1 2017-02-16 Richard Biener <rguenther@suse.de>
3 * graphite.h: Do not include isl/isl_val_gmp.h, instead include
5 * graphite-isl-ast-to-gimple.c (gmp_cst_to_tree): Remove.
6 (gcc_expression_from_isl_expr_int): Use generic isl_val interface.
7 * graphite-sese-to-poly.c: Do not include isl/isl_val_gmp.h.
8 (isl_val_int_from_wi): New function.
9 (extract_affine_gmp): Rename to ...
10 (extract_affine_wi): ... this, take a widest_int.
11 (extract_affine_int): Just wrap extract_affine_wi.
12 (add_param_constraints): Use isl_val_int_from_wi.
13 (add_loop_constraints): Likewise, and extract_affine_wi.
15 2017-02-15 Jeff Law <law@redhat.com>
18 * ira-costs.c (scan_one_insn): Check have_regs_of_mode before calling
19 ira_init_register_move_cost_if_necessary.
21 2017-02-15 Martin Sebor <msebor@redhat.com>
24 * doc/invoke.texi (-fdump-final-insns): Replace option accidentally
25 removed in a prior commit.
27 2017-02-15 Bin Cheng <bin.cheng@arm.com>
29 PR tree-optimization/79347
30 * tree-vect-loop-manip.c (vect_do_peeling): Maintain profile
31 counters during peeling.
33 2017-02-15 Thomas Schwinge <thomas@codesourcery.com>
35 * Makefile.in (site.exp): Remove "set ISLVER".
37 2017-02-15 Jakub Jelinek <jakub@redhat.com>
40 * real.c (real_from_integer): Call real_convert even for decimal.
42 2017-02-15 Dominik Vogt <vogt@linux.vnet.ibm.com>
45 * config/s390/s390.c: define TARGET_CUSTOM_FUNCTION_DESCRIPTORS.
47 2017-02-14 Andrew Pinski <apinski@cavium.com>
49 * config/aarch64/aarch64-cores.def (thunderx2t99): Move to under 'C"
50 cores and change the partno/implementer to be correct.
51 (thunderx2t99p1): New core which replaces thunderx2t99 and still has
52 the 'B" as the implementer.
53 * config/aarch64/aarch64-tune.md: Regenerate.
55 2017-02-14 Carl Love <cel@us.ibm.com>
57 * config/rs6000/rs6000.c: Add case statement entry to make the
58 xvcvuxdsp built-in argument unsigned.
59 * config/rs6000/vsx.md: Fix the source and return operand types so they
60 match the instruction definitions from the ISA document. Fix typo
61 in the instruction generation for the (define_insn "vsx_xvcvuxdsp"
64 2017-02-14 Vladimir Makarov <vmakarov@redhat.com>
67 * lra-int.h (struct lra_operand_data, struct lra_insn_reg): Add
68 member early_clobber_alts.
69 * lra-lives.c (reg_early_clobber_p): New.
70 (process_bb_lives): Use it.
71 * lra.c (new_insn_reg): New arg early_clobber_alts. Use it.
72 (debug_operand_data): Initialize early_clobber_alts.
73 (setup_operand_alternative): Set up early_clobber_alts.
74 (collect_non_operand_hard_regs): Ditto. Pass early clobber
75 alternatives to new_insn_reg.
76 (add_regs_to_insn_regno_info): Add arg early_clobber_alts. Use
78 (lra_update_insn_regno_info): Pass the new arg.
80 2017-02-14 Jakub Jelinek <jakub@redhat.com>
83 * omp-offload.c (free_oacc_loop): Release loop->ifns vector.
84 (new_oacc_loop_raw): Don't clear already cleared fields.
87 * config/i386/avx512pfintrin.h (_mm512_prefetch_i32gather_pd,
88 _mm512_prefetch_i32gather_ps, _mm512_prefetch_i64gather_pd,
89 _mm512_prefetch_i64gather_ps): New inline functions and macros.
91 2017-02-14 Uros Bizjak <ubizjak@gmail.com>
94 * config/i386/i386.md (*movxf_internal): Add (o,rC) alternative.
96 2017-02-14 H.J. Lu <hongjiu.lu@intel.com>
99 * config/i386/i386.c (timode_scalar_chain::convert_insn): Insert
100 the extra instruction to the right place to store 128-bit constant
103 2017-02-14 Martin Sebor <msebor@redhat.com>
106 * gimple-ssa-sprintf.c (format_directive): Avoid issuing INT_MAX
107 warning for strings of unknown length.
109 2017-02-13 Segher Boessenkool <segher@kernel.crashing.org>
111 * config.gcc (supported_defaults) [powerpc*-*-*]: Update.
113 2017-02-14 Jeff Law <law@redhat.com>
116 * ira-costs.c (scan_one_insn): Initialize register move costs
117 for pseudos seen in USE/CLOBBER insns.
119 PR tree-optimization/79095
120 * tree-vrp.c (extract_range_from_binary_expr_1): For EXACT_DIV_EXPR,
121 if the numerator has the range ~[0,0] make the resultant range ~[0,0].
122 (extract_range_from_binary_expr): For MINUS_EXPR with no derived range,
123 if the operands are known to be not equal, then the resulting range
125 (intersect_ranges): If the new range is ~[0,0] and the old range is
126 wide, then prefer ~[0,0].
127 * tree-vrp.c (overflow_comparison_p_1): New function.
128 (overflow_comparison_p): New function.
129 * tree-vrp.c (register_edge_assert_for_2): Register additional asserts
130 if NAME is used in an overflow test.
131 (vrp_evaluate_conditional_warnv_with_ops): If the ops represent an
132 overflow check that can be expressed as an equality test, then adjust
133 ops to be that equality test.
135 2017-02-14 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
137 * config/s390/s390-builtin-types.def: Remove flags argument.
138 * config/s390/s390.c (s390_init_builtins): Likewise.
140 2017-02-14 Martin Liska <mliska@suse.cz>
142 * tree-ssa-loop-unswitch.c (hoist_guard): Release get_loop_body
143 vector. Fix trailing white spaces.
145 2017-02-14 James Greenhalgh <james.greenhalgh@arm.com>
147 * config/aarch64/aarch64.c (aarch64_simd_container_mode): Handle
150 2017-02-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
152 PR rtl-optimization/68664
153 * config/arm/arm.c (arm_sched_can_speculate_insn):
154 New function. Declare prototype.
155 (TARGET_SCHED_CAN_SPECULATE_INSN): Define.
157 2017-02-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
159 PR rtl-optimization/68664
160 * config/aarch64/aarch64.c (aarch64_sched_can_speculate_insn):
162 (TARGET_SCHED_CAN_SPECULATE_INSN): Define.
164 2017-02-14 Amit Pawar <amit.pawar@amd.com>
166 * config/i386/i386.c (znver1_cost): Fix the alignment for function and
167 max skip bytes for function, loop and jump.
169 2017-02-14 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
171 * gimple-pretty-print.c (dump_unary_rhs): Adjust dump format for
172 ABS_EXPR for gimple dump.
174 2017-02-14 Jakub Jelinek <jakub@redhat.com>
177 * config/sh/sh.c (expand_cbranchdi4): Don't clear operands[4].
179 PR tree-optimization/79408
180 * tree-vrp.c (simplify_div_or_mod_using_ranges): Handle also the
181 case when on TRUNC_MOD_EXPR op0 is INTEGER_CST.
182 (simplify_stmt_using_ranges): Call simplify_div_or_mod_using_ranges
183 also if rhs1 is INTEGER_CST.
185 2017-02-14 Richard Biener <rguenther@suse.de>
188 * tree-into-ssa.c (insert_phi_nodes): When the function can
189 have abnormal edges rewrite SSA names with broken use-def
190 dominance out of SSA and register them for PHI insertion.
192 2017-02-13 Martin Sebor <msebor@redhat.com>
195 * gimple-ssa-sprintf.c (pass_sprintf_length::handle_gimple_call): Avoid
196 clearing info.nowrite flag when snprintf size argument is a range.
198 2017-02-13 Jakub Jelinek <jakub@redhat.com>
200 * cprop.c (cprop_jump): Add missing space in string literal.
201 * tree-ssa-structalias.c (rewrite_constraints): Likewise.
202 (get_constraint_for_component_ref): Likewise.
203 * df-core.c (df_worklist_dataflow_doublequeue): Likewise.
204 * tree-outof-ssa.c (insert_partition_copy_on_edge): Likewise.
205 * lra-constraints.c (process_alt_operands): Likewise.
206 * ipa-inline.c (inline_small_functions): Likewise.
207 * tree-ssa-sccvn.c (visit_reference_op_store): Likewise.
208 * cgraph.c (cgraph_edge::redirect_call_stmt_to_callee): Likewise.
209 * trans-mem.c (diagnose_tm_1_op): Likewise.
210 * omp-grid.c (grid_find_single_omp_among_assignments): Likewise.
211 (grid_parallel_clauses_gridifiable): Likewise.
213 * config/nvptx/mkoffload.c (process): Add space in between
216 * config/i386/i386.h (REG_CLASS_NAMES): Add , in between
217 "MOD4_SSE_REGS" and "ALL_REGS".
219 * spellcheck.c (test_data): Add , in between "foo" and "food".
221 2017-02-13 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
224 * config/rs6000/rs6000.c (expand_block_compare): Make sure runtime
225 boundary crossing check and subsequent code generation agree.
227 2017-02-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
229 * config/aarch64/aarch64.c (has_memory_op): Delete.
230 (aarch64_madd_needs_nop): Use contains_mem_rtx_p instead of
233 2017-02-13 Jakub Jelinek <jakub@redhat.com>
235 PR rtl-optimization/79388
236 PR rtl-optimization/79450
237 * combine.c (distribute_notes): When removing TEM_INSN for which
238 corresponding dest has last value recorded, invalidate that last
241 2017-02-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
243 * config/arm/arm.c (arm_print_tune_info): Use ASM_COMMENT_START instead
244 of explicit '@'. Add missing assembly comment marker on branch costs
247 2017-02-13 Nathan Sidwell <nathan@acm.org>
249 * gengtype-lex.l (<in_struct>): Add '/'.
251 2017-02-13 Martin Liska <mliska@suse.cz>
254 * calls.c (expand_call): Replace XALLOCAVEC with XCNEWVEC.
256 2017-02-13 Richard Biener <rguenther@suse.de>
258 * configure.ac (HAVE_ISL_OPTIONS_SET_SCHEDULE_SERIALIZE_SCCS):
260 * configure: Re-generate.
261 * config.in: Likewise.
262 * graphite-dependences.c: Simplify as if
263 HAVE_ISL_OPTIONS_SET_SCHEDULE_SERIALIZE_SCCS was defined.
264 * graphite-isl-ast-to-gimple.c: Likewise.
265 * graphite-optimize-isl.c: Likewise.
266 * graphite-poly.c: Likewise.
267 * graphite-sese-to-poly.c: Likewise.
268 * graphite.h: Likewise.
269 * toplev.c: Include isl/version.h and use isl_version () for
270 printing the ISL version.
271 * doc/install.texi: Update ISL requirement.
273 2017-02-12 Gerald Pfeifer <gerald@pfeifer.com>
275 * doc/standards.texi (Standards): Update reference to
278 2017-02-12 Gerald Pfeifer <gerald@pfeifer.com>
280 * doc/extend.texi (Named Address Spaces): sourceware.org now
282 * doc/install.texi (Binaries): Ditto.
285 2017-02-11 Sandra Loosemore <sandra@codesourcery.com>
287 * doc/cpp.texi: Replace "stringify"/"stringification" with C
288 standard terminology "stringize"/"stringizing" throughout.
289 * doc/cppinternals.texi: Likewise.
291 2017-02-11 Sandra Loosemore <sandra@codesourcery.com>
293 * doc/extend.texi: Fix some spelling mistakes and typos.
294 * doc/invoke.texi: Likewise.
296 2017-02-11 Jan Hubicka <hubicka@ucw.cz>
299 * params.def (inline-min-speedup) Change from 10 to 8.
301 2017-02-11 Jakub Jelinek <jakub@redhat.com>
303 * doc/invoke.texi (fopenmp): Bump OpenMP version from 4.0 to
306 2017-02-11 Jan Hubicka <hubicka@ucw.cz>
309 * ipa-inline-analysis.c (get_minimal_bb): New function.
310 (record_modified): Use it.
311 (remap_edge_change_prob): Handle also ancestor functions.
313 2017-02-11 Gerald Pfeifer <gerald@pfeifer.com>
315 * doc/contrib.texi (Contributors): Remove broken link into
316 the Mauve CVS repository.
318 2017-02-11 Jakub Jelinek <jakub@redhat.com>
321 * internal-fn.c (expand_vector_ubsan_overflow): Use piece-wise
322 result computation whenever lhs doesn't have vector mode, not
323 just when it has BLKmode.
325 2017-02-10 Gerald Pfeifer <gerald@pfeifer.com>
327 * doc/makefile.texi (profiledbootstrap): Refer to the
328 installation instructions only in textual form.
330 2017-02-10 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
333 * config/rs6000/altivec.md (bcd<bcd_add_sub>): Fix constraints.
335 2017-02-10 Gerald Pfeifer <gerald@pfeifer.com>
337 * doc/install.texi (Specific): Use https for blackfin.uclinux.org.
338 (Specific): Update mingw-w64 reference.
340 (Specific): Remove broken link to Renesas RX processor.
342 2017-02-10 Richard Biener <rguenther@suse.de>
344 * toplev.c (process_options): Do not mention obsolete graphite
345 options when printing sorry message about missing graphite support.
346 Mention -floop-nest-optimize.
348 2017-02-10 Christophe Lyon <christophe.lyon@linaro.org>
350 * config/aarch64/arm_neon.h (vtst_p8): Rewrite without asm.
351 (vtst_p16): Likewise.
352 (vtstq_p8): Likewise.
353 (vtstq_p16): Likewise.
355 (vtstq_p64): Likewise.
356 * config/arm/arm_neon.h (vgetq_lane_p64): New.
357 (vset_lane_p64): New.
358 (vsetq_lane_p64): New.
360 2017-02-10 Jakub Jelinek <jakub@redhat.com>
362 PR tree-optimization/79411
363 * tree-ssa-reassoc.c (is_reassociable_op): Return false if
364 stmt operands are SSA_NAMEs used in abnormal phis.
365 (can_reassociate_p): Return false if op is SSA_NAME used in abnormal
368 2017-02-09 Jan Hubicka <hubicka@ucw.cz>
371 * cgraphunit.c (cgraph_node::add_new_function): Set externally_visible
374 2017-02-09 Jan Hubicka <hubicka@ucw.cz>
376 * tree-ssa-loop-unswitch.c (hoist_guard): Update profile.
378 2017-02-09 Jakub Jelinek <jakub@redhat.com>
380 * omp-offload.c (oacc_loop_auto_partitions): Use || instead of |
384 * gimplify.h (is_gimple_sizepos): Only test for INTEGER_CST constants,
385 not arbitrary TREE_CONSTANT.
388 * gimplify.c (gimplify_adjust_omp_clauses): Ignore
389 "omp declare target link" attribute unless is_global_var.
390 * omp-offload.c (find_link_var_op): Likewise.
392 2017-02-09 Nathan Sidwell <nathan@codesourcery.com>
393 Chung-Lin Tang <cltang@codesourcery.com>
395 * gimplify.c (gimplify_scan_omp_clauses): No special handling for
397 (gimplify_adjust_omp_clauses): Don't delete TILE.
398 (gimplify_omp_for): Deal with TILE.
399 * internal-fn.c (expand_GOACC_TILE): New function.
400 * internal-fn.def (GOACC_DIM_POS): Comment may be overly conservative.
402 * omp-expand.c (struct oacc_collapse): Add tile and outer fields.
403 (expand_oacc_collapse_init): Add LOC paramter. Initialize tile
405 (expand_oacc_collapse_vars): Add INNER parm, adjust for tiling,
406 avoid DIV for outermost collapse var.
407 (expand_oacc_for): Insert tile element loop as needed. Adjust.
408 Remove out of date comments, fix whitespace.
409 * omp-general.c (omp_extract_for_data): Deal with tiling.
410 * omp-general.h (enum oacc_loop_flags): Add OLF_TILE flag,
411 adjust OLF_DIM_BASE value.
412 (struct omp_for_data): Add tiling field.
413 * omp-low.c (scan_sharing_clauses): Allow OMP_CLAUSE_TILE.
414 (lower_oacc_head_mark): Add OLF_TILE as appropriate. Ensure 2 levels
415 for auto loops. Remove default auto determining, moved to
416 oacc_loop_fixed_partitions.
417 * omp-offload.c (struct oacc_loop): Change 'ifns' to vector of call
418 stmts, add e_mask field.
419 (oacc_dim_call): New function, abstracted out from oacc_thread_numbers.
420 (oacc_thread_numbers): Use oacc_dim_call.
421 (oacc_xform_tile): New.
422 (new_oacc_loop_raw): Initialize e_mask, adjust for ifns vector.
423 (finish_oacc_loop): Adjust for ifns vector.
424 (oacc_loop_discover_walk): Append loop abstraction sites to list,
425 add case for GOACC_TILE fns.
426 (oacc_loop_xform_loop): Delete.
427 (oacc_loop_process): Iterate over call list directly, and add
428 handling for GOACC_TILE fns.
429 (oacc_loop_fixed_partitions): Determine default auto, deal with TILE,
431 (oacc_loop_auto_partitions): Add outer_assign parm. Assign all but
432 vector partitioning to outer loops. Assign 2 partitions to loops
433 when available. Add TILE handling.
434 (oacc_loop_partition): Adjust oacc_loop_auto_partitions call.
435 (execite_oacc_device_lower): Process GOACC_TILE fns, ignore unknown specs.
436 * tree-nested.c (convert_nonlocal_omp_clauses): Allow OMP_CLAUSE_TILE.
437 * tree.c (omp_clause_num_ops): Adjust TILE ops.
438 * tree.h (OMP_CLAUSE_TILE_ITERVAR, OMP_CLAUSE_TILE_COUNT): New.
440 2017-02-09 Gerald Pfeifer <gerald@pfeifer.com>
442 * configure.ac (ACX_BUGURL): Update.
443 * configure: Regenerate.
445 2017-02-09 Richard Biener <rguenther@suse.de>
447 PR tree-optimization/69823
448 * graphite-scop-detection.c (scop_detection::harmful_loop_in_region):
449 Properly enumerate all BBs in the region. Use auto_vec/auto_bitmap.
451 2017-02-09 Andrew Burgess <andrew.burgess@embecosm.com>
453 * config/arc/arc-c.def: Add __NPS400__ definition.
454 * config/arc/arc.h (CPP_SPEC): Don't define __NPS400__ here.
455 (TARGET_NPS400): Define.
457 2017-02-09 Andrew Burgess <andrew.burgess@embecosm.com>
459 * config/arc/arc-arch.h (arc_arch_t): Move unchanged to earlier in
461 (arc_cpu_t): Change base_architecture field, arch, to a arc_arc_t
463 (arc_cpu_types): Fill the arch_info field with a pointer into the
464 arc_arch_types table.
465 (arc_selected_cpu): Declare.
466 * config/arc/arc.c (arc_selected_cpu): Make global.
467 (arc_selected_arch): Delete.
468 (arc_base_cpu): Delete.
469 (arc_override_options): Remove references to deleted variables,
470 update access to arch information.
471 (ARC_OPT): Update access to arch information.
472 (ARC_OPTX): Likewise.
473 * config/arc/arc.h (arc_base_cpu): Remove declaration.
474 (TARGET_ARC600): Update access to arch information.
475 (TARGET_ARC601): Likewise.
476 (TARGET_ARC700): Likewise.
477 (TARGET_EM): Likewise.
478 (TARGET_HS): Likewise.
479 * config/arc/driver-arc.c (arc_cpu_to_as): Update access to arch
482 2017-02-08 Pat Haugen <pthaugen@us.ibm.com>
485 * config/rs6000/rs6000.c (rs6000_emit_vector_cond_expr): Invert
486 condition/operands for integer GE/LE/GEU/LEU operations.
488 2017-02-08 Segher Boessenkool <segher@kernel.crashing.org>
491 * config/rs6000/rs6000.opt (maltivec=le, maltivec=be): Fix spelling
494 2017-02-08 Martin Jambor <mjambor@suse.cz>
497 * ipa-prop.c (ipa_alloc_node_params): Make static, return bool
498 whether allocation happened.
499 (ipa_initialize_node_params): Do not call ipa_alloc_node_params if
500 nothing was allocated.
502 2017-02-08 Jakub Jelinek <jakub@redhat.com>
504 PR tree-optimization/79408
505 * tree-vrp.c (simplify_div_or_mod_using_ranges): If op1 is not
506 constant, but SSA_NAME with a known integer range, use the minimum
507 of that range instead of op1 to determine if modulo can be replaced
508 with its first operand.
510 2016-02-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
512 * config/riscv/riscv.c (riscv_build_integer_1): Avoid use of INT16_MAX.
514 2017-02-08 Richard Biener <rguenther@suse.de>
516 PR tree-optimization/71824
517 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
518 Check all loops contained in the merged region.
520 2017-02-07 Andrew Pinski <apinski@cavium.com>
522 * config/aarch64/aarch64.md (popcount<mode>2): New pattern.
524 2017-02-07 Andrew Pinski <apinski@cavium.com>
526 * config/aarch64/aarch64-cores.def (thunderx): Disable LSE.
527 (thunderxt88): Likewise.
528 (thunderxt81): Disable LSE and change v8.1 to v8.
529 (thunderxt83): Likewise.
531 2017-02-07 Jakub Jelinek <jakub@redhat.com>
532 Richard Biener <rguenther@suse.de>
535 * ira-int.h (struct target_ira_int): Change x_max_struct_costs_size
536 type from int to size_t.
537 * ira-costs.c (struct_costs_size): Change type from int to size_t.
539 2017-02-07 Jakub Jelinek <jakub@redhat.com>
541 PR rtl-optimization/79386
542 * cprop.c (bypass_conditional_jumps): Initialize
543 bypass_last_basic_block already before splitting bbs after
544 unconditional traps...
545 (bypass_conditional_jumps): ... rather than here.
548 * config/i386/sse.md (xtg_mode, gatherq_mode): New mode attrs.
549 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2,
550 *avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2): Use them,
551 fix -masm=intel patterns.
553 2017-02-07 Richard Biener <rguenther@suse.de>
555 PR tree-optimization/79256
557 * builtins.c (get_object_alignment_2): Use min_align_of_type
558 to extract alignment for MEM_REFs to honor BIGGEST_FIELD_ALIGNMENT
559 and ADJUST_FIELD_ALIGN.
561 * doc/tm.texi.in (ADJUST_FIELD_ALIGN): Adjust to take additional
563 * doc/tm.texi: Regenerate.
564 * stor-layout.c (layout_decl): Adjust.
565 (update_alignment_for_field): Likewise.
566 (place_field): Likewise.
567 (min_align_of_type): Likewise.
568 * config/arc/arc.h (ADJUST_FIELD_ALIGN): Adjust.
569 * config/epiphany/epiphany.h (ADJUST_FIELD_ALIGN): Likewise.
570 * config/epiphany/epiphany.c (epiphany_adjust_field_align): Likewise.
571 * config/frv/frv.h (ADJUST_FIELD_ALIGN): Likewise.
572 * config/frv/frv.c (frv_adjust_field_align): Likewise.
573 * config/i386/i386.h (ADJUST_FIELD_ALIGN): Likewise.
574 * config/i386/i386.c (x86_field_alignment): Likewise.
575 * config/rs6000/aix.h (ADJUST_FIELD_ALIGN): Likewise.
576 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Likewise.
577 * config/rs6000/freebsd64.h (ADJUST_FIELD_ALIGN): Likewise.
578 * config/rs6000/linux64.h (ADJUST_FIELD_ALIGN): Likewise.
579 * config/rs6000/sysv4.h (ADJUST_FIELD_ALIGN): Likewise.
580 * config/rs6000/rs6000.c (rs6000_special_adjust_field_align_p):
584 2017-01-30 Richard Biener <rguenther@suse.de>
586 PR tree-optimization/79256
587 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
588 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
591 2017-02-07 Toma Tabacu <toma.tabacu@imgtec.com>
593 * config/mips/mips.c (mips_expand_builtin_insn): Convert the QImode
594 argument of the pshufh, psllh, psllw, psrah, psraw, psrlh, psrlw
595 builtins to SImode and emit a zero-extend, if necessary.
597 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
599 * docs/invoke.texi (RISC-V Options): Alphabetize.
601 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
603 * doc/invoke.texi (RISC-V Options): Use two spaces to separate
606 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
608 * config/riscv/riscv.c: New file.
609 * gcc/common/config/riscv/riscv-common.c: Likewise.
610 * config.gcc: Likewise.
611 * config/riscv/constraints.md: Likewise.
612 * config/riscv/elf.h: Likewise.
613 * config/riscv/generic.md: Likewise.
614 * config/riscv/linux.h: Likewise.
615 * config/riscv/multilib-generator: Likewise.
616 * config/riscv/peephole.md: Likewise.
617 * config/riscv/pic.md: Likewise.
618 * config/riscv/predicates.md: Likewise.
619 * config/riscv/riscv-builtins.c: Likewise.
620 * config/riscv/riscv-c.c: Likewise.
621 * config/riscv/riscv-ftypes.def: Likewise.
622 * config/riscv/riscv-modes.def: Likewise.
623 * config/riscv/riscv-opts.h: Likewise.
624 * config/riscv/riscv-protos.h: Likewise.
625 * config/riscv/riscv.h: Likewise.
626 * config/riscv/riscv.md: Likewise.
627 * config/riscv/riscv.opt: Likewise.
628 * config/riscv/sync.md: Likewise.
629 * config/riscv/t-elf-multilib: Likewise.
630 * config/riscv/t-linux: Likewise.
631 * config/riscv/t-linux-multilib: Likewise.
632 * config/riscv/t-riscv: Likewise.
633 * configure.ac: Likewise.
634 * doc/contrib.texi: Add Kito Cheng, Palmer Dabbelt, and Andrew
635 Waterman as RISC-V maintainers.
636 * doc/install.texi: Add RISC-V entries.
637 * doc/invoke.texi: Add RISC-V options section.
638 * doc/md.texi: Add RISC-V constraints section.
639 * configure: Regenerated.
641 2017-02-06 Michael Meissner <meissner@linux.vnet.ibm.com>
644 * config/rs6000/vector.md (vcond<mode><mode>): Allow the true and
645 false values to be constant vectors with all 0 or all 1 bits set.
646 (vcondu<mode><mode>): Likewise.
647 * config/rs6000/predicates.md (vector_int_reg_or_same_bit): New
649 (fpmask_comparison_operator): Update comment.
650 (vecint_comparison_operator): New predicate.
651 * config/rs6000/rs6000.c (rs6000_emit_vector_cond_expr): Optimize
652 vector conditionals when the true and false values are constant
653 vectors with all 0 bits or all 1 bits set.
655 2017-02-06 Martin Sebor <msebor@redhat.com>
657 PR tree-optimization/79376
658 * gimple-fold.c (get_range_strlen): Set the minimum length to zero.
660 2017-02-06 Uros Bizjak <ubizjak@gmail.com>
662 * config/i386/sse.md (vector modes -> vec_extract* splitter): Use
663 explicit subreg RTX with operand 1. Use VECTOR_MODE_P predicate
664 to simplify split condition.
666 2017-02-06 Jakub Jelinek <jakub@redhat.com>
668 * omp-expand.c (oxpand_omp_atomic_fetch_op,
669 expand_omp_atomic_pipeline): Return false if can_atomic_load_p is
672 2017-02-06 Segher Boessenkool <segher@kernel.crashing.org>
674 PR rtl-optimization/68664
675 * target.def (can_speculate_insn): New hook.
676 * doc/tm.texi.in (TARGET_SCHED_CAN_SPECULATE_INSN): New hook.
677 * doc/tm.texi: Regenerate.
678 * sched-rgn.c (can_schedule_ready_p): Use the new hook.
679 * config/rs6000/rs6000.c (TARGET_SCHED_CAN_SPECULATE_INSN): New macro.
680 (rs6000_sched_can_speculate_insn): New function.
682 2017-02-06 Jakub Jelinek <jakub@redhat.com>
684 PR tree-optimization/79284
685 * tree-vectorizer.h (VECT_SCALAR_BOOLEAN_TYPE_P): Define.
686 * tree-vect-stmts.c (vect_get_vec_def_for_operand,
687 vectorizable_mask_load_store, vectorizable_operation,
688 vect_is_simple_cond, get_same_sized_vectype): Use it instead
689 of comparing TREE_CODE of a type against BOOLEAN_TYPE.
690 * tree-vect-patterns.c (check_bool_pattern, search_type_for_mask_1,
691 vect_recog_bool_pattern, vect_recog_mask_conversion_pattern): Likewise.
692 * tree-vect-slp.c (vect_get_constant_vectors): Likewise.
693 * tree-vect-loop.c (vect_determine_vectorization_factor): Likewise.
694 Remove redundant gimple_code (stmt) == GIMPLE_ASSIGN test after
695 is_gimple_assign (stmt). Replace another such test with
696 is_gimple_assign (stmt).
698 2017-02-06 Georg-Johann Lay <avr@gjlay.de>
701 * config/avr/avr.c (rtl-iter.h): Include it.
702 (TARGET_LEGITIMATE_COMBINED_INSN): New hook define...
703 (avr_legitimate_combined_insn): ...and implementation.
705 2017-02-06 Dominik Vogt <vogt@linux.vnet.ibm.com>
707 * config/s390/predicates.md ("larl_operand"): Use macros from hwint.h.
708 * config/s390/s390.c (s390_const_operand_ok)
709 (s390_canonicalize_comparison, s390_extract_part)
710 (s390_single_part, s390_contiguous_bitmask_nowrap_p)
711 (s390_contiguous_bitmask_p, s390_rtx_costs)
712 (legitimize_pic_address): Likewise.
713 * config/s390/s390.md ("clzdi2", "clztidi2"): Likewise.
714 * config/s390/vx-builtins.md ("vec_genbytemaskv16qi")
715 ("vec_permi<mode>", "vfae<mode>", "*vfaes<mode>", "vstrc<mode>")
716 ("*vstrcs<mode>"): Use UINTVAL() to set unsigned HOST_WIDE_INT.
717 * config/s390/vector.md ("vec_vfenes<mode>"): Likewise.
719 2017-02-06 Georg-Johann Lay <avr@gjlay.de>
721 * config/avr/avr.md (*addhi3_zero_extend): Add alternative where
722 REGNO($0) == REGNO($1).
724 2017-02-06 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
726 * config/s390/linux.h(SIZE_TYPE): Add comment.
728 2017-02-06 Julian Brown <julian@codesourcery.com>
729 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
730 Virendra Pathak <virendra.pathak@broadcom.com>
732 * config/aarch64/aarch64-cores.def: Change the scheduler
734 * config/aarch64/aarch64.md: Include thunderx2t99.md.
735 * config/aarch64/thunderx2t99.md: New file.
737 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
739 * doc/standards.texi (Go Language): Update link to language
742 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
744 * tree-eh.c (lower_resx): Sanitize profile.
745 (cleanup_empty_eh_move_lp): Likewise.
747 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
750 * cfgloopmanip.c (lv_adjust_loop_entry_edge, loop_version): Add
752 * cfgloopmanip.h (loop_version): Update prototype.
753 * modulo-sched.c (sms_schedule): Update call of loop_version.
754 * tree-if-conv.c(version_loop_for_if_conversion): Likewise.
755 * tree-parloops.c (gen_parallel_loop): Likewise.
756 * tree-ssa-loop-manip.c (tree_transform_and_unroll_loop): Likewise.
757 * tree-ssa-loop-split.c (split_loop): Likewise.
758 * tree-ssa-loop-unswitch.c (tree_unswitch_loop): Likewise.
759 * tree-vect-loop-manip.c (vect_loop_versioning): Likewise.
761 2017-02-05 Martin Liska <mliska@suse.cz>
764 * config/s390/s390.c (s390_gimplify_va_arg): Initialize local
766 (print_operand_address): Initialize a struct to zero.
768 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
770 * doc/contrib.texi (Contributors): Refer to Hans Boehm's
771 garbage collector only in textual form.
773 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
775 * doc/extend.texi (x86 specific memory model extensions for
776 transactional memory): Simplify a phrase.
778 2017-02-05 Eric Botcazou <ebotcazou@adacore.com>
781 * config/sparc/sync.md (atomic_loaddi_1): Replace 'U' constraint with
782 'r', 'm' constraint with 'T' and !TARGET_ARCH64 with TARGET_ARCH32.
783 (atomic_storedi_1): Likewise.
785 2017-02-04 Jakub Jelinek <jakub@redhat.com>
787 PR tree-optimization/79338
788 * tree-parloops.c (gather_scalar_reductions): Don't call
789 vect_analyze_loop_form for loop->inner before destroying loop's
792 2017-02-03 Martin Sebor <msebor@redhat.com>
794 PR tree-optimization/79327
795 * gimple-ssa-sprintf.c (tree_digits): Avoid adding the base prefix
796 when precision has resulted in leading zeros.
797 (format_integer): Adjust the likely counter to assume an unknown
798 argument that may be zero is non-zero.
800 2017-02-03 Jason Merrill <jason@redhat.com>
803 * tree-inline.c (copy_tree_body_r) [COND_EXPR]: Revert change to
804 avoid copying non-taken branch.
806 2017-02-03 Jakub Jelinek <jakub@redhat.com>
808 PR tree-optimization/79340
809 * tree-vect-loop.c (vectorizable_reduction): Release
810 vec_defs elements after safe_splicing them into other vectors.
813 PR tree-optimization/79327
814 * gimple-ssa-sprintf.c (adjust_range_for_overflow): If returning
815 true, always set *argmin and *argmax to TYPE_{MIN,MAX}_VALUE of
817 (format_integer): Use wide_int_to_tree instead of build_int_cst
818 + to_?hwi. If argmin is NULL, just set argmin and argmax to
819 TYPE_{MIN,MAX}_VALUE of argtype. Simplify and fix computation
820 of shortest and longest sequence.
822 2017-02-03 Uros Bizjak <ubizjak@gmail.com>
824 * config/i386/i386.c (dimode_scalar_chain::convert_reg):
825 Use pextrd for TARGET_SSE4_1 when creating scalar copy.
827 2017-02-03 Walter Lee <walt@tilera.com>
830 * config/tilegx/tilegx.md (tilegx_expand_prologue): Add blockage
831 after initial stackframe link reg save.
832 * config/tilepro/tilepro.md (tilepro_expand_prologue): Likewise.
834 2017-02-03 Jakub Jelinek <jakub@redhat.com>
837 * config/rs6000/rs6000.md (movsi_from_sf): Use wb constraint instead of
838 wu for stxssp alternative.
840 2017-02-03 Martin Sebor <msebor@redhat.com>
842 PR tree-optimization/79352
843 * gimple-fold.c (get_range_strlen): Add argument.
844 (get_range_strlen): Change return type to bool.
845 (get_maxval_strlen): Pass in a dummy argument.
846 * gimple-fold.h (get_range_strlen): Change return type to bool.
847 * gimple-ssa-sprintf.c (get_string_length): Set unlikely counter.
848 * tree.h (array_at_struct_end_p): Add argument.
849 * tree.c (array_at_struct_end_p): Handle it.
851 2017-02-03 Martin Liska <mliska@suse.cz>
854 * multiple_target.c (create_dispatcher_calls): Redirect edge
855 from a caller of a dispatcher.
856 (expand_target_clones): Make the clones local.
857 (ipa_target_clone): Do both target clones and resolvers.
858 (ipa_dispatcher_calls): Remove the pass.
859 (pass_dispatcher_calls::gate): Likewise.
860 (make_pass_dispatcher_calls): Likewise.
861 * passes.def (pass_target_clone): Put as very first IPA early
864 2017-02-03 Martin Liska <mliska@suse.cz>
866 * symtab.c (symtab_node::binds_to_current_def_p): Bail out
867 in case of a function with ifunc attribute.
869 2017-02-03 Martin Liska <mliska@suse.cz>
871 * cgraph.c (cgraph_node::dump): Dump function version info.
872 * symtab.c (symtab_node::dump_base): Add missing new line.
874 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
876 * tree-ssa-ifcombine.c (update_profile_after_ifcombine): New function.
877 (ifcombine_ifandif): Use it.
879 2017-02-03 Martin Liska <mliska@suse.cz>
881 * doc/invoke.texi: Document default value for
882 use-after-scope-direct-emission-threshold.
884 2017-02-03 Martin Liska <mliska@suse.cz>
886 PR tree-optimization/79339
887 * gimple-ssa-sprintf.c (format_floating_max): Call mpfr_clear.
888 (format_floating): Likewise.
890 2017-02-03 Martin Liska <mliska@suse.cz>
893 * ipa-prop.c (ipa_node_params_t::insert): Remove current
895 (ipa_node_params_t::remove): Likewise.
896 * ipa-prop.h (ipa_node_params::ipa_node_params): Make default
897 initialization from removed ipa_node_params_t::insert.
898 (ipa_node_params::~ipa_node_params): Move from removed
899 ipa_node_params_t::release.
900 * symbol-summary.h (symbol_summary::m_released): New member.
901 Do not release a summary twice. Do not allow to call finalizer
902 for types of a summary that live in GGC memory.
904 2017-02-02 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
906 * config/aarch64/aarch64.c (thunderx2t99_tunings): Enable AES and
909 2017-02-02 Martin Sebor <msebor@redhat.com>
912 * gimple-ssa-sprintf.c (get_string_length): Set lower bound to zero.
913 (format_string): Tighten up the range of output for non-constant
914 strings and correct the expected range for wide non-constant strings.
916 2017-02-02 Martin Sebor <msebor@redhat.com>
918 * doc/invoke.texi (-maccumulate-args): Fix bad grammar.
921 * doc/invoke.texi (-fdump-tree-): Remove pass-specific options from
923 (-fdump-tree-@var): Add to index and document how to come up
924 with pass-specific option and dump file names.
925 (-fdump-passes): Clarify where to look for output.
927 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
930 * tree-ssa-threadbackward.c (profitable_jump_thread_path): Dump
931 statistics of the analyzed path; allow threading for speed when
932 any of BBs along the path are optimized for speed.
934 2017-02-02 Eric Botcazou <ebotcazou@adacore.com>
937 * emit-rtl.c (init_emit): Add ??? comment for problematic alignment
938 settings of the virtual registers.
941 2016-08-23 Dominik Vogt <vogt@linux.vnet.ibm.com>
943 * explow.c (get_dynamic_stack_size): Take known alignment of stack
944 pointer + STACK_DYNAMIC_OFFSET into account when calculating the size
947 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
949 * config/s390/vx-builtins.md ("vec_ceil", "vec_floor")
950 ("vec_trunc", "vec_roundc", "vec_round"): Remove expanders.
952 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
954 * config/s390/s390.md: Add missing comments with the expanded
956 * config/s390/vector.md: Likewise.
957 * config/s390/vx-builtins.md: Likewise.
959 2017-02-02 Jakub Jelinek <jakub@redhat.com>
962 * config/rs6000/rs6000.md (*fixuns_trunc<mode>di2_fctiduz): Rename to ...
963 (fixuns_trunc<mode>di2): ... this, remove previous expander. Put all
964 conditions on a single line.
966 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
968 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
969 __S390_VX__ to __VX__.
971 2017-02-01 Andrew Pinski <apinski@cavium.com>
973 * tree-vect-loop.c (vect_compute_single_scalar_iteration_cost): Pass
974 stmt_info to record_stmt_cost.
975 (vect_get_known_peeling_cost): Pass stmt_info if known to
977 * config/aarch64/aarch64-protos.h (cpu_vector_cost): Split
978 cpu_vector_cost field into
979 scalar_int_stmt_cost and scalar_fp_stmt_cost. Split vec_stmt_cost
980 field into vec_int_stmt_cost and vec_fp_stmt_cost.
981 * config/aarch64/aarch64.c (generic_vector_cost): Update for the
982 splitting of scalar_stmt_cost and vec_stmt_cost.
983 (thunderx_vector_cost): Likewise.
984 (cortexa57_vector_cost): LIkewise.
985 (exynosm1_vector_cost): Likewise.
986 (xgene1_vector_cost): Likewise.
987 (thunderx2t99_vector_cost): Improve after the splitting of the two
989 (aarch64_builtin_vectorization_cost): Update for the splitting of
990 scalar_stmt_cost and vec_stmt_cost.
992 2017-02-01 Torvald Riegel <triegel@redhat.com>
993 Richard Henderson <rth@redhat.com>
995 * builtins.c (fold_builtin_atomic_always_lock_free): Make "lock-free"
996 conditional on existance of a fast atomic load.
997 * optabs-query.c (can_atomic_load_p): New function.
998 * optabs-query.h (can_atomic_load_p): Declare it.
999 * optabs.c (expand_atomic_exchange): Always delegate to libatomic if
1000 no fast atomic load is available for the particular size of access.
1001 (expand_atomic_compare_and_swap): Likewise.
1002 (expand_atomic_load): Likewise.
1003 (expand_atomic_store): Likewise.
1004 (expand_atomic_fetch_op): Likewise.
1005 * testsuite/lib/target-supports.exp
1006 (check_effective_target_sync_int_128): Remove x86 because it provides
1007 no fast atomic load.
1008 (check_effective_target_sync_int_128_runtime): Likewise.
1010 2017-02-01 Richard Biener <rguenther@suse.de>
1012 * graphite.c: Include tree-vectorizer.h for find_loop_location.
1013 (graphite_transform_loops): Provide opt-info for optimized nests.
1014 * tree-parloop.c (parallelize_loops): Provide opt-info for
1017 2017-02-01 Richard Biener <rguenther@suse.de>
1020 * tree-cfg.c (move_stmt_op): Never set TREE_BLOCK when it
1023 2017-02-01 Richard Biener <rguenther@suse.de>
1025 PR tree-optimization/71824
1026 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
1027 Verify the loops are valid in the merged SESE region.
1028 (scop_detection::can_represent_loop_1): Check analyzing the
1029 evolution of the number of iterations in the region succeeds.
1031 2017-01-31 Ian Lance Taylor <iant@golang.org>
1033 * config/i386/i386.c (ix86_expand_split_stack_prologue): Add
1034 REG_ARGS_SIZE note to 32-bit push insns and call insn.
1036 2017-01-31 David Malcolm <dmalcolm@redhat.com>
1038 PR preprocessor/79210
1039 * input.c (get_substring_ranges_for_loc): Replace line_width
1040 assertion with error-handling.
1042 2017-01-31 Richard Biener <rguenther@suse.de>
1044 PR tree-optimization/77318
1045 * graphite-sese-to-poly.c (extract_affine): Fix assert.
1046 (create_pw_aff_from_tree): Take loop parameter.
1047 (add_condition_to_pbb): Pass loop of the condition to
1048 create_pw_aff_from_tree.
1050 2017-01-31 Jakub Jelinek <jakub@redhat.com>
1052 * config/s390/s390.c (s390_asan_shadow_offset): New function.
1053 (TARGET_ASAN_SHADOW_OFFSET): Redefine.
1055 2017-01-31 Michael Meissner <meissner@linux.vnet.ibm.com>
1059 * config/rs6000/rs6000-protos.h (convert_float128_to_int): Delete,
1061 (convert_int_to_float128): Likewise.
1062 * config/rs6000/rs6000.c (convert_float128_to_int): Likewise.
1063 (convert_int_to_float128): Likewise.
1064 * config/rs6000/rs6000.md (UNSPEC_IEEE128_MOVE): Likewise.
1065 (UNSPEC_IEEE128_CONVERT): Likewise.
1066 (floatsi<mode>2, FLOAT128 iterator): Bypass calling
1067 rs6000_expand_float128_convert if we have IEEE 128-bit hardware.
1068 Use local variables for IBM extended format.
1069 (fix_trunc<mode>si2, FLOAT128 iterator): Likewise.
1070 (fix_trunc<mode>si2_fprs): Likewise.
1071 (fixuns_trunc<IEEE128:mode><SDI:mode>2): Likewise.
1072 (floatuns<IEEE128:mode>2, IEEE128 iterator): Likewise.
1073 (fix<uns>_<mode>si2_hw): Rework the IEEE 128-bt hardware support
1074 to know that we can now have integers of all sizes in vector
1076 (fix<uns>_<mode>di2_hw): Likewise.
1077 (float<uns>_<mode>si2_hw): Likewise.
1078 (fix_<mode>si2_hw): Likewise.
1079 (fixuns_<mode>si2_hw): Likewise.
1080 (float<uns>_<mode>di2_hw): Likewise.
1081 (float_<mode>di2_hw): Likewise.
1082 (float_<mode>si2_hw): Likewise.
1083 (floatuns_<mode>di2_hw): Likewise.
1084 (floatuns_<mode>si2_hw): Likewise.
1085 (xscvqp<su>wz_<mode>): Delete, no longer used.
1086 (xscvqp<su>dz_<mode>): Likewise.
1087 (xscv<su>dqp_<mode>): Likewise.
1088 (ieee128_mfvsrd_64bit): Likewise.
1089 (ieee128_mfvsrd_32bit): Likewise.
1090 (ieee128_mfvsrwz): Likewise.
1091 (ieee128_mtvsrw): Likewise.
1092 (ieee128_mtvsrd_64bit): Likewise.
1093 (ieee128_mtvsrd_32bit): Likewise.
1095 2017-01-31 Martin Liska <mliska@suse.cz>
1098 * ipa-prop.c (ipa_free_all_node_params): Call release method
1099 instead of ~sumbol_summary to not to trigger double times
1102 2017-01-31 Aldy Hernandez <aldyh@redhat.com>
1104 PR tree-optimization/71691
1105 * bitmap.h (class auto_bitmap): New.
1106 * tree-ssa-loop-unswitch.c (tree_may_unswitch_on): Call
1107 is_maybe_undefined instead of ssa_undefined_value_p.
1109 2017-01-31 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1111 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
1112 __S390_ARCH_LEVEL__ to __ARCH__.
1114 2017-01-31 Jakub Jelinek <jakub@redhat.com>
1116 PR tree-optimization/79267
1117 * value-prof.c (gimple_ic): Only drop lhs for noreturn calls
1118 if should_remove_lhs_p is true.
1120 2017-01-30 Alexandre Oliva <aoliva@redhat.com>
1123 * dwarf2out.c (clone_as_declaration): Drop DW_AT_alignment.
1124 (add_alignment_attribute): New.
1125 (base_type_die): Add alignment attribute.
1126 (subrange_type_die): Likewise.
1127 (modified_type_die): Likewise.
1128 (gen_array_type_die): Likewise.
1129 (gen_descr_array_type_die: Likewise.
1130 (gen_enumeration_type_die): Likewise.
1131 (gen_subprogram_die): Likewise.
1132 (gen_variable_die): Likewise.
1133 (gen_field_die): Likewise.
1134 (gen_ptr_to_mbr_type_die): Likewise.
1135 (gen_struct_or_union_type_die): Likewise.
1136 (gen_subroutine_type_die): Likewise.
1137 (gen_typedef_die): Likewise.
1138 (base_type_cmp): Compare alignment attribute.
1140 2017-01-30 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
1143 * config/rs6000/altivec.md (*setb_internal): Rename to setb_signed.
1144 (setb_unsigned) New pattern for setb with CCUNS.
1145 * config/rs6000/rs6000.c (expand_block_compare): Use a different
1146 subfc./subfe sequence to avoid overflow problems. Generate a
1147 shorter sequence with cmpld/setb for power9.
1148 * config/rs6000/rs6000.md (subf<mode>3_carry_dot2): Add a new pattern
1149 for generating subfc. instruction.
1150 (cmpstrsi): Add TARGET_POPCNTD predicate as the generate sequence
1151 now uses this instruction.
1153 2017-01-30 Ian Lance Taylor <iant@google.com>
1156 * dwarf2out.c (gen_type_die_with_usage): When picking a variant
1157 for FUNCTION_TYPE/METHOD_TYPE, use the first matching one.
1159 2017-01-30 Martin Sebor <msebor@redhat.com>
1161 * gimple-ssa-sprintf.c (fmtresult::adjust_for_width_or_precision):
1162 Move constant to the right of a relational operator.
1163 (get_mpfr_format_length, format_character, format_string): Ditto.
1164 (should_warn_p, maybe_warn): Same.
1166 * doc/invoke.texi (-Wformat-truncation=1): Fix typo.
1168 2017-01-30 Maxim Ostapenko <m.ostapenko@samsung.com>
1171 * asan.c (get_translation_unit_decl): Remove function.
1172 (asan_add_global): Force has_dynamic_init to zero in LTO mode.
1174 2017-01-30 Martin Liska <mliska@suse.cz>
1176 PR gcov-profile/79259
1177 * opts.c (common_handle_option): Enable flag_ipa_bit_cp w/
1180 2017-01-30 Martin Liska <mliska@suse.cz>
1183 * config/aarch64/cortex-a57-fma-steering.c (func_fma_steering::analyze):
1184 Initialize variables with NULL value.
1186 2017-01-30 Richard Earnshaw <rearnsh@arm.com>
1189 * config.gcc (arm*-*-*): Add arm/arm-flags.h and arm/arm-isa.h to
1191 * arm/arm-protos.h: Don't directly include arm-flags.h and arm-isa.h.
1193 2017-01-30 Richard Biener <rguenther@suse.de>
1195 PR tree-optimization/79276
1196 * tree-vrp.c (process_assert_insertions): Properly adjust common
1197 when removing a duplicate.
1199 * gcc.dg/torture/pr79276.c: New testcase.
1201 2017-01-30 Richard Biener <rguenther@suse.de>
1203 PR tree-optimization/79256
1204 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
1205 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
1207 * tree.c (build_aligned_type): Set TYPE_USER_ALIGN.
1209 2017-01-30 Dominik Vogt <vogt@linux.vnet.ibm.com>
1212 * config/s390/s390.md ("*r<noxa>sbg_<mode>_srl_bitmask")
1213 ("*r<noxa>sbg_<mode>_sll_bitmask")
1214 ("*extzv_<mode>_srl<clobbercc_or_nocc>")
1215 ("*extzv_<mode>_sll<clobbercc_or_nocc>"):
1216 Use contiguous_bitmask_nowrap_operand.
1218 2017-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1221 * config/rs6000/altivec.h (vec_xl): Revise #define.
1222 (vec_xst): Likewise.
1224 2017-01-27 Uros Bizjak <ubizjak@gmail.com>
1226 * config/i386/i386.c (print_reg): Use REGNO instead of true_regnum.
1228 2017-01-27 Bernd Schmidt <bschmidt@redhat.com>
1230 PR rtl-optimization/79194
1231 * cprop.c (one_cprop_pass): Move deletion of code after unconditional
1232 traps before call to bypass_conditional_jumps.
1234 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
1236 PR tree-optimization/71374
1237 * lra-constraints.c (check_conflict_input_operands): New.
1238 (match_reload): Use it.
1240 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
1243 * lra-assigns.c (find_hard_regno_for_1): Take endianess for into
1244 account to calculate conflict_set.
1246 2017-01-27 Bin Cheng <bin.cheng@arm.com>
1248 PR rtl-optimization/78559
1249 * combine.c (try_combine): Discard REG_EQUAL and REG_EQUIV for
1250 other_insn in combine.
1252 2017-01-27 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
1254 * builtin-types.def: Use unsigned_char_type_node for BT_UINT8. Use
1255 uint16_type_node for BT_UINT16.
1257 2017-01-27 David Malcolm <dmalcolm@redhat.com>
1259 * doc/sourcebuild.texi (Testsuites): Add "GIMPLE Tests" and
1260 "RTL Tests" to menu.
1261 (GIMPLE Tests): New node.
1262 (RTL Tests): New node.
1264 2017-01-27 Richard Biener <rguenther@suse.de>
1266 PR tree-optimization/79245
1267 * tree-loop-distribution.c (distribute_loop): Apply cost
1268 modeling also to detected patterns.
1270 2017-01-27 Richard Biener <rguenther@suse.de>
1272 PR tree-optimization/71433
1273 * tree-vrp.c (register_new_assert_for): Revert earlier changes.
1274 (compare_assert_loc): New function.
1275 (process_assert_insertions): Sort and optimize assert locations
1276 to remove duplicates and push down identical assertions on
1277 edges to their destination block.
1279 2017-01-27 Richard Biener <rguenther@suse.de>
1281 PR tree-optimization/79244
1282 * tree-vrp.c (remove_range_assertions): Forcefully propagate
1283 out SSA names even if abnormal.
1285 2017-01-27 Jakub Jelinek <jakub@redhat.com>
1287 * realmpfr.h: Poison MPFR_RND{N,Z,U,D}.
1288 * gimple-ssa-sprintf.c (format_floating_max): Use GMP_RNDN
1289 instead of MPFR_RNDN.
1291 2017-01-27 Richard Earnshaw <rearnsha@arm.com>
1294 * arm.c (arm_option_override): Don't call build_target_option_node
1295 until after doing all option overrides.
1296 (arm_valid_target_attribute_tree): Likewise.
1298 2017-01-27 Martin Liska <mliska@suse.cz>
1300 * doc/invoke.texi (-fprofile-arcs): Document profiling support
1301 for {cd}tors and C++ {cd}tors.
1303 2017-01-27 Dominik Vogt <vogt@linux.vnet.ibm.com>
1305 * config/s390/s390.md ("*setmem_long_and")
1306 ("*setmem_long_and_31z"): Use zero_extend instead of and.
1308 2017-01-26 Martin Sebor <msebor@redhat.com>
1310 * gimple-ssa-sprintf.c (format_floating): Simplify the computation
1313 2017-01-26 Martin Sebor <msebor@redhat.com>
1315 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
1316 HAVE_DFmode before using XFmode or DFmode.
1317 (parse_directive): Avoid using the z length modifier to avoid
1318 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
1321 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
1322 to accept adjustment as an array.
1323 (get_int_range): New function.
1324 (struct directive): Make width and prec arrays.
1325 (directive::set_width, directive::set_precision): Call get_int_range.
1326 (format_integer, format_floating): Handle width and precision ranges.
1327 (format_string, parse_directive): Same.
1329 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1332 * dwarf2out.c (generate_skeleton_bottom_up): For children with
1333 comdat_type_p set, just clone them, but keep the children in the
1337 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
1338 which have direct callers with -fvar-tracking-assignments enabled
1340 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
1341 inside of type units.
1343 2017-01-26 Martin Sebor <msebor@redhat.com>
1346 * gimple-ssa-sprintf.c (struct result_range): Add likely and
1348 (struct format_result): Replace number_chars, number_chars_min,
1349 and number_chars_max with a single member of struct result_range.
1351 (format_result::operator+=): Adjust.
1352 (struct fmtresult): Remove bounded. Handle likely and unlikely
1354 (fmtresult::adjust_for_width_or_precision): New function.
1355 (fmtresult:type_max_digits): New function.
1356 (bytes_remaining): Handle likely and unlikely counters.
1357 (min_bytes_remaining): Remove.
1358 (format_percent): Simplify.
1359 (format_integer, format_floating): Set likely and unlikely counters.
1360 (get_string_length, format_character, format_string): Same.
1361 (format_plain, should_warn_p): New function.
1362 (maybe_warn): Call should_warn_p. Update diagnostic messages
1363 and handle those for all directives, including plain strings.
1364 (format_directive): Handle likely and unlikely counters.
1365 Remove unnecessary quoting from diagnostics. Add an informational
1367 (add_bytes): Remove.
1368 (pass_sprintf_length::compute_format_length): Simplify.
1369 (try_substitute_return_value): Handle likely and unlikely counters.
1371 2017-01-26 Carl Love <cel@us.ibm.com>
1373 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
1374 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
1376 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
1379 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
1380 endianess for subregs into account.
1381 * lra-constraints.c (lra_constraints): Do risky transformations
1382 always on the first iteration.
1383 * lra-lives.c (check_pseudos_live_through_calls): Add arg
1384 last_call_used_reg_set.
1385 (process_bb_lives): Define and use last_call_used_reg_set.
1386 * lra.c (lra): Always continue after lra_constraints on the first
1389 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
1391 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
1393 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
1395 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1397 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
1398 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
1399 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
1400 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
1401 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
1402 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
1403 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
1404 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
1405 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
1407 2017-01-26 Marek Polacek <polacek@redhat.com>
1410 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
1411 for the third operand.
1413 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1416 * omp-low.c (struct omp_context): Add simt_stmt field.
1417 (scan_omp_for): Return omp_context *.
1418 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
1419 context to the _simt_ SIMD stmt.
1420 (lower_omp_for): For combined SIMD with sibling _simt_
1421 SIMD, make sure to use the same decls in _looptemp_
1422 clauses as in the sibling.
1424 2017-01-26 David Sherwood <david.sherwood@arm.com>
1427 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
1430 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1433 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
1434 emit fld b; fld a; if possible.
1436 * brig-builtins.def: Update copyright years.
1437 * config/arm/arm_acle_builtins.def: Update copyright years.
1439 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
1442 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
1443 constraint instead of o for the stxsd instruction.
1445 2017-01-25 Carl Love <cel@us.ibm.com>
1447 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
1448 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
1450 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
1452 * doc/invoke.texi (C++ Dialect Options): Fix typo.
1454 2017-01-25 Richard Biener <rguenther@suse.de>
1456 PR tree-optimization/69264
1457 * target.def (vector_alignment_reachable): Improve documentation.
1458 * doc/tm.texi: Regenerate.
1459 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
1461 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
1462 earlier changes with respect to TYPE_USER_ALIGN.
1463 (vector_alignment_reachable_p): Likewise. Improve dumping.
1465 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1468 * config/arm/arm.md (xordi3): Force constant operand into a register
1471 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1473 * doc/invoke.texi (-fstore-merging): Correct default optimization
1474 levels at which it is enabled.
1475 (-O): Move -fstore-merging from list to...
1478 2017-01-25 Richard Biener <rguenther@suse.de>
1481 * omp-expand.c: Include debug.h.
1482 (expand_omp_taskreg): Make sure to generate early debug before
1483 outlining anything from a function.
1484 (expand_omp_target): Likewise.
1485 (grid_expand_target_grid_body): Likewise.
1487 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
1490 * asan.c (get_translation_unit_decl): New function.
1491 (asan_add_global): Extract modules file name from globals
1492 TRANSLATION_UNIT_DECL name.
1494 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
1497 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
1498 for long calls with APCS frame and VFP.
1500 2017-01-24 David Malcolm <dmalcolm@redhat.com>
1502 * cfg.c (original_copy_tables_initialized_p): New function.
1503 * cfg.h (original_copy_tables_initialized_p): New decl.
1504 * cfgrtl.c (relink_block_chain): Guard the call to
1505 free_original_copy_tables with a call to
1506 original_copy_tables_initialized_p.
1507 * cgraph.h (symtab_node::native_rtl_p): New decl.
1508 * cgraphunit.c (symtab_node::native_rtl_p): New function.
1509 (symtab_node::needed_p): Don't assert for early assembly output
1510 for __RTL functions.
1511 (cgraph_node::finalize_function): Set "force_output" for __RTL
1513 (cgraph_node::analyze): Bail out early for __RTL functions.
1514 (analyze_functions): Update assertion to support __RTL functions.
1515 (cgraph_node::expand): Bail out early for __RTL functions.
1516 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
1518 * function.h (struct function): Update comment for field
1520 * gimple-expr.c: Include "tree-pass.h".
1521 (gimple_has_body_p): Return false for __RTL functions.
1522 * Makefile.in (OBJS): Add run-rtl-passes.o.
1523 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
1525 (gcc::pass_manager::get_clean_slate): New accessor.
1526 * passes.c: Include "insn-addr.h".
1527 (should_skip_pass_p): Add logging. Update logic for running
1528 "expand" to be compatible with both __GIMPLE and __RTL. Guard
1529 property-provider override so it is only done for gimple passes.
1531 (skip_pass): New function.
1532 (execute_one_pass): Call skip_pass when skipping passes.
1533 * read-md.c (md_reader::read_char): Support filtering
1534 the input to a subset of line numbers.
1535 (md_reader::md_reader): Initialize fields
1536 m_first_line and m_last_line.
1537 (md_reader::read_file_fragment): New function.
1538 * read-md.h (md_reader::read_file_fragment): New decl.
1539 (md_reader::m_first_line): New field.
1540 (md_reader::m_last_line): New field.
1541 * read-rtl-function.c (function_reader::create_function): Only
1542 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
1543 curr_properties. Set DECL_INITIAL to a dummy block.
1544 (read_rtl_function_body_from_file_range): New function.
1545 * read-rtl-function.h (read_rtl_function_body_from_file_range):
1547 * run-rtl-passes.c: New file.
1548 * run-rtl-passes.h: New file.
1550 2017-01-24 Jeff Law <law@redhat.com>
1552 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
1555 2017-01-24 Bin Cheng <bin.cheng@arm.com>
1557 PR tree-optimization/79159
1558 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
1559 (record_nonwrapping_iv): Improve boundary using above function if no
1560 value range information.
1562 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
1563 Martin Jambor <mjambor@suse.cz>
1565 * brig-builtins.def: New file.
1566 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
1567 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
1568 (DEF_HSAIL_SAT_BUILTIN): Likewise.
1569 (DEF_HSAIL_INTR_BUILTIN): Likewise.
1570 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
1571 * builtin-types.def (BT_INT8): New.
1572 (BT_INT16): Likewise.
1573 (BT_UINT8): Likewise.
1574 (BT_UINT16): Likewise.
1575 (BT_FN_ULONG): Likewise.
1576 (BT_FN_UINT_INT): Likewise.
1577 (BT_FN_UINT_ULONG): Likewise.
1578 (BT_FN_UINT_LONG): Likewise.
1579 (BT_FN_UINT_PTR): Likewise.
1580 (BT_FN_ULONG_PTR): Likewise.
1581 (BT_FN_INT8_FLOAT): Likewise.
1582 (BT_FN_INT16_FLOAT): Likewise.
1583 (BT_FN_UINT32_FLOAT): Likewise.
1584 (BT_FN_UINT16_FLOAT): Likewise.
1585 (BT_FN_UINT8_FLOAT): Likewise.
1586 (BT_FN_UINT64_FLOAT): Likewise.
1587 (BT_FN_UINT16_UINT32): Likewise.
1588 (BT_FN_UINT32_UINT16): Likewise.
1589 (BT_FN_UINT16_UINT16_UINT16): Likewise.
1590 (BT_FN_INT_PTR_INT): Likewise.
1591 (BT_FN_UINT_PTR_UINT): Likewise.
1592 (BT_FN_LONG_PTR_LONG): Likewise.
1593 (BT_FN_ULONG_PTR_ULONG): Likewise.
1594 (BT_FN_VOID_UINT64_UINT64): Likewise.
1595 (BT_FN_UINT8_UINT8_UINT8): Likewise.
1596 (BT_FN_INT8_INT8_INT8): Likewise.
1597 (BT_FN_INT16_INT16_INT16): Likewise.
1598 (BT_FN_INT_INT_INT): Likewise.
1599 (BT_FN_UINT_FLOAT_UINT): Likewise.
1600 (BT_FN_FLOAT_UINT_UINT): Likewise.
1601 (BT_FN_ULONG_UINT_UINT): Likewise.
1602 (BT_FN_ULONG_UINT_PTR): Likewise.
1603 (BT_FN_ULONG_ULONG_ULONG): Likewise.
1604 (BT_FN_UINT_UINT_UINT): Likewise.
1605 (BT_FN_VOID_UINT_PTR): Likewise.
1606 (BT_FN_UINT_UINT_PTR: Likewise.
1607 (BT_FN_UINT32_UINT64_PTR): Likewise.
1608 (BT_FN_INT_INT_UINT_UINT): Likewise.
1609 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
1610 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
1611 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
1612 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
1613 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
1614 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
1615 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
1616 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
1617 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
1618 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
1619 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
1620 * doc/frontends.texi: List BRIG FE.
1621 * doc/install.texi (Testing): Add BRIG tesring requirements.
1622 * doc/invoke.texi (Overall Options): Mention BRIG.
1623 * doc/standards.texi (Standards): Doucment BRIG HSA version.
1625 2017-01-24 Richard Biener <rguenther@suse.de>
1627 PR translation/79208
1628 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
1630 2017-01-24 Martin Jambor <mjambor@suse.cz>
1633 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
1634 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
1637 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
1640 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
1641 casts from signed to unsigned really don't have a range.
1643 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
1645 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
1646 GMP_RNDx for compatiblity.
1648 2017-01-24 Martin Liska <mliska@suse.cz>
1651 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
1652 that would prevent us to call alloca with -1 as argument.
1654 2017-01-24 Jakub Jelinek <jakub@redhat.com>
1656 * dwarf2out.c (output_compilation_unit_header, output_file_names):
1657 Avoid -Wformat-security warning.
1659 2017-01-23 Andrew Pinski <apinski@cavium.com>
1661 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
1664 2017-01-23 Martin Sebor <msebor@redhat.com>
1667 * gimple-ssa-sprintf.c (warn_level): New global.
1668 (format_integer): Use it here and throughout the rest of the file.
1669 Use the same switch to compute sign as base.
1670 (maybe_warn): New function.
1671 (format_directive): Factor out warnings into maybe_warn.
1672 Add debugging output. Use warn_level.
1673 (add_bytes): Use warn_level.
1674 (pass_sprintf_length::compute_format_length): Add debugging output.
1675 (try_substitute_return_value): Same.
1676 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
1679 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
1680 (struct fmtresult, format_integer, format_floating): Adjust.
1681 (fmtresult::fmtresult): Set max correctly in two argument ctor.
1682 (get_string_length, format_string,format_directive): Same.
1683 (pass_sprintf_length::compute_format_length): Same.
1684 (try_substitute_return_value): Simplify slightly.
1687 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
1688 (fmtresult::operator+=): Outlined.
1689 (struct fmtresult): Add ctors.
1690 (struct conversion_spec): Rename...
1691 (struct directive): ...to this. Add and remove data members.
1692 (directive::set_width, directive::set_precision): New functions.
1693 (format_percent): Use fmtresult ctor.
1694 (get_width_and_precision): Remove.
1695 (format_integer): Make naming changes. Avoid computing width and
1697 (format_floating): Same. Adjust indentation.
1698 (format_character, format_none): New functions.
1699 (format_string): Moved character handling to format_character.
1700 (format_directive): Remove arguments, change return type.
1701 (parse_directive): New function.
1702 (pass_sprintf_length::compute_format_length): Move directive
1703 parsing to parse_directive.
1705 2017-01-23 Jakub Jelinek <jakub@redhat.com>
1707 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
1708 (assign_assembler_name_if_needed): ... this.
1709 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
1710 (assign_assembler_name_if_needed): ... this.
1711 (free_lang_data_in_cgraph): Adjust callers.
1712 * cgraphunit.c (cgraph_node::analyze): Likewise.
1713 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
1716 2017-01-23 Richard Biener <rguenther@suse.de>
1718 PR tree-optimization/79088
1719 PR tree-optimization/79188
1720 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
1721 resetting loop bounds after last path deletion. Reset loop
1722 bounds of the target loop, make code match the comments.
1723 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
1724 Make sure loops need no fixups.
1726 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
1728 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
1729 exponent support with double type for first argument.
1730 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
1731 type returned by __builtin_vec_extract_sig,
1732 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
1733 functions from "vector int" to "vector unsigned int" or from
1734 "vector long long int" to "vector unsigned long long int".
1735 Changed type returned by __builtin_vec_extract_exp,
1736 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
1737 functions from "vector int" to "vector unsigned int" or from
1738 "vector long long int" to "vector unsigned long long int".
1739 Changed return type of __builtin_vec_test_data_class,
1740 __builtin_vec_test_data_class_sp, and
1741 __builtin_vec_test_data_class_dp from "vector int" to
1742 "vector bool int" or from "vector long long int" to "vector bool
1743 long long int" and changed second argument type from "unsigned
1744 int" to "int". Added new overloaded function forms "vector float
1745 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
1746 "vector float __builtin_vec_insert_exp_sp (vector float, vector
1747 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
1748 double, vector unsigned long long int)" and "vector double
1749 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
1750 long int)". Changed return type of
1751 __builtin_scalar_test_data_class and
1752 __builtin_scalar_test_data_class_sp and
1753 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
1754 int" and changed second argument from "unsigned int" to "int".
1755 Changed type returned by __builtin_scalar_test_neg,
1756 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
1757 from "int" to "bool int". Added new overloaded function form
1758 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
1759 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
1760 exponent double-precision with floating point first argument.
1761 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
1762 documentation of scalar_test_data_class, scalar_test_neg,
1763 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
1764 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
1765 vec_test_data_class built-in functions to reflect refinements in
1766 their type signatures.
1768 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
1770 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
1772 (aarch64_elf_asm_destructor): Likewise.
1774 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
1776 PR rtl-optimization/78634
1777 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
1778 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
1779 * ifcvt.c (noce_try_cmove): Add missing cost check.
1781 PR rtl-optimization/71724
1782 * combine.c (if_then_else_cond): Look for situations where it is
1783 beneficial to undo the work of one of the recursive calls.
1785 2017-01-23 Bin Cheng <bin.cheng@arm.com>
1787 PR tree-optimization/70754
1788 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
1789 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
1790 combined stmt before it if not NULL.
1791 (combine_chains): Process refs reversely and compute dominance point
1794 2017-01-23 Martin Liska <mliska@suse.cz>
1796 PR tree-optimization/79196
1797 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
1798 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
1800 (strlen_optimize_stmt): Call the renamed function.
1802 2017-01-23 Michael Matz <matz@suse.de>
1804 PR tree-optimization/78384
1805 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
1807 2017-01-23 Richard Biener <rguenther@suse.de>
1809 PR tree-optimization/79186
1810 * tree-vrp.c (register_new_assert_for): Make sure we've seen
1811 both incoming edges before moving an assert.
1813 2017-01-23 Martin Jambor <mjambor@suse.cz>
1815 * ipa-prop.c (load_from_param_1): Removed.
1816 (load_from_unmodified_param): Bits from load_from_param_1 put back
1818 (load_from_param): Removed.
1819 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
1820 with stmt. Reverted back to use of load_from_unmodified_param.
1822 2017-01-23 Martin Jambor <mjambor@suse.cz>
1825 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
1826 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
1827 field a pointer to garbage collected vector, mark lattices and
1828 ipcp_orig_node with GTY((skip)).
1829 (ipa_get_param_count): Adjust to descriptors being a pointer.
1830 (ipa_get_param): Likewise.
1831 (ipa_get_type): Likewise.
1832 (ipa_get_param_move_cost): Likewise.
1833 (ipa_set_param_used): Likewise.
1834 (ipa_get_controlled_uses): Likewise.
1835 (ipa_set_controlled_uses): Likewise.
1836 (ipa_is_param_used): Likewise.
1837 (ipa_node_params_t): Move into garbage collector. New methods insert
1839 (ipa_node_params_sum): Annotate wth GTY(()).
1840 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
1842 (ipa_load_from_parm_agg): Adjust declaration.
1843 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
1844 * ipa-profile.c (ipa_profile): Likewise.
1845 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
1846 (ipa_populate_param_decls): Make descriptors parameter garbage
1848 (ipa_dump_param): Adjust to descriptors being a pointer.
1849 (ipa_alloc_node_params): Likewise.
1850 (ipa_initialize_node_params): Likewise.
1851 (load_from_param_1): Make descriptors parameter garbage collected.
1852 (load_from_unmodified_param): Likewise.
1853 (load_from_param): Likewise.
1854 (ipa_load_from_parm_agg): Likewise.
1855 (ipa_node_params::~ipa_node_params): Removed.
1856 (ipa_free_all_node_params): Remove call to delete operator.
1857 (ipa_node_params_t::insert): New.
1858 (ipa_node_params_t::remove): Likewise.
1859 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
1860 copy known_csts and known_contexts vectors.
1861 (ipa_read_node_info): Adjust to descriptors being a pointer.
1862 (ipcp_modif_dom_walker): Make m_descriptors field garbage
1864 (ipcp_transform_function): Make descriptors variable garbage
1867 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
1869 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
1870 * config/i386/avx512dqintrin.h: Ditto.
1871 * config/i386/avx512fintrin.h: Ditto.
1872 * gcc/config/i386/i386.c: Handle new builtins.
1873 * config/i386/i386-builtin.def: Add new builtins.
1874 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
1875 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
1877 2017-01-23 Jakub Jelinek <jakub@redhat.com>
1878 Martin Liska <mliska@suse.cz>
1880 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
1881 * asan.c (asan_expand_poison_ifn): Support stores and use
1882 appropriate ASAN report function.
1883 * internal-fn.c (expand_ASAN_POISON_USE): New function.
1884 * internal-fn.def (ASAN_POISON_USE): Declare.
1885 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
1886 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
1887 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
1888 ASAN_POISON calls w/o LHS.
1889 * tree-ssa.c (execute_update_addresses_taken): Create clobber
1890 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
1891 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
1892 * gimplify.c (asan_poison_variables): Add attribute
1893 use_after_scope_memory to variables that really needs to live
1895 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
1896 having the attribute.
1898 2017-01-23 Martin Liska <mliska@suse.cz>
1900 * asan.c (create_asan_shadow_var): New function.
1901 (asan_expand_poison_ifn): Likewise.
1902 * asan.h (asan_expand_poison_ifn): New declaration.
1903 * internal-fn.c (expand_ASAN_POISON): Likewise.
1904 * internal-fn.def (ASAN_POISON): New builtin.
1905 * sanopt.c (pass_sanopt::execute): Expand
1906 asan_expand_poison_ifn.
1907 * tree-inline.c (copy_decl_for_dup_finish): Make function
1909 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
1910 * tree-ssa.c (is_asan_mark_p): New function.
1911 (execute_update_addresses_taken): Rewrite local variables
1912 (identified just by use-after-scope as addressable) into SSA.
1914 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
1916 * doc/install.texi (Specific): opensource.apple.com uses https
1917 now. Remove trailing slash.
1919 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
1921 * README.Portability: Remove note on an Irix compatibility issue.
1923 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
1925 * gcov.c (INCLUDE_ALGORITHM): Define.
1926 (INCLUDE_VECTOR): Define.
1927 No longer include <vector> and <algorithm> directly.
1929 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
1931 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
1933 * doc/invoke.texi (Code Gen Options): Ditto.
1935 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
1938 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
1940 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
1942 rtl-optimization/79125
1943 * cprop.c (local_cprop_pass): Handle cases where we make an
1946 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
1950 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
1951 read from, for big endian.
1953 2017-01-20 Jiong Wang <jiong.wang@arm.com>
1955 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
1956 register pauth builtins for LP64 only.
1958 2017-01-20 Marek Polacek <polacek@redhat.com>
1961 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
1964 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1966 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
1968 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
1969 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
1970 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
1972 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1975 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
1976 in big-endian mode when they are not a single duplicated value.
1978 2017-01-20 Richard Biener <rguenther@suse.de>
1980 * BASE-VER: Bump to 7.0.1.
1982 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1984 * omp-low.c (omplow_simd_context): New struct. Use it...
1985 (lower_rec_simd_input_clauses): ...here and...
1986 (lower_rec_input_clauses): ...here to hold common data. Adjust all
1987 references to idx, lane, max_vf, is_simt.
1989 2017-01-20 Graham Markall <graham.markall@embecosm.com>
1991 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
1994 2017-01-20 Martin Jambor <mjambor@suse.cz>
1996 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
1997 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
1999 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
2000 (GTFILES): Rename hsa.c to hsa-common.c.
2001 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
2002 * hsa-dump.c: Likewise.
2003 * hsa-gen.c: Likewise.
2004 * hsa-regalloc.c: Likewise.
2005 * ipa-hsa.c: Likewise.
2006 * omp-expand.c: Likewise.
2007 * omp-low.c: Likewise.
2008 * toplev.c: Likewise.
2010 2017-01-20 Marek Polacek <polacek@redhat.com>
2013 * doc/invoke.texi: Document -Wduplicated-branches.
2014 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
2015 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
2016 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
2017 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
2018 return 0 only when not OEP_LEXICOGRAPHIC.
2019 (fold_build_cleanup_point_expr): Use the expression
2020 location when building CLEANUP_POINT_EXPR.
2021 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
2022 * tree.c (add_expr): Handle error_mark_node.
2024 2017-01-20 Martin Liska <mliska@suse.cz>
2027 * tree-profile.c (init_ic_make_global_vars): Do not call
2029 (gimple_init_gcov_profiler): Likewise.
2031 2017-01-20 Martin Liska <mliska@suse.cz>
2034 * cgraph.h (maybe_create_reference): Remove argument and
2036 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
2038 * ipa-cp.c (create_specialized_node): Likewise.
2039 * symtab.c (symtab_node::maybe_create_reference): Handle
2040 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
2042 2017-01-20 Martin Liska <mliska@suse.cz>
2044 * read-rtl-function.c (function_reader::create_function): Use
2045 build_decl instread of build_decl_stat.
2047 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
2049 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
2050 * config/i386/avx512dqintrin.h: Ditto.
2051 * config/i386/avx512fintrin.h: Ditto.
2052 * config/i386/i386-builtin-types.def: Add new types.
2053 * gcc/config/i386/i386.c: Handle new types.
2054 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
2055 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
2056 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
2057 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
2058 (__builtin_ia32_kshiftridi): New.
2059 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
2061 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
2065 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
2066 define to rs6000_init_stack_protect_guard.
2067 (rs6000_init_stack_protect_guard): New function.
2069 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
2070 Yunqiang Su <yunqiang.su@imgtec.com>
2072 * config.gcc (supported_defaults): Add madd4.
2073 (with_madd4): Add validation.
2074 (all_defaults): Add madd4.
2075 * config/mips/mips.opt (mmadd4): New option.
2076 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
2078 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
2080 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
2081 (ISA_HAS_FUSED_MADD4): Likewise.
2082 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
2083 * gcc/doc/install.texi (--with-madd4): Document the new option.
2085 2017-01-19 Jiong Wang <jiong.wang@arm.com>
2087 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
2088 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
2089 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
2090 (aarch64_init_pauth_hint_builtins): New.
2091 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
2092 (aarch64_expand_builtin): Expand new builtins.
2094 2017-01-19 Jiong Wang <jiong.wang@arm.com>
2096 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
2097 * combine-stack-adj.c (no_unhandled_cfa): Handle
2098 REG_CFA_TOGGLE_RA_MANGLE.
2099 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
2100 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
2101 info for return address signing.
2102 (aarch64_expand_epilogue): Likewise.
2104 2017-01-19 Jiong Wang <jiong.wang@arm.com>
2106 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
2107 * config/aarch64/aarch64-protos.h
2108 (aarch64_return_address_signing_enabled): New declaration.
2109 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
2111 (aarch64_expand_prologue): Sign return address before it's pushed onto
2113 (aarch64_expand_epilogue): Authenticate return address fetched from
2115 (aarch64_override_options): Sanity check for ILP32 and ISA level.
2116 (aarch64_attributes): New function attributes for "sign-return-address".
2117 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
2118 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
2119 ("*do_return"): Generate combined instructions according to key index.
2120 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
2121 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
2123 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
2124 * config/aarch64/aarch64.opt (msign-return-address=): New.
2125 * doc/extend.texi (AArch64 Function Attributes): Documents
2126 "sign-return-address=".
2127 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
2129 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
2131 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
2132 overall option summary.
2134 2017-01-19 Jiong Wang <jiong.wang@arm.com>
2136 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
2137 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
2138 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
2139 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
2141 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
2143 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
2144 -mpower9-minmax by default for -mcpu=power9.
2145 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
2146 128-bit floating point.
2148 2017-01-20 Alan Modra <amodra@gmail.com>
2150 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
2151 optimizing for size.
2153 2017-01-20 Alan Modra <amodra@gmail.com>
2156 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
2157 for strcmp and strncmp from corresponding builtin decl.
2159 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
2161 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
2162 instead of i386/rtems-64.h.
2163 * config/i386/rtems-64.h: Remove.
2165 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
2169 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
2171 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
2173 2017-01-19 Tamar Christina <tamar.christina@arm.com>
2175 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
2176 Change int to HOST_WIDE_INT.
2177 * config/aarch64/aarch64-protos.h
2178 (aarch64_simd_gen_const_vector_dup): Likewise.
2179 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
2181 2017-01-19 David Malcolm <dmalcolm@redhat.com>
2183 * langhooks-def.h (lhd_type_for_size): New decl.
2184 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
2185 * langhooks.c (lhd_type_for_size): New function, taken from
2188 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
2190 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
2191 define_bypass for CR latency.
2192 (power9-cracked-alu): Update bypass latency and remove power9-branch.
2193 (power9-alu2): Add define_bypass for CR latency.
2195 (power9-mul): Update insn latency.
2196 (power9-mul-compare): Update insn latency, bypass latency and remove
2199 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2201 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
2203 * config/aarch64/aarch64.md
2204 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
2205 aarch64_nopcrelative_literal_loads.
2206 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
2208 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
2210 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
2212 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
2214 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
2217 * config.gcc (supported_defaults): Add lxc1-sxc1.
2218 (with_lxc1_sxc1): Add validation.
2219 (all_defaults): Add lxc1-sxc1.
2220 * config/mips/mips.opt (mlxc1-sxc1): New option.
2221 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
2223 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
2224 __mips_no_lxc1_sxc1.
2225 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
2226 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
2227 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
2229 2017-01-19 Richard Biener <rguenther@suse.de>
2231 PR tree-optimization/72488
2232 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
2233 sure to restore SSA info.
2234 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
2236 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
2238 PR rtl-optimization/79121
2239 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
2240 of the inner type when shifting an extended value.
2242 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
2245 * symtab.c (symtab_node::equal_address_to): Fix comparing of
2246 interposable aliases.
2248 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
2251 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
2252 Use the evmergelohi instruction.
2253 (mov_si<mode>_e500_subreg4_2_le): Likewise.
2254 (mov_sitf_e500_subreg8_2_be): Likewise.
2255 (mov_sitf_e500_subreg12_2_le): Likewise.
2256 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
2257 (mov_si<mode>_e500_subreg4_2_be): Likewise.
2258 (mov_sitf_e500_subreg8_2_le): Likewise.
2259 (mov_sitf_e500_subreg12_2_be): Likewise.
2261 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2263 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
2264 attribute from vecsimple to vecperm.
2265 (altivec_vbpermq2): Likewise.
2267 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2270 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
2272 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
2273 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
2274 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
2275 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
2276 case where N arg is SIZE_MAX.
2277 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
2278 (cmpstrsi): Add pattern.
2280 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
2282 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2283 __builtin_vec_revb builtins.
2284 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
2285 built-in functions to support generation of the ISA 3.0 XXBR<x>
2286 vector byte reverse instructions.
2287 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
2288 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
2289 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
2290 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
2291 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
2292 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
2293 (P9V_BUILTIN_VEC_REVB): Likewise.
2294 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
2295 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
2296 (p9_xxbrq_v16qi): Likewise.
2297 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
2298 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
2299 (p9_xxbrh_v8hi): Likewise.
2300 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
2301 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
2302 vec_revb built-in functions.
2304 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
2306 PR rtl-optimization/78952
2307 * config/i386/i386.md (any_extract): New code iterator.
2308 (*insvqi_2): Use any_extract for source operand.
2309 (*insvqi_3): Use any_shiftrt for source operand.
2311 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
2313 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
2315 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
2317 2017-01-18 Matthias Klose <doko@ubuntu.com>
2319 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
2321 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2323 * config/rs6000/altivec.h (vec_bperm): Change #define.
2324 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
2325 (altivec_vbpermq2): New define_insn.
2326 (altivec_vbpermd): Likewise.
2327 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
2329 (VBPERMD): Likewise.
2330 (VBPERM): New polymorphic function interface.
2331 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
2332 Add entries for P9V_BUILTIN_VEC_VBPERM.
2333 * doc/extend.texi: Add interfaces for vec_bperm.
2335 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2337 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
2338 first letter of error messages.
2339 (s390_resolve_overloaded_builtin): Likewise.
2340 * config/s390/s390.c (s390_expand_builtin): Likewise.
2341 (s390_invalid_arg_for_unprototyped_fn): Likewise.
2342 (s390_valid_target_attribute_inner_p): Likewise.
2343 * config/s390/s390.md ("tabort"): Likewise.
2345 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
2347 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
2348 (ISA_AVOID_DIV_HILO): New macro.
2349 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
2350 (ISA_HAS_DDIV): Likewise.
2352 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2354 * doc/invoke.texi (fabi-version): Correct number of occurrences.
2356 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2358 * doc/invoke.texi (fabi-version): Spelling fix.
2360 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2363 * doc/invoke.texi (fabi-version): Mention mangling fix for
2366 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2369 * doc/invoke.texi (fabi-version): Document discriminator mangling.
2371 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
2374 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
2375 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
2377 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
2379 (stack_protect_test): Ditto.
2380 * config/rs6000/rs6000.opt (mstack-protector-guard=,
2381 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
2383 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
2384 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
2385 -mstack-protector-guard-offset=.
2386 (RS/6000 and PowerPC Options): Ditto.
2388 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
2390 * config/i386/i386.h (MASK_CLASS_P): New define.
2391 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
2392 there are no registers from different register sets also when
2393 mask registers are used. Update function comment.
2394 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
2395 to (*k/*r) and (*k/*km) alternatives.
2397 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
2399 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
2400 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
2401 (EH_RETURN_HANDLER_RTX): New define.
2402 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
2403 Force frame pointer in EH return functions.
2404 (aarch64_expand_epilogue): Add barrier for eh_return.
2405 (aarch64_final_eh_return_addr): Remove.
2406 (aarch64_eh_return_handler_rtx): New function.
2407 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
2409 (aarch64_eh_return_handler_rtx): New prototype.
2411 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2413 * config/rs6000/altivec.h (vec_rlmi): New #define.
2414 (vec_vrlnm): Likewise.
2415 (vec_rlnm): Likewise.
2416 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
2417 (UNSPEC_VRLNM): Likewise.
2418 (VIlong): New mode iterator.
2419 (altivec_vrl<VI_char>mi): New define_insn.
2420 (altivec_vrl<VI_char>nm): Likewise.
2421 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
2424 (RLNM): New polymorphic function entry.
2425 (VRLWMI): New monomorphic function entry.
2427 (RLMI): New polymorphic function entry.
2428 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
2429 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
2430 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
2433 2017-01-17 Jakub Jelinek <jakub@redhat.com>
2436 * dwarf2out.c (field_byte_offset): Restore the
2437 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
2438 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
2439 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
2442 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
2445 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
2447 2017-01-17 Jakub Jelinek <jakub@redhat.com>
2450 * dwarf2out.c (add_data_member_location_attribute): For constant
2451 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
2452 instead of DW_AT_data_member_location, DW_AT_bit_offset and
2453 DW_AT_byte_size attributes.
2455 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
2457 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
2458 after forcing to constant memory when the code model is medium.
2460 2017-01-17 Julia Koval <julia.koval@intel.com>
2463 * config/i386/avx512fintrin.h
2464 (_mm512_i32gather_ps): Change __addr type to void const*.
2465 (_mm512_mask_i32gather_ps): Ditto.
2466 (_mm512_i32gather_pd): Ditto.
2467 (_mm512_mask_i32gather_pd): Ditto.
2468 (_mm512_i64gather_ps): Ditto.
2469 (_mm512_mask_i64gather_ps): Ditto.
2470 (_mm512_i64gather_pd): Ditto.
2471 (_mm512_mask_i64gather_pd): Ditto.
2472 (_mm512_i32gather_epi32): Ditto.
2473 (_mm512_mask_i32gather_epi32): Ditto.
2474 (_mm512_i32gather_epi64): Ditto.
2475 (_mm512_mask_i32gather_epi64): Ditto.
2476 (_mm512_i64gather_epi32): Ditto.
2477 (_mm512_mask_i64gather_epi32): Ditto.
2478 (_mm512_i64gather_epi64): Ditto.
2479 (_mm512_mask_i64gather_epi64): Ditto.
2480 (_mm512_i32scatter_ps): Change __addr type to void*.
2481 (_mm512_mask_i32scatter_ps): Ditto.
2482 (_mm512_i32scatter_pd): Ditto.
2483 (_mm512_mask_i32scatter_pd): Ditto.
2484 (_mm512_i64scatter_ps): Ditto.
2485 (_mm512_mask_i64scatter_ps): Ditto.
2486 (_mm512_i64scatter_pd): Ditto.
2487 (_mm512_mask_i64scatter_pd): Ditto.
2488 (_mm512_i32scatter_epi32): Ditto.
2489 (_mm512_mask_i32scatter_epi32): Ditto.
2490 (_mm512_i32scatter_epi64): Ditto.
2491 (_mm512_mask_i32scatter_epi64): Ditto.
2492 (_mm512_i64scatter_epi32): Ditto.
2493 (_mm512_mask_i64scatter_epi32): Ditto.
2494 (_mm512_i64scatter_epi64): Ditto.
2495 (_mm512_mask_i64scatter_epi64): Ditto.
2496 * config/i386/avx512pfintrin.h
2497 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
2498 (_mm512_mask_prefetch_i32gather_ps): Ditto.
2499 (_mm512_mask_prefetch_i64gather_pd): Ditto.
2500 (_mm512_mask_prefetch_i64gather_ps): Ditto.
2501 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
2502 (_mm512_prefetch_i32scatter_ps): Ditto.
2503 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
2504 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
2505 (_mm512_prefetch_i64scatter_pd): Ditto.
2506 (_mm512_prefetch_i64scatter_ps): Ditto.
2507 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
2508 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
2509 * config/i386/avx512vlintrin.h
2510 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
2511 (_mm_mmask_i32gather_ps): Ditto.
2512 (_mm256_mmask_i32gather_pd): Ditto.
2513 (_mm_mmask_i32gather_pd): Ditto.
2514 (_mm256_mmask_i64gather_ps): Ditto.
2515 (_mm_mmask_i64gather_ps): Ditto.
2516 (_mm256_mmask_i64gather_pd): Ditto.
2517 (_mm_mmask_i64gather_pd): Ditto.
2518 (_mm256_mmask_i32gather_epi32): Ditto.
2519 (_mm_mmask_i32gather_epi32): Ditto.
2520 (_mm256_mmask_i32gather_epi64): Ditto.
2521 (_mm_mmask_i32gather_epi64): Ditto.
2522 (_mm256_mmask_i64gather_epi32): Ditto.
2523 (_mm_mmask_i64gather_epi32): Ditto.
2524 (_mm256_mmask_i64gather_epi64): Ditto.
2525 (_mm_mmask_i64gather_epi64): Ditto.
2526 (_mm256_i32scatter_ps): Change __addr type to void*.
2527 (_mm256_mask_i32scatter_ps): Ditto.
2528 (_mm_i32scatter_ps): Ditto.
2529 (_mm_mask_i32scatter_ps): Ditto.
2530 (_mm256_i32scatter_pd): Ditto.
2531 (_mm256_mask_i32scatter_pd): Ditto.
2532 (_mm_i32scatter_pd): Ditto.
2533 (_mm_mask_i32scatter_pd): Ditto.
2534 (_mm256_i64scatter_ps): Ditto.
2535 (_mm256_mask_i64scatter_ps): Ditto.
2536 (_mm_i64scatter_ps): Ditto.
2537 (_mm_mask_i64scatter_ps): Ditto.
2538 (_mm256_i64scatter_pd): Ditto.
2539 (_mm256_mask_i64scatter_pd): Ditto.
2540 (_mm_i64scatter_pd): Ditto.
2541 (_mm_mask_i64scatter_pd): Ditto.
2542 (_mm256_i32scatter_epi32): Ditto.
2543 (_mm256_mask_i32scatter_epi32): Ditto.
2544 (_mm_i32scatter_epi32): Ditto.
2545 (_mm_mask_i32scatter_epi32): Ditto.
2546 (_mm256_i32scatter_epi64): Ditto.
2547 (_mm256_mask_i32scatter_epi64): Ditto.
2548 (_mm_i32scatter_epi64): Ditto.
2549 (_mm_mask_i32scatter_epi64): Ditto.
2550 (_mm256_i64scatter_epi32): Ditto.
2551 (_mm256_mask_i64scatter_epi32): Ditto.
2552 (_mm_i64scatter_epi32): Ditto.
2553 (_mm_mask_i64scatter_epi32): Ditto.
2554 (_mm256_i64scatter_epi64): Ditto.
2555 (_mm256_mask_i64scatter_epi64): Ditto.
2556 (_mm_i64scatter_epi64): Ditto.
2557 (_mm_mask_i64scatter_epi64): Ditto.
2558 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
2559 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
2560 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
2561 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
2562 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
2563 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
2564 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
2565 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
2566 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
2567 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
2568 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
2569 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
2570 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
2571 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
2572 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
2573 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
2574 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
2575 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
2576 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
2577 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
2578 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
2579 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
2580 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
2581 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
2582 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
2583 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
2584 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
2585 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
2586 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
2587 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
2588 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
2589 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
2590 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
2591 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
2592 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
2593 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
2594 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
2595 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
2596 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
2597 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
2598 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
2599 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
2600 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
2601 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
2602 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
2603 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
2604 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
2605 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
2606 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
2607 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
2608 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
2609 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
2610 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
2611 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
2612 definitions accordingly.
2614 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
2615 Kuan-Lin Chen <kuanlinchentw@gmail.com>
2618 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
2621 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
2624 * ira-conflicts.c (ira_build_conflicts): Update total conflict
2625 hard regs for inner regno.
2627 2017-01-17 Martin Liska <mliska@suse.cz>
2630 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
2631 assumption and add comment.
2633 2017-01-17 Nathan Sidwell <nathan@acm.org>
2635 * ipa-visibility.c (localize_node): New function, broken out of ...
2636 (function_and_variable_visibility): ... here. Call it.
2638 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
2641 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
2642 correctly set frequency of oudgoing edge.
2643 (duplicate_thread_path): Fix profile updating.
2645 2017-01-17 Jakub Jelinek <jakub@redhat.com>
2648 * configure.ac: Add GCC_BASE_VER.
2649 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
2650 version from BASE-VER file.
2651 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
2652 (gcc.o): Depend on $(BASEVER).
2653 * common.opt (dumpfullversion): New option.
2654 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
2655 * doc/invoke.texi: Document -dumpfullversion.
2656 * doc/install.texi: Document --with-gcc-major-version-only.
2657 * configure: Regenerated.
2659 2017-01-17 Richard Biener <rguenther@suse.de>
2661 PR tree-optimization/71433
2662 * tree-vrp.c (register_new_assert_for): Merge same asserts
2663 on all incoming edges.
2664 (process_assert_insertions_for): Handle insertions at the
2667 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
2669 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
2670 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
2672 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
2675 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
2678 2017-01-17 Alan Modra <amodra@gmail.com>
2681 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
2682 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
2683 symbolic stack limit when pic.
2685 2017-01-16 Martin Sebor <msebor@redhat.com>
2687 PR tree-optimization/78608
2688 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
2690 2017-01-16 Jeff Law <law@redhat.com>
2693 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
2694 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
2695 for several include directories that may be relative to sysroot.
2696 * config/i386/x-mingw32 (gplus_includedir): Define.
2697 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
2698 (native_system_includedir): Likewise.
2699 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
2700 override if TARGET_SYSTEM_ROOT is defined.
2701 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
2703 PR tree-optimization/79090
2704 PR tree-optimization/33562
2705 PR tree-optimization/61912
2706 PR tree-optimization/77485
2707 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
2708 and computed trims into the dump file.
2710 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
2712 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
2714 2017-01-16 Jakub Jelinek <jakub@redhat.com>
2717 * gimplify.c (gimplify_init_constructor): If want_value and
2718 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
2722 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
2723 sequence. Formatting fixes.
2724 (doloop_optimize): Formatting fixes.
2727 * gcc.c (debug_level_greater_than_spec_func): New function.
2728 (static_spec_functions): Add debug-level-gt spec function.
2729 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
2731 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
2732 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
2733 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
2734 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
2735 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
2736 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
2738 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
2740 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
2741 QImode fixups to general and mask registers only.
2743 2017-01-16 Carl Love <cel@us.ibm.com>
2745 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
2746 for built-in functions
2747 vector signed char vec_nabs (vector signed char)
2748 vector signed short vec_nabs (vector signed short)
2749 vector signed int vec_nabs (vector signed int)
2750 vector signed long long vec_nabs (vector signed long long)
2751 vector float vec_nabs (vector float)
2752 vector double vec_nabs (vector double)
2753 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
2755 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
2756 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
2757 * doc/extend.texi: Update the documentation file for the new built-in
2760 2017-01-16 Martin Sebor <msebor@redhat.com>
2762 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
2765 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2767 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
2768 UNSPEC_VSX__XXSPLTD to require special splat handling.
2770 2017-01-16 David Malcolm <dmalcolm@redhat.com>
2773 * system.h: Poison strndup.
2775 2017-01-16 Alan Modra <amodra@gmail.com>
2778 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
2781 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
2783 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
2785 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
2787 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
2788 call recog here. Assert that INSN_CODE (insn) is non-negative.
2790 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
2793 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
2795 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
2796 in the currently scheduled RTL fragment.
2798 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
2800 PR rtl-optimization/78751
2801 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
2804 2017-01-14 Jeff Law <law@redhat.com>
2806 PR tree-optimization/79090
2807 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
2808 variable length stores.
2809 (compute_trims): Delete dead assignment to *trim_tail.
2810 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
2813 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
2815 PR rtl-optimization/78626
2816 PR rtl-optimization/78727
2817 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
2818 of a block, and split such blocks after everything else is finished.
2820 2017-01-14 Alan Modra <amodra@gmail.com>
2823 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
2824 target legitimate_combined_insn.
2825 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
2826 (rs6000_legitimate_combined_insn): New function.
2827 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
2829 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
2830 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
2831 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
2833 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
2835 * doc/frontends.texi (G++ and GCC): Remove references to Java.
2837 2017-01-13 Jeff Law <law@redhat.com>
2839 PR tree-optimization/33562
2840 PR tree-optimization/61912
2841 PR tree-optimization/77485
2842 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
2844 (delete_dead_assignment): Likewise.
2845 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
2846 statement to delete_dead_call and delete_dead_assignment.
2848 2017-01-13 David Malcolm <dmalcolm@redhat.com>
2851 * substring-locations.c (format_warning_va): Strengthen case 1 so
2852 that both endpoints of the substring must be within the format
2853 range for just the substring to be printed.
2855 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
2857 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
2858 * config/i386/i386.c (ix86_target_string): Add missing options
2859 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
2860 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
2861 flags_other and ix86_target_other to flags2_other. Display unknown
2863 (ix86_valid_target_attribute_inner_p): Add missing options and
2864 reorder options by implied ISAs, as in ix86_target_string.
2866 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
2868 * hash-table.h (hash_table::too_empty_p): New function.
2869 (hash_table::expand): Use it.
2870 (hash_table::traverse): Likewise.
2871 (hash_table::empty_slot): Use sizeof (value_type) instead of
2872 sizeof (PTR) to convert bytes to elements. Shrink the table
2873 if the current size is excessive for the current number of
2876 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
2878 * ira-costs.c (record_reg_classes): Break from the inner loop
2879 early once alt_fail is known to be true. Update outer loop
2880 handling accordingly.
2882 2017-01-13 Jeff Law <law@redhat.com>
2884 * tree-ssa-dse.c (decrement_count): New function.
2885 (increment_start_addr, maybe_trim_memstar_call): Likewise.
2886 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
2887 when we know the partially dead statement is a mem* function.
2889 PR tree-optimization/61912
2890 PR tree-optimization/77485
2891 * tree-ssa-dse.c: Include expr.h.
2892 (maybe_trim_constructor_store): New function.
2893 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
2895 PR tree-optimization/33562
2896 PR tree-optimization/61912
2897 PR tree-optimization/77485
2898 * doc/invoke.texi: Document new dse-max-object-size param.
2899 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
2900 * tree-ssa-dse.c: Include params.h.
2901 (dse_store_status): New enum.
2902 (initialize_ao_ref_for_dse): New, partially extracted from
2904 (valid_ao_ref_for_dse, normalize_ref): New.
2905 (setup_live_bytes_from_ref, compute_trims): Likewise.
2906 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
2907 (maybe_trim_partially_dead_store): Likewise.
2908 (maybe_trim_complex_store): Likewise.
2909 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
2910 Track what bytes live from the original store. Return tri-state
2911 for dead, partially dead or live.
2912 (dse_dom_walker): Add constructor, destructor and new private members.
2913 (delete_dead_call, delete_dead_assignment): New extracted from
2915 (dse_optimize_stmt): Make a member of dse_dom_walker.
2916 Use initialize_ao_ref_for_dse.
2918 PR tree-optimization/33562
2919 PR tree-optimization/61912
2920 PR tree-optimization/77485
2921 * sbitmap.h (bitmap_count_bits): Prototype.
2922 (bitmap_clear_range, bitmap_set_range): Likewise.
2923 * sbitmap.c (bitmap_clear_range): New function.
2924 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
2926 2017-01-13 Martin Liska <mliska@suse.cz>
2929 * function.c (set_cfun): Add new argument force.
2930 * function.h (set_cfun): Likewise.
2931 * ipa-inline-transform.c (inline_call): Use the function when
2932 strict alising from is dropped for function we inline to.
2934 2017-01-13 Richard Biener <rguenther@suse.de>
2936 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
2937 for dumping GIMPLE INTEGER_CSTs.
2939 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2941 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
2942 to 201112L since C++17.
2944 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
2947 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
2948 if -fsanitize=kernel-address is present.
2950 2017-01-13 Richard Biener <rguenther@suse.de>
2952 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
2953 as _Literal ( type ) number in case usual suffixes do not
2954 preserve all information.
2956 2017-01-13 Richard Biener <rguenther@suse.de>
2958 PR tree-optimization/77283
2959 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
2960 and ssa-iterators.h.
2961 (is_feasible_trace): Implement a cost model based on joiner
2964 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
2967 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
2968 char or short to __float128/_Float128 directly.
2970 2017-01-12 Martin Sebor <msebor@redhat.com>
2972 to -Wformat-overflow.
2973 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
2974 (min_bytes_remaining): Same.
2975 (get_string_length): Same.
2976 (format_string): Same.
2977 (format_directive): Same.
2979 (pass_sprintf_length::handle_gimple_call): Same.
2981 2017-01-12 Jakub Jelinek <jakub@redhat.com>
2983 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
2984 info.nowrite calls with no lhs that can't throw. Return bool
2985 whether gsi_remove has been called or not.
2986 (pass_sprintf_length::handle_gimple_call): Return bool whether
2987 try_substitute_return_value called gsi_remove. Formatting fix.
2988 (pass_sprintf_length::execute): Don't use gsi_remove if
2989 handle_gimple_call returned true.
2992 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
2993 be removed due to side-effects, don't remove following barrier nor
2994 turn the successor edge into fallthru edge.
2996 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2999 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
3000 element-reversing loads and stores as not swappable.
3002 2017-01-12 Nathan Sidwell <nathan@acm.org>
3003 Nicolai Stange <nicstange@gmail.com>
3005 * combine.c (try_combine): Don't ignore result of overlap checking
3006 loop. Combine overlap & asm check into single loop.
3008 2017-01-12 Richard Biener <rguenther@suse.de>
3010 * tree-pretty-print.c (dump_generic_node): Provide -gimple
3011 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
3013 2017-01-12 Richard Biener <rguenther@suse.de>
3015 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
3016 and TS_TARGET_OPTION directly derive from TS_BASE.
3017 * tree-core.h (tree_optimization_option): Derive from tree_base.
3018 (tree_target_option): Likewise.
3020 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
3022 * config/i386/i386.c (memory_address_length): Increase len
3023 only when rip_relative_addr_p returns false.
3025 2017-01-11 Julia Koval <julia.koval@intel.com>
3027 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
3028 (OPTION_MASK_ISA_SGX_SET): New.
3029 (ix86_handle_option): Handle OPT_msgx.
3030 * config.gcc: Added sgxintrin.h.
3031 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
3032 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
3033 * config/i386/i386.c (ix86_target_string): Add -msgx.
3035 (ix86_option_override_internal): Handle new options.
3036 (ix86_valid_target_attribute_inner_p): Add sgx.
3037 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
3038 * config/i386/i386.opt: Add msgx.
3039 * config/i386/sgxintrin.h: New file.
3040 * config/i386/x86intrin.h: Add sgxintrin.h.
3042 2017-01-11 Jakub Jelinek <jakub@redhat.com>
3045 * fold-const.c (maybe_nonzero_address): Return 1 for function
3047 (tree_single_nonzero_warnv_p): Don't handle function local objects
3051 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
3054 2017-01-11 David Malcolm <dmalcolm@redhat.com>
3057 * opts.c: Include "spellcheck.h"
3058 (struct string_fragment): New struct.
3059 (struct edit_distance_traits<const string_fragment &>): New
3061 (get_closest_sanitizer_option): New function.
3062 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
3064 2017-01-11 Jakub Jelinek <jakub@redhat.com>
3066 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
3068 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
3069 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
3070 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
3071 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
3072 for initial die_offset if dwarf_split_debug_info.
3073 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
3074 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
3076 (output_skeleton_debug_sections): Formatting fix. Use
3077 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
3078 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
3080 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
3082 * config/arm/cortex-a53.md: Add bypasses for
3084 (cortex_a53_r2f): Only use for transfers.
3085 (cortex_a53_f2r): Likewise.
3086 (cortex_a53_r2f_cvt): Add reservation for conversions.
3087 (cortex_a53_f2r_cvt): Likewise.
3089 2017-01-11 Tamar Christina <tamar.christina@arm.com>
3091 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
3092 to all inlined functions, change static to extern.
3094 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
3097 * config/arm/arm.c (legitimize_pic_address): Handle reference to
3099 (arm_assemble_integer): Likewise.
3101 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
3103 * config.gcc: Use new awk script to check CPU, FPU and architecture
3104 parameters for --with-... options.
3105 * config/arm/parsecpu.awk: New file
3106 * config/arm/arm-cpus.in: New file.
3107 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
3109 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
3111 * config/arm/t-arm: Update dependency rules.
3112 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
3113 of processing .def files.
3114 * config/arm/genopt.sh: Deleted.
3115 * config/arm/gentune.sh: Deleted.
3116 * config/arm/arm-cores.def: Deleted.
3117 * config/arm/arm-arches.def: Deleted.
3118 * config/arm/arm-fpus.def: Deleted.
3119 * config/arm/arm-tune.md: Regenerated.
3120 * config/arm/arm-tables.opt: Regenerated.
3121 * config/arm/arm-cpu.h: New generated file.
3122 * config/arm/arm-cpu-data.h: New generated file.
3123 * config/arm/arm-cpu-cdata.h: New generated file.
3125 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
3128 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
3130 (input_varpool_node): Unpack dynamically_initialized bit.
3132 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
3134 PR rtl-optimization/79032
3135 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
3136 the alignment of the adjusted memory reference against that of MODE,
3137 instead of the alignment of the original memory reference.
3139 2017-01-11 Martin Jambor <mjambor@suse.cz>
3141 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
3143 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
3144 decorated functions.
3146 2017-01-11 Richard Biener <rguenther@suse.de>
3148 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
3149 set range/nonnull info for PHI results. Do not set it on
3150 stmts marked for removal.
3152 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
3154 * expr.c (store_field): In the bitfield case, fetch the return value
3155 from the registers before applying a single big-endian adjustment.
3156 Always do a final load for a BLKmode value not larger than a word.
3158 2017-01-10 David Malcolm <dmalcolm@redhat.com>
3161 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
3162 that we correctly handle column numbers greater than
3163 LINE_MAP_MAX_COLUMN_NUMBER.
3165 2017-01-10 Martin Sebor <msebor@redhat.com>
3168 * gimple-ssa-sprintf.c (get_destination_size): Call
3169 {init,fini}object_sizes.
3170 * tree-object-size.c (addr_object_size): Adjust.
3171 (pass_through_call): Adjust.
3172 (pass_object_sizes::execute): Adjust.
3173 * tree-object-size.h (fini_object_sizes): Declare.
3175 2017-01-10 Martin Sebor <msebor@redhat.com>
3177 PR tree-optimization/78775
3178 * builtins.c (get_size_range): Move...
3179 * calls.c: ...to here.
3180 (alloc_max_size): Accept zero argument.
3181 (operand_signed_p): Remove.
3182 (maybe_warn_alloc_args_overflow): Call get_size_range.
3183 * calls.h (get_size_range): Declare.
3185 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
3187 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
3188 from TI's devices.csv file as of September 2016.
3189 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
3191 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
3193 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
3194 * doc/invoke.texi: Likewise.
3195 * doc/md.texi: Likewise.
3196 * doc/objc.texi: Likewise.
3198 2017-01-10 Joshua Conner <joshconner@google.com>
3200 * config/arm/fuchsia-elf.h: New file.
3201 * config/fuchsia.h: New file.
3202 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
3203 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
3205 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
3207 2016-01-10 Richard Biener <rguenther@suse.de>
3209 PR tree-optimization/79034
3210 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
3211 Propagate out degenerate PHIs in the joiner.
3213 2017-01-10 Martin Liska <mliska@suse.cz>
3215 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
3216 (sort_congruence_classes_by_decl_uid): Likewise.
3217 (sort_congruence_class_groups_by_decl_uid): Likewise.
3218 (sem_item_optimizer::merge_classes): Sort class, groups in these
3219 classes and members in the groups by DECL_UID of declarations.
3220 This would make merge operations stable.
3222 2017-01-10 Martin Liska <mliska@suse.cz>
3224 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
3225 usage of m_classes_vec.
3226 (sem_item_optimizer::~sem_item_optimizer): Likewise.
3227 (sem_item_optimizer::get_group_by_hash): Likewise.
3228 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
3229 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
3230 (sem_item_optimizer::verify_classes): Likewise.
3231 (sem_item_optimizer::process_cong_reduction): Likewise.
3232 (sem_item_optimizer::dump_cong_classes): Likewise.
3233 (sem_item_optimizer::merge_classes): Likewise.
3234 * ipa-icf.h (congruence_class_hash): Rename from
3235 congruence_class_group_hash. Remove declaration of m_classes_vec.
3237 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
3239 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
3240 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
3241 * config.gcc: Add avx512vpopcntdqintrin.h.
3242 * config/i386/avx512vpopcntdqintrin.h: New.
3243 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
3244 * config/i386/i386-builtin-types.def: Add new types.
3245 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
3246 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
3247 __builtin_ia32_vpopcountq_v8di_mask): New.
3248 * config/i386/i386-c.c (ix86_target_macros_internal): Define
3249 __AVX512VPOPCNTDQ__.
3250 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
3251 (PTA_AVX512VPOPCNTDQ): Define.
3252 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
3253 TARGET_AVX512VPOPCNTDQ_P): Define.
3254 * config/i386/i386.opt: Add mavx512vpopcntdq.
3255 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
3256 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
3258 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3261 * predict.def (PRED_CALL): Set to 67.
3263 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
3265 * expr.c (store_field): In the bitfield case, if the value comes from
3266 a function call and is of an aggregate type returned in registers, do
3267 not modify the field mode; extract the value in all cases if the mode
3268 is BLKmode and the size is not larger than a word.
3270 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
3273 * config/i386/cpuid.h: Fix undefined behavior.
3275 2017-01-04 Jeff Law <law@redhat.com>
3277 PR tree-optimization/79007
3278 PR tree-optimization/67955
3279 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
3280 conservative for pt.null when flag_non_call_exceptions is on.
3282 2017-01-09 Jakub Jelinek <jakub@redhat.com>
3284 PR translation/79019
3285 PR translation/79020
3286 * params.def (PARAM_INLINE_MIN_SPEEDUP,
3287 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
3288 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
3290 * config/avr/avr.opt (maccumulate-args): Likewise.
3291 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
3292 * common.opt (freport-bug): Likewise.
3293 * cif-code.def (CIF_FINAL_ERROR): Likewise.
3294 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
3295 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
3296 translatable string.
3297 * config/i386/i386.c (function_value_32): Likewise.
3298 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
3299 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
3301 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
3302 * common/config/msp430/msp430-common.c (msp430_handle_option):
3304 * symtab.c (symtab_node::verify_base): Likewise.
3305 * opts.c (set_debug_level): Likewise.
3306 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
3307 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
3308 missing whitespace to translatable strings.
3309 * config/avr/avr.md (bswapsi2): Fix typo in comment.
3310 * config/sh/superh.h: Likewise.
3311 * config/i386/xopintrin.h: Likewise.
3312 * config/i386/znver1.md: Likewise.
3313 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
3314 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
3315 * double-int.h (struct double_int): Likewise.
3316 * double-int.c (div_and_round_double): Likewise.
3317 * wide-int.cc: Likewise.
3318 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
3319 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
3320 * cfgcleanup.c (crossjumps_occured): Renamed to ...
3321 (crossjumps_occurred): ... this.
3322 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
3325 PR tree-optimization/78899
3326 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
3327 returning bool return struct loop *, NULL for failure and the new
3329 (versionable_outer_loop_p): Don't version outer loop if it has
3330 dont_vectorized bit set.
3331 (tree_if_conversion): When versioning outer loop, ensure
3332 tree_if_conversion is performed also on the inner loop of the
3333 non-vectorizable outer loop copy.
3334 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
3335 LOOP_VECTORIZED in inner loop of the scalar outer loop and
3336 prevent vectorization of it.
3337 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
3338 the outer loop vectorization of the non-scalar version is attempted
3339 before vectorization of the inner loop in scalar version. If
3340 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
3341 vectorization of its inner loop.
3342 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
3343 has 2 inner loops, rename also on edges from bb whose single pred
3344 is outer_loop->header. Fix typo in function comment.
3346 2017-01-09 Martin Sebor <msebor@redhat.com>
3349 * asan.c (asan_emit_stack_protection): Increase local buffer size
3350 to avoid snprintf truncation warning.
3352 2017-01-09 Andrew Pinski <apinski@cavium.com>
3354 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
3355 to reference thunderx2t99 for the tuning structure
3356 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
3358 (thunderx2t99_extra_costs): This.
3359 * config/aarch64/aarch64-tune.md: Regenerate.
3360 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
3361 (vulcan_addrcost_table): This.
3362 (vulcan_regmove_cost): Rename to ...
3363 (thunderx2t99_regmove_cost): This.
3364 (vulcan_vector_cost): Rename to ...
3365 (thunderx2t99_vector_cost): this.
3366 (vulcan_branch_cost): Rename to ...
3367 (thunderx2t99_branch_cost): This.
3368 (vulcan_tunings): Rename to ...
3369 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
3370 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
3372 2017-01-09 Martin Jambor <mjambor@suse.cz>
3376 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
3377 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
3378 (propagate_vr_accross_jump_function): Use the above function for all
3379 value range computations for pass-through jump functions and type
3380 converasion from explicit value range values.
3381 (ipcp_propagate_stage): Do not attempt to deduce types of formal
3382 parameters from TYPE_ARG_TYPES.
3383 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
3384 (ipa_write_node_info): Stream type of the actual argument.
3385 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
3387 2017-01-09 Martin Liska <mliska@suse.cz>
3390 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
3391 (lookup_compiler): Do not show error message with have_E.
3393 2017-01-09 Jakub Jelinek <jakub@redhat.com>
3395 PR tree-optimization/78938
3396 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
3397 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
3398 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
3399 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
3402 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3404 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
3407 2017-01-09 Richard Biener <rguenther@suse.de>
3409 PR tree-optimization/78997
3410 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
3411 name condition properly.
3413 2017-01-09 Richard Biener <rguenther@suse.de>
3416 * dwarf2out.c (is_cxx): New overload with context.
3417 (is_naming_typedef_decl): Use it.
3419 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
3421 * invoke.texi (Option Summary): Correct spacing in option lists
3422 and add line breaks to fix over-long lines.
3424 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
3428 * extend.texi (Common Variable Attributes): Add xref to GCC
3429 Internals manual to explain mode attribute keywords.
3431 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
3434 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
3435 and Preprocessor Options.
3436 (Options for Linking): Document -pthread here....
3437 (RS/6000 and PowerPC Options): ...not here.
3438 (Solaris 2 Options): ...or here.
3439 * doc/cppopts.texi: Document -pthread.
3441 2017-01-08 Martin Sebor <msebor@redhat.com>
3444 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
3445 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
3446 New member functions.
3447 (format_directive): Used them.
3449 (pass_sprintf_length::handle_gimple_call): Same.
3450 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
3451 to avoid truncation for any argument.
3452 (extract_affine_mul): Same.
3453 * tree.c (get_file_function_name): Same.
3455 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3458 * predict.def (PRED_INDIR_CALL): Set to 86.
3460 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
3462 PR preprocessor/54124
3463 * doc/cppopts.texi: Reformat -d subtable to list the full name
3464 of the options. Add cross-reference to the docs for the general
3465 compiler -d options.
3466 * doc/invoke.texi (Developer Options): Add cross-reference to the
3467 preprocessor-specific -d option documentation.
3469 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
3471 PR preprocessor/13498
3472 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
3473 redudant material, and reflect new command-line options.
3474 (System Headers): Likewise.
3476 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
3478 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
3479 -isystem, and -idirafter. Copy-edit.
3480 * doc/cppopts.texi: Copy-edit. Remove contradiction about
3481 default for -ftrack-macro-expansion. Delete obsolete and
3482 badly-formatted implementation details about -fdebug-cpp output.
3483 * doc/cppwarnopts.texi: Copy-edit.
3485 2017-01-07 David Malcolm <dmalcolm@redhat.com>
3488 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
3489 that the transition from a max line width >= 1<<10 to narrower
3490 lines works correctly.
3492 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
3494 * doc/options.texi (PerFunction): New.
3495 * opt-functions.awk (switch_flags): Map both Optimization and
3496 PerFunction to CL_OPTIMIZATION.
3497 * opth-gen.awk: Test for PerFunction flag along with
3499 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
3500 it only when the latter is present. Skip those that don't in
3501 the hash function generator.
3502 * common.opt (fvar-tracking): Mark as PerFunction instead of
3504 (fvar-tracking-assignments): Likewise.
3505 (fvar-tracking-assignments-toggle): Likewise.
3506 (fvar-tracking-uninit): Likewise.
3508 2017-01-07 Jakub Jelinek <jakub@redhat.com>
3510 PR translation/79018
3511 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
3514 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
3517 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
3518 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
3519 TARGET_LONG_JUMP_TABLE_OFFSETS.
3520 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
3521 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
3522 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
3523 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
3524 * config/m68k/m68k.md (tablejump expander): Likewise.
3525 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
3526 TARGET_LONG_JUMP_TABLE_OFFSETS.
3527 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
3528 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
3530 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
3531 David Holsgrove <david.holsgrove@xilinx.com>
3533 * common/config/microblaze/microblaze-common.c
3534 (TARGET_EXCEPT_UNWIND_INFO): Remove.
3535 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
3537 * config/microblaze/microblaze.c (microblaze_must_save_register)
3538 (microblaze_expand_epilogue, microblaze_return_addr): Handle
3540 (microblaze_eh_return): New function.
3541 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
3542 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
3543 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
3544 * config/microblaze/microblaze.md (eh_return): New pattern.
3546 2017-01-06 Jakub Jelinek <jakub@redhat.com>
3548 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
3549 GCC_DIAGNOSTIC_STRINGIFY): Define.
3551 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
3553 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3555 * config/arm/arm.md (<mcrr>): New.
3557 * config/arm/arm.c (arm_arch5te): New.
3558 (arm_option_override): Set arm_arch5te.
3559 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
3561 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
3562 (arm_mcrr_qualifiers): ... this. New.
3563 (MRRC_QUALIFIERS): Define to...
3564 (arm_mrrc_qualifiers): ... this. New.
3565 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
3567 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
3568 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
3569 (MRRCI, mrrc, MRRC): New.
3570 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
3571 VUNSPEC_MRRC2): New.
3573 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3575 * config/arm/arm.md (<mcr>): New.
3577 * config/arm/arm.c (arm_coproc_builtin_available): Add
3578 support for mcr, mrc, mcr2 and mrc2.
3579 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
3580 (arm_mcr_qualifiers): ... this. New.
3581 (MRC_QUALIFIERS): Define to ...
3582 (arm_mrc_qualifiers): ... this. New.
3583 (MCR_QUALIFIERS): Define to ...
3584 (arm_mcr_qualifiers): ... this. New.
3585 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
3587 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
3588 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
3589 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
3592 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3594 * config/arm/arm.md (*ldc): New.
3598 * config/arm/arm.c (arm_coproc_builtin_available): Add
3599 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
3600 (arm_coproc_ldc_stc_legitimate_address): New.
3601 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
3602 'qualifier_const_pointer'.
3603 (LDC_QUALIFIERS): Define to...
3604 (arm_ldc_qualifiers): ... this. New.
3605 (STC_QUALIFIERS): Define to...
3606 (arm_stc_qualifiers): ... this. New.
3607 * config/arm/arm-protos.h
3608 (arm_coproc_ldc_stc_legitimate_address): New.
3609 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
3610 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
3611 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
3612 stc2, stcl, stc2l): New.
3613 * config/arm/constraints.md (Uz): New.
3614 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
3615 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
3616 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
3617 VUNSPEC_STC2L): New.
3619 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3621 * config/arm/arm.md (<cdp>): New.
3622 * config/arm/arm.c (neon_const_bounds): Rename this ...
3623 (arm_const_bounds): ... this.
3624 (arm_coproc_builtin_available): New.
3625 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
3626 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
3627 (CDP_QUALIFIERS): Define to...
3628 (arm_cdp_qualifiers): ... this. New.
3630 (arm_expand_builtin_args): Add case for 6 arguments.
3631 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
3632 (arm_const_bounds): ... this.
3633 (arm_coproc_builtin_available): New.
3634 * config/arm/arm_acle.h (__arm_cdp): New.
3636 * config/arm/arm_acle_builtins.def (cdp): New.
3638 * config/arm/iterators.md (CDPI,CDP,cdp): New.
3639 * config/arm/neon.md: Rename all 'neon_const_bounds' to
3641 * config/arm/types.md (coproc): New.
3642 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
3643 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
3644 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
3645 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
3647 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3649 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
3650 (UBINOP_QUALIFIERS): New.
3652 (acle_builtin_data): New. Change comment.
3653 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
3654 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
3655 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
3656 arm_acle_builtins.def.
3657 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
3658 (arm_init_acle_builtins): New.
3659 (CRC32_BUILTIN): Remove.
3660 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
3661 crc32cb, crc32ch and crc32cw.
3662 (arm_init_crc32_builtins): Remove.
3663 (arm_init_builtins): Use arm_init_acle_builtins rather
3664 than arm_init_crc32_builtins.
3665 (arm_expand_acle_builtin): New.
3666 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
3667 * config/arm/arm_acle_builtins.def: New.
3669 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3671 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
3672 (arm_builtin_datum): ... this.
3673 (arm_init_neon_builtin): Rename to ...
3674 (arm_init_builtin): ... this. Add a new parameters PREFIX
3675 and USE_SIG_IN_NAME.
3676 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
3677 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
3678 'arm_builtin_datum'.
3679 (arm_init_vfp_builtins): Likewise.
3680 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
3681 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
3682 (arm_expand_neon_args): Rename to ...
3683 (arm_expand_builtin_args): ... this. Rename builtin_arg
3684 enum values and differentiate between ARG_BUILTIN_MEMORY
3685 and ARG_BUILTIN_NEON_MEMORY.
3686 (arm_expand_neon_builtin_1): Rename to ...
3687 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
3688 values, arm_expand_builtin_args and add bool parameter NEON.
3689 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
3690 (arm_expand_vfp_builtin): Likewise.
3691 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
3693 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3696 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
3697 * predict.c (tree_estimate_probability_bb): Reverse direction of
3698 polymorphic call predictor.
3700 2017-01-06 David Malcolm <dmalcolm@redhat.com>
3702 * passes.c (execute_one_pass): Split out pass-skipping logic into...
3703 (determine_pass_name_match): ...this new function and...
3704 (should_skip_pass_p): ...this new function.
3706 2017-01-06 Nathan Sidwell <nathan@acm.org>
3708 * ipa-visibility.c (function_and_variable_visibility): Reformat
3709 comments and long lines. Remove extrneous if.
3710 * symtab.c (symtab_node::make_decl_local): Fix code format.
3711 (symtab_node::set_section_for_node): Fix comment typo.
3713 2017-01-06 Martin Liska <mliska@suse.cz>
3716 * lra-constraints.c: Rename invariant to lra_invariant.
3717 * predict.c (set_even_probabilities): Initialize e to NULL.
3719 2017-01-05 Martin Sebor <msebor@redhat.com>
3721 PR tree-optimization/78910
3722 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
3723 (format_integer): Correct off-by-one error in the handling
3724 of precision with negative numbers in signed conversions..
3726 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
3728 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
3730 2017-01-05 Jakub Jelinek <jakub@redhat.com>
3732 PR tree-optimization/71016
3733 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
3734 factor_out_conditional_conversion. Formatting fix.
3735 (factor_out_conditional_conversion): Add cond_stmt argument.
3736 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
3737 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
3740 2017-01-05 David Malcolm <dmalcolm@redhat.com>
3742 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
3743 read-rtl-function.o, and selftest-rtl.o.
3744 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
3745 (selftest::aarch64_test_loading_full_dump): New function.
3746 (selftest::aarch64_run_selftests): New function.
3747 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
3748 selftest::aarch64_run_selftests.
3749 * config/i386/i386.c
3750 (selftest::ix86_test_loading_dump_fragment_1): New function.
3751 (selftest::ix86_test_loading_call_insn): New function.
3752 (selftest::ix86_test_loading_full_dump): New function.
3753 (selftest::ix86_test_loading_unspec): New function.
3754 (selftest::ix86_run_selftests): Call the new functions.
3755 * emit-rtl.c (maybe_set_max_label_num): New function.
3756 * emit-rtl.h (maybe_set_max_label_num): New decl.
3757 * function.c (instantiate_decls): Guard call to
3758 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
3759 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
3761 * gensupport.c (gen_reader::gen_reader): Pass "false"
3762 for new "compact" param of rtx_reader.
3763 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
3764 rather than an empty string for NULL strings.
3765 * read-md.c: Potentially include config.h rather than bconfig.h.
3766 Wrap include of errors.h with #ifdef GENERATOR_FILE.
3767 (have_error): New global, copied from errors.c.
3768 (md_reader::read_name): Rename to...
3769 (md_reader::read_name_1): ...this, adding "out_loc" param,
3770 and converting "missing name or number" to returning false, rather
3772 (md_reader::read_name): Reimplement in terms of read_name_1.
3773 (md_reader::read_name_or_nil): New function.
3774 (md_reader::read_string): Handle "(nil)" by returning NULL.
3775 (md_reader::md_reader): Add new param "compact".
3776 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
3777 (md_reader::read_file): New method.
3778 * read-md.h (md_reader::md_reader): Add new param "compact".
3779 (md_reader::read_file): New method.
3780 (md_reader::is_compact): New accessor.
3781 (md_reader::read_name): Convert return type from void to file_location.
3782 (md_reader::read_name_or_nil): New decl.
3783 (md_reader::read_name_1): New decl.
3784 (md_reader::m_compact): New field.
3785 (noop_reader::noop_reader): Pass "false" for new "compact" param
3787 (rtx_reader::rtx_reader): Add new "compact" param.
3788 (rtx_reader::read_rtx_operand): Make virtual and convert return
3789 type from void to rtx.
3790 (rtx_reader::read_until): New decl.
3791 (rtx_reader::handle_any_trailing_information): New virtual function.
3792 (rtx_reader::postprocess): New virtual function.
3793 (rtx_reader::finalize_string): New virtual function.
3794 (rtx_reader::m_in_call_function_usage): New field.
3795 (rtx_reader::m_reuse_rtx_by_id): New field.
3796 * read-rtl-function.c: New file.
3797 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
3798 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
3799 (selftest::verify_three_block_rtl_cfg): New decl.
3800 * read-rtl-function.h: New file.
3801 * read-rtl.c: Potentially include config.h rather than bconfig.h.
3802 For host, include function.h, memmodel.h, and emit-rtl.h.
3803 (one_time_initialization): New function.
3804 (struct compact_insn_name): New struct.
3805 (compact_insn_names): New array.
3806 (find_code): Handle insn codes in compact dumps.
3807 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
3808 (bind_subst_iter_and_attr): Likewise.
3809 (add_condition_to_string): Likewise.
3810 (add_condition_to_rtx): Likewise.
3811 (apply_attribute_uses): Likewise.
3812 (add_current_iterators): Likewise.
3813 (apply_iterators): Likewise.
3814 (initialize_iterators): Guard usage of apply_subst_iterator with
3815 #ifdef GENERATOR_FILE.
3816 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
3817 (md_reader::read_mapping): Likewise.
3818 (add_define_attr_for_define_subst): Likewise.
3819 (add_define_subst_attr): Likewise.
3820 (read_subst_mapping): Likewise.
3821 (check_code_iterator): Likewise.
3822 (rtx_reader::read_rtx): Likewise. Move one-time initialization
3824 (one_time_initialization): New function.
3825 (rtx_reader::read_until): New method.
3826 (read_flags): New function.
3827 (parse_reg_note_name): New function.
3828 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
3829 Handle reuse_rtx ids.
3830 Wrap iterator lookup within #ifdef GENERATOR_FILE.
3831 Add parsing support for RTL dumps, mirroring the special-cases in
3832 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
3833 values, and calling handle_any_trailing_information.
3834 (rtx_reader::read_rtx_operand): Convert return type from void
3835 to rtx, returning return_rtx. Handle case 'e'. Call
3836 finalize_string on XSTR and XTMPL fields.
3837 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
3838 "(nil)" values were omitted. Call the postprocess vfunc on the
3840 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
3841 class ctor. Initialize m_in_call_function_usage. Call
3842 one_time_initialization.
3843 * rtl-tests.c (selftest::test_uncond_jump): Call
3844 set_new_first_and_last_insn.
3845 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
3846 * selftest-rtl.c: New file.
3847 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
3848 (selftest::get_insn_by_uid): New decl.
3849 * selftest-run-tests.c (selftest::run_tests): Call
3850 read_rtl_function_c_tests.
3851 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
3852 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
3855 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
3857 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
3858 operands in a special way. Assert that pos+len <= mode precision.
3860 2017-01-05 Jakub Jelinek <jakub@redhat.com>
3862 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
3863 3 argument Alias with unlimited for the negative form.
3864 (fno-vect-cost-model): Removed.
3866 2017-01-05 Martin Liska <mliska@suse.cz>
3868 * hsa-gen.c (gen_hsa_divmod): New function.
3869 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
3871 2017-01-05 Martin Liska <mliska@suse.cz>
3874 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
3877 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
3879 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
3880 small constant length operands.
3882 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
3884 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
3885 between loop iterations.
3887 2017-01-05 Martin Liska <mliska@suse.cz>
3890 * gimplify.c (gimplify_decl_expr): Compare to
3891 asan_poisoned_variables instread of checking flags.
3892 (gimplify_target_expr): Likewise.
3893 (gimplify_expr): Likewise.
3894 (gimplify_function_tree): Conditionally initialize
3895 asan_poisoned_variables.
3897 2017-01-04 Jeff Law <law@redhat.com>
3899 PR tree-optimizatin/78812
3900 * rtl.h (contains_mem_rtx_p): Prototype.
3901 * ifcvt.c (containts_mem_rtx_p): Move from here to...
3902 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
3903 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
3904 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
3905 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
3907 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3909 * input.c (assert_char_at_range): Default-initialize actual_range.
3911 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3913 * df-scan.c (df_ref_create_structure): Make regno unsigned,
3914 to match the caller.
3916 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3918 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
3919 insns after final jump in test to emit dummy move.
3921 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3923 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
3924 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
3926 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3928 * multiple_target.c (create_dispatcher_calls): Init e_next.
3929 * tree-ssa-loop-split.c (split_loop): Init border.
3930 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
3933 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
3938 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
3939 (altivec_register_operand): Do not return true if the operand
3940 contains a SUBREG mixing SImode and SFmode.
3941 (vsx_register_operand): Likewise.
3942 (vsx_reg_sfsubreg_ok): New predicate.
3943 (vfloat_operand): Do not return true if the operand contains a
3944 SUBREG mixing SImode and SFmode.
3945 (vint_operand): Likewise.
3946 (vlogical_operand): Likewise.
3947 (gpc_reg_operand): Likewise.
3948 (int_reg_operand): Likewise.
3949 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
3950 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
3951 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
3953 (rs6000_emit_move_si_sf_subreg): New helper function.
3954 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
3955 fixup SUBREGs involving SImode and SFmode.
3956 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
3957 numbers for the new peephole2 optimization.
3958 (peephole2 for SFmode unions): New peephole2 to optimize cases in
3959 the GLIBC math library that do AND/IOR/XOR operations on single
3960 precision floating point.
3961 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
3962 target macros to say whether we need to avoid SUBREGs mixing
3964 (TARGET_ALLOW_SF_SUBREG): Likewise.
3965 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
3966 (UNSPEC_SI_FROM_SF): Likewise.
3967 (iorxor): Change spacing.
3968 (and_ior_xor): New iterator for AND, IOR, and XOR.
3969 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
3970 (movdi_from_sf_zero_ext): Likewise.
3971 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
3972 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
3973 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
3974 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
3975 (fms<mode>4): Likewise.
3976 (fnma<mode>4): Likewise.
3977 (fnms<mode>4): Likewise.
3978 (nfma<mode>4): Likewise.
3979 (nfms<mode>4): Likewise.
3981 2017-01-04 Marek Polacek <polacek@redhat.com>
3984 * doc/invoke.texi: Document -Wpointer-compare.
3986 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3988 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
3991 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
3992 descriptions for -gdwarf-5 and emit them as uleb128 instead of
3995 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
3998 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
3999 documentation of the powerpc_popcntb_ok attribute.
4000 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
4001 code to issue warning messages if a requested CPU configuration is
4002 not supported by the binary (assembler and loader) toolchain.
4003 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
4004 made to define a built-in function that has been disabled.
4005 (paired_init_builtins): Add assertion to prevent ICE if attempt is
4006 made to define a built-in function that has been disabled.
4007 (altivec_init_builtins): Add comment explaining why definition
4008 of the DST built-in functions is not preceded by an assertion
4009 check. Add assertions to prevent ICE if attempts are made to
4010 define an altivec predicate or an abs* built-in function that has
4012 (htm_init_builtins): Add comment explaining why definition of the
4013 htm built-in functions is not preceded by an assertion check.
4015 2017-01-04 Jeff Law <law@redhat.com>
4017 PR tree-optimizatin/67955
4018 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
4019 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
4020 the points-to solution does not include pt_null. Use DECL_PT_UID
4023 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
4025 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
4026 Use gen_int_mode instead of gen_lopwart for const_int operands.
4028 2017-01-04 Jakub Jelinek <jakub@redhat.com>
4030 PR tree-optimization/71563
4031 * match.pd: Simplify X << Y into X if Y is known to be 0 or
4032 out of range value - has low bits known to be zero.
4034 2017-01-04 Alan Modra <amodra@gmail.com>
4036 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
4037 * configure: Regenerate.
4038 * config.in: Regenerate.
4040 2017-01-04 Jakub Jelinek <jakub@redhat.com>
4043 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
4044 a substring of the message, but strcmp with the whole message. Ifdef
4045 ENABLE_NLS, translate the message first using dgettext.
4047 2017-01-03 Jeff Law <law@redhat.com>
4049 PR tree-optimizatin/78856
4050 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
4051 (mark_threaded_blocks): Remove code to truncate thread paths that
4052 cross multiple loop headers. Instead invalidate the cached loop
4053 iteration information and handle case of a thread path walking
4054 into an irreducible region.
4056 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
4059 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
4060 assertions. Add support for doing the signbit if the IEEE 128-bit
4061 floating point value is in a GPR.
4062 * config/rs6000/rs6000.md (Fsignbit): Delete.
4063 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
4064 Update the length attribute if the value is in a GPR.
4065 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
4066 the sign or zero extension instruction, since the value is always 0/1.
4067 (signbit<mode>2_dm2): Delete using <Fsignbit>.
4070 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
4071 extracting SImode to a GPR register so that we can generate a
4072 store, limit the vector to be in a traditional Altivec register
4073 for the vextuwrx instruction.
4075 2017-01-03 Ian Lance Taylor <iant@google.com>
4077 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
4079 2017-01-03 Martin Sebor <msebor@redhat.com>
4081 PR tree-optimization/78696
4082 * gimple-ssa-sprintf.c (format_floating): Correct handling of
4083 precision. Use MPFR for %f for greater fidelity. Correct handling
4085 (pass_sprintf_length::compute_format_length): Set width and precision
4086 specified by asrerisk to void_node for vararg functions.
4087 (try_substitute_return_value): Adjust dump output.
4089 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
4091 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
4093 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
4095 * doc/invoke.texi (SPARC options): Document -mlra as the default.
4096 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
4097 -mlra/-mno-lra was passed to the compiler.
4099 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
4101 PR rtl-optimization/65618
4102 * emit-rtl.c (try_split): Move initialization of "before" and
4103 "after" to just before the call to emit_insn_after_setloc.
4105 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
4107 * doc/md.texi (Standard Names): Remove reference to Java frontend.
4109 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
4111 * dwarf2out.c (gen_enumeration_type_die): When
4112 -gno-strict-dwarf, add a DW_AT_encoding attribute.
4114 2017-01-03 Jakub Jelinek <jakub@redhat.com>
4116 PR tree-optimization/78965
4117 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
4118 Change first argument from const call_info & to call_info &. For %n
4119 set info.nowrite to false.
4122 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
4123 possibly throwing calls.
4125 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
4126 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
4127 and fns handling, rather than in a separate case SSA_NAME.
4129 2017-01-02 Jeff Law <law@redhat.com>
4131 * config/darwin-driver.c (darwin_driver_init): Const-correctness
4132 fixes for first_period and second_period variables.
4134 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
4137 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
4138 (*insvqi_1): New insn pattern.
4139 (*insvqi_1_mem_rex64): Ditto.
4141 (*insvqi_3): Rename from *insvqi.
4143 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
4145 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
4147 * doc/cfg.texi (Edges): Remove reference to Java.
4148 (Maintaining the CFG): Ditto.
4150 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
4153 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
4154 transparent aliases.
4156 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
4159 * predict.def (PRED_CALL): Update hitrate.
4160 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
4161 * predict.c (tree_estimate_probability_bb): Split CALL predictor
4162 into direct/indirect/polymorphic variants.
4164 2017-01-01 Jakub Jelinek <jakub@redhat.com>
4166 Update copyright years.
4168 * gcc.c (process_command): Update copyright notice dates.
4169 * gcov-dump.c (print_version): Ditto.
4170 * gcov.c (print_version): Ditto.
4171 * gcov-tool.c (print_version): Ditto.
4172 * gengtype.c (create_file): Ditto.
4173 * doc/cpp.texi: Bump @copying's copyright year.
4174 * doc/cppinternals.texi: Ditto.
4175 * doc/gcc.texi: Ditto.
4176 * doc/gccint.texi: Ditto.
4177 * doc/gcov.texi: Ditto.
4178 * doc/install.texi: Ditto.
4179 * doc/invoke.texi: Ditto.
4181 Copyright (C) 2017 Free Software Foundation, Inc.
4183 Copying and distribution of this file, with or without modification,
4184 are permitted in any medium without royalty provided the copyright
4185 notice and this notice are preserved.