (pass_through_call): Adjust.
2959 (pass_object_sizes::execute): Adjust.
2960 * tree-object-size.h (fini_object_sizes): Declare.
2962 2017-01-10 Martin Sebor <msebor@redhat.com>
2964 PR tree-optimization/78775
2965 * builtins.c (get_size_range): Move...
2966 * calls.c: ...to here.
2967 (alloc_max_size): Accept zero argument.
2968 (operand_signed_p): Remove.
2969 (maybe_warn_alloc_args_overflow): Call get_size_range.
2970 * calls.h (get_size_range): Declare.
2972 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
2974 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
2975 from TI's devices.csv file as of September 2016.
2976 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
2978 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
2980 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
2981 * doc/invoke.texi: Likewise.
2982 * doc/md.texi: Likewise.
2983 * doc/objc.texi: Likewise.
2985 2017-01-10 Joshua Conner <joshconner@google.com>
2987 * config/arm/fuchsia-elf.h: New file.
2988 * config/fuchsia.h: New file.
2989 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
2990 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
2992 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
2994 2016-01-10 Richard Biener <rguenther@suse.de>
2996 PR tree-optimization/79034
2997 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
2998 Propagate out degenerate PHIs in the joiner.
3000 2017-01-10 Martin Liska <mliska@suse.cz>
3002 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
3003 (sort_congruence_classes_by_decl_uid): Likewise.
3004 (sort_congruence_class_groups_by_decl_uid): Likewise.
3005 (sem_item_optimizer::merge_classes): Sort class, groups in these
3006 classes and members in the groups by DECL_UID of declarations.
3007 This would make merge operations stable.
3009 2017-01-10 Martin Liska <mliska@suse.cz>
3011 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
3012 usage of m_classes_vec.
3013 (sem_item_optimizer::~sem_item_optimizer): Likewise.
3014 (sem_item_optimizer::get_group_by_hash): Likewise.
3015 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
3016 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
3017 (sem_item_optimizer::verify_classes): Likewise.
3018 (sem_item_optimizer::process_cong_reduction): Likewise.
3019 (sem_item_optimizer::dump_cong_classes): Likewise.
3020 (sem_item_optimizer::merge_classes): Likewise.
3021 * ipa-icf.h (congruence_class_hash): Rename from
3022 congruence_class_group_hash. Remove declaration of m_classes_vec.
3024 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
3026 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
3027 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
3028 * config.gcc: Add avx512vpopcntdqintrin.h.
3029 * config/i386/avx512vpopcntdqintrin.h: New.
3030 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
3031 * config/i386/i386-builtin-types.def: Add new types.
3032 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
3033 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
3034 __builtin_ia32_vpopcountq_v8di_mask): New.
3035 * config/i386/i386-c.c (ix86_target_macros_internal): Define
3036 __AVX512VPOPCNTDQ__.
3037 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
3038 (PTA_AVX512VPOPCNTDQ): Define.
3039 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
3040 TARGET_AVX512VPOPCNTDQ_P): Define.
3041 * config/i386/i386.opt: Add mavx512vpopcntdq.
3042 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
3043 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
3045 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3048 * predict.def (PRED_CALL): Set to 67.
3050 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
3052 * expr.c (store_field): In the bitfield case, if the value comes from
3053 a function call and is of an aggregate type returned in registers, do
3054 not modify the field mode; extract the value in all cases if the mode
3055 is BLKmode and the size is not larger than a word.
3057 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
3060 * config/i386/cpuid.h: Fix undefined behavior.
3062 2017-01-04 Jeff Law <law@redhat.com>
3064 PR tree-optimization/79007
3065 PR tree-optimization/67955
3066 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
3067 conservative for pt.null when flag_non_call_exceptions is on.
3069 2017-01-09 Jakub Jelinek <jakub@redhat.com>
3071 PR translation/79019
3072 PR translation/79020
3073 * params.def (PARAM_INLINE_MIN_SPEEDUP,
3074 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
3075 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
3077 * config/avr/avr.opt (maccumulate-args): Likewise.
3078 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
3079 * common.opt (freport-bug): Likewise.
3080 * cif-code.def (CIF_FINAL_ERROR): Likewise.
3081 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
3082 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
3083 translatable string.
3084 * config/i386/i386.c (function_value_32): Likewise.
3085 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
3086 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
3088 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
3089 * common/config/msp430/msp430-common.c (msp430_handle_option):
3091 * symtab.c (symtab_node::verify_base): Likewise.
3092 * opts.c (set_debug_level): Likewise.
3093 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
3094 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
3095 missing whitespace to translatable strings.
3096 * config/avr/avr.md (bswapsi2): Fix typo in comment.
3097 * config/sh/superh.h: Likewise.
3098 * config/i386/xopintrin.h: Likewise.
3099 * config/i386/znver1.md: Likewise.
3100 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
3101 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
3102 * double-int.h (struct double_int): Likewise.
3103 * double-int.c (div_and_round_double): Likewise.
3104 * wide-int.cc: Likewise.
3105 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
3106 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
3107 * cfgcleanup.c (crossjumps_occured): Renamed to ...
3108 (crossjumps_occurred): ... this.
3109 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
3112 PR tree-optimization/78899
3113 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
3114 returning bool return struct loop *, NULL for failure and the new
3116 (versionable_outer_loop_p): Don't version outer loop if it has
3117 dont_vectorized bit set.
3118 (tree_if_conversion): When versioning outer loop, ensure
3119 tree_if_conversion is performed also on the inner loop of the
3120 non-vectorizable outer loop copy.
3121 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
3122 LOOP_VECTORIZED in inner loop of the scalar outer loop and
3123 prevent vectorization of it.
3124 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
3125 the outer loop vectorization of the non-scalar version is attempted
3126 before vectorization of the inner loop in scalar version. If
3127 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
3128 vectorization of its inner loop.
3129 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
3130 has 2 inner loops, rename also on edges from bb whose single pred
3131 is outer_loop->header. Fix typo in function comment.
3133 2017-01-09 Martin Sebor <msebor@redhat.com>
3136 * asan.c (asan_emit_stack_protection): Increase local buffer size
3137 to avoid snprintf truncation warning.
3139 2017-01-09 Andrew Pinski <apinski@cavium.com>
3141 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
3142 to reference thunderx2t99 for the tuning structure
3143 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
3145 (thunderx2t99_extra_costs): This.
3146 * config/aarch64/aarch64-tune.md: Regenerate.
3147 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
3148 (vulcan_addrcost_table): This.
3149 (vulcan_regmove_cost): Rename to ...
3150 (thunderx2t99_regmove_cost): This.
3151 (vulcan_vector_cost): Rename to ...
3152 (thunderx2t99_vector_cost): this.
3153 (vulcan_branch_cost): Rename to ...
3154 (thunderx2t99_branch_cost): This.
3155 (vulcan_tunings): Rename to ...
3156 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
3157 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
3159 2017-01-09 Martin Jambor <mjambor@suse.cz>
3163 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
3164 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
3165 (propagate_vr_accross_jump_function): Use the above function for all
3166 value range computations for pass-through jump functions and type
3167 converasion from explicit value range values.
3168 (ipcp_propagate_stage): Do not attempt to deduce types of formal
3169 parameters from TYPE_ARG_TYPES.
3170 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
3171 (ipa_write_node_info): Stream type of the actual argument.
3172 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
3174 2017-01-09 Martin Liska <mliska@suse.cz>
3177 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
3178 (lookup_compiler): Do not show error message with have_E.
3180 2017-01-09 Jakub Jelinek <jakub@redhat.com>
3182 PR tree-optimization/78938
3183 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
3184 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
3185 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
3186 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
3189 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3191 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
3194 2017-01-09 Richard Biener <rguenther@suse.de>
3196 PR tree-optimization/78997
3197 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
3198 name condition properly.
3200 2017-01-09 Richard Biener <rguenther@suse.de>
3203 * dwarf2out.c (is_cxx): New overload with context.
3204 (is_naming_typedef_decl): Use it.
3206 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
3208 * invoke.texi (Option Summary): Correct spacing in option lists
3209 and add line breaks to fix over-long lines.
3211 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
3215 * extend.texi (Common Variable Attributes): Add xref to GCC
3216 Internals manual to explain mode attribute keywords.
3218 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
3221 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
3222 and Preprocessor Options.
3223 (Options for Linking): Document -pthread here....
3224 (RS/6000 and PowerPC Options): ...not here.
3225 (Solaris 2 Options): ...or here.
3226 * doc/cppopts.texi: Document -pthread.
3228 2017-01-08 Martin Sebor <msebor@redhat.com>
3231 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
3232 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
3233 New member functions.
3234 (format_directive): Used them.
3236 (pass_sprintf_length::handle_gimple_call): Same.
3237 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
3238 to avoid truncation for any argument.
3239 (extract_affine_mul): Same.
3240 * tree.c (get_file_function_name): Same.
3242 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3245 * predict.def (PRED_INDIR_CALL): Set to 86.
3247 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
3249 PR preprocessor/54124
3250 * doc/cppopts.texi: Reformat -d subtable to list the full name
3251 of the options. Add cross-reference to the docs for the general
3252 compiler -d options.
3253 * doc/invoke.texi (Developer Options): Add cross-reference to the
3254 preprocessor-specific -d option documentation.
3256 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
3258 PR preprocessor/13498
3259 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
3260 redudant material, and reflect new command-line options.
3261 (System Headers): Likewise.
3263 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
3265 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
3266 -isystem, and -idirafter. Copy-edit.
3267 * doc/cppopts.texi: Copy-edit. Remove contradiction about
3268 default for -ftrack-macro-expansion. Delete obsolete and
3269 badly-formatted implementation details about -fdebug-cpp output.
3270 * doc/cppwarnopts.texi: Copy-edit.
3272 2017-01-07 David Malcolm <dmalcolm@redhat.com>
3275 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
3276 that the transition from a max line width >= 1<<10 to narrower
3277 lines works correctly.
3279 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
3281 * doc/options.texi (PerFunction): New.
3282 * opt-functions.awk (switch_flags): Map both Optimization and
3283 PerFunction to CL_OPTIMIZATION.
3284 * opth-gen.awk: Test for PerFunction flag along with
3286 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
3287 it only when the latter is present. Skip those that don't in
3288 the hash function generator.
3289 * common.opt (fvar-tracking): Mark as PerFunction instead of
3291 (fvar-tracking-assignments): Likewise.
3292 (fvar-tracking-assignments-toggle): Likewise.
3293 (fvar-tracking-uninit): Likewise.
3295 2017-01-07 Jakub Jelinek <jakub@redhat.com>
3297 PR translation/79018
3298 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
3301 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
3304 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
3305 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
3306 TARGET_LONG_JUMP_TABLE_OFFSETS.
3307 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
3308 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
3309 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
3310 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
3311 * config/m68k/m68k.md (tablejump expander): Likewise.
3312 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
3313 TARGET_LONG_JUMP_TABLE_OFFSETS.
3314 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
3315 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
3317 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
3318 David Holsgrove <david.holsgrove@xilinx.com>
3320 * common/config/microblaze/microblaze-common.c
3321 (TARGET_EXCEPT_UNWIND_INFO): Remove.
3322 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
3324 * config/microblaze/microblaze.c (microblaze_must_save_register)
3325 (microblaze_expand_epilogue, microblaze_return_addr): Handle
3327 (microblaze_eh_return): New function.
3328 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
3329 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
3330 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
3331 * config/microblaze/microblaze.md (eh_return): New pattern.
3333 2017-01-06 Jakub Jelinek <jakub@redhat.com>
3335 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
3336 GCC_DIAGNOSTIC_STRINGIFY): Define.
3338 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
3340 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3342 * config/arm/arm.md (<mcrr>): New.
3344 * config/arm/arm.c (arm_arch5te): New.
3345 (arm_option_override): Set arm_arch5te.
3346 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
3348 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
3349 (arm_mcrr_qualifiers): ... this. New.
3350 (MRRC_QUALIFIERS): Define to...
3351 (arm_mrrc_qualifiers): ... this. New.
3352 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
3354 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
3355 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
3356 (MRRCI, mrrc, MRRC): New.
3357 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
3358 VUNSPEC_MRRC2): New.
3360 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3362 * config/arm/arm.md (<mcr>): New.
3364 * config/arm/arm.c (arm_coproc_builtin_available): Add
3365 support for mcr, mrc, mcr2 and mrc2.
3366 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
3367 (arm_mcr_qualifiers): ... this. New.
3368 (MRC_QUALIFIERS): Define to ...
3369 (arm_mrc_qualifiers): ... this. New.
3370 (MCR_QUALIFIERS): Define to ...
3371 (arm_mcr_qualifiers): ... this. New.
3372 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
3374 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
3375 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
3376 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
3379 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3381 * config/arm/arm.md (*ldc): New.
3385 * config/arm/arm.c (arm_coproc_builtin_available): Add
3386 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
3387 (arm_coproc_ldc_stc_legitimate_address): New.
3388 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
3389 'qualifier_const_pointer'.
3390 (LDC_QUALIFIERS): Define to...
3391 (arm_ldc_qualifiers): ... this. New.
3392 (STC_QUALIFIERS): Define to...
3393 (arm_stc_qualifiers): ... this. New.
3394 * config/arm/arm-protos.h
3395 (arm_coproc_ldc_stc_legitimate_address): New.
3396 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
3397 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
3398 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
3399 stc2, stcl, stc2l): New.
3400 * config/arm/constraints.md (Uz): New.
3401 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
3402 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
3403 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
3404 VUNSPEC_STC2L): New.
3406 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3408 * config/arm/arm.md (<cdp>): New.
3409 * config/arm/arm.c (neon_const_bounds): Rename this ...
3410 (arm_const_bounds): ... this.
3411 (arm_coproc_builtin_available): New.
3412 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
3413 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
3414 (CDP_QUALIFIERS): Define to...
3415 (arm_cdp_qualifiers): ... this. New.
3417 (arm_expand_builtin_args): Add case for 6 arguments.
3418 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
3419 (arm_const_bounds): ... this.
3420 (arm_coproc_builtin_available): New.
3421 * config/arm/arm_acle.h (__arm_cdp): New.
3423 * config/arm/arm_acle_builtins.def (cdp): New.
3425 * config/arm/iterators.md (CDPI,CDP,cdp): New.
3426 * config/arm/neon.md: Rename all 'neon_const_bounds' to
3428 * config/arm/types.md (coproc): New.
3429 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
3430 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
3431 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
3432 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
3434 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3436 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
3437 (UBINOP_QUALIFIERS): New.
3439 (acle_builtin_data): New. Change comment.
3440 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
3441 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
3442 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
3443 arm_acle_builtins.def.
3444 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
3445 (arm_init_acle_builtins): New.
3446 (CRC32_BUILTIN): Remove.
3447 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
3448 crc32cb, crc32ch and crc32cw.
3449 (arm_init_crc32_builtins): Remove.
3450 (arm_init_builtins): Use arm_init_acle_builtins rather
3451 than arm_init_crc32_builtins.
3452 (arm_expand_acle_builtin): New.
3453 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
3454 * config/arm/arm_acle_builtins.def: New.
3456 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3458 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
3459 (arm_builtin_datum): ... this.
3460 (arm_init_neon_builtin): Rename to ...
3461 (arm_init_builtin): ... this. Add a new parameters PREFIX
3462 and USE_SIG_IN_NAME.
3463 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
3464 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
3465 'arm_builtin_datum'.
3466 (arm_init_vfp_builtins): Likewise.
3467 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
3468 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
3469 (arm_expand_neon_args): Rename to ...
3470 (arm_expand_builtin_args): ... this. Rename builtin_arg
3471 enum values and differentiate between ARG_BUILTIN_MEMORY
3472 and ARG_BUILTIN_NEON_MEMORY.
3473 (arm_expand_neon_builtin_1): Rename to ...
3474 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
3475 values, arm_expand_builtin_args and add bool parameter NEON.
3476 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
3477 (arm_expand_vfp_builtin): Likewise.
3478 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
3480 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3483 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
3484 * predict.c (tree_estimate_probability_bb): Reverse direction of
3485 polymorphic call predictor.
3487 2017-01-06 David Malcolm <dmalcolm@redhat.com>
3489 * passes.c (execute_one_pass): Split out pass-skipping logic into...
3490 (determine_pass_name_match): ...this new function and...
3491 (should_skip_pass_p): ...this new function.
3493 2017-01-06 Nathan Sidwell <nathan@acm.org>
3495 * ipa-visibility.c (function_and_variable_visibility): Reformat
3496 comments and long lines. Remove extrneous if.
3497 * symtab.c (symtab_node::make_decl_local): Fix code format.
3498 (symtab_node::set_section_for_node): Fix comment typo.
3500 2017-01-06 Martin Liska <mliska@suse.cz>
3503 * lra-constraints.c: Rename invariant to lra_invariant.
3504 * predict.c (set_even_probabilities): Initialize e to NULL.
3506 2017-01-05 Martin Sebor <msebor@redhat.com>
3508 PR tree-optimization/78910
3509 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
3510 (format_integer): Correct off-by-one error in the handling
3511 of precision with negative numbers in signed conversions..
3513 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
3515 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
3517 2017-01-05 Jakub Jelinek <jakub@redhat.com>
3519 PR tree-optimization/71016
3520 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
3521 factor_out_conditional_conversion. Formatting fix.
3522 (factor_out_conditional_conversion): Add cond_stmt argument.
3523 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
3524 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
3527 2017-01-05 David Malcolm <dmalcolm@redhat.com>
3529 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
3530 read-rtl-function.o, and selftest-rtl.o.
3531 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
3532 (selftest::aarch64_test_loading_full_dump): New function.
3533 (selftest::aarch64_run_selftests): New function.
3534 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
3535 selftest::aarch64_run_selftests.
3536 * config/i386/i386.c
3537 (selftest::ix86_test_loading_dump_fragment_1): New function.
3538 (selftest::ix86_test_loading_call_insn): New function.
3539 (selftest::ix86_test_loading_full_dump): New function.
3540 (selftest::ix86_test_loading_unspec): New function.
3541 (selftest::ix86_run_selftests): Call the new functions.
3542 * emit-rtl.c (maybe_set_max_label_num): New function.
3543 * emit-rtl.h (maybe_set_max_label_num): New decl.
3544 * function.c (instantiate_decls): Guard call to
3545 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
3546 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
3548 * gensupport.c (gen_reader::gen_reader): Pass "false"
3549 for new "compact" param of rtx_reader.
3550 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
3551 rather than an empty string for NULL strings.
3552 * read-md.c: Potentially include config.h rather than bconfig.h.
3553 Wrap include of errors.h with #ifdef GENERATOR_FILE.
3554 (have_error): New global, copied from errors.c.
3555 (md_reader::read_name): Rename to...
3556 (md_reader::read_name_1): ...this, adding "out_loc" param,
3557 and converting "missing name or number" to returning false, rather
3559 (md_reader::read_name): Reimplement in terms of read_name_1.
3560 (md_reader::read_name_or_nil): New function.
3561 (md_reader::read_string): Handle "(nil)" by returning NULL.
3562 (md_reader::md_reader): Add new param "compact".
3563 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
3564 (md_reader::read_file): New method.
3565 * read-md.h (md_reader::md_reader): Add new param "compact".
3566 (md_reader::read_file): New method.
3567 (md_reader::is_compact): New accessor.
3568 (md_reader::read_name): Convert return type from void to file_location.
3569 (md_reader::read_name_or_nil): New decl.
3570 (md_reader::read_name_1): New decl.
3571 (md_reader::m_compact): New field.
3572 (noop_reader::noop_reader): Pass "false" for new "compact" param
3574 (rtx_reader::rtx_reader): Add new "compact" param.
3575 (rtx_reader::read_rtx_operand): Make virtual and convert return
3576 type from void to rtx.
3577 (rtx_reader::read_until): New decl.
3578 (rtx_reader::handle_any_trailing_information): New virtual function.
3579 (rtx_reader::postprocess): New virtual function.
3580 (rtx_reader::finalize_string): New virtual function.
3581 (rtx_reader::m_in_call_function_usage): New field.
3582 (rtx_reader::m_reuse_rtx_by_id): New field.
3583 * read-rtl-function.c: New file.
3584 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
3585 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
3586 (selftest::verify_three_block_rtl_cfg): New decl.
3587 * read-rtl-function.h: New file.
3588 * read-rtl.c: Potentially include config.h rather than bconfig.h.
3589 For host, include function.h, memmodel.h, and emit-rtl.h.
3590 (one_time_initialization): New function.
3591 (struct compact_insn_name): New struct.
3592 (compact_insn_names): New array.
3593 (find_code): Handle insn codes in compact dumps.
3594 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
3595 (bind_subst_iter_and_attr): Likewise.
3596 (add_condition_to_string): Likewise.
3597 (add_condition_to_rtx): Likewise.
3598 (apply_attribute_uses): Likewise.
3599 (add_current_iterators): Likewise.
3600 (apply_iterators): Likewise.
3601 (initialize_iterators): Guard usage of apply_subst_iterator with
3602 #ifdef GENERATOR_FILE.
3603 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
3604 (md_reader::read_mapping): Likewise.
3605 (add_define_attr_for_define_subst): Likewise.
3606 (add_define_subst_attr): Likewise.
3607 (read_subst_mapping): Likewise.
3608 (check_code_iterator): Likewise.
3609 (rtx_reader::read_rtx): Likewise. Move one-time initialization
3611 (one_time_initialization): New function.
3612 (rtx_reader::read_until): New method.
3613 (read_flags): New function.
3614 (parse_reg_note_name): New function.
3615 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
3616 Handle reuse_rtx ids.
3617 Wrap iterator lookup within #ifdef GENERATOR_FILE.
3618 Add parsing support for RTL dumps, mirroring the special-cases in
3619 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
3620 values, and calling handle_any_trailing_information.
3621 (rtx_reader::read_rtx_operand): Convert return type from void
3622 to rtx, returning return_rtx. Handle case 'e'. Call
3623 finalize_string on XSTR and XTMPL fields.
3624 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
3625 "(nil)" values were omitted. Call the postprocess vfunc on the
3627 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
3628 class ctor. Initialize m_in_call_function_usage. Call
3629 one_time_initialization.
3630 * rtl-tests.c (selftest::test_uncond_jump): Call
3631 set_new_first_and_last_insn.
3632 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
3633 * selftest-rtl.c: New file.
3634 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
3635 (selftest::get_insn_by_uid): New decl.
3636 * selftest-run-tests.c (selftest::run_tests): Call
3637 read_rtl_function_c_tests.
3638 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
3639 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
3642 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
3644 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
3645 operands in a special way. Assert that pos+len <= mode precision.
3647 2017-01-05 Jakub Jelinek <jakub@redhat.com>
3649 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
3650 3 argument Alias with unlimited for the negative form.
3651 (fno-vect-cost-model): Removed.
3653 2017-01-05 Martin Liska <mliska@suse.cz>
3655 * hsa-gen.c (gen_hsa_divmod): New function.
3656 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
3658 2017-01-05 Martin Liska <mliska@suse.cz>
3661 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
3664 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
3666 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
3667 small constant length operands.
3669 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
3671 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
3672 between loop iterations.
3674 2017-01-05 Martin Liska <mliska@suse.cz>
3677 * gimplify.c (gimplify_decl_expr): Compare to
3678 asan_poisoned_variables instread of checking flags.
3679 (gimplify_target_expr): Likewise.
3680 (gimplify_expr): Likewise.
3681 (gimplify_function_tree): Conditionally initialize
3682 asan_poisoned_variables.
3684 2017-01-04 Jeff Law <law@redhat.com>
3686 PR tree-optimizatin/78812
3687 * rtl.h (contains_mem_rtx_p): Prototype.
3688 * ifcvt.c (containts_mem_rtx_p): Move from here to...
3689 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
3690 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
3691 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
3692 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
3694 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3696 * input.c (assert_char_at_range): Default-initialize actual_range.
3698 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3700 * df-scan.c (df_ref_create_structure): Make regno unsigned,
3701 to match the caller.
3703 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3705 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
3706 insns after final jump in test to emit dummy move.
3708 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3710 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
3711 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
3713 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3715 * multiple_target.c (create_dispatcher_calls): Init e_next.
3716 * tree-ssa-loop-split.c (split_loop): Init border.
3717 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
3720 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
3725 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
3726 (altivec_register_operand): Do not return true if the operand
3727 contains a SUBREG mixing SImode and SFmode.
3728 (vsx_register_operand): Likewise.
3729 (vsx_reg_sfsubreg_ok): New predicate.
3730 (vfloat_operand): Do not return true if the operand contains a
3731 SUBREG mixing SImode and SFmode.
3732 (vint_operand): Likewise.
3733 (vlogical_operand): Likewise.
3734 (gpc_reg_operand): Likewise.
3735 (int_reg_operand): Likewise.
3736 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
3737 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
3738 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
3740 (rs6000_emit_move_si_sf_subreg): New helper function.
3741 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
3742 fixup SUBREGs involving SImode and SFmode.
3743 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
3744 numbers for the new peephole2 optimization.
3745 (peephole2 for SFmode unions): New peephole2 to optimize cases in
3746 the GLIBC math library that do AND/IOR/XOR operations on single
3747 precision floating point.
3748 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
3749 target macros to say whether we need to avoid SUBREGs mixing
3751 (TARGET_ALLOW_SF_SUBREG): Likewise.
3752 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
3753 (UNSPEC_SI_FROM_SF): Likewise.
3754 (iorxor): Change spacing.
3755 (and_ior_xor): New iterator for AND, IOR, and XOR.
3756 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
3757 (movdi_from_sf_zero_ext): Likewise.
3758 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
3759 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
3760 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
3761 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
3762 (fms<mode>4): Likewise.
3763 (fnma<mode>4): Likewise.
3764 (fnms<mode>4): Likewise.
3765 (nfma<mode>4): Likewise.
3766 (nfms<mode>4): Likewise.
3768 2017-01-04 Marek Polacek <polacek@redhat.com>
3771 * doc/invoke.texi: Document -Wpointer-compare.
3773 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3775 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
3778 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
3779 descriptions for -gdwarf-5 and emit them as uleb128 instead of
3782 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
3785 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
3786 documentation of the powerpc_popcntb_ok attribute.
3787 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
3788 code to issue warning messages if a requested CPU configuration is
3789 not supported by the binary (assembler and loader) toolchain.
3790 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
3791 made to define a built-in function that has been disabled.
3792 (paired_init_builtins): Add assertion to prevent ICE if attempt is
3793 made to define a built-in function that has been disabled.
3794 (altivec_init_builtins): Add comment explaining why definition
3795 of the DST built-in functions is not preceded by an assertion
3796 check. Add assertions to prevent ICE if attempts are made to
3797 define an altivec predicate or an abs* built-in function that has
3799 (htm_init_builtins): Add comment explaining why definition of the
3800 htm built-in functions is not preceded by an assertion check.
3802 2017-01-04 Jeff Law <law@redhat.com>
3804 PR tree-optimizatin/67955
3805 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
3806 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
3807 the points-to solution does not include pt_null. Use DECL_PT_UID
3810 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
3812 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
3813 Use gen_int_mode instead of gen_lopwart for const_int operands.
3815 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3817 PR tree-optimization/71563
3818 * match.pd: Simplify X << Y into X if Y is known to be 0 or
3819 out of range value - has low bits known to be zero.
3821 2017-01-04 Alan Modra <amodra@gmail.com>
3823 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
3824 * configure: Regenerate.
3825 * config.in: Regenerate.
3827 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3830 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
3831 a substring of the message, but strcmp with the whole message. Ifdef
3832 ENABLE_NLS, translate the message first using dgettext.
3834 2017-01-03 Jeff Law <law@redhat.com>
3836 PR tree-optimizatin/78856
3837 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
3838 (mark_threaded_blocks): Remove code to truncate thread paths that
3839 cross multiple loop headers. Instead invalidate the cached loop
3840 iteration information and handle case of a thread path walking
3841 into an irreducible region.
3843 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
3846 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
3847 assertions. Add support for doing the signbit if the IEEE 128-bit
3848 floating point value is in a GPR.
3849 * config/rs6000/rs6000.md (Fsignbit): Delete.
3850 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
3851 Update the length attribute if the value is in a GPR.
3852 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
3853 the sign or zero extension instruction, since the value is always 0/1.
3854 (signbit<mode>2_dm2): Delete using <Fsignbit>.
3857 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
3858 extracting SImode to a GPR register so that we can generate a
3859 store, limit the vector to be in a traditional Altivec register
3860 for the vextuwrx instruction.
3862 2017-01-03 Ian Lance Taylor <iant@google.com>
3864 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
3866 2017-01-03 Martin Sebor <msebor@redhat.com>
3868 PR tree-optimization/78696
3869 * gimple-ssa-sprintf.c (format_floating): Correct handling of
3870 precision. Use MPFR for %f for greater fidelity. Correct handling
3872 (pass_sprintf_length::compute_format_length): Set width and precision
3873 specified by asrerisk to void_node for vararg functions.
3874 (try_substitute_return_value): Adjust dump output.
3876 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
3878 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
3880 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
3882 * doc/invoke.texi (SPARC options): Document -mlra as the default.
3883 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
3884 -mlra/-mno-lra was passed to the compiler.
3886 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
3888 PR rtl-optimization/65618
3889 * emit-rtl.c (try_split): Move initialization of "before" and
3890 "after" to just before the call to emit_insn_after_setloc.
3892 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
3894 * doc/md.texi (Standard Names): Remove reference to Java frontend.
3896 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
3898 * dwarf2out.c (gen_enumeration_type_die): When
3899 -gno-strict-dwarf, add a DW_AT_encoding attribute.
3901 2017-01-03 Jakub Jelinek <jakub@redhat.com>
3903 PR tree-optimization/78965
3904 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
3905 Change first argument from const call_info & to call_info &. For %n
3906 set info.nowrite to false.
3909 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
3910 possibly throwing calls.
3912 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
3913 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
3914 and fns handling, rather than in a separate case SSA_NAME.
3916 2017-01-02 Jeff Law <law@redhat.com>
3918 * config/darwin-driver.c (darwin_driver_init): Const-correctness
3919 fixes for first_period and second_period variables.
3921 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
3924 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
3925 (*insvqi_1): New insn pattern.
3926 (*insvqi_1_mem_rex64): Ditto.
3928 (*insvqi_3): Rename from *insvqi.
3930 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
3932 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
3934 * doc/cfg.texi (Edges): Remove reference to Java.
3935 (Maintaining the CFG): Ditto.
3937 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3940 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
3941 transparent aliases.
3943 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3946 * predict.def (PRED_CALL): Update hitrate.
3947 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
3948 * predict.c (tree_estimate_probability_bb): Split CALL predictor
3949 into direct/indirect/polymorphic variants.
3951 2017-01-01 Jakub Jelinek <jakub@redhat.com>
3953 Update copyright years.
3955 * gcc.c (process_command): Update copyright notice dates.
3956 * gcov-dump.c (print_version): Ditto.
3957 * gcov.c (print_version): Ditto.
3958 * gcov-tool.c (print_version): Ditto.
3959 * gengtype.c (create_file): Ditto.
3960 * doc/cpp.texi: Bump @copying's copyright year.
3961 * doc/cppinternals.texi: Ditto.
3962 * doc/gcc.texi: Ditto.
3963 * doc/gccint.texi: Ditto.
3964 * doc/gcov.texi: Ditto.
3965 * doc/install.texi: Ditto.
3966 * doc/invoke.texi: Ditto.
3968 Copyright (C) 2017 Free Software Foundation, Inc.
3970 Copying and distribution of this file, with or without modification,
3971 are permitted in any medium without royalty provided the copyright
3972 notice and this notice are preserved.