1 2023-08-26 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3 * config/riscv/autovec.md (len_fold_extract_last_<mode>): New pattern.
4 * config/riscv/riscv-protos.h (enum insn_type): New enum.
5 (expand_fold_extract_last): New function.
6 * config/riscv/riscv-v.cc (emit_nonvlmax_slide_insn): Ditto.
7 (emit_cpop_insn): Ditto.
8 (emit_nonvlmax_compress_insn): Ditto.
9 (expand_fold_extract_last): Ditto.
10 * config/riscv/vector.md: Fix vcpop.m ratio demand.
12 2023-08-25 Edwin Lu <ewlu@rivosinc.com>
14 * config/riscv/sync-rvwmo.md: updated types to "multi" or
15 "atomic" based on number of assembly lines generated
16 * config/riscv/sync-ztso.md: likewise
17 * config/riscv/sync.md: likewise
19 2023-08-25 Jin Ma <jinma@linux.alibaba.com>
21 * common/config/riscv/riscv-common.cc: Add zfa extension version, which depends on
23 * config/riscv/constraints.md (zfli): Constrain the floating point number that the
24 instructions FLI.H/S/D can load.
25 * config/riscv/iterators.md (ceil): New.
26 * config/riscv/riscv-opts.h (MASK_ZFA): New.
28 * config/riscv/riscv-protos.h (riscv_float_const_rtx_index_for_fli): New.
29 * config/riscv/riscv.cc (riscv_float_const_rtx_index_for_fli): New.
30 (riscv_cannot_force_const_mem): If instruction FLI.H/S/D can be used, memory is
32 (riscv_const_insns): Likewise.
33 (riscv_legitimize_const_move): Likewise.
34 (riscv_split_64bit_move_p): If instruction FLI.H/S/D can be used, no split is
36 (riscv_split_doubleword_move): Likewise.
37 (riscv_output_move): Output the mov instructions in zfa extension.
38 (riscv_print_operand): Output the floating-point value of the FLI.H/S/D immediate
40 (riscv_secondary_memory_needed): Likewise.
41 * config/riscv/riscv.md (fminm<mode>3): New.
43 (movsidf2_low_rv32): New.
44 (movsidf2_high_rv32): New.
45 (movdfsisi3_rv32): New.
46 (f<quiet_pattern>_quiet<ANYF:mode><X:mode>4_zfa): New.
47 * config/riscv/riscv.opt: New.
49 2023-08-25 Sandra Loosemore <sandra@codesourcery.com>
52 * omp-general.cc (omp_runtime_api_procname): New.
53 (omp_runtime_api_call): Moved here from omp-low.cc, and make
55 * omp-general.h: Include omp-api.h.
56 * omp-low.cc (omp_runtime_api_call): Delete this copy.
58 2023-08-25 Sandra Loosemore <sandra@codesourcery.com>
60 * doc/generic.texi (OpenMP): Document OMP_STRUCTURED_BLOCK.
61 * doc/gimple.texi (GIMPLE instruction set): Add
62 GIMPLE_OMP_STRUCTURED_BLOCK.
63 (GIMPLE_OMP_STRUCTURED_BLOCK): New subsection.
64 * gimple-low.cc (lower_stmt): Error on GIMPLE_OMP_STRUCTURED_BLOCK.
65 * gimple-pretty-print.cc (dump_gimple_omp_block): Handle
66 GIMPLE_OMP_STRUCTURED_BLOCK.
67 (pp_gimple_stmt_1): Likewise.
68 * gimple-walk.cc (walk_gimple_stmt): Likewise.
69 * gimple.cc (gimple_build_omp_structured_block): New.
70 * gimple.def (GIMPLE_OMP_STRUCTURED_BLOCK): New.
71 * gimple.h (gimple_build_omp_structured_block): Declare.
72 (gimple_has_substatements): Handle GIMPLE_OMP_STRUCTURED_BLOCK.
73 (CASE_GIMPLE_OMP): Likewise.
74 * gimplify.cc (is_gimple_stmt): Handle OMP_STRUCTURED_BLOCK.
75 (gimplify_expr): Likewise.
76 * omp-expand.cc (GIMPLE_OMP_STRUCTURED_BLOCK): Error on
77 GIMPLE_OMP_STRUCTURED_BLOCK.
78 * omp-low.cc (scan_omp_1_stmt): Handle GIMPLE_OMP_STRUCTURED_BLOCK.
79 (lower_omp_1): Likewise.
80 (diagnose_sb_1): Likewise.
81 (diagnose_sb_2): Likewise.
82 * tree-inline.cc (remap_gimple_stmt): Handle
83 GIMPLE_OMP_STRUCTURED_BLOCK.
84 (estimate_num_insns): Likewise.
85 * tree-nested.cc (convert_nonlocal_reference_stmt): Likewise.
86 (convert_local_reference_stmt): Likewise.
87 (convert_gimple_call): Likewise.
88 * tree-pretty-print.cc (dump_generic_node): Handle
90 * tree.def (OMP_STRUCTURED_BLOCK): New.
91 * tree.h (OMP_STRUCTURED_BLOCK_BODY): New.
93 2023-08-25 Vineet Gupta <vineetg@rivosinc.com>
95 * config/riscv/riscv.cc (riscv_rtx_costs): Adjust const_int
96 cost. Add some comments about different constants handling.
98 2023-08-25 Andrew Pinski <apinski@marvell.com>
100 * match.pd (`a ? one_zero : one_zero`): Move
101 below detection of minmax.
103 2023-08-25 Andrew Pinski <apinski@marvell.com>
105 * match.pd (`a | C -> C`): New pattern.
107 2023-08-25 Uros Bizjak <ubizjak@gmail.com>
109 * caller-save.cc (new_saved_hard_reg):
110 Rename TRUE/FALSE to true/false.
111 (setup_save_areas): Ditto.
112 * gcc.cc (set_collect_gcc_options): Ditto.
113 (driver::build_multilib_strings): Ditto.
114 (print_multilib_info): Ditto.
115 * genautomata.cc (gen_cpu_unit): Ditto.
116 (gen_query_cpu_unit): Ditto.
118 (gen_excl_set): Ditto.
119 (gen_presence_absence_set): Ditto.
120 (gen_presence_set): Ditto.
121 (gen_final_presence_set): Ditto.
122 (gen_absence_set): Ditto.
123 (gen_final_absence_set): Ditto.
124 (gen_automaton): Ditto.
125 (gen_regexp_repeat): Ditto.
126 (gen_regexp_allof): Ditto.
127 (gen_regexp_oneof): Ditto.
128 (gen_regexp_sequence): Ditto.
129 (process_decls): Ditto.
130 (reserv_sets_are_intersected): Ditto.
131 (initiate_excl_sets): Ditto.
132 (form_reserv_sets_list): Ditto.
133 (check_presence_pattern_sets): Ditto.
134 (check_absence_pattern_sets): Ditto.
135 (check_regexp_units_distribution): Ditto.
136 (check_unit_distributions_to_automata): Ditto.
137 (create_ainsns): Ditto.
138 (output_insn_code_cases): Ditto.
139 (output_internal_dead_lock_func): Ditto.
140 (form_important_insn_automata_lists): Ditto.
141 * gengtype-state.cc (read_state_files_list): Ditto.
142 * gengtype.cc (main): Ditto.
143 * gimple-array-bounds.cc (array_bounds_checker::check_array_bounds):
145 * gimple.cc (gimple_build_call_from_tree): Ditto.
146 (preprocess_case_label_vec_for_gimple): Ditto.
147 * gimplify.cc (gimplify_call_expr): Ditto.
148 * ordered-hash-map-tests.cc (test_map_of_int_to_strings): Ditto.
150 2023-08-25 Richard Biener <rguenther@suse.de>
152 PR tree-optimization/111137
153 * tree-vect-data-refs.cc (vect_slp_analyze_load_dependences):
154 Properly handle grouped stores from other SLP instances.
156 2023-08-25 Richard Biener <rguenther@suse.de>
158 * tree-vect-data-refs.cc (vect_slp_analyze_store_dependences):
159 Split out from vect_slp_analyze_node_dependences, remove
161 (vect_slp_analyze_load_dependences): Split out from
162 vect_slp_analyze_node_dependences, adjust comments. Process
163 queued stores before any disambiguation.
164 (vect_slp_analyze_node_dependences): Remove.
165 (vect_slp_analyze_instance_dependence): Adjust.
167 2023-08-25 Aldy Hernandez <aldyh@redhat.com>
169 * range-op-float.cc (frelop_early_resolve): Rewrite for better NAN
171 (operator_not_equal::fold_range): Adjust for relations.
172 (operator_lt::fold_range): Same.
173 (operator_gt::fold_range): Same.
174 (foperator_unordered_equal::fold_range): Same.
175 (foperator_unordered_lt::fold_range): Same.
176 (foperator_unordered_le::fold_range): Same.
177 (foperator_unordered_gt::fold_range): Same.
178 (foperator_unordered_ge::fold_range): Same.
180 2023-08-25 Richard Biener <rguenther@suse.de>
182 PR tree-optimization/111136
183 * tree-vect-loop.cc (vect_dissolve_slp_only_groups): For
184 stores force STMT_VINFO_STRIDED_P and also duplicate that
187 2023-08-25 Juzhe-Zhong <juzhe.zhong@rivai.ai>
189 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::compute_local_properties):
192 2023-08-25 liuhongt <hongtao.liu@intel.com>
194 * config/i386/sse.md (vec_set<mode>): Removed.
195 (V_128H): Merge into ..
197 (V_256H): Merge into ..
199 (V_512): Add V32HF, V32BF.
200 (*ssse3_palignr<mode>_perm): Adjust mode iterator from V_128H
202 (vcond<mode><sseintvecmodelower>): Removed
203 (vcondu<mode><sseintvecmodelower>): Removed.
204 (avx_vbroadcastf128_<mode>): Refator from V_256H to V_256.
206 2023-08-25 Hongyu Wang <hongyu.wang@intel.com>
209 * config/i386/sse.md (avx512f_cvtne2ps2bf16_<mode>_maskz):
210 Adjust paramter order.
212 2023-08-24 Uros Bizjak <ubizjak@gmail.com>
215 * config/i386/sse.md (*sse2_movq128_<mode>_1): New insn pattern.
217 2023-08-24 David Malcolm <dmalcolm@redhat.com>
220 * doc/invoke.texi (Static Analyzer Options): Add "strcat" to the
221 list of functions known to the analyzer.
223 2023-08-24 Richard Biener <rguenther@suse.de>
225 PR tree-optimization/111123
226 * tree-ssa-ccp.cc (pass_fold_builtins::execute): Do not
227 remove indirect clobbers here ...
228 * tree-outof-ssa.cc (rewrite_out_of_ssa): ... but here.
229 (remove_indirect_clobbers): New function.
231 2023-08-24 Jan Hubicka <jh@suse.cz>
233 * cfg.h (struct control_flow_graph): New field full_profile.
234 * auto-profile.cc (afdo_annotate_cfg): Set full_profile to true.
235 * cfg.cc (init_flow): Set full_profile to false.
236 * graphite.cc (graphite_transform_loops): Set full_profile to false.
237 * lto-streamer-in.cc (input_cfg): Initialize full_profile flag.
238 * predict.cc (pass_profile::execute): Set full_profile to true.
239 * symtab-thunks.cc (expand_thunk): Set full_profile to true.
240 * tree-cfg.cc (gimple_verify_flow_info): Verify that profile is full
241 if full_profile is set.
242 * tree-inline.cc (initialize_cfun): Initialize full_profile.
243 (expand_call_inline): Combine full_profile.
245 2023-08-24 Richard Biener <rguenther@suse.de>
247 * tree-vect-slp.cc (vect_build_slp_tree_1): Rename
248 load_p to ldst_p, fix mistakes and rely on
251 2023-08-24 Jan Hubicka <jh@suse.cz>
253 * gimple-harden-conditionals.cc (insert_check_and_trap): Set count
254 of newly build trap bb.
256 2023-08-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
258 * config/riscv/riscv.cc (riscv_preferred_else_value): Remove it since
259 it forbid COND_LEN_FMS/COND_LEN_FNMS STMT fold.
260 (TARGET_PREFERRED_ELSE_VALUE): Ditto.
262 2023-08-24 Robin Dapp <rdapp.gcc@gmail.com>
264 * common/config/riscv/riscv-common.cc: Add -fsched-pressure.
265 * config/riscv/riscv.cc (riscv_option_override): Set sched
268 2023-08-24 Robin Dapp <rdapp@ventanamicro.com>
270 * config/riscv/riscv.cc (riscv_print_operand): Allow vk operand.
272 2023-08-24 Richard Biener <rguenther@suse.de>
274 PR tree-optimization/111125
275 * tree-vect-slp.cc (vect_slp_function): Split at novector
276 loop entry, do not push blocks in novector loops.
278 2023-08-24 Richard Sandiford <richard.sandiford@arm.com>
280 * doc/extend.texi: Document the C [[__extension__ ...]] construct.
282 2023-08-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
284 * genmatch.cc (decision_tree::gen): Support
285 COND_LEN_FNMA/COND_LEN_FMS/COND_LEN_FNMS gimple fold.
286 * gimple-match-exports.cc (gimple_simplify): Ditto.
287 (gimple_resimplify6): New function.
288 (gimple_resimplify7): New function.
289 (gimple_match_op::resimplify): Support
290 COND_LEN_FNMA/COND_LEN_FMS/COND_LEN_FNMS gimple fold.
291 (convert_conditional_op): Ditto.
292 (build_call_internal): Ditto.
293 (try_conditional_simplification): Ditto.
294 (gimple_extract): Ditto.
295 * gimple-match.h (gimple_match_cond::gimple_match_cond): Ditto.
296 * internal-fn.cc (CASE): Ditto.
298 2023-08-24 Richard Biener <rguenther@suse.de>
300 PR tree-optimization/111115
301 * tree-vectorizer.h (vect_slp_child_index_for_operand): New.
302 * tree-vect-data-refs.cc (can_group_stmts_p): Also group
304 * tree-vect-slp.cc (arg3_arg2_map): New.
305 (vect_get_operand_map): Handle IFN_MASK_STORE.
306 (vect_slp_child_index_for_operand): New function.
307 (vect_build_slp_tree_1): Handle statements with no LHS,
309 (vect_remove_slp_scalar_calls): Likewise.
310 * tree-vect-stmts.cc (vect_check_store_rhs): Lookup the
311 SLP child corresponding to the ifn value index.
312 (vectorizable_store): Likewise for the mask index. Support
314 (vectorizable_load): Lookup the SLP child corresponding to the
317 2023-08-24 Richard Biener <rguenther@suse.de>
319 PR tree-optimization/111125
320 * tree-vect-slp.cc (vectorizable_bb_reduc_epilogue): Account
321 for the remain_defs processing.
323 2023-08-24 Richard Sandiford <richard.sandiford@arm.com>
325 * config/aarch64/aarch64.cc: Include ssa.h.
326 (aarch64_multiply_add_p): Require the second operand of an
327 Advanced SIMD subtraction to be a multiplication. Assume that
328 such an operation won't be fused if the second operand is used
329 multiple times and if the first operand is also a multiplication.
331 2023-08-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
333 * tree-vect-loop.cc (vectorizable_reduction): Apply
334 LEN_FOLD_EXTRACT_LAST.
335 * tree-vect-stmts.cc (vectorizable_condition): Ditto.
337 2023-08-24 Richard Biener <rguenther@suse.de>
339 PR tree-optimization/111128
340 * tree-vect-patterns.cc (vect_recog_over_widening_pattern):
341 Emit external shift operand inline if we promoted it with
342 another pattern stmt.
344 2023-08-24 Pan Li <pan2.li@intel.com>
346 * config/riscv/autovec.md: Fix typo.
348 2023-08-24 Pan Li <pan2.li@intel.com>
350 * config/riscv/riscv-vector-builtins-bases.cc
351 (class binop_frm): Removed.
352 (class reverse_binop_frm): Ditto.
353 (class widen_binop_frm): Ditto.
354 (class vfmacc_frm): Ditto.
355 (class vfnmacc_frm): Ditto.
356 (class vfmsac_frm): Ditto.
357 (class vfnmsac_frm): Ditto.
358 (class vfmadd_frm): Ditto.
359 (class vfnmadd_frm): Ditto.
360 (class vfmsub_frm): Ditto.
361 (class vfnmsub_frm): Ditto.
362 (class vfwmacc_frm): Ditto.
363 (class vfwnmacc_frm): Ditto.
364 (class vfwmsac_frm): Ditto.
365 (class vfwnmsac_frm): Ditto.
366 (class unop_frm): Ditto.
367 (class vfrec7_frm): Ditto.
368 (class binop): Add frm_op_type template arg.
370 (class widen_binop): Ditto.
371 (class widen_binop_fp): Ditto.
372 (class reverse_binop): Ditto.
373 (class vfmacc): Ditto.
374 (class vfnmsac): Ditto.
375 (class vfmadd): Ditto.
376 (class vfnmsub): Ditto.
377 (class vfnmacc): Ditto.
378 (class vfmsac): Ditto.
379 (class vfnmadd): Ditto.
380 (class vfmsub): Ditto.
381 (class vfwmacc): Ditto.
382 (class vfwnmacc): Ditto.
383 (class vfwmsac): Ditto.
384 (class vfwnmsac): Ditto.
385 (class float_misc): Ditto.
387 2023-08-24 Andrew Pinski <apinski@marvell.com>
389 PR tree-optimization/111109
390 * match.pd (ior(cond,cond), ior(vec_cond,vec_cond)):
391 Add check to make sure cmp and icmp are inverse.
393 2023-08-24 Andrew Pinski <apinski@marvell.com>
395 PR tree-optimization/95929
396 * match.pd (convert?(-a)): New pattern
397 for 1bit integer types.
399 2023-08-24 Haochen Jiang <haochen.jiang@intel.com>
402 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
404 * common/config/i386/cpuinfo.h (get_available_features):
405 Add avx10_set and version and detect avx10.1.
406 (cpu_indicator_init): Handle avx10.1-512.
407 * common/config/i386/i386-common.cc
408 (OPTION_MASK_ISA2_AVX10_512BIT_SET): New.
409 (OPTION_MASK_ISA2_AVX10_1_SET): Ditto.
410 (OPTION_MASK_ISA2_AVX10_512BIT_UNSET): Ditto.
411 (OPTION_MASK_ISA2_AVX10_1_UNSET): Ditto.
412 (OPTION_MASK_ISA2_AVX2_UNSET): Modify for AVX10_1.
413 (ix86_handle_option): Handle -mavx10.1, -mavx10.1-256 and
415 * common/config/i386/i386-cpuinfo.h (enum processor_features):
416 Add FEATURE_AVX10_512BIT, FEATURE_AVX10_1 and
417 FEATURE_AVX10_512BIT.
418 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
419 AVX10_512BIT, AVX10_1 and AVX10_1_512.
420 * config/i386/constraints.md (Yk): Add AVX10_1.
423 * config/i386/cpuid.h (bit_AVX10): New.
424 (bit_AVX10_256): Ditto.
425 (bit_AVX10_512): Ditto.
426 * config/i386/i386-c.cc (ix86_target_macros_internal):
427 Define AVX10_512BIT and AVX10_1.
428 * config/i386/i386-isa.def
429 (AVX10_512BIT): Add DEF_PTA(AVX10_512BIT).
430 (AVX10_1): Add DEF_PTA(AVX10_1).
431 * config/i386/i386-options.cc (isa2_opts): Add -mavx10.1.
432 (ix86_valid_target_attribute_inner_p): Handle avx10-512bit, avx10.1
434 (ix86_option_override_internal): Enable AVX512{F,VL,BW,DQ,CD,BF16,
435 FP16,VBMI,VBMI2,VNNI,IFMA,BITALG,VPOPCNTDQ} features for avx10.1-512.
436 (ix86_valid_target_attribute_inner_p): Handle AVX10_1.
437 * config/i386/i386.cc (ix86_get_ssemov): Add AVX10_1.
438 (ix86_conditional_register_usage): Ditto.
439 (ix86_hard_regno_mode_ok): Ditto.
440 (ix86_rtx_costs): Ditto.
441 * config/i386/i386.h (VALID_MASK_AVX10_MODE): New macro.
442 * config/i386/i386.opt: Add option -mavx10.1, -mavx10.1-256 and
444 * doc/extend.texi: Document avx10.1, avx10.1-256 and avx10.1-512.
445 * doc/invoke.texi: Document -mavx10.1, -mavx10.1-256 and -mavx10.1-512.
446 * doc/sourcebuild.texi: Document target avx10.1, avx10.1-256
449 2023-08-24 Haochen Jiang <haochen.jiang@intel.com>
452 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
454 * common/config/i386/i386-common.cc
455 (ix86_check_avx10): New function to check isa_flags and
456 isa_flags_explicit to emit warning when AVX10 is enabled
458 (ix86_check_avx512): New function to check isa_flags and
459 isa_flags_explicit to emit warning when AVX512 is enabled
461 (ix86_handle_option): Do not change the flags when warning
463 * config/i386/driver-i386.cc (host_detect_local_cpu):
464 Do not append -mno-avx10.1 for -march=native.
466 2023-08-24 Haochen Jiang <haochen.jiang@intel.com>
469 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
471 * common/config/i386/i386-common.cc
472 (ix86_check_avx10_vector_width): New function to check isa_flags
473 to emit a warning when there is a conflict in AVX10 options for
475 (ix86_handle_option): Add check for avx10.1-256 and avx10.1-512.
476 * config/i386/driver-i386.cc (host_detect_local_cpu):
477 Do not append -mno-avx10-max-512bit for -march=native.
479 2023-08-24 Haochen Jiang <haochen.jiang@intel.com>
482 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
484 * config/i386/avx512vldqintrin.h: Remove target attribute.
485 * config/i386/i386-builtin.def (BDESC):
486 Add OPTION_MASK_ISA2_AVX10_1.
487 * config/i386/i386-builtins.cc (def_builtin): Handle AVX10_1.
488 * config/i386/i386-expand.cc
489 (ix86_check_builtin_isa_match): Ditto.
490 (ix86_expand_sse2_mulvxdi3): Add TARGET_AVX10_1.
491 * config/i386/i386.md: Add new isa attribute avx10_1_or_avx512dq
492 and avx10_1_or_avx512vl.
493 * config/i386/sse.md: (VF2_AVX512VLDQ_AVX10_1): New.
494 (VF1_128_256VLDQ_AVX10_1): Ditto.
495 (VI8_AVX512VLDQ_AVX10_1): Ditto.
496 (<sse>_andnot<mode>3<mask_name>):
497 Add TARGET_AVX10_1 and change isa attr from avx512dq to
499 (*andnot<mode>3): Add TARGET_AVX10_1 and change isa attr from
500 avx512vl to avx10_1_or_avx512vl.
501 (fix<fixunssuffix>_trunc<mode><sseintvecmodelower>2<mask_name><round_saeonly_name>):
502 Change iterator to VF2_AVX512VLDQ_AVX10_1. Remove target check.
503 (fix_notrunc<mode><sseintvecmodelower>2<mask_name><round_name>):
505 (ufix_notrunc<mode><sseintvecmodelower>2<mask_name><round_name>):
507 (fix<fixunssuffix>_trunc<mode><sselongvecmodelower>2<mask_name><round_saeonly_name>):
508 Change iterator to VF1_128_256VLDQ_AVX10_1. Remove target check.
509 (avx512dq_fix<fixunssuffix>_truncv2sfv2di2<mask_name>):
511 (fix<fixunssuffix>_truncv2sfv2di2): Ditto.
512 (cond_mul<mode>): Change iterator to VI8_AVX10_1_AVX512DQVL.
514 (avx512dq_mul<mode>3<mask_name>): Ditto.
515 (*avx512dq_mul<mode>3<mask_name>): Ditto.
516 (VI4F_BRCST32x2): Add TARGET_AVX512DQ and TARGET_AVX10_1.
517 (<mask_codefor>avx512dq_broadcast<mode><mask_name>):
519 (VI8F_BRCST64x2): Add TARGET_AVX512DQ and TARGET_AVX10_1.
520 (<mask_codefor>avx512dq_broadcast<mode><mask_name>_1):
522 * config/i386/subst.md (mask_mode512bit_condition): Add TARGET_AVX10_1.
523 (mask_avx512vl_condition): Ditto.
526 2023-08-24 Haochen Jiang <haochen.jiang@intel.com>
529 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
531 * config/i386/avx512vldqintrin.h: Remove target attribute.
532 * config/i386/i386-builtin.def (BDESC):
533 Add OPTION_MASK_ISA2_AVX10_1.
534 * config/i386/i386.cc (standard_sse_constant_opcode): Add TARGET_AVX10_1.
535 * config/i386/sse.md: (VI48_AVX512VL_AVX10_1): New.
536 (VI48_AVX512VLDQ_AVX10_1): Ditto.
537 (VF2_AVX512VL): Remove.
538 (VI8_256_512VLDQ_AVX10_1): Rename from VI8_256_512.
540 (*<code><mode>3<mask_name>): Change isa attribute to
541 avx10_1_or_avx512dq. Add TARGET_AVX10_1.
542 (<code><mode>3): Add TARGET_AVX10_1. Change isa attr
543 to avx10_1_or_avx512vl.
544 (<mask_codefor>avx512dq_cvtps2qq<mode><mask_name><round_name>):
545 Change iterator to VI8_256_512VLDQ_AVX10_1. Remove target check.
546 (<mask_codefor>avx512dq_cvtps2qqv2di<mask_name>):
548 (<mask_codefor>avx512dq_cvtps2uqq<mode><mask_name><round_name>):
549 Change iterator to VI8_256_512VLDQ_AVX10_1. Remove target check.
550 (<mask_codefor>avx512dq_cvtps2uqqv2di<mask_name>):
552 (float<floatunssuffix><sseintvecmodelower><mode>2<mask_name><round_name>):
553 Change iterator to VF2_AVX512VLDQ_AVX10_1. Remove target check.
554 (float<floatunssuffix><sselongvecmodelower><mode>2<mask_name><round_name>):
555 Change iterator to VF1_128_256VLDQ_AVX10_1. Remove target check.
556 (float<floatunssuffix>v4div4sf2<mask_name>):
558 (avx512dq_float<floatunssuffix>v2div2sf2): Ditto.
559 (*avx512dq_float<floatunssuffix>v2div2sf2): Ditto.
560 (float<floatunssuffix>v2div2sf2): Ditto.
561 (float<floatunssuffix>v2div2sf2_mask): Ditto.
562 (*float<floatunssuffix>v2div2sf2_mask): Ditto.
563 (*float<floatunssuffix>v2div2sf2_mask_1): Ditto.
564 (<avx512>_cvt<ssemodesuffix>2mask<mode>):
565 Change iterator to VI48_AVX512VLDQ_AVX10_1. Remove target check.
566 (<avx512>_cvtmask2<ssemodesuffix><mode>): Ditto.
567 (*<avx512>_cvtmask2<ssemodesuffix><mode>):
568 Change iterator to VI48_AVX512VL_AVX10_1. Remove target check.
569 Change when constraint is enabled.
571 2023-08-24 Haochen Jiang <haochen.jiang@intel.com>
574 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
576 * config/i386/avx512vldqintrin.h: Remove target attribute.
577 * config/i386/i386-builtin.def (BDESC):
578 Add OPTION_MASK_ISA2_AVX10_1.
579 * config/i386/sse.md (VF_AVX512VLDQ_AVX10_1): New.
580 (VFH_AVX512VLDQ_AVX10_1): Ditto.
581 (VF1_AVX512VLDQ_AVX10_1): Ditto.
582 (<mask_codefor>reducep<mode><mask_name><round_saeonly_name>):
583 Change iterator to VFH_AVX512VLDQ_AVX10_1. Remove target check.
584 (vec_pack<floatprefix>_float_<mode>): Change iterator to
585 VI8_AVX512VLDQ_AVX10_1. Remove target check.
586 (vec_unpack_<fixprefix>fix_trunc_lo_<mode>): Change iterator to
587 VF1_AVX512VLDQ_AVX10_1. Remove target check.
588 (vec_unpack_<fixprefix>fix_trunc_hi_<mode>): Ditto.
589 (VI48F_256_DQVL_AVX10_1): Rename from VI48F_256_DQ.
590 (avx512vl_vextractf128<mode>): Change iterator to
591 VI48F_256_DQVL_AVX10_1. Remove target check.
592 (vec_extract_hi_<mode>_mask): Add TARGET_AVX10_1.
593 (vec_extract_hi_<mode>): Ditto.
594 (avx512vl_vinsert<mode>): Ditto.
595 (vec_set_lo_<mode><mask_name>): Ditto.
596 (vec_set_hi_<mode><mask_name>): Ditto.
597 (avx512dq_rangep<mode><mask_name><round_saeonly_name>): Change
598 iterator to VF_AVX512VLDQ_AVX10_1. Remove target check.
599 (avx512dq_fpclass<mode><mask_scalar_merge_name>): Change
600 iterator to VFH_AVX512VLDQ_AVX10_1. Remove target check.
601 * config/i386/subst.md (mask_avx512dq_condition): Add
603 (mask_scalar_merge): Ditto.
605 2023-08-24 Haochen Jiang <haochen.jiang@intel.com>
608 2023-08-18 Haochen Jiang <haochen.jiang@intel.com>
611 * config/i386/avx512vldqintrin.h: Push AVX2 when AVX2 is
614 2023-08-24 Richard Biener <rguenther@suse.de>
617 * dwarf2out.cc (prune_unused_types_walk): Handle
618 DW_TAG_restrict_type, DW_TAG_shared_type, DW_TAG_atomic_type,
619 DW_TAG_immutable_type, DW_TAG_coarray_type, DW_TAG_unspecified_type
620 and DW_TAG_dynamic_type as to only output them when referenced.
622 2023-08-24 liuhongt <hongtao.liu@intel.com>
624 * config/i386/i386.cc (ix86_invalid_conversion): Adjust GCC
627 2023-08-24 liuhongt <hongtao.liu@intel.com>
629 * common/config/i386/i386-common.cc (processor_names): Add new
630 member graniterapids-s and arrowlake-s.
631 * config/i386/i386-options.cc (processor_alias_table): Update
632 table with PROCESSOR_ARROWLAKE_S and
633 PROCESSOR_GRANITERAPIDS_D.
634 (m_GRANITERAPID_D): New macro.
635 (m_ARROWLAKE_S): Ditto.
636 (m_CORE_AVX512): Add m_GRANITERAPIDS_D.
637 (processor_cost_table): Add icelake_cost for
638 PROCESSOR_GRANITERAPIDS_D and alderlake_cost for
639 PROCESSOR_ARROWLAKE_S.
640 * config/i386/x86-tune.def: Hanlde m_ARROWLAKE_S same as
642 * config/i386/i386.h (enum processor_type): Add new member
643 PROCESSOR_GRANITERAPIDS_D and PROCESSOR_ARROWLAKE_S.
644 * config/i386/i386-c.cc (ix86_target_macros_internal): Handle
645 PROCESSOR_GRANITERAPIDS_D and PROCESSOR_ARROWLAKE_S
647 2023-08-23 Jivan Hakobyan <jivanhakobyan9@gmail.com>
649 * lra-eliminations.cc (eliminate_regs_in_insn): Use equivalences to
650 to help simplify code further.
652 2023-08-23 Andrew MacLeod <amacleod@redhat.com>
654 * gimple-range-fold.cc (fold_using_range::range_of_phi): Tweak output.
655 * gimple-range-phi.cc (phi_group::phi_group): Remove unused members.
656 Initialize using a range instead of value and edge.
657 (phi_group::calculate_using_modifier): Use initializer value and
658 process for relations after trying for iteration convergence.
659 (phi_group::refine_using_relation): Use initializer range.
660 (phi_group::dump): Rework the dump output.
661 (phi_analyzer::process_phi): Allow multiple constant initilizers.
662 Dump groups immediately as created.
663 (phi_analyzer::dump): Tweak output.
664 * gimple-range-phi.h (phi_group::phi_group): Adjust prototype.
665 (phi_group::initial_value): Delete.
666 (phi_group::refine_using_relation): Adjust prototype.
667 (phi_group::m_initial_value): Delete.
668 (phi_group::m_initial_edge): Delete.
669 (phi_group::m_vr): Use int_range_max.
670 * tree-vrp.cc (execute_ranger_vrp): Don't dump phi groups.
672 2023-08-23 Andrew MacLeod <amacleod@redhat.com>
674 * gimple-range-phi.cc (phi_analyzer::operator[]): Return NULL if
675 no group was created.
676 (phi_analyzer::process_phi): Do not create groups of one phi node.
678 2023-08-23 Richard Earnshaw <rearnsha@arm.com>
680 * target.def (gen_ccmp_first, gen_ccmp_next): Use rtx_code for
681 CODE, CMP_CODE and BIT_CODE arguments.
682 * config/aarch64/aarch64.cc (aarch64_gen_ccmp_first): Likewise.
683 (aarch64_gen_ccmp_next): Likewise.
684 * doc/tm.texi: Regenerated.
686 2023-08-23 Richard Earnshaw <rearnsha@arm.com>
688 * coretypes.h (rtx_code): Add forward declaration.
689 * rtl.h (rtx_code): Make compatible with forward declaration.
691 2023-08-23 Uros Bizjak <ubizjak@gmail.com>
694 * config/i386/i386.md (*concat<any_or_plus:mode><dwi>3_3):
695 Merge pattern from *concatditi3_3 and *concatsidi3_3 using
696 DWIH mode iterator. Disable (=&r,m,m) alternative for
698 (*concat<any_or_plus:mode><dwi>3_3): Disable (=&r,m,m)
699 alternative for 32-bit targets.
701 2023-08-23 Zhangjin Liao <liaozhangjin@eswincomputing.com>
703 * config/riscv/bitmanip.md (*<bitmanip_optab>disi2_sext): Add a more
704 appropriate type attribute.
706 2023-08-23 Lehua Ding <lehua.ding@rivai.ai>
708 * config/riscv/autovec-opt.md (*cond_abs<mode>): New combine pattern.
709 (*copysign<mode>_neg): Ditto.
710 * config/riscv/autovec.md (@vcond_mask_<mode><vm>): Adjust.
711 (<optab><mode>2): Ditto.
712 (cond_<optab><mode>): New.
713 (cond_len_<optab><mode>): Ditto.
714 * config/riscv/riscv-protos.h (enum insn_type): New.
715 (expand_cond_len_unop): New helper func.
716 * config/riscv/riscv-v.cc (shuffle_merge_patterns): Adjust.
717 (expand_cond_len_unop): New helper func.
719 2023-08-23 Jan Hubicka <jh@suse.cz>
721 * tree-ssa-loop-ch.cc (enum ch_decision): Fix comment.
722 (should_duplicate_loop_header_p): Fix return value for static exits.
723 (ch_base::copy_headers): Improve handling of ch_possible_zero_cost.
725 2023-08-23 Kewen Lin <linkw@linux.ibm.com>
727 * tree-vect-stmts.cc (vectorizable_store): Move the handlings on
728 VMAT_GATHER_SCATTER in the final loop nest to its own loop,
729 and update the final nest accordingly.
731 2023-08-23 Kewen Lin <linkw@linux.ibm.com>
733 * tree-vect-stmts.cc (vectorizable_store): Move the handlings on
734 VMAT_LOAD_STORE_LANES in the final loop nest to its own loop,
735 and update the final nest accordingly.
737 2023-08-23 Kewen Lin <linkw@linux.ibm.com>
739 * tree-vect-stmts.cc (vectorizable_store): Remove vec oprnds,
740 adjust vec result_chain, vec_oprnd with auto_vec, and adjust
741 gvec_oprnds with auto_delete_vec.
743 2023-08-23 Juzhe-Zhong <juzhe.zhong@rivai.ai>
745 * config/riscv/riscv-vsetvl.cc
746 (pass_vsetvl::global_eliminate_vsetvl_insn): Fix potential ICE.
748 2023-08-23 Juzhe-Zhong <juzhe.zhong@rivai.ai>
750 * config/riscv/riscv-vsetvl.cc (ge_sew_ratio_unavailable_p):
752 * config/riscv/riscv-vsetvl.def (DEF_SEW_LMUL_FUSE_RULE): Ditto.
754 2023-08-23 Juzhe-Zhong <juzhe.zhong@rivai.ai>
756 * config/riscv/vector.md: Add attribute.
758 2023-08-22 Juzhe-Zhong <juzhe.zhong@rivai.ai>
760 * config/riscv/riscv-vsetvl.cc (change_insn): Clang format.
761 (vector_infos_manager::all_same_ratio_p): Ditto.
762 (vector_infos_manager::all_same_avl_p): Ditto.
763 (pass_vsetvl::refine_vsetvls): Ditto.
764 (pass_vsetvl::cleanup_vsetvls): Ditto.
765 (pass_vsetvl::commit_vsetvls): Ditto.
766 (pass_vsetvl::local_eliminate_vsetvl_insn): Ditto.
767 (pass_vsetvl::global_eliminate_vsetvl_insn): Ditto.
768 (pass_vsetvl::compute_probabilities): Ditto.
770 2023-08-22 Juzhe-Zhong <juzhe.zhong@rivai.ai>
772 * config/riscv/t-riscv: Add riscv-vsetvl.def
774 2023-08-22 Vineet Gupta <vineetg@rivosinc.com>
776 * config/riscv/riscv.opt: Add --param names
777 riscv-autovec-preference and riscv-autovec-lmul
779 2023-08-22 Raphael Moreira Zinsly <rzinsly@ventanamicro.com>
781 * config/riscv/t-linux: Add MULTIARCH_DIRNAME.
783 2023-08-22 Tobias Burnus <tobias@codesourcery.com>
785 * tree-core.h (enum omp_clause_defaultmap_kind): Add
786 OMP_CLAUSE_DEFAULTMAP_CATEGORY_ALL.
787 * gimplify.cc (gimplify_scan_omp_clauses): Handle it.
788 * tree-pretty-print.cc (dump_omp_clause): Likewise.
790 2023-08-22 Jakub Jelinek <jakub@redhat.com>
793 * doc/extend.texi (_Float<n>): Drop obsolete sentence that the
794 types aren't supported in C++.
796 2023-08-22 Juzhe-Zhong <juzhe.zhong@rivai.ai>
798 * doc/md.texi: Add LEN_FOLD_EXTRACT_LAST pattern.
799 * internal-fn.cc (fold_len_extract_direct): Ditto.
800 (expand_fold_len_extract_optab_fn): Ditto.
801 (direct_fold_len_extract_optab_supported_p): Ditto.
802 * internal-fn.def (LEN_FOLD_EXTRACT_LAST): Ditto.
803 * optabs.def (OPTAB_D): Ditto.
805 2023-08-22 Richard Biener <rguenther@suse.de>
807 * tree-vect-stmts.cc (vectorizable_store): Do not bump
808 DR_GROUP_STORE_COUNT here. Remove early out.
809 (vect_transform_stmt): Only call vectorizable_store on
810 the last element of an interleaving chain.
812 2023-08-22 Richard Biener <rguenther@suse.de>
814 PR tree-optimization/94864
815 PR tree-optimization/94865
816 PR tree-optimization/93080
817 * match.pd (bit_insert @0 (BIT_FIELD_REF @1 ..) ..): New pattern
818 for vector insertion from vector extraction.
820 2023-08-22 Juzhe-Zhong <juzhe.zhong@rivai.ai>
821 Kewen.Lin <linkw@linux.ibm.com>
823 * tree-vect-loop.cc (vect_verify_loop_lens): Add exists check.
824 (vectorizable_live_operation): Add live vectorization for length loop
827 2023-08-22 David Malcolm <dmalcolm@redhat.com>
830 * doc/invoke.texi: Remove -Wanalyzer-unterminated-string.
832 2023-08-22 Pan Li <pan2.li@intel.com>
834 * config/riscv/riscv-vector-builtins-bases.cc
835 (vfwredusum_frm_obj): New declaration.
837 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
838 * config/riscv/riscv-vector-builtins-functions.def
839 (vfwredusum_frm): New intrinsic function def.
841 2023-08-21 David Faust <david.faust@oracle.com>
843 * config/bpf/bpf.md (neg): Second operand must be a register.
845 2023-08-21 Edwin Lu <ewlu@rivosinc.com>
847 * config/riscv/bitmanip.md: Added bitmanip type to insns
848 that are missing types.
850 2023-08-21 Jeff Law <jlaw@ventanamicro.com>
852 * config/riscv/sync-ztso.md (atomic_load_ztso<mode>): Avoid extraenous
855 2023-08-21 Francois-Xavier Coudert <fxcoudert@gcc.gnu.org>
857 * config/aarch64/falkor-tag-collision-avoidance.cc (dump_insn_list):
858 Fix format specifier.
860 2023-08-21 Aldy Hernandez <aldyh@redhat.com>
862 * value-range.cc (frange::union_nans): Return false if nothing
864 (range_tests_floats): New test.
866 2023-08-21 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
868 PR tree-optimization/111048
869 * fold-const.cc (valid_mask_for_fold_vec_perm_cst_p): Set arg_npatterns
871 (fold_vec_perm_cst): Remove workaround and again call
872 valid_mask_fold_vec_perm_cst_p for both VLS and VLA vectors.
873 (test_fold_vec_perm_cst::test_nunits_min_4): Add test-case.
875 2023-08-21 Richard Biener <rguenther@suse.de>
877 PR tree-optimization/111082
878 * tree-vect-slp.cc (vectorize_slp_instance_root_stmt): Only
879 pun operations that can overflow.
881 2023-08-21 Juzhe-Zhong <juzhe.zhong@rivai.ai>
883 * lcm.cc (compute_antinout_edge): Export as global use.
884 (compute_earliest): Ditto.
885 (compute_rev_insert_delete): Ditto.
886 * lcm.h (compute_antinout_edge): Ditto.
887 (compute_earliest): Ditto.
889 2023-08-21 Richard Biener <rguenther@suse.de>
891 PR tree-optimization/111070
892 * tree-ssa-ifcombine.cc (ifcombine_ifandif): Check we have
893 an SSA name before checking SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
895 2023-08-21 Andrew Pinski <apinski@marvell.com>
897 PR tree-optimization/111002
898 * match.pd (view_convert(vec_cond(a,b,c))): New pattern.
900 2023-08-21 liuhongt <hongtao.liu@intel.com>
902 * common/config/i386/cpuinfo.h (get_intel_cpu): Detect
904 * common/config/i386/i386-common.cc (alias_table): Support
905 -march=gracemont as an alias of -march=alderlake.
907 2023-08-20 Uros Bizjak <ubizjak@gmail.com>
909 * config/i386/i386-expand.cc (ix86_expand_sse_extend): Use ops[1]
910 instead of src in the call to ix86_expand_sse_cmp.
911 * config/i386/sse.md (<any_extend:insn>v8qiv8hi2): Do not
912 force operands[1] to a register.
913 (<any_extend:insn>v4hiv4si2): Ditto.
914 (<any_extend:insn>v2siv2di2): Ditto.
916 2023-08-20 Andrew Pinski <apinski@marvell.com>
918 PR tree-optimization/111006
919 PR tree-optimization/110986
920 * match.pd: (op(vec_cond(a,b,c))): Handle convert for op.
922 2023-08-20 Eric Gallager <egallager@gcc.gnu.org>
925 * Makefile.in: improve error message when /usr/include is
928 2023-08-19 Tobias Burnus <tobias@codesourcery.com>
931 * omp-expand.cc (expand_omp_for_init_vars): Pass after=true
932 to expand_omp_build_cond for 'factor != 0' condition, resulting
933 in pre-r12-5295-g47de0b56ee455e code for the gimple insert.
935 2023-08-19 Guo Jie <guojie@loongson.cn>
936 Lulu Cheng <chenglulu@loongson.cn>
938 * config/loongarch/t-loongarch: Add loongarch-driver.h into
939 TM_H. Add loongarch-def.h and loongarch-tune.h into
942 2023-08-18 Uros Bizjak <ubizjak@gmail.com>
945 * config/i386/i386-expand.cc (ix86_split_mmx_punpck):
946 Also handle V2QImode.
947 (ix86_expand_sse_extend): New function.
948 * config/i386/i386-protos.h (ix86_expand_sse_extend): New prototype.
949 * config/i386/mmx.md (<any_extend:insn>v4qiv4hi2): Enable for
950 TARGET_SSE2. Expand through ix86_expand_sse_extend for !TARGET_SSE4_1.
951 (<any_extend:insn>v2hiv2si2): Ditto.
952 (<any_extend:insn>v2qiv2hi2): Ditto.
953 * config/i386/sse.md (<any_extend:insn>v8qiv8hi2): Ditto.
954 (<any_extend:insn>v4hiv4si2): Ditto.
955 (<any_extend:insn>v2siv2di2): Ditto.
957 2023-08-18 Aldy Hernandez <aldyh@redhat.com>
960 * value-range.cc (irange::union_bitmask): Return FALSE if updated
961 bitmask is semantically equivalent to the original mask.
962 (irange::intersect_bitmask): Same.
963 (irange::get_bitmask): Add comment.
965 2023-08-18 Richard Biener <rguenther@suse.de>
967 PR tree-optimization/111019
968 * tree-ssa-loop-im.cc (gather_mem_refs_stmt): When canonicalizing
969 also scrap base and offset in case the ref is indirect.
971 2023-08-18 Jose E. Marchesi <jose.marchesi@oracle.com>
973 * config/bpf/bpf.opt (mframe-limit): Set default to 32767.
975 2023-08-18 Kewen Lin <linkw@linux.ibm.com>
978 * Makefile.in (TM_P_H): Add $(TREE_H) as dependence.
980 2023-08-18 Kewen Lin <linkw@linux.ibm.com>
982 * tree-vect-stmts.cc (vect_build_scatter_store_calls): New, factor
984 (vectorizable_store): ... here.
986 2023-08-18 Richard Biener <rguenther@suse.de>
988 PR tree-optimization/111048
989 * fold-const.cc (fold_vec_perm_cst): Check for non-VLA
992 2023-08-18 Haochen Jiang <haochen.jiang@intel.com>
995 * config/i386/avx512vldqintrin.h: Push AVX2 when AVX2 is
998 2023-08-18 Kewen Lin <linkw@linux.ibm.com>
1000 * tree-vect-stmts.cc (vectorizable_load): Move the handlings on
1001 VMAT_GATHER_SCATTER in the final loop nest to its own loop,
1002 and update the final nest accordingly.
1004 2023-08-18 Andrew Pinski <apinski@marvell.com>
1006 * doc/md.texi (Standard patterns): Document cond_neg, cond_one_cmpl,
1007 cond_len_neg and cond_len_one_cmpl.
1009 2023-08-18 Lehua Ding <lehua.ding@rivai.ai>
1011 * config/riscv/iterators.md (TARGET_HARD_FLOAT || TARGET_ZFINX): New.
1012 * config/riscv/pic.md (*local_pic_load<ANYF:mode>): Change ANYF.
1013 (*local_pic_load<ANYLSF:mode>): To ANYLSF.
1014 (*local_pic_load_32d<ANYF:mode>): Ditto.
1015 (*local_pic_load_32d<ANYLSF:mode>): Ditto.
1016 (*local_pic_store<ANYF:mode>): Ditto.
1017 (*local_pic_store<ANYLSF:mode>): Ditto.
1018 (*local_pic_store_32d<ANYF:mode>): Ditto.
1019 (*local_pic_store_32d<ANYLSF:mode>): Ditto.
1021 2023-08-18 Lehua Ding <lehua.ding@rivai.ai>
1022 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
1024 * config/riscv/predicates.md (vector_const_0_operand): New.
1025 * config/riscv/vector.md (*pred_broadcast<mode>_zero): Ditto.
1027 2023-08-18 Lehua Ding <lehua.ding@rivai.ai>
1029 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::backward_demand_fusion):
1032 2023-08-17 Andrew MacLeod <amacleod@redhat.com>
1034 PR tree-optimization/111009
1035 * range-op.cc (operator_addr_expr::op1_range): Be more restrictive.
1037 2023-08-17 Vladimir N. Makarov <vmakarov@redhat.com>
1039 * lra-spills.cc (assign_stack_slot_num_and_sort_pseudos): Moving
1040 slots_num initialization from here ...
1041 (lra_spill): ... to here before the 1st call of
1042 assign_stack_slot_num_and_sort_pseudos. Add the 2nd call after
1045 2023-08-17 Jose E. Marchesi <jose.marchesi@oracle.com>
1048 * doc/invoke.texi (Option Summary): Mention
1049 -Wcompare-distinct-pointer-types under `Warning Options'.
1050 (Warning Options): Document -Wcompare-distinct-pointer-types.
1052 2023-08-17 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1054 * recog.cc (memory_address_addr_space_p): Mark possibly unused
1057 2023-08-17 Richard Biener <rguenther@suse.de>
1059 PR tree-optimization/111039
1060 * tree-ssa-ifcombine.cc (ifcombine_ifandif): Check for
1061 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
1063 2023-08-17 Alex Coplan <alex.coplan@arm.com>
1065 * doc/rtl.texi: Fix up sample code for RTL-SSA insn changes.
1067 2023-08-17 Jose E. Marchesi <jose.marchesi@oracle.com>
1070 * config/bpf/bpf.cc (bpf_attribute_table): Add entry for the
1071 `naked' function attribute.
1072 (bpf_warn_func_return): New function.
1073 (TARGET_WARN_FUNC_RETURN): Define.
1074 (bpf_expand_prologue): Add preventive comment.
1075 (bpf_expand_epilogue): Likewise.
1076 * doc/extend.texi (BPF Function Attributes): Document the `naked'
1079 2023-08-17 Richard Biener <rguenther@suse.de>
1081 * tree-vect-slp.cc (vect_slp_check_for_roots): Use
1082 !needs_fold_left_reduction_p to decide whether we can
1083 handle the reduction with association.
1084 (vectorize_slp_instance_root_stmt): For TYPE_OVERFLOW_UNDEFINED
1085 reductions perform all arithmetic in an unsigned type.
1087 2023-08-17 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1089 * configure.ac (gcc_cv_ld64_version): Allow for dyld in ld -v
1091 * configure: Regenerate.
1093 2023-08-17 Pan Li <pan2.li@intel.com>
1095 * config/riscv/riscv-vector-builtins-bases.cc
1096 (widen_freducop): Add frm_opt_type template arg.
1097 (vfwredosum_frm_obj): New declaration.
1099 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1100 * config/riscv/riscv-vector-builtins-functions.def
1101 (vfwredosum_frm): New intrinsic function def.
1103 2023-08-17 Pan Li <pan2.li@intel.com>
1105 * config/riscv/riscv-vector-builtins-bases.cc
1106 (vfredosum_frm_obj): New declaration.
1108 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1109 * config/riscv/riscv-vector-builtins-functions.def
1110 (vfredosum_frm): New intrinsic function def.
1112 2023-08-17 Pan Li <pan2.li@intel.com>
1114 * config/riscv/riscv-vector-builtins-bases.cc
1115 (class freducop): Add frm_op_type template arg.
1116 (vfredusum_frm_obj): New declaration.
1118 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1119 * config/riscv/riscv-vector-builtins-functions.def
1120 (vfredusum_frm): New intrinsic function def.
1121 * config/riscv/riscv-vector-builtins-shapes.cc
1122 (struct reduc_alu_frm_def): New class for frm shape.
1123 (SHAPE): New declaration.
1124 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
1126 2023-08-17 Pan Li <pan2.li@intel.com>
1128 * config/riscv/riscv-vector-builtins-bases.cc
1129 (class vfncvt_f): Add frm_op_type template arg.
1130 (vfncvt_f_frm_obj): New declaration.
1132 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1133 * config/riscv/riscv-vector-builtins-functions.def
1134 (vfncvt_f_frm): New intrinsic function def.
1136 2023-08-17 Pan Li <pan2.li@intel.com>
1138 * config/riscv/riscv-vector-builtins-bases.cc
1139 (vfncvt_xu_frm_obj): New declaration.
1141 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1142 * config/riscv/riscv-vector-builtins-functions.def
1143 (vfncvt_xu_frm): New intrinsic function def.
1145 2023-08-17 Pan Li <pan2.li@intel.com>
1147 * config/riscv/riscv-vector-builtins-bases.cc
1148 (class vfncvt_x): Add frm_op_type template arg.
1149 (BASE): New declaration.
1150 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1151 * config/riscv/riscv-vector-builtins-functions.def
1152 (vfncvt_x_frm): New intrinsic function def.
1153 * config/riscv/riscv-vector-builtins-shapes.cc
1154 (struct narrow_alu_frm_def): New shape function for frm.
1155 (SHAPE): New declaration.
1156 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
1158 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
1160 * config/i386/avx512vldqintrin.h: Remove target attribute.
1161 * config/i386/i386-builtin.def (BDESC):
1162 Add OPTION_MASK_ISA2_AVX10_1.
1163 * config/i386/sse.md (VF_AVX512VLDQ_AVX10_1): New.
1164 (VFH_AVX512VLDQ_AVX10_1): Ditto.
1165 (VF1_AVX512VLDQ_AVX10_1): Ditto.
1166 (<mask_codefor>reducep<mode><mask_name><round_saeonly_name>):
1167 Change iterator to VFH_AVX512VLDQ_AVX10_1. Remove target check.
1168 (vec_pack<floatprefix>_float_<mode>): Change iterator to
1169 VI8_AVX512VLDQ_AVX10_1. Remove target check.
1170 (vec_unpack_<fixprefix>fix_trunc_lo_<mode>): Change iterator to
1171 VF1_AVX512VLDQ_AVX10_1. Remove target check.
1172 (vec_unpack_<fixprefix>fix_trunc_hi_<mode>): Ditto.
1173 (VI48F_256_DQVL_AVX10_1): Rename from VI48F_256_DQ.
1174 (avx512vl_vextractf128<mode>): Change iterator to
1175 VI48F_256_DQVL_AVX10_1. Remove target check.
1176 (vec_extract_hi_<mode>_mask): Add TARGET_AVX10_1.
1177 (vec_extract_hi_<mode>): Ditto.
1178 (avx512vl_vinsert<mode>): Ditto.
1179 (vec_set_lo_<mode><mask_name>): Ditto.
1180 (vec_set_hi_<mode><mask_name>): Ditto.
1181 (avx512dq_rangep<mode><mask_name><round_saeonly_name>): Change
1182 iterator to VF_AVX512VLDQ_AVX10_1. Remove target check.
1183 (avx512dq_fpclass<mode><mask_scalar_merge_name>): Change
1184 iterator to VFH_AVX512VLDQ_AVX10_1. Remove target check.
1185 * config/i386/subst.md (mask_avx512dq_condition): Add
1187 (mask_scalar_merge): Ditto.
1189 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
1191 * config/i386/avx512vldqintrin.h: Remove target attribute.
1192 * config/i386/i386-builtin.def (BDESC):
1193 Add OPTION_MASK_ISA2_AVX10_1.
1194 * config/i386/i386.cc (standard_sse_constant_opcode): Add TARGET_AVX10_1.
1195 * config/i386/sse.md: (VI48_AVX512VL_AVX10_1): New.
1196 (VI48_AVX512VLDQ_AVX10_1): Ditto.
1197 (VF2_AVX512VL): Remove.
1198 (VI8_256_512VLDQ_AVX10_1): Rename from VI8_256_512.
1200 (*<code><mode>3<mask_name>): Change isa attribute to
1201 avx10_1_or_avx512dq. Add TARGET_AVX10_1.
1202 (<code><mode>3): Add TARGET_AVX10_1. Change isa attr
1203 to avx10_1_or_avx512vl.
1204 (<mask_codefor>avx512dq_cvtps2qq<mode><mask_name><round_name>):
1205 Change iterator to VI8_256_512VLDQ_AVX10_1. Remove target check.
1206 (<mask_codefor>avx512dq_cvtps2qqv2di<mask_name>):
1208 (<mask_codefor>avx512dq_cvtps2uqq<mode><mask_name><round_name>):
1209 Change iterator to VI8_256_512VLDQ_AVX10_1. Remove target check.
1210 (<mask_codefor>avx512dq_cvtps2uqqv2di<mask_name>):
1212 (float<floatunssuffix><sseintvecmodelower><mode>2<mask_name><round_name>):
1213 Change iterator to VF2_AVX512VLDQ_AVX10_1. Remove target check.
1214 (float<floatunssuffix><sselongvecmodelower><mode>2<mask_name><round_name>):
1215 Change iterator to VF1_128_256VLDQ_AVX10_1. Remove target check.
1216 (float<floatunssuffix>v4div4sf2<mask_name>):
1218 (avx512dq_float<floatunssuffix>v2div2sf2): Ditto.
1219 (*avx512dq_float<floatunssuffix>v2div2sf2): Ditto.
1220 (float<floatunssuffix>v2div2sf2): Ditto.
1221 (float<floatunssuffix>v2div2sf2_mask): Ditto.
1222 (*float<floatunssuffix>v2div2sf2_mask): Ditto.
1223 (*float<floatunssuffix>v2div2sf2_mask_1): Ditto.
1224 (<avx512>_cvt<ssemodesuffix>2mask<mode>):
1225 Change iterator to VI48_AVX512VLDQ_AVX10_1. Remove target check.
1226 (<avx512>_cvtmask2<ssemodesuffix><mode>): Ditto.
1227 (*<avx512>_cvtmask2<ssemodesuffix><mode>):
1228 Change iterator to VI48_AVX512VL_AVX10_1. Remove target check.
1229 Change when constraint is enabled.
1231 2023-08-17 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1234 * config/riscv/riscv-vsetvl.cc (float_insn_valid_sew_p): New function.
1235 (second_sew_less_than_first_sew_p): Fix bug.
1236 (first_sew_less_than_second_sew_p): Ditto.
1238 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
1240 * config/i386/avx512vldqintrin.h: Remove target attribute.
1241 * config/i386/i386-builtin.def (BDESC):
1242 Add OPTION_MASK_ISA2_AVX10_1.
1243 * config/i386/i386-builtins.cc (def_builtin): Handle AVX10_1.
1244 * config/i386/i386-expand.cc
1245 (ix86_check_builtin_isa_match): Ditto.
1246 (ix86_expand_sse2_mulvxdi3): Add TARGET_AVX10_1.
1247 * config/i386/i386.md: Add new isa attribute avx10_1_or_avx512dq
1248 and avx10_1_or_avx512vl.
1249 * config/i386/sse.md: (VF2_AVX512VLDQ_AVX10_1): New.
1250 (VF1_128_256VLDQ_AVX10_1): Ditto.
1251 (VI8_AVX512VLDQ_AVX10_1): Ditto.
1252 (<sse>_andnot<mode>3<mask_name>):
1253 Add TARGET_AVX10_1 and change isa attr from avx512dq to
1254 avx10_1_or_avx512dq.
1255 (*andnot<mode>3): Add TARGET_AVX10_1 and change isa attr from
1256 avx512vl to avx10_1_or_avx512vl.
1257 (fix<fixunssuffix>_trunc<mode><sseintvecmodelower>2<mask_name><round_saeonly_name>):
1258 Change iterator to VF2_AVX512VLDQ_AVX10_1. Remove target check.
1259 (fix_notrunc<mode><sseintvecmodelower>2<mask_name><round_name>):
1261 (ufix_notrunc<mode><sseintvecmodelower>2<mask_name><round_name>):
1263 (fix<fixunssuffix>_trunc<mode><sselongvecmodelower>2<mask_name><round_saeonly_name>):
1264 Change iterator to VF1_128_256VLDQ_AVX10_1. Remove target check.
1265 (avx512dq_fix<fixunssuffix>_truncv2sfv2di2<mask_name>):
1267 (fix<fixunssuffix>_truncv2sfv2di2): Ditto.
1268 (cond_mul<mode>): Change iterator to VI8_AVX10_1_AVX512DQVL.
1269 Remove target check.
1270 (avx512dq_mul<mode>3<mask_name>): Ditto.
1271 (*avx512dq_mul<mode>3<mask_name>): Ditto.
1272 (VI4F_BRCST32x2): Add TARGET_AVX512DQ and TARGET_AVX10_1.
1273 (<mask_codefor>avx512dq_broadcast<mode><mask_name>):
1274 Remove target check.
1275 (VI8F_BRCST64x2): Add TARGET_AVX512DQ and TARGET_AVX10_1.
1276 (<mask_codefor>avx512dq_broadcast<mode><mask_name>_1):
1277 Remove target check.
1278 * config/i386/subst.md (mask_mode512bit_condition): Add TARGET_AVX10_1.
1279 (mask_avx512vl_condition): Ditto.
1282 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
1284 * common/config/i386/i386-common.cc
1285 (ix86_check_avx10_vector_width): New function to check isa_flags
1286 to emit a warning when there is a conflict in AVX10 options for
1288 (ix86_handle_option): Add check for avx10.1-256 and avx10.1-512.
1289 * config/i386/driver-i386.cc (host_detect_local_cpu):
1290 Do not append -mno-avx10-max-512bit for -march=native.
1292 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
1294 * common/config/i386/i386-common.cc
1295 (ix86_check_avx10): New function to check isa_flags and
1296 isa_flags_explicit to emit warning when AVX10 is enabled
1298 (ix86_check_avx512): New function to check isa_flags and
1299 isa_flags_explicit to emit warning when AVX512 is enabled
1301 (ix86_handle_option): Do not change the flags when warning
1303 * config/i386/driver-i386.cc (host_detect_local_cpu):
1304 Do not append -mno-avx10.1 for -march=native.
1306 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
1308 * common/config/i386/cpuinfo.h (get_available_features):
1309 Add avx10_set and version and detect avx10.1.
1310 (cpu_indicator_init): Handle avx10.1-512.
1311 * common/config/i386/i386-common.cc
1312 (OPTION_MASK_ISA2_AVX10_512BIT_SET): New.
1313 (OPTION_MASK_ISA2_AVX10_1_SET): Ditto.
1314 (OPTION_MASK_ISA2_AVX10_512BIT_UNSET): Ditto.
1315 (OPTION_MASK_ISA2_AVX10_1_UNSET): Ditto.
1316 (OPTION_MASK_ISA2_AVX2_UNSET): Modify for AVX10_1.
1317 (ix86_handle_option): Handle -mavx10.1, -mavx10.1-256 and
1319 * common/config/i386/i386-cpuinfo.h (enum processor_features):
1320 Add FEATURE_AVX10_512BIT, FEATURE_AVX10_1 and
1321 FEATURE_AVX10_512BIT.
1322 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
1323 AVX10_512BIT, AVX10_1 and AVX10_1_512.
1324 * config/i386/constraints.md (Yk): Add AVX10_1.
1327 * config/i386/cpuid.h (bit_AVX10): New.
1328 (bit_AVX10_256): Ditto.
1329 (bit_AVX10_512): Ditto.
1330 * config/i386/i386-c.cc (ix86_target_macros_internal):
1331 Define AVX10_512BIT and AVX10_1.
1332 * config/i386/i386-isa.def
1333 (AVX10_512BIT): Add DEF_PTA(AVX10_512BIT).
1334 (AVX10_1): Add DEF_PTA(AVX10_1).
1335 * config/i386/i386-options.cc (isa2_opts): Add -mavx10.1.
1336 (ix86_valid_target_attribute_inner_p): Handle avx10-512bit, avx10.1
1338 (ix86_option_override_internal): Enable AVX512{F,VL,BW,DQ,CD,BF16,
1339 FP16,VBMI,VBMI2,VNNI,IFMA,BITALG,VPOPCNTDQ} features for avx10.1-512.
1340 (ix86_valid_target_attribute_inner_p): Handle AVX10_1.
1341 * config/i386/i386.cc (ix86_get_ssemov): Add AVX10_1.
1342 (ix86_conditional_register_usage): Ditto.
1343 (ix86_hard_regno_mode_ok): Ditto.
1344 (ix86_rtx_costs): Ditto.
1345 * config/i386/i386.h (VALID_MASK_AVX10_MODE): New macro.
1346 * config/i386/i386.opt: Add option -mavx10.1, -mavx10.1-256 and
1348 * doc/extend.texi: Document avx10.1, avx10.1-256 and avx10.1-512.
1349 * doc/invoke.texi: Document -mavx10.1, -mavx10.1-256 and -mavx10.1-512.
1350 * doc/sourcebuild.texi: Document target avx10.1, avx10.1-256
1353 2023-08-17 Sergei Trofimovich <siarheit@google.com>
1355 * flag-types.h (vrp_mode): Remove unused.
1357 2023-08-17 Yanzhang Wang <yanzhang.wang@intel.com>
1359 * simplify-rtx.cc (simplify_context::simplify_binary_operation_1): Use
1362 2023-08-17 Andrew Pinski <apinski@marvell.com>
1364 * internal-fn.def (COND_NOT): New internal function.
1365 * match.pd (UNCOND_UNARY, COND_UNARY): Add bit_not/not
1367 (`vec (a ? -1 : 0) ^ b`): New pattern to convert
1368 into conditional not.
1369 * optabs.def (cond_one_cmpl): New optab.
1370 (cond_len_one_cmpl): Likewise.
1372 2023-08-16 Surya Kumari Jangala <jskumari@linux.ibm.com>
1374 PR rtl-optimization/110254
1375 * ira-color.cc (improve_allocation): Update array
1376 allocated_hard_reg_p.
1378 2023-08-16 Vladimir N. Makarov <vmakarov@redhat.com>
1380 * lra-int.h (lra_update_fp2sp_elimination): Change the prototype.
1381 * lra-eliminations.cc (spill_pseudos): Record spilled pseudos.
1382 (lra_update_fp2sp_elimination): Ditto.
1383 (update_reg_eliminate): Adjust spill_pseudos call.
1384 * lra-spills.cc (lra_spill): Assign stack slots to pseudos spilled
1385 in lra_update_fp2sp_elimination.
1387 2023-08-16 Richard Ball <richard.ball@arm.com>
1389 * config/aarch64/aarch64-cores.def (AARCH64_CORE): Add Cortex-A720 CPU.
1390 * config/aarch64/aarch64-tune.md: Regenerate.
1391 * doc/invoke.texi: Document Cortex-A720 CPU.
1393 2023-08-16 Robin Dapp <rdapp@ventanamicro.com>
1395 * config/riscv/autovec.md (<u>avg<v_double_trunc>3_floor):
1397 (<u>avg<v_double_trunc>3_ceil): Ditto.
1398 * config/riscv/vector-iterators.md (ashiftrt): New iterator.
1401 2023-08-16 Robin Dapp <rdapp@ventanamicro.com>
1403 * internal-fn.cc (vec_extract_direct): Change type argument
1405 (expand_vec_extract_optab_fn): Call convert_optab_fn.
1406 (direct_vec_extract_optab_supported_p): Use
1407 convert_optab_supported_p.
1409 2023-08-16 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
1410 Richard Sandiford <richard.sandiford@arm.com>
1412 * fold-const.cc (INCLUDE_ALGORITHM): Add Include.
1413 (valid_mask_for_fold_vec_perm_cst_p): New function.
1414 (fold_vec_perm_cst): Likewise.
1415 (fold_vec_perm): Adjust assert and call fold_vec_perm_cst.
1416 (test_fold_vec_perm_cst): New namespace.
1417 (test_fold_vec_perm_cst::build_vec_cst_rand): New function.
1418 (test_fold_vec_perm_cst::validate_res): Likewise.
1419 (test_fold_vec_perm_cst::validate_res_vls): Likewise.
1420 (test_fold_vec_perm_cst::builder_push_elems): Likewise.
1421 (test_fold_vec_perm_cst::test_vnx4si_v4si): Likewise.
1422 (test_fold_vec_perm_cst::test_v4si_vnx4si): Likewise.
1423 (test_fold_vec_perm_cst::test_all_nunits): Likewise.
1424 (test_fold_vec_perm_cst::test_nunits_min_2): Likewise.
1425 (test_fold_vec_perm_cst::test_nunits_min_4): Likewise.
1426 (test_fold_vec_perm_cst::test_nunits_min_8): Likewise.
1427 (test_fold_vec_perm_cst::test_nunits_max_4): Likewise.
1428 (test_fold_vec_perm_cst::is_simple_vla_size): Likewise.
1429 (test_fold_vec_perm_cst::test): Likewise.
1430 (fold_const_cc_tests): Call test_fold_vec_perm_cst::test.
1432 2023-08-16 Pan Li <pan2.li@intel.com>
1434 * config/riscv/riscv-vector-builtins-bases.cc
1435 (BASE): New declaration.
1436 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1437 * config/riscv/riscv-vector-builtins-functions.def
1438 (vfwcvt_xu_frm): New intrinsic function def.
1440 2023-08-16 Pan Li <pan2.li@intel.com>
1442 * config/riscv/riscv-vector-builtins-bases.cc: Use explicit argument.
1444 2023-08-16 Pan Li <pan2.li@intel.com>
1446 * config/riscv/riscv-vector-builtins-bases.cc
1447 (BASE): New declaration.
1448 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1449 * config/riscv/riscv-vector-builtins-functions.def
1450 (vfwcvt_x_frm): New intrinsic function def.
1452 2023-08-16 Pan Li <pan2.li@intel.com>
1454 * config/riscv/riscv-vector-builtins-bases.cc (BASE): New declaration.
1455 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1456 * config/riscv/riscv-vector-builtins-functions.def
1457 (vfcvt_f_frm): New intrinsic function def.
1459 2023-08-16 Pan Li <pan2.li@intel.com>
1461 * config/riscv/riscv-vector-builtins-bases.cc
1462 (BASE): New declaration.
1463 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1464 * config/riscv/riscv-vector-builtins-functions.def
1465 (vfcvt_xu_frm): New intrinsic function def..
1467 2023-08-16 Haochen Gui <guihaoc@gcc.gnu.org>
1470 * config/rs6000/vsx.md (*vsx_extract_<mode>_store_p9): Skip vector
1471 extract when the element is 7 on BE while 8 on LE for byte or 3 on
1472 BE while 4 on LE for halfword.
1474 2023-08-16 Haochen Gui <guihaoc@gcc.gnu.org>
1477 * config/rs6000/vsx.md (expand vsx_extract_<mode>): Set it only
1479 (vsx_extract_v4si): New expand for V4SI extraction.
1480 (vsx_extract_v4si_w1): New insn pattern for V4SI extraction on
1481 word 1 from BE order.
1482 (*mfvsrwz): New insn pattern for mfvsrwz.
1483 (*vsx_extract_<mode>_di_p9): Assert that it won't be generated on
1484 word 1 from BE order.
1485 (*vsx_extract_si): Remove.
1486 (*vsx_extract_v4si_w023): New insn and split pattern on word 0, 2,
1489 2023-08-16 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1491 * config/riscv/autovec.md (vec_mask_len_load_lanes<mode><vsingle>):
1493 (vec_mask_len_store_lanes<mode><vsingle>): Ditto.
1494 * config/riscv/riscv-protos.h (expand_lanes_load_store): New function.
1495 * config/riscv/riscv-v.cc (get_mask_mode): Add tuple mask mode.
1496 (expand_lanes_load_store): New function.
1497 * config/riscv/vector-iterators.md: New iterator.
1499 2023-08-16 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1501 * internal-fn.cc (internal_load_fn_p): Apply
1502 MASK_LEN_{LOAD_LANES,STORE_LANES} into vectorizer.
1503 (internal_store_fn_p): Ditto.
1504 (internal_fn_len_index): Ditto.
1505 (internal_fn_mask_index): Ditto.
1506 (internal_fn_stored_value_index): Ditto.
1507 * tree-vect-data-refs.cc (vect_store_lanes_supported): Ditto.
1508 (vect_load_lanes_supported): Ditto.
1509 * tree-vect-loop.cc: Ditto.
1510 * tree-vect-slp.cc (vect_slp_prefer_store_lanes_p): Ditto.
1511 * tree-vect-stmts.cc (check_load_store_for_partial_vectors): Ditto.
1512 (get_group_load_store_type): Ditto.
1513 (vectorizable_store): Ditto.
1514 (vectorizable_load): Ditto.
1515 * tree-vectorizer.h (vect_store_lanes_supported): Ditto.
1516 (vect_load_lanes_supported): Ditto.
1518 2023-08-16 Pan Li <pan2.li@intel.com>
1520 * config/riscv/riscv-vector-builtins-bases.cc
1521 (enum frm_op_type): New type for frm.
1522 (BASE): New declaration.
1523 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1524 * config/riscv/riscv-vector-builtins-functions.def
1525 (vfcvt_x_frm): New intrinsic function def.
1527 2023-08-16 liuhongt <hongtao.liu@intel.com>
1529 * config/i386/i386-builtins.cc
1530 (ix86_vectorize_builtin_gather): Adjust for use_gather_8parts.
1531 * config/i386/i386-options.cc (parse_mtune_ctrl_str):
1532 Set/Clear tune features use_{gather,scatter}_{2parts, 4parts,
1533 8parts} for -mtune-crtl={,^}{use_gather,use_scatter}.
1534 * config/i386/i386.cc (ix86_vectorize_builtin_scatter): Adjust
1535 for use_scatter_8parts
1536 * config/i386/i386.h (TARGET_USE_GATHER): Rename to ..
1537 (TARGET_USE_GATHER_8PARTS): .. this.
1538 (TARGET_USE_SCATTER): Rename to ..
1539 (TARGET_USE_SCATTER_8PARTS): .. this.
1540 * config/i386/x86-tune.def (X86_TUNE_USE_GATHER): Rename to
1541 (X86_TUNE_USE_GATHER_8PARTS): .. this.
1542 (X86_TUNE_USE_SCATTER): Rename to
1543 (X86_TUNE_USE_SCATTER_8PARTS): .. this.
1544 * config/i386/i386.opt: Add new options mgather, mscatter.
1546 2023-08-16 liuhongt <hongtao.liu@intel.com>
1548 * config/i386/i386-options.cc (m_GDS): New macro.
1549 * config/i386/x86-tune.def (X86_TUNE_USE_GATHER_2PARTS): Don't
1551 (X86_TUNE_USE_GATHER_4PARTS): Ditto.
1552 (X86_TUNE_USE_GATHER): Ditto.
1554 2023-08-16 liuhongt <hongtao.liu@intel.com>
1556 * config/i386/i386.md (movdf_internal): Generate vmovapd instead of
1557 vmovsd when moving DFmode between SSE_REGS.
1558 (movhi_internal): Generate vmovdqa instead of vmovsh when
1559 moving HImode between SSE_REGS.
1560 (mov<mode>_internal): Use vmovaps instead of vmovsh when
1561 moving HF/BFmode between SSE_REGS.
1563 2023-08-15 David Faust <david.faust@oracle.com>
1565 * config/bpf/bpf.md (extendsisi2): Delete useless define_insn.
1567 2023-08-15 David Faust <david.faust@oracle.com>
1570 * config/bpf/bpf.cc (bpf_print_register): Print 'w' registers
1571 for any mode 32-bits or smaller, not just SImode.
1573 2023-08-15 Martin Jambor <mjambor@suse.cz>
1577 * ipa-prop.h (ipcp_get_aggregate_const): Declare.
1578 * ipa-prop.cc (ipcp_get_aggregate_const): New function.
1579 (ipcp_transform_function): Do not deallocate transformation info.
1580 * tree-ssa-sccvn.cc: Include alloc-pool.h, symbol-summary.h and
1582 (vn_reference_lookup_2): When hitting default-def vuse, query
1583 IPA-CP transformation info for any known constants.
1585 2023-08-15 Chung-Lin Tang <cltang@codesourcery.com>
1586 Thomas Schwinge <thomas@codesourcery.com>
1588 * gimplify.cc (oacc_region_type_name): New function.
1589 (oacc_default_clause): If no 'default' clause appears on this
1590 compute construct, see if one appears on a lexically containing
1592 (gimplify_scan_omp_clauses): Upon OMP_CLAUSE_DEFAULT case, set
1593 ctx->oacc_default_clause_ctx to current context.
1595 2023-08-15 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1598 * config/riscv/predicates.md: Fix predicate.
1600 2023-08-15 Richard Biener <rguenther@suse.de>
1602 * tree-vect-slp.cc (vect_analyze_slp_instance): Remove
1603 slp_inst_kind_ctor handling.
1604 (vect_analyze_slp): Simplify.
1605 (vect_build_slp_instance): Dump when we analyze a CTOR.
1606 (vect_slp_check_for_constructors): Rename to ...
1607 (vect_slp_check_for_roots): ... this. Register a
1608 slp_root for CONSTRUCTORs instead of shoving them to
1609 the set of grouped stores.
1610 (vect_slp_analyze_bb_1): Adjust.
1612 2023-08-15 Richard Biener <rguenther@suse.de>
1614 * tree-vectorizer.h (_slp_instance::remain_stmts): Change
1616 (_slp_instance::remain_defs): ... this.
1617 (SLP_INSTANCE_REMAIN_STMTS): Rename to ...
1618 (SLP_INSTANCE_REMAIN_DEFS): ... this.
1619 (slp_root::remain): New.
1620 (slp_root::slp_root): Adjust.
1621 * tree-vect-slp.cc (vect_free_slp_instance): Adjust.
1622 (vect_build_slp_instance): Get extra remain parameter,
1623 adjust former handling of a cut off stmt.
1624 (vect_analyze_slp_instance): Adjust.
1625 (vect_analyze_slp): Likewise.
1626 (_bb_vec_info::~_bb_vec_info): Likewise.
1627 (vectorizable_bb_reduc_epilogue): Dump something if we fail.
1628 (vect_slp_check_for_constructors): Handle non-internal
1629 defs as remain defs of a reduction.
1630 (vectorize_slp_instance_root_stmt): Adjust.
1632 2023-08-15 Richard Biener <rguenther@suse.de>
1634 * tree-ssa-loop-ivcanon.cc: Include tree-vectorizer.h
1635 (canonicalize_loop_induction_variables): Use find_loop_location.
1637 2023-08-15 Hans-Peter Nilsson <hp@axis.com>
1640 * config/cris/cris-protos.h: Revert recent change.
1641 * config/cris/cris.cc (cris_legitimate_address_p): Remove
1642 code_helper unused parameter.
1643 (cris_legitimate_address_p_hook): New wrapper function.
1644 (TARGET_LEGITIMATE_ADDRESS_P): Change to
1645 cris_legitimate_address_p_hook.
1647 2023-08-15 Richard Biener <rguenther@suse.de>
1649 PR tree-optimization/110963
1650 * tree-ssa-pre.cc (do_pre_regular_insertion): Also insert
1651 a PHI node when the expression is available on all edges
1652 and we insert at most one copy from a constant.
1654 2023-08-15 Richard Biener <rguenther@suse.de>
1656 PR tree-optimization/110991
1657 * tree-ssa-loop-ivcanon.cc (constant_after_peeling): Handle
1658 VIEW_CONVERT_EXPR <op>, handle more simple IV-like SSA cycles
1659 that will end up constant.
1661 2023-08-15 Kewen Lin <linkw@linux.ibm.com>
1664 * Makefile.in (RECOG_H): Add $(TREE_H) as dependence.
1666 2023-08-15 Kewen Lin <linkw@linux.ibm.com>
1668 * tree-vect-stmts.cc (vectorizable_load): Move the handlings on
1669 VMAT_LOAD_STORE_LANES in the final loop nest to its own loop,
1670 and update the final nest accordingly.
1672 2023-08-15 Kewen Lin <linkw@linux.ibm.com>
1674 * tree-vect-stmts.cc (vectorizable_load): Remove some useless checks
1677 2023-08-15 Pan Li <pan2.li@intel.com>
1679 * mode-switching.cc (create_pre_exit): Add SET insn check.
1681 2023-08-15 Pan Li <pan2.li@intel.com>
1683 * config/riscv/riscv-vector-builtins-bases.cc
1684 (class vfrec7_frm): New class for frm.
1685 (vfrec7_frm_obj): New declaration.
1687 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1688 * config/riscv/riscv-vector-builtins-functions.def
1689 (vfrec7_frm): New intrinsic function definition.
1690 * config/riscv/vector-iterators.md
1691 (VFMISC): Remove VFREC7.
1693 (float_insn_type): Ditto.
1694 (VFMISC_FRM): New int iterator.
1695 (misc_frm_op): New op for frm.
1696 (float_frm_insn_type): New type for frm.
1697 * config/riscv/vector.md (@pred_<misc_frm_op><mode>):
1698 New pattern for misc frm.
1700 2023-08-14 Vladimir N. Makarov <vmakarov@redhat.com>
1702 * lra-constraints.cc (curr_insn_transform): Process output stack
1703 pointer reloads before emitting reload insns.
1705 2023-08-14 benjamin priour <vultkayn@gcc.gnu.org>
1708 * doc/invoke.texi: Add documentation of
1709 fanalyzer-show-events-in-system-headers
1711 2023-08-14 Jan Hubicka <jh@suse.cz>
1713 PR gcov-profile/110988
1714 * tree-cfg.cc (fold_loop_internal_call): Avoid division by zero.
1716 2023-08-14 Jiawei <jiawei@iscas.ac.cn>
1718 * config/riscv/riscv-c.cc (riscv_cpu_cpp_builtins):
1719 Enable compressed builtins when ZC* extensions enabled.
1720 * config/riscv/riscv-shorten-memrefs.cc:
1721 Enable shorten_memrefs pass when ZC* extensions enabled.
1722 * config/riscv/riscv.cc (riscv_compressed_reg_p):
1723 Enable compressible registers when ZC* extensions enabled.
1724 (riscv_rtx_costs): Allow adjusting rtx costs when ZC* extensions enabled.
1725 (riscv_address_cost): Allow adjusting address cost when ZC* extensions enabled.
1726 (riscv_first_stack_step): Allow compression of the register saves
1727 without adding extra instructions.
1728 * config/riscv/riscv.h (FUNCTION_BOUNDARY): Adjusts function boundary
1729 to 16 bits when ZC* extensions enabled.
1731 2023-08-14 Jiawei <jiawei@iscas.ac.cn>
1733 * common/config/riscv/riscv-common.cc (riscv_subset_list::parse): New extensions.
1734 * config/riscv/riscv-opts.h (MASK_ZCA): New mask.
1741 (TARGET_ZCA): New target.
1742 (TARGET_ZCB): Ditto.
1743 (TARGET_ZCE): Ditto.
1744 (TARGET_ZCF): Ditto.
1745 (TARGET_ZCD): Ditto.
1746 (TARGET_ZCMP): Ditto.
1747 (TARGET_ZCMT): Ditto.
1748 * config/riscv/riscv.opt: New target variable.
1750 2023-08-14 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1753 2023-05-17 Jin Ma <jinma@linux.alibaba.com>
1755 * genrecog.cc (print_nonbool_test): Fix type error of
1756 switch (SUBREG_BYTE (op))'.
1758 2023-08-14 Richard Biener <rguenther@suse.de>
1760 * tree-cfg.cc (print_loop_info): Dump to 'file', not 'dump_file'.
1762 2023-08-14 Pan Li <pan2.li@intel.com>
1764 * config/riscv/riscv-vector-builtins-bases.cc
1765 (class unop_frm): New class for frm.
1766 (vfsqrt_frm_obj): New declaration.
1768 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1769 * config/riscv/riscv-vector-builtins-functions.def
1770 (vfsqrt_frm): New intrinsic function definition.
1772 2023-08-14 Pan Li <pan2.li@intel.com>
1774 * config/riscv/riscv-vector-builtins-bases.cc
1775 (class vfwnmsac_frm): New class for frm.
1776 (vfwnmsac_frm_obj): New declaration.
1778 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1779 * config/riscv/riscv-vector-builtins-functions.def
1780 (vfwnmsac_frm): New intrinsic function definition.
1782 2023-08-14 Pan Li <pan2.li@intel.com>
1784 * config/riscv/riscv-vector-builtins-bases.cc
1785 (class vfwmsac_frm): New class for frm.
1786 (vfwmsac_frm_obj): New declaration.
1788 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1789 * config/riscv/riscv-vector-builtins-functions.def
1790 (vfwmsac_frm): New intrinsic function definition.
1792 2023-08-14 Pan Li <pan2.li@intel.com>
1794 * config/riscv/riscv-vector-builtins-bases.cc
1795 (class vfwnmacc_frm): New class for frm.
1796 (vfwnmacc_frm_obj): New declaration.
1798 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1799 * config/riscv/riscv-vector-builtins-functions.def
1800 (vfwnmacc_frm): New intrinsic function definition.
1802 2023-08-14 Cui, Lili <lili.cui@intel.com>
1804 * common/config/i386/cpuinfo.h (get_intel_cpu): Add model value 0xba
1807 2023-08-14 Hans-Peter Nilsson <hp@axis.com>
1809 * config/mmix/predicates.md (mmix_address_operand): Use
1810 lra_in_progress, not reload_in_progress.
1812 2023-08-14 Hans-Peter Nilsson <hp@axis.com>
1814 * config/mmix/mmix.cc: Re-enable LRA.
1816 2023-08-14 Hans-Peter Nilsson <hp@axis.com>
1818 * config/mmix/predicates.md (frame_pointer_operand): Handle FP+offset
1819 when lra_in_progress.
1821 2023-08-14 Hans-Peter Nilsson <hp@axis.com>
1823 * config/mmix/mmix.cc: Disable LRA for MMIX.
1825 2023-08-14 Pan Li <pan2.li@intel.com>
1827 * config/riscv/riscv-vector-builtins-bases.cc
1828 (class vfwmacc_frm): New class for vfwmacc frm.
1829 (vfwmacc_frm_obj): New declaration.
1831 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1832 * config/riscv/riscv-vector-builtins-functions.def
1833 (vfwmacc_frm): Function definition for vfwmacc.
1834 * config/riscv/riscv-vector-builtins.cc
1835 (function_expander::use_widen_ternop_insn): Add frm support.
1837 2023-08-14 Pan Li <pan2.li@intel.com>
1839 * config/riscv/riscv-vector-builtins-bases.cc
1840 (class vfnmsub_frm): New class for vfnmsub frm.
1841 (vfnmsub_frm): New declaration.
1843 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1844 * config/riscv/riscv-vector-builtins-functions.def
1845 (vfnmsub_frm): New function declaration.
1847 2023-08-14 Vladimir N. Makarov <vmakarov@redhat.com>
1849 * lra-constraints.cc (curr_insn_transform): Set done_p up and
1850 check it on true after processing output stack pointer reload.
1852 2023-08-12 Jakub Jelinek <jakub@redhat.com>
1854 * Makefile.in (USER_H): Add stdckdint.h.
1855 * ginclude/stdckdint.h: New file.
1857 2023-08-12 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1860 * config/riscv/riscv-opts.h (TARGET_VECTOR_VLS): Add TARGET_VETOR.
1862 2023-08-12 Patrick Palka <ppalka@redhat.com>
1864 * tree-pretty-print.cc (dump_generic_node) <case TREE_VEC>:
1865 Delimit output with braces.
1867 2023-08-12 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1870 * config/riscv/riscv-v.cc (expand_vec_series): Refactor the expander.
1872 2023-08-12 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1874 * config/riscv/autovec.md: Add VLS CONST_VECTOR.
1875 * config/riscv/riscv.cc (riscv_const_insns): Ditto.
1876 * config/riscv/vector.md: Ditto.
1878 2023-08-11 David Malcolm <dmalcolm@redhat.com>
1881 * doc/analyzer.texi (__analyzer_get_strlen): New.
1882 * doc/invoke.texi: Add -Wanalyzer-unterminated-string.
1884 2023-08-11 Jeff Law <jlaw@ventanamicro.com>
1886 * config/rx/rx.md (subdi3): Fix test for borrow.
1888 2023-08-11 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1890 PR middle-end/110989
1891 * tree-vect-stmts.cc (vectorizable_store): Replace iv_type with sizetype.
1892 (vectorizable_load): Ditto.
1894 2023-08-11 Jose E. Marchesi <jose.marchesi@oracle.com>
1896 * config/bpf/bpf.md (allocate_stack): Define.
1897 * config/bpf/bpf.h (FIRST_PSEUDO_REGISTER): Make room for fake
1898 stack pointer register.
1899 (FIXED_REGISTERS): Adjust accordingly.
1900 (CALL_USED_REGISTERS): Likewise.
1901 (REG_CLASS_CONTENTS): Likewise.
1902 (REGISTER_NAMES): Likewise.
1903 * config/bpf/bpf.cc (bpf_compute_frame_layout): Do not reserve
1904 space for callee-saved registers.
1905 (bpf_expand_prologue): Do not save callee-saved registers in xbpf.
1906 (bpf_expand_epilogue): Do not restore callee-saved registers in
1909 2023-08-11 Jose E. Marchesi <jose.marchesi@oracle.com>
1911 * config/bpf/bpf.cc (bpf_function_arg_advance): Do not complain
1912 about too many arguments if function is always inlined.
1914 2023-08-11 Patrick Palka <ppalka@redhat.com>
1916 * tree-pretty-print.cc (dump_generic_node) <case COMPONENT_REF>:
1917 Don't call component_ref_field_offset if the RHS isn't a decl.
1919 2023-08-11 John David Anglin <danglin@gcc.gnu.org>
1922 * gensupport.cc(class conlist): Use strtol instead of std::stoi.
1924 2023-08-11 Vladimir N. Makarov <vmakarov@redhat.com>
1926 * lra-constraints.cc (goal_alt_out_sp_reload_p): New flag.
1927 (process_alt_operands): Set the flag.
1928 (curr_insn_transform): Modify stack pointer offsets if output
1929 stack pointer reload is generated.
1931 2023-08-11 Joseph Myers <joseph@codesourcery.com>
1933 * configure: Regenerate.
1935 2023-08-11 Richard Biener <rguenther@suse.de>
1937 PR tree-optimization/110979
1938 * tree-vect-loop.cc (vectorizable_reduction): For
1939 FOLD_LEFT_REDUCTION without target support make sure
1940 we don't need to honor signed zeros and sign dependent rounding.
1942 2023-08-11 Richard Biener <rguenther@suse.de>
1944 * tree-vect-slp.cc (vect_slp_region): Provide opt-info for all SLP
1945 subgraph entries. Dump the used vector size based on the
1946 SLP subgraph entry root vector type.
1948 2023-08-11 Pan Li <pan2.li@intel.com>
1950 * config/riscv/riscv-vector-builtins-bases.cc
1951 (class vfmsub_frm): New class for vfmsub frm.
1952 (vfmsub_frm): New declaration.
1954 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1955 * config/riscv/riscv-vector-builtins-functions.def
1956 (vfmsub_frm): New function declaration.
1958 2023-08-11 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1960 * doc/md.texi: Add vec_mask_len_{load_lanes,store_lanes} patterns.
1961 * internal-fn.cc (expand_partial_load_optab_fn): Ditto.
1962 (expand_partial_store_optab_fn): Ditto.
1963 * internal-fn.def (MASK_LEN_LOAD_LANES): Ditto.
1964 (MASK_LEN_STORE_LANES): Ditto.
1965 * optabs.def (OPTAB_CD): Ditto.
1967 2023-08-11 Pan Li <pan2.li@intel.com>
1969 * config/riscv/riscv-vector-builtins-bases.cc
1970 (class vfnmadd_frm): New class for vfnmadd frm.
1971 (vfnmadd_frm): New declaration.
1973 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1974 * config/riscv/riscv-vector-builtins-functions.def
1975 (vfnmadd_frm): New function declaration.
1977 2023-08-11 Drew Ross <drross@redhat.com>
1978 Jakub Jelinek <jakub@redhat.com>
1980 PR tree-optimization/109938
1981 * match.pd (((x ^ y) & z) | x -> (z & y) | x): New simplification.
1983 2023-08-11 Pan Li <pan2.li@intel.com>
1985 * config/riscv/riscv-vector-builtins-bases.cc
1986 (class vfmadd_frm): New class for vfmadd frm.
1987 (vfmadd_frm_obj): New declaration.
1989 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1990 * config/riscv/riscv-vector-builtins-functions.def
1991 (vfmadd_frm): New function definition.
1993 2023-08-11 Pan Li <pan2.li@intel.com>
1995 * config/riscv/riscv-vector-builtins-bases.cc
1996 (class vfnmsac_frm): New class for vfnmsac frm.
1997 (vfnmsac_frm_obj): New declaration.
1999 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
2000 * config/riscv/riscv-vector-builtins-functions.def
2001 (vfnmsac_frm): New function definition.
2003 2023-08-11 Jakub Jelinek <jakub@redhat.com>
2005 * doc/extend.texi (Typeof): Document typeof_unqual
2006 and __typeof_unqual__.
2008 2023-08-11 Andrew Pinski <apinski@marvell.com>
2010 PR tree-optimization/110954
2011 * generic-match-head.cc (bitwise_inverted_equal_p): Add
2012 wascmp argument and set it accordingly.
2013 * gimple-match-head.cc (bitwise_inverted_equal_p): Add
2014 wascmp argument to the macro.
2015 (gimple_bitwise_inverted_equal_p): Add
2016 wascmp argument and set it accordingly.
2017 * match.pd (`a & ~a`, `a ^| ~a`): Update call
2018 to bitwise_inverted_equal_p and handle wascmp case.
2019 (`(~x | y) & x`, `(~x | y) & x`, `a?~t:t`): Update
2020 call to bitwise_inverted_equal_p and check to see
2021 if was !wascmp or if precision was 1.
2023 2023-08-11 Martin Uecker <uecker@tugraz.at>
2026 * doc/invoke.texi: Update.
2028 2023-08-11 Pan Li <pan2.li@intel.com>
2030 * config/riscv/riscv-vector-builtins-bases.cc
2031 (class vfmsac_frm): New class for vfmsac frm.
2032 (vfmsac_frm_obj): New declaration.
2034 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
2035 * config/riscv/riscv-vector-builtins-functions.def
2036 (vfmsac_frm): New function definition
2038 2023-08-10 Jan Hubicka <jh@suse.cz>
2040 PR middle-end/110923
2041 * tree-ssa-loop-split.cc (split_loop): Watch for division by zero.
2043 2023-08-10 Patrick O'Neill <patrick@rivosinc.com>
2045 * common/config/riscv/riscv-common.cc: Add Ztso and mark Ztso as
2046 dependent on 'a' extension.
2047 * config/riscv/riscv-opts.h (MASK_ZTSO): New mask.
2048 (TARGET_ZTSO): New target.
2049 * config/riscv/riscv.cc (riscv_memmodel_needs_amo_acquire): Add
2051 (riscv_memmodel_needs_amo_release): Add Ztso case.
2052 (riscv_print_operand): Add Ztso case for LR/SC annotations.
2053 * config/riscv/riscv.md: Import sync-rvwmo.md and sync-ztso.md.
2054 * config/riscv/riscv.opt: Add Ztso target variable.
2055 * config/riscv/sync.md (mem_thread_fence_1): Expand to RVWMO or
2057 (atomic_load<mode>): Expand to RVWMO or Ztso specific insn.
2058 (atomic_store<mode>): Expand to RVWMO or Ztso specific insn.
2059 * config/riscv/sync-rvwmo.md: New file. Seperate out RVWMO
2060 specific load/store/fence mappings.
2061 * config/riscv/sync-ztso.md: New file. Seperate out Ztso
2062 specific load/store/fence mappings.
2064 2023-08-10 Jan Hubicka <jh@suse.cz>
2066 * cfgloopmanip.cc (duplicate_loop_body_to_header_edge): Special case loops with
2069 2023-08-10 Jan Hubicka <jh@suse.cz>
2071 * tree-ssa-threadupdate.cc (ssa_fix_duplicate_block_edges): Fix profile update.
2073 2023-08-10 Jan Hubicka <jh@suse.cz>
2075 * profile-count.cc (profile_count::differs_from_p): Fix overflow and
2076 handling of undefined values.
2078 2023-08-10 Jakub Jelinek <jakub@redhat.com>
2081 * tree-ssa-phiopt.cc (single_non_singleton_phi_for_edges): Never
2082 return virtual phis and return NULL if there is a virtual phi
2083 where the arguments from E0 and E1 edges aren't equal.
2085 2023-08-10 Richard Biener <rguenther@suse.de>
2087 * internal-fn.def (VCOND, VCONDU, VCONDEQ, VCOND_MASK,
2088 VEC_SET, VEC_EXTRACT): Make ECF_CONST | ECF_NOTHROW.
2090 2023-08-10 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2093 * config/riscv/autovec.md (vec_duplicate<mode>): New pattern.
2095 2023-08-10 Pan Li <pan2.li@intel.com>
2097 * config/riscv/riscv-vector-builtins-bases.cc
2098 (class vfnmacc_frm): New class for vfnmacc.
2099 (vfnmacc_frm_obj): New declaration.
2101 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
2102 * config/riscv/riscv-vector-builtins-functions.def
2103 (vfnmacc_frm): New function definition.
2105 2023-08-10 Pan Li <pan2.li@intel.com>
2107 * config/riscv/riscv-vector-builtins-bases.cc
2108 (class vfmacc_frm): New class for vfmacc frm.
2109 (vfmacc_frm_obj): New declaration.
2111 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
2112 * config/riscv/riscv-vector-builtins-functions.def
2113 (vfmacc_frm): New function definition.
2115 2023-08-10 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2118 * config/riscv/riscv-v.cc (expand_cond_len_ternop): Add integer ternary.
2120 2023-08-10 Richard Biener <rguenther@suse.de>
2122 * tree-vectorizer.h (vectorizable_live_operation): Remove
2123 gimple_stmt_iterator * argument.
2124 * tree-vect-loop.cc (vectorizable_live_operation): Likewise.
2125 Adjust plumbing around vect_get_loop_mask.
2126 (vect_analyze_loop_operations): Adjust.
2127 * tree-vect-slp.cc (vect_slp_analyze_node_operations_1): Likewise.
2128 (vect_bb_slp_mark_live_stmts): Likewise.
2129 (vect_schedule_slp_node): Likewise.
2130 * tree-vect-stmts.cc (can_vectorize_live_stmts): Likewise.
2131 Remove gimple_stmt_iterator * argument.
2132 (vect_transform_stmt): Adjust.
2134 2023-08-10 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2136 * config/riscv/vector-iterators.md: Add missing modes.
2138 2023-08-10 Jakub Jelinek <jakub@redhat.com>
2141 * lto-streamer-in.cc (lto_input_tree_1): Assert TYPE_PRECISION
2142 is up to WIDE_INT_MAX_PRECISION rather than MAX_BITSIZE_MODE_ANY_INT.
2144 2023-08-10 Jakub Jelinek <jakub@redhat.com>
2147 * expr.cc (expand_expr_real_1) <case MEM_REF>: Add an early return for
2148 EXPAND_WRITE or EXPAND_MEMORY modifiers to avoid testing it multiple
2151 2023-08-10 liuhongt <hongtao.liu@intel.com>
2154 * config/i386/mmx.md: (movq_<mode>_to_sse): Also do not
2155 sanitize upper part of V4HFmode register with
2157 (<insn>v4hf3): Enable for ix86_partial_vec_fp_math.
2159 (<insn>v2hf3): Ditto.
2161 (movd_v2hf_to_sse): Do not sanitize upper part of V2HFmode
2162 register with -fno-trapping-math.
2164 2023-08-10 Pan Li <pan2.li@intel.com>
2165 Kito Cheng <kito.cheng@sifive.com>
2167 * config/riscv/riscv-protos.h
2168 (enum floating_point_rounding_mode): Add NONE, DYN_EXIT and DYN_CALL.
2169 (get_frm_mode): New declaration.
2170 * config/riscv/riscv-v.cc (get_frm_mode): New function to get frm mode.
2171 * config/riscv/riscv-vector-builtins.cc
2172 (function_expander::use_ternop_insn): Take care of frm reg.
2173 * config/riscv/riscv.cc (riscv_static_frm_mode_p): Migrate to FRM_XXX.
2174 (riscv_emit_frm_mode_set): Ditto.
2175 (riscv_emit_mode_set): Ditto.
2176 (riscv_frm_adjust_mode_after_call): Ditto.
2177 (riscv_frm_mode_needed): Ditto.
2178 (riscv_frm_mode_after): Ditto.
2179 (riscv_mode_entry): Ditto.
2180 (riscv_mode_exit): Ditto.
2181 * config/riscv/riscv.h (NUM_MODES_FOR_MODE_SWITCHING): Ditto.
2182 * config/riscv/vector.md
2183 (rne,rtz,rdn,rup,rmm,dyn,dyn_exit,dyn_call,none): Removed
2184 (symbol_ref): * config/riscv/vector.md: Set frm_mode attr explicitly.
2186 2023-08-09 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2188 * config/riscv/riscv-vsetvl.cc (anticipatable_occurrence_p): Fix
2189 incorrect anticipate info.
2191 2023-08-09 Tsukasa OI <research_trasio@irq.a4lg.com>
2193 * common/config/riscv/riscv-common.cc (riscv_ext_version_table):
2194 Remove 'Zve32d' from the version list.
2196 2023-08-09 Jin Ma <jinma@linux.alibaba.com>
2198 * config/riscv/riscv.cc (riscv_sched_variable_issue): New function.
2199 (TARGET_SCHED_VARIABLE_ISSUE): New macro.
2200 Co-authored-by: Philipp Tomsich <philipp.tomsich@vrull.eu>
2201 Co-authored-by: Jeff Law <jlaw@ventanamicro.com>
2203 2023-08-09 Jivan Hakobyan <jivanhakobyan9@gmail.com>
2205 * config/riscv/riscv.cc (riscv_legitimize_address): Handle folding.
2206 (mem_shadd_or_shadd_rtx_p): New function.
2208 2023-08-09 Andrew Pinski <apinski@marvell.com>
2210 PR tree-optimization/110937
2211 PR tree-optimization/100798
2212 * match.pd (`a ? ~b : b`): Handle this
2215 2023-08-09 Uros Bizjak <ubizjak@gmail.com>
2217 * config/i386/i386.opt (mpartial-vector-fp-math): Add dot.
2219 2023-08-09 Richard Ball <richard.ball@arm.com>
2221 * config/aarch64/aarch64-cores.def (AARCH64_CORE): Add Cortex-A520 CPU.
2222 * config/aarch64/aarch64-tune.md: Regenerate.
2223 * doc/invoke.texi: Document Cortex-A520 CPU.
2225 2023-08-09 Carl Love <cel@us.ibm.com>
2227 * config/rs6000/rs6000-builtins.def (vcmpneb, vcmpneh, vcmpnew):
2228 Move definitions to Altivec stanza.
2229 * config/rs6000/altivec.md (vcmpneb, vcmpneh, vcmpnew): New
2232 2023-08-09 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2235 * config/riscv/riscv-v.cc (expand_const_vector): Add NPATTERNS = 1
2236 stepped vector support.
2238 2023-08-09 liuhongt <hongtao.liu@intel.com>
2240 * common/config/i386/cpuinfo.h (get_available_features):
2241 Rename local variable subleaf_level to max_subleaf_level.
2243 2023-08-09 Richard Biener <rguenther@suse.de>
2245 PR rtl-optimization/110587
2246 * lra-assigns.cc (find_hard_regno_for_1): Re-order checks.
2248 2023-08-09 Kewen Lin <linkw@linux.ibm.com>
2250 PR tree-optimization/110248
2251 * config/rs6000/rs6000.cc (rs6000_legitimate_address_p): Check if
2252 the given code is for ifn LEN_{LOAD,STORE}, if yes then make it not
2253 legitimate when outer code is PLUS.
2255 2023-08-09 Kewen Lin <linkw@linux.ibm.com>
2257 PR tree-optimization/110248
2258 * recog.cc (memory_address_addr_space_p): Add one more argument ch of
2259 type code_helper and pass it to targetm.addr_space.legitimate_address_p
2260 instead of ERROR_MARK.
2261 (offsettable_address_addr_space_p): Update one function pointer with
2262 one more argument of type code_helper as its assignees
2263 memory_address_addr_space_p and strict_memory_address_addr_space_p
2264 have been adjusted, and adjust some call sites with ERROR_MARK.
2265 * recog.h (tree.h): New include header file for tree_code ERROR_MARK.
2266 (memory_address_addr_space_p): Adjust with one more unnamed argument
2267 of type code_helper with default ERROR_MARK.
2268 (strict_memory_address_addr_space_p): Likewise.
2269 * reload.cc (strict_memory_address_addr_space_p): Add one unnamed
2270 argument of type code_helper.
2271 * tree-ssa-address.cc (valid_mem_ref_p): Add one more argument ch of
2272 type code_helper and pass it to memory_address_addr_space_p.
2273 * tree-ssa-address.h (valid_mem_ref_p): Adjust the declaration with
2274 one more unnamed argument of type code_helper with default value
2276 * tree-ssa-loop-ivopts.cc (get_address_cost): Use ERROR_MARK as code
2277 by default, change it with ifn code for USE_PTR_ADDRESS type use, and
2278 pass it to all valid_mem_ref_p calls.
2280 2023-08-09 Kewen Lin <linkw@linux.ibm.com>
2282 PR tree-optimization/110248
2283 * coretypes.h (class code_helper): Add forward declaration.
2284 * doc/tm.texi: Regenerate.
2285 * lra-constraints.cc (valid_address_p): Call target hook
2286 targetm.addr_space.legitimate_address_p with an extra parameter
2287 ERROR_MARK as its prototype changes.
2288 * recog.cc (memory_address_addr_space_p): Likewise.
2289 * reload.cc (strict_memory_address_addr_space_p): Likewise.
2290 * target.def (legitimate_address_p, addr_space.legitimate_address_p):
2291 Extend with one more argument of type code_helper, update the
2292 documentation accordingly.
2293 * targhooks.cc (default_legitimate_address_p): Adjust for the
2294 new code_helper argument.
2295 (default_addr_space_legitimate_address_p): Likewise.
2296 * targhooks.h (default_legitimate_address_p): Likewise.
2297 (default_addr_space_legitimate_address_p): Likewise.
2298 * config/aarch64/aarch64.cc (aarch64_legitimate_address_hook_p): Adjust
2299 with extra unnamed code_helper argument with default ERROR_MARK.
2300 * config/alpha/alpha.cc (alpha_legitimate_address_p): Likewise.
2301 * config/arc/arc.cc (arc_legitimate_address_p): Likewise.
2302 * config/arm/arm-protos.h (arm_legitimate_address_p): Likewise.
2303 (tree.h): New include for tree_code ERROR_MARK.
2304 * config/arm/arm.cc (arm_legitimate_address_p): Adjust with extra
2305 unnamed code_helper argument with default ERROR_MARK.
2306 * config/avr/avr.cc (avr_addr_space_legitimate_address_p): Likewise.
2307 * config/bfin/bfin.cc (bfin_legitimate_address_p): Likewise.
2308 * config/bpf/bpf.cc (bpf_legitimate_address_p): Likewise.
2309 * config/c6x/c6x.cc (c6x_legitimate_address_p): Likewise.
2310 * config/cris/cris-protos.h (cris_legitimate_address_p): Likewise.
2311 (tree.h): New include for tree_code ERROR_MARK.
2312 * config/cris/cris.cc (cris_legitimate_address_p): Adjust with extra
2313 unnamed code_helper argument with default ERROR_MARK.
2314 * config/csky/csky.cc (csky_legitimate_address_p): Likewise.
2315 * config/epiphany/epiphany.cc (epiphany_legitimate_address_p):
2317 * config/frv/frv.cc (frv_legitimate_address_p): Likewise.
2318 * config/ft32/ft32.cc (ft32_addr_space_legitimate_address_p): Likewise.
2319 * config/gcn/gcn.cc (gcn_addr_space_legitimate_address_p): Likewise.
2320 * config/h8300/h8300.cc (h8300_legitimate_address_p): Likewise.
2321 * config/i386/i386.cc (ix86_legitimate_address_p): Likewise.
2322 * config/ia64/ia64.cc (ia64_legitimate_address_p): Likewise.
2323 * config/iq2000/iq2000.cc (iq2000_legitimate_address_p): Likewise.
2324 * config/lm32/lm32.cc (lm32_legitimate_address_p): Likewise.
2325 * config/loongarch/loongarch.cc (loongarch_legitimate_address_p):
2327 * config/m32c/m32c.cc (m32c_legitimate_address_p): Likewise.
2328 (m32c_addr_space_legitimate_address_p): Likewise.
2329 * config/m32r/m32r.cc (m32r_legitimate_address_p): Likewise.
2330 * config/m68k/m68k.cc (m68k_legitimate_address_p): Likewise.
2331 * config/mcore/mcore.cc (mcore_legitimate_address_p): Likewise.
2332 * config/microblaze/microblaze-protos.h (tree.h): New include for
2333 tree_code ERROR_MARK.
2334 (microblaze_legitimate_address_p): Adjust with extra unnamed
2335 code_helper argument with default ERROR_MARK.
2336 * config/microblaze/microblaze.cc (microblaze_legitimate_address_p):
2338 * config/mips/mips.cc (mips_legitimate_address_p): Likewise.
2339 * config/mmix/mmix.cc (mmix_legitimate_address_p): Likewise.
2340 * config/mn10300/mn10300.cc (mn10300_legitimate_address_p): Likewise.
2341 * config/moxie/moxie.cc (moxie_legitimate_address_p): Likewise.
2342 * config/msp430/msp430.cc (msp430_legitimate_address_p): Likewise.
2343 (msp430_addr_space_legitimate_address_p): Adjust with extra code_helper
2344 argument with default ERROR_MARK and adjust the call to function
2345 msp430_legitimate_address_p.
2346 * config/nds32/nds32.cc (nds32_legitimate_address_p): Adjust with extra
2347 unnamed code_helper argument with default ERROR_MARK.
2348 * config/nios2/nios2.cc (nios2_legitimate_address_p): Likewise.
2349 * config/nvptx/nvptx.cc (nvptx_legitimate_address_p): Likewise.
2350 * config/or1k/or1k.cc (or1k_legitimate_address_p): Likewise.
2351 * config/pa/pa.cc (pa_legitimate_address_p): Likewise.
2352 * config/pdp11/pdp11.cc (pdp11_legitimate_address_p): Likewise.
2353 * config/pru/pru.cc (pru_addr_space_legitimate_address_p): Likewise.
2354 * config/riscv/riscv.cc (riscv_legitimate_address_p): Likewise.
2355 * config/rl78/rl78-protos.h (rl78_as_legitimate_address): Likewise.
2356 (tree.h): New include for tree_code ERROR_MARK.
2357 * config/rl78/rl78.cc (rl78_as_legitimate_address): Adjust with
2358 extra unnamed code_helper argument with default ERROR_MARK.
2359 * config/rs6000/rs6000.cc (rs6000_legitimate_address_p): Likewise.
2360 (rs6000_debug_legitimate_address_p): Adjust with extra code_helper
2361 argument and adjust the call to function rs6000_legitimate_address_p.
2362 * config/rx/rx.cc (rx_is_legitimate_address): Adjust with extra
2363 unnamed code_helper argument with default ERROR_MARK.
2364 * config/s390/s390.cc (s390_legitimate_address_p): Likewise.
2365 * config/sh/sh.cc (sh_legitimate_address_p): Likewise.
2366 * config/sparc/sparc.cc (sparc_legitimate_address_p): Likewise.
2367 * config/v850/v850.cc (v850_legitimate_address_p): Likewise.
2368 * config/vax/vax.cc (vax_legitimate_address_p): Likewise.
2369 * config/visium/visium.cc (visium_legitimate_address_p): Likewise.
2370 * config/xtensa/xtensa.cc (xtensa_legitimate_address_p): Likewise.
2371 * config/stormy16/stormy16-protos.h (xstormy16_legitimate_address_p):
2373 (tree.h): New include for tree_code ERROR_MARK.
2374 * config/stormy16/stormy16.cc (xstormy16_legitimate_address_p):
2375 Adjust with extra unnamed code_helper argument with default
2378 2023-08-09 liuhongt <hongtao.liu@intel.com>
2380 * common/config/i386/cpuinfo.h (get_available_features): Check
2381 EAX for valid subleaf before use CPUID.
2383 2023-08-08 Jeff Law <jlaw@ventanamicro.com>
2385 * config/riscv/riscv.cc (riscv_expand_conditional_move): Use word_mode
2386 for the temporary when canonicalizing the condition.
2388 2023-08-08 Cupertino Miranda <cupertino.miranda@oracle.com>
2390 * config/bpf/core-builtins.cc: Cleaned include headers.
2391 (struct cr_builtins): Added GTY.
2392 (cr_builtins_ref): Created.
2393 (builtins_data) Changed to GC root.
2394 (allocate_builtin_data): Changed.
2395 Included gt-core-builtins.h.
2396 * config/bpf/coreout.cc: (bpf_core_extra) Added GTY.
2397 (bpf_core_extra_ref): Created.
2398 (bpf_comment_info): Changed to GC root.
2399 (bpf_core_reloc_add, output_btfext_header, btf_ext_init): Changed.
2401 2023-08-08 Uros Bizjak <ubizjak@gmail.com>
2404 * config/i386/i386.opt (mpartial-vector-fp-math): New option.
2405 * config/i386/mmx.md (movq_<mode>_to_sse): Do not sanitize
2406 upper part of V2SFmode register with -fno-trapping-math.
2407 (<plusminusmult:insn>v2sf3): Enable for ix86_partial_vec_fp_math.
2409 (<smaxmin:code>v2sf3): Ditto.
2411 (*mmx_haddv2sf3_low): Ditto.
2412 (*mmx_hsubv2sf3_low): Ditto.
2413 (vec_addsubv2sf3): Ditto.
2414 (vec_cmpv2sfv2si): Ditto.
2415 (vcond<V2FI:mode>v2sf): Ditto.
2420 (fix_truncv2sfv2si2): Ditto.
2421 (fixuns_truncv2sfv2si2): Ditto.
2422 (floatv2siv2sf2): Ditto.
2423 (floatunsv2siv2sf2): Ditto.
2424 (nearbyintv2sf2): Ditto.
2426 (lrintv2sfv2si2): Ditto.
2428 (lceilv2sfv2si2): Ditto.
2429 (floorv2sf2): Ditto.
2430 (lfloorv2sfv2si2): Ditto.
2431 (btruncv2sf2): Ditto.
2432 (roundv2sf2): Ditto.
2433 (lroundv2sfv2si2): Ditto.
2434 * doc/invoke.texi (x86 Options): Document
2435 -mpartial-vector-fp-math option.
2437 2023-08-08 Andrew Pinski <apinski@marvell.com>
2439 PR tree-optimization/103281
2440 PR tree-optimization/28794
2441 * vr-values.cc (simplify_using_ranges::simplify_cond_using_ranges_1): Split out
2443 (simplify_using_ranges::simplify_compare_using_ranges_1): Here.
2444 (simplify_using_ranges::simplify_casted_cond): Rename to ...
2445 (simplify_using_ranges::simplify_casted_compare): This
2446 and change arguments to take op0 and op1.
2447 (simplify_using_ranges::simplify_compare_assign_using_ranges_1): New method.
2448 (simplify_using_ranges::simplify): For tcc_comparison assignments call
2449 simplify_compare_assign_using_ranges_1.
2450 * vr-values.h (simplify_using_ranges): Add
2451 new methods, simplify_compare_using_ranges_1 and simplify_compare_assign_using_ranges_1.
2452 Rename simplify_casted_cond and simplify_casted_compare and
2453 update argument types.
2455 2023-08-08 Andrzej Turko <andrzej.turko@gmail.com>
2457 * genmatch.cc: Log line numbers indirectly.
2459 2023-08-08 Andrzej Turko <andrzej.turko@gmail.com>
2461 * genmatch.cc: Make sinfo map ordered.
2462 * Makefile.in: Require the ordered map header for genmatch.o.
2464 2023-08-08 Andrzej Turko <andrzej.turko@gmail.com>
2466 * ordered-hash-map.h: Add get_or_insert.
2467 * ordered-hash-map-tests.cc: Use get_or_insert in tests.
2469 2023-08-08 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2471 * config/riscv/autovec.md (cond_<optab><mode>): New pattern.
2472 (cond_len_<optab><mode>): Ditto.
2473 (cond_fma<mode>): Ditto.
2474 (cond_len_fma<mode>): Ditto.
2475 (cond_fnma<mode>): Ditto.
2476 (cond_len_fnma<mode>): Ditto.
2477 (cond_fms<mode>): Ditto.
2478 (cond_len_fms<mode>): Ditto.
2479 (cond_fnms<mode>): Ditto.
2480 (cond_len_fnms<mode>): Ditto.
2481 * config/riscv/riscv-protos.h (riscv_get_v_regno_alignment): Export
2483 (enum insn_type): Add new enum type.
2484 (prepare_ternary_operands): New function.
2485 * config/riscv/riscv-v.cc (emit_vlmax_masked_fp_mu_insn): Ditto.
2486 (emit_nonvlmax_tumu_insn): Ditto.
2487 (emit_nonvlmax_fp_tumu_insn): Ditto.
2488 (expand_cond_len_binop): Add condtional operations.
2489 (expand_cond_len_ternop): Ditto.
2490 (prepare_ternary_operands): New function.
2491 * config/riscv/riscv.cc (riscv_memmodel_needs_amo_release): Export
2492 riscv_get_v_regno_alignment as global scope.
2493 * config/riscv/vector.md: Fix ternary bugs.
2495 2023-08-08 Richard Biener <rguenther@suse.de>
2497 PR tree-optimization/49955
2498 * tree-vectorizer.h (_slp_instance::remain_stmts): New.
2499 (SLP_INSTANCE_REMAIN_STMTS): Likewise.
2500 * tree-vect-slp.cc (vect_free_slp_instance): Release
2501 SLP_INSTANCE_REMAIN_STMTS.
2502 (vect_build_slp_instance): Make the number of lanes of
2503 a BB reduction even.
2504 (vectorize_slp_instance_root_stmt): Handle unvectorized
2505 defs of a BB reduction.
2507 2023-08-08 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
2509 * internal-fn.cc (get_len_internal_fn): New function.
2510 (DEF_INTERNAL_COND_FN): Ditto.
2511 (DEF_INTERNAL_SIGNED_COND_FN): Ditto.
2512 * internal-fn.h (get_len_internal_fn): Ditto.
2513 * tree-vect-stmts.cc (vectorizable_call): Add CALL auto-vectorization.
2515 2023-08-08 Richard Biener <rguenther@suse.de>
2517 PR tree-optimization/110924
2518 * tree-ssa-live.h (virtual_operand_live): Update comment.
2519 * tree-ssa-live.cc (virtual_operand_live::get_live_in): Remove
2520 optimization, look at each predecessor.
2521 * tree-ssa-sink.cc (pass_sink_code::execute): Mark backedges.
2523 2023-08-08 yulong <shiyulong@iscas.ac.cn>
2525 * config/riscv/riscv-v.cc (slide1_sew64_helper): Modify.
2527 2023-08-08 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2529 * config/riscv/autovec-vls.md (<optab><mode>2): Add VLS neg.
2530 * config/riscv/vector.md: Ditto.
2532 2023-08-08 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2534 * config/riscv/autovec.md: Add VLS shift.
2536 2023-08-07 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2538 * config/riscv/autovec-vls.md (<optab><mode>3): Add VLS modes.
2539 * config/riscv/vector-iterators.md: Ditto.
2540 * config/riscv/vector.md: Ditto.
2542 2023-08-07 Jonathan Wakely <jwakely@redhat.com>
2544 * config/i386/i386.cc (ix86_invalid_conversion): Fix grammar.
2546 2023-08-07 Nick Alcock <nick.alcock@oracle.com>
2548 * configure: Regenerate.
2550 2023-08-07 John Ericson <git@JohnEricson.me>
2552 * configure: Regenerate.
2554 2023-08-07 Alan Modra <amodra@gmail.com>
2556 * configure: Regenerate.
2558 2023-08-07 Alexander von Gluck IV <kallisti5@unixzen.com>
2560 * configure: Regenerate.
2562 2023-08-07 Nick Alcock <nick.alcock@oracle.com>
2564 * configure: Regenerate.
2566 2023-08-07 Nick Alcock <nick.alcock@oracle.com>
2568 * configure: Regenerate.
2570 2023-08-07 H.J. Lu <hjl.tools@gmail.com>
2572 * configure: Regenerate.
2574 2023-08-07 H.J. Lu <hjl.tools@gmail.com>
2576 * configure: Regenerate.
2578 2023-08-07 Jeff Law <jlaw@ventanamicro.com>
2580 * config/riscv/riscv.cc (riscv_expand_conditional_move): Allow
2581 VOIDmode operands to conditional before canonicalization.
2583 2023-08-07 Manolis Tsamis <manolis.tsamis@vrull.eu>
2585 * regcprop.cc (maybe_copy_reg_attrs): Remove unnecessary function.
2586 (find_oldest_value_reg): Inline stack_pointer_rtx check.
2587 (copyprop_hardreg_forward_1): Inline stack_pointer_rtx check.
2589 2023-08-07 Martin Jambor <mjambor@suse.cz>
2592 * ipa-param-manipulation.h (class ipa_param_body_adjustments): New
2593 members get_ddef_if_exists_and_is_used and mark_clobbers_dead.
2594 * ipa-sra.cc (isra_track_scalar_value_uses): Ignore clobbers.
2595 (ptr_parm_has_nonarg_uses): Likewise.
2596 * ipa-param-manipulation.cc
2597 (ipa_param_body_adjustments::get_ddef_if_exists_and_is_used): New.
2598 (ipa_param_body_adjustments::mark_dead_statements): Move initial
2599 checks to get_ddef_if_exists_and_is_used.
2600 (ipa_param_body_adjustments::mark_clobbers_dead): New.
2601 (ipa_param_body_adjustments::common_initialization): Call
2602 mark_clobbers_dead when splitting.
2604 2023-08-07 Raphael Zinsly <rzinsly@ventanamicro.com>
2606 * config/riscv/riscv.cc (riscv_expand_int_scc): Add invert_ptr
2607 as an argument and pass it to riscv_emit_int_order_test.
2608 (riscv_expand_conditional_move): Handle cases where the condition
2609 is not EQ/NE or the second argument to the conditional is not
2611 * config/riscv/riscv-protos.h (riscv_expand_int_scc): Update prototype.
2612 Co-authored-by: Jeff Law <jlaw@ventanamicro.com>
2614 2023-08-07 Andrew Pinski <apinski@marvell.com>
2616 PR tree-optimization/109959
2617 * match.pd (`(a > 1) ? 0 : (cast)a`, `(a <= 1) & (cast)a`):
2620 2023-08-07 Richard Biener <rguenther@suse.de>
2622 * tree-ssa-sink.cc (pass_sink_code::execute): Do not
2623 calculate post-dominators. Calculate RPO on the inverted
2624 graph and process blocks in that order.
2626 2023-08-07 liuhongt <hongtao.liu@intel.com>
2629 * config/i386/i386-protos.h
2630 (vpternlog_redundant_operand_mask): Adjust parameter type.
2631 * config/i386/i386.cc (vpternlog_redundant_operand_mask): Use
2632 INTVAL instead of XINT, also adjust parameter type from rtx*
2633 to rtx since the function only needs operands[4] in vpternlog
2635 (substitute_vpternlog_operands): Pass operands[4] instead of
2636 operands to vpternlog_redundant_operand_mask.
2637 * config/i386/sse.md: Ditto.
2639 2023-08-07 Richard Biener <rguenther@suse.de>
2641 * tree-vect-slp.cc (vect_slp_region): Save/restore vect_location
2642 around dumping code.
2644 2023-08-07 liuhongt <hongtao.liu@intel.com>
2647 * config/i386/mmx.md (<insn><mode>3): Changed from define_insn
2648 to define_expand and break into ..
2649 (<insn>v4hf3): .. this.
2650 (divv4hf3): .. this.
2651 (<insn>v2hf3): .. this.
2652 (divv2hf3): .. this.
2653 (movd_v2hf_to_sse): New define_expand.
2654 (movq_<mode>_to_sse): Extend to V4HFmode.
2655 (mmxdoublevecmode): Ditto.
2656 (V2FI_V4HF): New mode iterator.
2657 * config/i386/sse.md (*vec_concatv4sf): Extend to hanlde V8HF
2658 by using mode iterator V4SF_V8HF, renamed to ..
2659 (*vec_concat<mode>): .. this.
2660 (*vec_concatv4sf_0): Extend to handle V8HF by using mode
2661 iterator V4SF_V8HF, renamed to ..
2662 (*vec_concat<mode>_0): .. this.
2663 (*vec_concatv8hf_movss): New define_insn.
2664 (V4SF_V8HF): New mode iterator.
2666 2023-08-07 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2668 * tree-vect-patterns.cc (vect_recog_over_widening_pattern): Add op vectype.
2670 2023-08-07 Jan Beulich <jbeulich@suse.com>
2672 * config/i386/mmx.md (*mmx_pinsrd): Drop "prefix_data16".
2673 (*mmx_pinsrb): Likewise.
2674 (*mmx_pextrb): Likewise.
2675 (*mmx_pextrb_zext): Likewise.
2676 (mmx_pshufbv8qi3): Likewise.
2677 (mmx_pshufbv4qi3): Likewise.
2678 (mmx_pswapdv2si2): Likewise.
2679 (*pinsrb): Likewise.
2680 (*pextrb): Likewise.
2681 (*pextrb_zext): Likewise.
2682 * config/i386/sse.md (*sse4_1_mulv2siv2di3<mask_name>): Likewise.
2683 (*sse2_eq<mode>3): Likewise.
2684 (*sse2_gt<mode>3): Likewise.
2685 (<sse2p4_1>_pinsr<ssemodesuffix>): Likewise.
2686 (*vec_extract<mode>): Likewise.
2687 (*vec_extract<PEXTR_MODE12:mode>_zext): Likewise.
2688 (*vec_extractv16qi_zext): Likewise.
2689 (ssse3_ph<plusminus_mnemonic>wv8hi3): Likewise.
2690 (ssse3_pmaddubsw128): Likewise.
2691 (*<ssse3_avx2>_pmulhrsw<mode>3<mask_name>): Likewise.
2692 (<ssse3_avx2>_pshufb<mode>3<mask_name>): Likewise.
2693 (<ssse3_avx2>_psign<mode>3): Likewise.
2694 (<ssse3_avx2>_palignr<mode>): Likewise.
2695 (*abs<mode>2): Likewise.
2696 (sse4_2_pcmpestr): Likewise.
2697 (sse4_2_pcmpestri): Likewise.
2698 (sse4_2_pcmpestrm): Likewise.
2699 (sse4_2_pcmpestr_cconly): Likewise.
2700 (sse4_2_pcmpistr): Likewise.
2701 (sse4_2_pcmpistri): Likewise.
2702 (sse4_2_pcmpistrm): Likewise.
2703 (sse4_2_pcmpistr_cconly): Likewise.
2704 (vgf2p8affineinvqb_<mode><mask_name>): Likewise.
2705 (vgf2p8affineqb_<mode><mask_name>): Likewise.
2706 (vgf2p8mulb_<mode><mask_name>): Likewise.
2707 (*<code>v8hi3 [smaxmin]): Drop "prefix_data16" and
2709 (*<code>v16qi3 [umaxmin]): Likewise.
2711 2023-08-07 Jan Beulich <jbeulich@suse.com>
2713 * config/i386/i386.md (sse4_1_round<mode>2): Make
2714 "length_immediate" uniformly 1.
2715 * config/i386/mmx.md (mmx_pblendvb_v8qi): Likewise.
2716 (mmx_pblendvb_<mode>): Likewise.
2718 2023-08-07 Jan Beulich <jbeulich@suse.com>
2720 * config/i386/sse.md
2721 (<avx512>_<complexopname>_<mode><maskc_name><round_name>): Add
2723 (avx512fp16_<complexopname>sh_v8hf<mask_scalarc_name><round_scalarcz_name>):
2726 2023-08-07 Jan Beulich <jbeulich@suse.com>
2728 * config/i386/sse.md (xop_phadd<u>bw): Add "prefix",
2729 "prefix_extra", and "mode" attributes.
2730 (xop_phadd<u>bd): Likewise.
2731 (xop_phadd<u>bq): Likewise.
2732 (xop_phadd<u>wd): Likewise.
2733 (xop_phadd<u>wq): Likewise.
2734 (xop_phadd<u>dq): Likewise.
2735 (xop_phsubbw): Likewise.
2736 (xop_phsubwd): Likewise.
2737 (xop_phsubdq): Likewise.
2738 (xop_rotl<mode>3): Add "prefix" and "prefix_extra" attributes.
2739 (xop_rotr<mode>3): Likewise.
2740 (xop_frcz<mode>2): Likewise.
2741 (*xop_vmfrcz<mode>2): Likewise.
2742 (xop_vrotl<mode>3): Add "prefix" attribute. Change
2743 "prefix_extra" to 1.
2744 (xop_sha<mode>3): Likewise.
2745 (xop_shl<mode>3): Likewise.
2747 2023-08-07 Jan Beulich <jbeulich@suse.com>
2749 * config/i386/sse.md
2750 (*<avx512>_eq<mode>3<mask_scalar_merge_name>_1): Drop
2752 (avx512dq_vextract<shuffletype>64x2_1_mask): Likewise.
2753 (*avx512dq_vextract<shuffletype>64x2_1): Likewise.
2754 (avx512f_vextract<shuffletype>32x4_1_mask): Likewise.
2755 (*avx512f_vextract<shuffletype>32x4_1): Likewise.
2756 (vec_extract_lo_<mode>_mask [AVX512 forms]): Likewise.
2757 (vec_extract_lo_<mode> [AVX512 forms]): Likewise.
2758 (vec_extract_hi_<mode>_mask [AVX512 forms]): Likewise.
2759 (vec_extract_hi_<mode> [AVX512 forms]): Likewise.
2760 (@vec_extract_lo_<mode> [AVX512 forms]): Likewise.
2761 (@vec_extract_hi_<mode> [AVX512 forms]): Likewise.
2762 (vec_extract_lo_v64qi): Likewise.
2763 (vec_extract_hi_v64qi): Likewise.
2764 (*vec_widen_umult_even_v16si<mask_name>): Likewise.
2765 (*vec_widen_smult_even_v16si<mask_name>): Likewise.
2766 (*avx512f_<code><mode>3<mask_name>): Likewise.
2767 (*vec_extractv4ti): Likewise.
2768 (avx512bw_<code>v32qiv32hi2<mask_name>): Likewise.
2769 (<mask_codefor>avx512dq_broadcast<mode><mask_name>_1): Likewise.
2770 Add "length_immediate".
2772 2023-08-07 Jan Beulich <jbeulich@suse.com>
2774 * config/i386/i386.md (@rdrand<mode>): Add "prefix_0f". Drop
2776 (@rdseed<mode>): Likewise.
2777 * config/i386/mmx.md (<code><mode>3 [smaxmin and umaxmin cases]):
2778 Adjust "prefix_extra".
2779 * config/i386/sse.md (@vec_set<mode>_0): Likewise.
2780 (*sse4_1_<code><mode>3<mask_name>): Likewise.
2781 (*avx2_eq<mode>3): Likewise.
2782 (avx2_gt<mode>3): Likewise.
2783 (<sse2p4_1>_pinsr<ssemodesuffix>): Likewise.
2784 (*vec_extract<mode>): Likewise.
2785 (<vi8_sse4_1_avx2_avx512>_movntdqa): Likewise.
2787 2023-08-07 Jan Beulich <jbeulich@suse.com>
2789 * config/i386/i386.md (rd<fsgs>base<mode>): Add "prefix_0f" and
2790 "prefix_rep". Drop "prefix_extra".
2791 (wr<fsgs>base<mode>): Likewise.
2792 (ptwrite<mode>): Likewise.
2794 2023-08-07 Jan Beulich <jbeulich@suse.com>
2796 * config/i386/i386.md (isa): Move up.
2797 (length_immediate): Handle "fma4".
2798 (prefix): Handle "ssemuladd".
2799 * config/i386/sse.md (*fma_fmadd_<mode>): Add "prefix" attribute.
2800 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2802 (<avx512>_fmadd_<mode>_mask<round_name>): Likewise.
2803 (<avx512>_fmadd_<mode>_mask3<round_name>): Likewise.
2804 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>):
2806 (<avx512>_fmsub_<mode>_mask<round_name>): Likewise.
2807 (<avx512>_fmsub_<mode>_mask3<round_name>): Likewise.
2808 (*fma_fnmadd_<mode>): Likewise.
2809 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>):
2811 (<avx512>_fnmadd_<mode>_mask<round_name>): Likewise.
2812 (<avx512>_fnmadd_<mode>_mask3<round_name>): Likewise.
2813 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>):
2815 (<avx512>_fnmsub_<mode>_mask<round_name>): Likewise.
2816 (<avx512>_fnmsub_<mode>_mask3<round_name>): Likewise.
2817 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2819 (<avx512>_fmaddsub_<mode>_mask<round_name>): Likewise.
2820 (<avx512>_fmaddsub_<mode>_mask3<round_name>): Likewise.
2821 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2823 (<avx512>_fmsubadd_<mode>_mask<round_name>): Likewise.
2824 (<avx512>_fmsubadd_<mode>_mask3<round_name>): Likewise.
2825 (*fmai_fmadd_<mode>): Likewise.
2826 (*fmai_fmsub_<mode>): Likewise.
2827 (*fmai_fnmadd_<mode><round_name>): Likewise.
2828 (*fmai_fnmsub_<mode><round_name>): Likewise.
2829 (avx512f_vmfmadd_<mode>_mask<round_name>): Likewise.
2830 (avx512f_vmfmadd_<mode>_mask3<round_name>): Likewise.
2831 (avx512f_vmfmadd_<mode>_maskz_1<round_name>): Likewise.
2832 (*avx512f_vmfmsub_<mode>_mask<round_name>): Likewise.
2833 (avx512f_vmfmsub_<mode>_mask3<round_name>): Likewise.
2834 (*avx512f_vmfmsub_<mode>_maskz_1<round_name>): Likewise.
2835 (avx512f_vmfnmadd_<mode>_mask<round_name>): Likewise.
2836 (avx512f_vmfnmadd_<mode>_mask3<round_name>): Likewise.
2837 (avx512f_vmfnmadd_<mode>_maskz_1<round_name>): Likewise.
2838 (*avx512f_vmfnmsub_<mode>_mask<round_name>): Likewise.
2839 (*avx512f_vmfnmsub_<mode>_mask3<round_name>): Likewise.
2840 (*avx512f_vmfnmsub_<mode>_maskz_1<round_name>): Likewise.
2841 (*fma4i_vmfmadd_<mode>): Likewise.
2842 (*fma4i_vmfmsub_<mode>): Likewise.
2843 (*fma4i_vmfnmadd_<mode>): Likewise.
2844 (*fma4i_vmfnmsub_<mode>): Likewise.
2845 (fma_<complexopname>_<mode><sdc_maskz_name><round_name>): Likewise.
2846 (<avx512>_<complexopname>_<mode>_mask<round_name>): Likewise.
2847 (avx512fp16_fma_<complexopname>sh_v8hf<mask_scalarcz_name><round_scalarcz_name>):
2849 (avx512fp16_<complexopname>sh_v8hf_mask<round_name>): Likewise.
2850 (xop_p<macs><ssemodesuffix><ssemodesuffix>): Likewise.
2851 (xop_p<macs>dql): Likewise.
2852 (xop_p<macs>dqh): Likewise.
2853 (xop_p<macs>wd): Likewise.
2854 (xop_p<madcs>wd): Likewise.
2855 (fma_<complexpairopname>_<mode>_pair): Likewise. Add "mode" attribute.
2857 2023-08-07 Jan Beulich <jbeulich@suse.com>
2859 * config/i386/i386.md (length_immediate): Handle "sse4arg".
2861 (*xop_pcmov_<mode>): Add "mode" attribute.
2862 * config/i386/mmx.md (*xop_maskcmp<mode>3): Drop "prefix_data16",
2863 "prefix_rep", "prefix_extra", and "length_immediate" attributes.
2864 (*xop_maskcmp_uns<mode>3): Likewise. Switch "type" to "sse4arg".
2865 (*xop_pcmov_<mode>): Add "mode" attribute.
2866 * config/i386/sse.md (xop_pcmov_<mode><avxsizesuffix>): Add "mode"
2868 (xop_maskcmp<mode>3): Drop "prefix_data16", "prefix_rep",
2869 "prefix_extra", and "length_immediate" attributes.
2870 (xop_maskcmp_uns<mode>3): Likewise. Switch "type" to "sse4arg".
2871 (xop_maskcmp_uns2<mode>3): Drop "prefix_data16", "prefix_extra",
2872 and "length_immediate" attributes. Switch "type" to "sse4arg".
2873 (xop_pcom_tf<mode>3): Likewise.
2874 (xop_vpermil2<mode>3): Drop "length_immediate" attribute.
2876 2023-08-07 Jan Beulich <jbeulich@suse.com>
2878 * config/i386/i386.md (prefix_extra): Correct comment. Fold
2879 cases yielding 2 into ones yielding 1.
2881 2023-08-07 Jan Hubicka <jh@suse.cz>
2883 PR tree-optimization/106293
2884 * tree-vect-loop-manip.cc (vect_loop_versioning): Fix profile update.
2885 * tree-vect-loop.cc (vect_transform_loop): Likewise.
2887 2023-08-07 Andrew Pinski <apinski@marvell.com>
2889 PR tree-optimization/96695
2890 * match.pd (min_value, max_value): Extend to
2893 2023-08-06 Jan Hubicka <jh@suse.cz>
2895 * config/i386/cpuid.h (__get_cpuid_count, __get_cpuid_max): Add
2896 __builtin_expect that CPU likely supports cpuid.
2898 2023-08-06 Jan Hubicka <jh@suse.cz>
2900 * tree-loop-distribution.cc (loop_distribution::execute): Disable
2901 distribution for loops with estimated iterations 0.
2903 2023-08-06 Jan Hubicka <jh@suse.cz>
2905 * tree-vect-loop-manip.cc (vect_do_peeling): Fix profile update of peeled epilogues.
2907 2023-08-04 Xiao Zeng <zengxiao@eswincomputing.com>
2909 * config/riscv/riscv.cc (riscv_expand_conditional_move): Recognize
2910 more Zicond patterns. Fix whitespace typo.
2911 (riscv_rtx_costs): Remove accidental code duplication.
2912 Co-authored-by: Jeff Law <jlaw@ventanamicro.com>
2914 2023-08-04 Yan Simonaytes <simonaytes.yan@ispras.ru>
2917 * config/i386/i386-protos.h
2918 (vpternlog_redundant_operand_mask): Declare.
2919 (substitute_vpternlog_operands): Declare.
2920 * config/i386/i386.cc
2921 (vpternlog_redundant_operand_mask): New helper.
2922 (substitute_vpternlog_operands): New function. Use them...
2923 * config/i386/sse.md: ... here in new VPTERNLOG define_splits.
2925 2023-08-04 Roger Sayle <roger@nextmovesoftware.com>
2927 * expmed.cc (extract_bit_field_1): Document that an UNSIGNEDP
2928 value of -1 is equivalent to don't care.
2929 (extract_integral_bit_field): Indicate that we don't require
2930 the most significant word to be zero extended, if we're about
2932 (extract_fixed_bit_field_1): Document that an UNSIGNEDP value
2933 of -1 is equivalent to don't care. Don't clear the most
2934 significant bits with AND mask when UNSIGNEDP is -1.
2936 2023-08-04 Roger Sayle <roger@nextmovesoftware.com>
2938 * config/i386/sse.md (define_split): Convert highpart:DF extract
2939 from V2DFmode register into a sse2_storehpd instruction.
2940 (define_split): Likewise, convert lowpart:DF extract from V2DF
2941 register into a sse2_storelpd instruction.
2943 2023-08-04 Qing Zhao <qing.zhao@oracle.com>
2945 * doc/invoke.texi (-Wflex-array-member-not-at-end): Document
2948 2023-08-04 Vladimir N. Makarov <vmakarov@redhat.com>
2950 * lra-lives.cc (process_bb_lives): Check input insn pattern hard regs
2951 against early clobber hard regs.
2953 2023-08-04 Tamar Christina <tamar.christina@arm.com>
2955 * doc/extend.texi: Document it.
2957 2023-08-04 Tamar Christina <tamar.christina@arm.com>
2960 * config/aarch64/aarch64-simd.md (vec_widen_<sur>shiftl_lo_<mode>,
2961 vec_widen_<sur>shiftl_hi_<mode>): Remove.
2962 (aarch64_<sur>shll<mode>_internal): Renamed to...
2963 (aarch64_<su>shll<mode>): .. This.
2964 (aarch64_<sur>shll2<mode>_internal): Renamed to...
2965 (aarch64_<su>shll2<mode>): .. This.
2966 (aarch64_<sur>shll_n<mode>, aarch64_<sur>shll2_n<mode>): Re-use new
2968 * config/aarch64/constraints.md (D2, DL): New.
2969 * config/aarch64/predicates.md (aarch64_simd_shll_imm_vec): New.
2971 2023-08-04 Tamar Christina <tamar.christina@arm.com>
2973 * gensupport.cc (conlist): Support length 0 attribute.
2975 2023-08-04 Tamar Christina <tamar.christina@arm.com>
2977 * config/aarch64/aarch64.cc (aarch64_bool_compound_p): New.
2978 (aarch64_adjust_stmt_cost, aarch64_vector_costs::count_ops): Use it.
2980 2023-08-04 Tamar Christina <tamar.christina@arm.com>
2982 * config/aarch64/aarch64.cc (aarch64_multiply_add_p): Update handling
2984 (aarch64_adjust_stmt_cost): Use it.
2985 (aarch64_vector_costs::count_ops): Likewise.
2986 (aarch64_vector_costs::add_stmt_cost): Pass vinfo to
2987 aarch64_adjust_stmt_cost.
2989 2023-08-04 Richard Biener <rguenther@suse.de>
2991 PR tree-optimization/110838
2992 * tree-vect-patterns.cc (vect_recog_over_widening_pattern):
2993 Fix right-shift value sanitizing. Properly emit external
2994 def mangling in the preheader rather than in the pattern
2995 def sequence where it will fail vectorizing.
2997 2023-08-04 Matthew Malcomson <matthew.malcomson@arm.com>
2999 PR middle-end/110316
3001 * timevar.cc (NANOSEC_PER_SEC, TICKS_TO_NANOSEC,
3002 CLOCKS_TO_NANOSEC, nanosec_to_floating_sec, percent_of): New.
3003 (TICKS_TO_MSEC, CLOCKS_TO_MSEC): Remove these macros.
3004 (timer::validate_phases): Use integral arithmetic to check
3006 (timer::print_row, timer::print): Convert from integral
3007 nanoseconds to floating point seconds before printing.
3008 (timer::all_zero): Change limit to nanosec count instead of
3009 fractional count of seconds.
3010 (make_json_for_timevar_time_def): Convert from integral
3011 nanoseconds to floating point seconds before recording.
3012 * timevar.h (struct timevar_time_def): Update all measurements
3013 to use uint64_t nanoseconds rather than seconds stored in a
3016 2023-08-04 Richard Biener <rguenther@suse.de>
3018 PR tree-optimization/110838
3019 * match.pd (([rl]shift @0 out-of-bounds) -> zero): Restrict
3020 the arithmetic right-shift case to non-negative operands.
3022 2023-08-04 Pan Li <pan2.li@intel.com>
3025 2023-08-04 Pan Li <pan2.li@intel.com>
3027 * config/riscv/riscv-vector-builtins-bases.cc
3028 (class vfmacc_frm): New class for vfmacc frm.
3029 (vfmacc_frm_obj): New declaration.
3031 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
3032 * config/riscv/riscv-vector-builtins-functions.def
3033 (vfmacc_frm): New function definition.
3034 * config/riscv/riscv-vector-builtins.cc
3035 (function_expander::use_ternop_insn): Add frm operand support.
3036 * config/riscv/vector.md: Add vfmuladd to frm_mode.
3038 2023-08-04 Pan Li <pan2.li@intel.com>
3041 2023-08-04 Pan Li <pan2.li@intel.com>
3043 * config/riscv/riscv-vector-builtins-bases.cc
3044 (class vfnmacc_frm): New class for vfnmacc.
3045 (vfnmacc_frm_obj): New declaration.
3047 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
3048 * config/riscv/riscv-vector-builtins-functions.def
3049 (vfnmacc_frm): New function definition.
3051 2023-08-04 Pan Li <pan2.li@intel.com>
3054 2023-08-04 Pan Li <pan2.li@intel.com>
3056 * config/riscv/riscv-vector-builtins-bases.cc
3057 (class vfmsac_frm): New class for vfmsac frm.
3058 (vfmsac_frm_obj): New declaration.
3060 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
3061 * config/riscv/riscv-vector-builtins-functions.def
3062 (vfmsac_frm): New function definition.
3064 2023-08-04 Pan Li <pan2.li@intel.com>
3067 2023-08-04 Pan Li <pan2.li@intel.com>
3069 * config/riscv/riscv-vector-builtins-bases.cc
3070 (class vfnmsac_frm): New class for vfnmsac frm.
3071 (vfnmsac_frm_obj): New declaration.
3073 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
3074 * config/riscv/riscv-vector-builtins-functions.def
3075 (vfnmsac_frm): New function definition.
3077 2023-08-04 Georg-Johann Lay <avr@gjlay.de>
3079 * config/avr/avr-mcus.def (avr64dd14, avr64dd20, avr64dd28, avr64dd32)
3080 (avr64ea28, avr64ea32, avr64ea48, attiny424, attiny426, attiny427)
3081 (attiny824, attiny826, attiny827, attiny1624, attiny1626, attiny1627)
3082 (attiny3224, attiny3226, attiny3227, avr16dd14, avr16dd20, avr16dd28)
3083 (avr16dd32, avr32dd14, avr32dd20, avr32dd28, avr32dd32)
3084 (attiny102, attiny104): New devices.
3085 * doc/avr-mmcu.texi: Regenerate.
3087 2023-08-04 Georg-Johann Lay <avr@gjlay.de>
3089 * config/avr/avr-mcus.def (avr128d*, avr64d*): Fix their FLASH_SIZE
3090 and PM_OFFSET entries.
3092 2023-08-04 Andrew Pinski <apinski@marvell.com>
3094 PR tree-optimization/110874
3095 * gimple-match-head.cc (gimple_bit_not_with_nop): New declaration.
3096 (gimple_maybe_cmp): Likewise.
3097 (gimple_bitwise_inverted_equal_p): Rewrite to use gimple_bit_not_with_nop
3098 and gimple_maybe_cmp instead of being recursive.
3099 * match.pd (bit_not_with_nop): New match pattern.
3100 (maybe_cmp): Likewise.
3102 2023-08-04 Drew Ross <drross@redhat.com>
3104 PR middle-end/101955
3105 * match.pd ((signed x << c) >> c): New canonicalization.
3107 2023-08-04 Pan Li <pan2.li@intel.com>
3109 * config/riscv/riscv-vector-builtins-bases.cc
3110 (class vfnmsac_frm): New class for vfnmsac frm.
3111 (vfnmsac_frm_obj): New declaration.
3113 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
3114 * config/riscv/riscv-vector-builtins-functions.def
3115 (vfnmsac_frm): New function definition.
3117 2023-08-04 Pan Li <pan2.li@intel.com>
3119 * config/riscv/riscv-vector-builtins-bases.cc
3120 (class vfmsac_frm): New class for vfmsac frm.
3121 (vfmsac_frm_obj): New declaration.
3123 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
3124 * config/riscv/riscv-vector-builtins-functions.def
3125 (vfmsac_frm): New function definition.
3127 2023-08-04 Pan Li <pan2.li@intel.com>
3129 * config/riscv/riscv-vector-builtins-bases.cc
3130 (class vfnmacc_frm): New class for vfnmacc.
3131 (vfnmacc_frm_obj): New declaration.
3133 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
3134 * config/riscv/riscv-vector-builtins-functions.def
3135 (vfnmacc_frm): New function definition.
3137 2023-08-04 Hao Liu <hliu@os.amperecomputing.com>
3140 * config/aarch64/aarch64.cc (aarch64_force_single_cycle): check
3141 STMT_VINFO_REDUC_DEF to avoid failures in info_for_reduction.
3143 2023-08-04 Pan Li <pan2.li@intel.com>
3145 * config/riscv/riscv-vector-builtins-bases.cc
3146 (class vfmacc_frm): New class for vfmacc frm.
3147 (vfmacc_frm_obj): New declaration.
3149 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
3150 * config/riscv/riscv-vector-builtins-functions.def
3151 (vfmacc_frm): New function definition.
3152 * config/riscv/riscv-vector-builtins.cc
3153 (function_expander::use_ternop_insn): Add frm operand support.
3154 * config/riscv/vector.md: Add vfmuladd to frm_mode.
3156 2023-08-04 Pan Li <pan2.li@intel.com>
3158 * config/riscv/riscv-vector-builtins-bases.cc
3159 (vfwmul_frm_obj): New declaration.
3160 (vfwmul_frm): Ditto.
3161 * config/riscv/riscv-vector-builtins-bases.h:
3162 (vfwmul_frm): Ditto.
3163 * config/riscv/riscv-vector-builtins-functions.def
3164 (vfwmul_frm): New function definition.
3165 * config/riscv/vector.md: (frm_mode) Add vfwmul to frm_mode.
3167 2023-08-04 Pan Li <pan2.li@intel.com>
3169 * config/riscv/riscv-vector-builtins-bases.cc
3170 (binop_frm): New declaration.
3171 (reverse_binop_frm): Likewise.
3173 * config/riscv/riscv-vector-builtins-bases.h:
3174 (vfdiv_frm): New extern declaration.
3175 (vfrdiv_frm): Likewise.
3176 * config/riscv/riscv-vector-builtins-functions.def
3177 (vfdiv_frm): New function definition.
3178 (vfrdiv_frm): Likewise.
3179 * config/riscv/vector.md: Add vfdiv to frm_mode.
3181 2023-08-03 Jan Hubicka <jh@suse.cz>
3183 * tree-cfg.cc (print_loop_info): Print entry count.
3185 2023-08-03 Jan Hubicka <jh@suse.cz>
3187 * tree-ssa-loop-split.cc (split_loop): Update estimated iteration counts.
3189 2023-08-03 Jan Hubicka <jh@suse.cz>
3192 * cfgloopmanip.cc (scale_loop_profile): (Un)initialize
3193 unadjusted_exit_count.
3195 2023-08-03 Aldy Hernandez <aldyh@redhat.com>
3197 * ipa-prop.cc (ipa_compute_jump_functions_for_edge): Read global
3200 2023-08-03 Xiao Zeng <zengxiao@eswincomputing.com>
3202 * config/riscv/riscv.cc (riscv_expand_conditional_move): Recognize
3203 various Zicond patterns.
3204 * config/riscv/riscv.md (mov<mode>cc): Allow TARGET_ZICOND. Use
3205 sfb_alu_operand for both arms of the conditional move.
3206 Co-authored-by: Jeff Law <jlaw@ventanamicro.com>
3208 2023-08-03 Cupertino Miranda <cupertino.miranda@oracle.com>
3214 * config.gcc: Added core-builtins.cc and .o files.
3215 * config/bpf/bpf-passes.def: Removed file.
3216 * config/bpf/bpf-protos.h (bpf_add_core_reloc,
3217 bpf_replace_core_move_operands): New prototypes.
3218 * config/bpf/bpf.cc (enum bpf_builtins, is_attr_preserve_access,
3219 maybe_make_core_relo, bpf_core_field_info, bpf_core_compute,
3220 bpf_core_get_index, bpf_core_new_decl, bpf_core_walk,
3221 bpf_is_valid_preserve_field_info_arg, is_attr_preserve_access,
3222 handle_attr_preserve, pass_data_bpf_core_attr, pass_bpf_core_attr):
3224 (def_builtin, bpf_expand_builtin, bpf_resolve_overloaded_builtin): Changed.
3225 * config/bpf/bpf.md (define_expand mov<MM:mode>): Changed.
3226 (mov_reloc_core<mode>): Added.
3227 * config/bpf/core-builtins.cc (struct cr_builtin, enum
3228 cr_decision struct cr_local, struct cr_final, struct
3229 core_builtin_helpers, enum bpf_plugin_states): Added types.
3230 (builtins_data, core_builtin_helpers, core_builtin_type_defs):
3232 (allocate_builtin_data, get_builtin-data, search_builtin_data,
3233 remove_parser_plugin, compare_same_kind, compare_same_ptr_expr,
3234 compare_same_ptr_type, is_attr_preserve_access, core_field_info,
3235 bpf_core_get_index, compute_field_expr,
3236 pack_field_expr_for_access_index, pack_field_expr_for_preserve_field,
3237 process_field_expr, pack_enum_value, process_enum_value, pack_type,
3238 process_type, bpf_require_core_support, make_core_relo, read_kind,
3239 kind_access_index, kind_preserve_field_info, kind_enum_value,
3240 kind_type_id, kind_preserve_type_info, get_core_builtin_fndecl_for_type,
3241 bpf_handle_plugin_finish_type, bpf_init_core_builtins,
3242 construct_builtin_core_reloc, bpf_resolve_overloaded_core_builtin,
3243 bpf_expand_core_builtin, bpf_add_core_reloc,
3244 bpf_replace_core_move_operands): Added functions.
3245 * config/bpf/core-builtins.h (enum bpf_builtins): Added.
3246 (bpf_init_core_builtins, bpf_expand_core_builtin,
3247 bpf_resolve_overloaded_core_builtin): Added functions.
3248 * config/bpf/coreout.cc (struct bpf_core_extra): Added.
3249 (bpf_core_reloc_add, output_asm_btfext_core_reloc): Changed.
3250 * config/bpf/coreout.h (bpf_core_reloc_add) Changed prototype.
3251 * config/bpf/t-bpf: Added core-builtins.o.
3252 * doc/extend.texi: Added documentation for new BPF builtins.
3254 2023-08-03 Andrew MacLeod <amacleod@redhat.com>
3256 * gimple-range-fold.cc (fold_using_range::range_of_range_op): Add
3257 ranges to the call to relation_fold_and_or.
3258 (fold_using_range::relation_fold_and_or): Add op1 and op2 ranges.
3259 (fur_source::register_outgoing_edges): Add op1 and op2 ranges.
3260 * gimple-range-fold.h (relation_fold_and_or): Adjust params.
3261 * gimple-range-gori.cc (gori_compute::compute_operand_range): Add
3262 a varying op1 and op2 to call.
3263 * range-op-float.cc (range_operator::op1_op2_relation): New dafaults.
3264 (operator_equal::op1_op2_relation): New float version.
3265 (operator_not_equal::op1_op2_relation): Ditto.
3266 (operator_lt::op1_op2_relation): Ditto.
3267 (operator_le::op1_op2_relation): Ditto.
3268 (operator_gt::op1_op2_relation): Ditto.
3269 (operator_ge::op1_op2_relation) Ditto.
3270 * range-op-mixed.h (operator_equal::op1_op2_relation): New float
3272 (operator_not_equal::op1_op2_relation): Ditto.
3273 (operator_lt::op1_op2_relation): Ditto.
3274 (operator_le::op1_op2_relation): Ditto.
3275 (operator_gt::op1_op2_relation): Ditto.
3276 (operator_ge::op1_op2_relation): Ditto.
3277 * range-op.cc (range_op_handler::op1_op2_relation): Dispatch new
3279 (range_operator::op1_op2_relation): Add extra params.
3280 (operator_equal::op1_op2_relation): Ditto.
3281 (operator_not_equal::op1_op2_relation): Ditto.
3282 (operator_lt::op1_op2_relation): Ditto.
3283 (operator_le::op1_op2_relation): Ditto.
3284 (operator_gt::op1_op2_relation): Ditto.
3285 (operator_ge::op1_op2_relation): Ditto.
3286 * range-op.h (range_operator): New prototypes.
3287 (range_op_handler): Ditto.
3289 2023-08-03 Andrew MacLeod <amacleod@redhat.com>
3291 * gimple-range-gori.cc (gori_compute::compute_operand1_range):
3292 Use identity relation.
3293 (gori_compute::compute_operand2_range): Ditto.
3294 * value-relation.cc (get_identity_relation): New.
3295 * value-relation.h (get_identity_relation): New prototype.
3297 2023-08-03 Andrew MacLeod <amacleod@redhat.com>
3299 * value-range.h (Value_Range::set_varying): Set the type.
3300 (Value_Range::set_zero): Ditto.
3301 (Value_Range::set_nonzero): Ditto.
3303 2023-08-03 Jeff Law <jeffreyalaw@gmail.com>
3305 * config/riscv/riscv.cc (riscv_rtx_costs): Remove errant hunk from
3308 2023-08-03 Pan Li <pan2.li@intel.com>
3310 * config/riscv/riscv-vector-builtins-bases.cc: Add vfsub.
3312 2023-08-03 Richard Sandiford <richard.sandiford@arm.com>
3314 * poly-int.h (can_div_trunc_p): Succeed for more boundary conditions.
3316 2023-08-03 Richard Biener <rguenther@suse.de>
3318 PR tree-optimization/110838
3319 * tree-vect-patterns.cc (vect_recog_over_widening_pattern):
3320 Adjust the shift operand of RSHIFT_EXPRs.
3322 2023-08-03 Richard Biener <rguenther@suse.de>
3324 PR tree-optimization/110702
3325 * tree-ssa-loop-ivopts.cc (rewrite_use_address): When
3326 we created a NULL pointer based access rewrite that to
3329 2023-08-03 Richard Biener <rguenther@suse.de>
3331 * tree-ssa-sink.cc: Include tree-ssa-live.h.
3332 (pass_sink_code::execute): Instantiate virtual_operand_live
3334 (sink_code_in_bb): Pass down virtual_operand_live.
3335 (statement_sink_location): Get virtual_operand_live and
3336 verify we are not sinking loads across stores by looking up
3337 the live virtual operand at the sink location.
3339 2023-08-03 Richard Biener <rguenther@suse.de>
3341 * tree-ssa-live.h (class virtual_operand_live): New.
3342 * tree-ssa-live.cc (virtual_operand_live::init): New.
3343 (virtual_operand_live::get_live_in): Likewise.
3344 (virtual_operand_live::get_live_out): Likewise.
3346 2023-08-03 Richard Biener <rguenther@suse.de>
3348 * passes.def: Exchange loop splitting and final value
3351 2023-08-03 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
3353 * config/s390/s390.cc (expand_perm_as_a_vlbr_vstbr_candidate):
3354 New function which handles bswap patterns for vec_perm_const.
3355 (vectorize_vec_perm_const_1): Call new function.
3356 * config/s390/vector.md (*bswap<mode>): Fix operands in output
3358 (*vstbr<mode>): New insn.
3360 2023-08-03 Alexandre Oliva <oliva@adacore.com>
3362 * config/vxworks-smp.opt: New. Introduce -msmp.
3363 * config.gcc: Enable it on powerpc* vxworks prior to 7r*.
3364 * config/rs6000/vxworks.h (STARTFILE_PREFIX_SPEC): Choose
3365 lib_smp when -msmp is present in the command line.
3366 * doc/invoke.texi: Document it.
3368 2023-08-03 Yanzhang Wang <yanzhang.wang@intel.com>
3370 * config/riscv/riscv.cc (riscv_save_reg_p): Save ra for leaf
3371 when enabling -mno-omit-leaf-frame-pointer
3372 (riscv_option_override): Override omit-frame-pointer.
3373 (riscv_frame_pointer_required): Save s0 for non-leaf function
3374 (TARGET_FRAME_POINTER_REQUIRED): Override defination
3375 * config/riscv/riscv.opt: Add option support.
3377 2023-08-03 Roger Sayle <roger@nextmovesoftware.com>
3380 * config/i386/i386.md (<any_rotate>ti3): For rotations by 64 bits
3381 place operand in a register before gen_<insn>64ti2_doubleword.
3382 (<any_rotate>di3): Likewise, for rotations by 32 bits, place
3383 operand in a register before gen_<insn>32di2_doubleword.
3384 (<any_rotate>32di2_doubleword): Constrain operand to be in register.
3385 (<any_rotate>64ti2_doubleword): Likewise.
3387 2023-08-03 Pan Li <pan2.li@intel.com>
3389 * config/riscv/riscv-vector-builtins-bases.cc
3390 (vfmul_frm_obj): New declaration.
3392 * config/riscv/riscv-vector-builtins-bases.h: Likewise.
3393 * config/riscv/riscv-vector-builtins-functions.def
3394 (vfmul_frm): New function definition.
3395 * config/riscv/vector.md: Add vfmul to frm_mode.
3397 2023-08-03 Andrew Pinski <apinski@marvell.com>
3399 * match.pd (`~X & X`): Check that the types match.
3400 (`~x | x`, `~x ^ x`): Likewise.
3402 2023-08-03 Pan Li <pan2.li@intel.com>
3404 * config/riscv/riscv-vector-builtins-bases.h: Remove
3405 redudant declaration.
3407 2023-08-03 Pan Li <pan2.li@intel.com>
3409 * config/riscv/riscv-vector-builtins-bases.cc (BASE): Add
3411 * config/riscv/riscv-vector-builtins-bases.h: Add declaration.
3412 * config/riscv/riscv-vector-builtins-functions.def (vfwsub_frm):
3413 Add vfwsub function definitions.
3415 2023-08-02 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
3417 PR rtl-optimization/110867
3418 * combine.cc (simplify_compare_const): Try the optimization only
3419 in case the constant fits into the comparison mode.
3421 2023-08-02 Jeff Law <jlaw@ventanamicro.com>
3423 * config/riscv/zicond.md: Remove incorrect zicond patterns and
3424 renumber/rename them.
3425 (zero.nez.<GPR:MODE><X:mode>.opt2): Fix output string.
3427 2023-08-02 Richard Biener <rguenther@suse.de>
3429 * tree-phinodes.h (add_phi_node_to_bb): Remove.
3430 * tree-phinodes.cc (add_phi_node_to_bb): Make static.
3432 2023-08-02 Jan Beulich <jbeulich@suse.com>
3434 * config/i386/sse.md (vec_dupv2df<mask_name>): Fold the middle
3435 two of the alternatives.
3437 2023-08-02 Richard Biener <rguenther@suse.de>
3439 PR tree-optimization/92335
3440 * tree-ssa-sink.cc (select_best_block): Before loop
3441 optimizations avoid sinking unconditional loads/stores
3442 in innermost loops to conditional executed places.
3444 2023-08-02 Andrew Pinski <apinski@marvell.com>
3446 * gimple-match-head.cc (gimple_bitwise_inverted_equal_p): Valueize
3447 the comparison operands before comparing them.
3449 2023-08-02 Andrew Pinski <apinski@marvell.com>
3451 * match.pd (`~X & X`, `~X | X`): Move over to
3452 use bitwise_inverted_equal_p, removing :c as bitwise_inverted_equal_p
3453 handles that already.
3454 Remove range test simplifications to true/false as they
3455 are now handled by these patterns.
3457 2023-08-02 Andrew Pinski <apinski@marvell.com>
3459 * tree-ssa-phiopt.cc (match_simplify_replacement): Mark's cond
3460 statement's lhs and rhs to check if trivial dead.
3461 Rename inserted_exprs to exprs_maybe_dce; also move it so
3462 bitmap is not allocated if not needed.
3464 2023-08-02 Pan Li <pan2.li@intel.com>
3466 * config/riscv/riscv-vector-builtins-bases.cc
3467 (class widen_binop_frm): New class for binop frm.
3468 (BASE): Add vfwadd_frm.
3469 * config/riscv/riscv-vector-builtins-bases.h: New declaration.
3470 * config/riscv/riscv-vector-builtins-functions.def
3471 (vfwadd_frm): New function definition.
3472 * config/riscv/riscv-vector-builtins-shapes.cc
3473 (BASE_NAME_MAX_LEN): New macro.
3474 (struct alu_frm_def): Leverage new base class.
3475 (struct build_frm_base): New build base for frm.
3476 (struct widen_alu_frm_def): New struct for widen alu frm.
3477 (SHAPE): Add widen_alu_frm shape.
3478 * config/riscv/riscv-vector-builtins-shapes.h: New declaration.
3479 * config/riscv/vector.md (frm_mode): Add vfwalu type.
3481 2023-08-02 Jan Hubicka <jh@suse.cz>
3483 * cfgloop.h (loop_count_in): Declare.
3484 * cfgloopanal.cc (expected_loop_iterations_by_profile): Use count_in.
3485 (loop_count_in): Move here from ...
3486 * cfgloopmanip.cc (loop_count_in): ... here.
3487 (scale_loop_profile): Improve dumping; cast iteration bound to sreal.
3489 2023-08-02 Jan Hubicka <jh@suse.cz>
3491 * cfg.cc (scale_strictly_dominated_blocks): New function.
3492 * cfg.h (scale_strictly_dominated_blocks): Declare.
3493 * tree-cfg.cc (fold_loop_internal_call): Fixup CFG profile.
3495 2023-08-02 Richard Biener <rguenther@suse.de>
3497 PR rtl-optimization/110587
3498 * lra-spills.cc (return_regno_p): Remove.
3499 (regno_in_use_p): Likewise.
3500 (lra_final_code_change): Do not remove noop moves
3501 between hard registers.
3503 2023-08-02 liuhongt <hongtao.liu@intel.com>
3506 * config/i386/sse.md (vec_fmaddsub<mode>4): Extend to vector
3507 HFmode, use mode iterator VFH instead.
3508 (vec_fmsubadd<mode>4): Ditto.
3509 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
3510 Remove scalar mode from iterator, use VFH_AVX512VL instead.
3511 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
3514 2023-08-02 liuhongt <hongtao.liu@intel.com>
3516 * config/i386/sse.md (*avx2_lddqu_inserti_to_bcasti): New
3517 pre_reload define_insn_and_split.
3519 2023-08-02 Xiao Zeng <zengxiao@eswincomputing.com>
3521 * config/riscv/riscv.cc (riscv_rtx_costs): Add costing for
3522 using Zicond to implement some conditional moves.
3524 2023-08-02 Jeff Law <jlaw@ventanamicro.com>
3526 * config/riscv/zicond.md: Use the X iterator instead of ANYI
3527 on the comparison input operands.
3529 2023-08-02 Xiao Zeng <zengxiao@eswincomputing.com>
3531 * config/riscv/riscv.cc (riscv_rtx_costs, case IF_THEN_ELSE): Add
3533 (case SET): For INSNs that just set a REG, take the cost from the
3535 Co-authored-by: Jeff Law <jlaw@ventanamicro.com>
3537 2023-08-02 Hu, Lin1 <lin1.hu@intel.com>
3539 * common/config/i386/i386-common.cc (OPTION_MASK_ISA2_AMX_INT8_SET):
3540 Change OPTION_MASK_ISA2_AMX_TILE to OPTION_MASK_ISA2_AMX_TILE_SET.
3541 (OPTION_MASK_ISA2_AMX_BF16_SET): Ditto
3542 (OPTION_MASK_ISA2_AMX_FP16_SET): Ditto
3543 (OPTION_MASK_ISA2_AMX_COMPLEX_SET): Ditto
3544 (OPTION_MASK_ISA_ABM_SET):
3545 Change OPTION_MASK_ISA_POPCNT to OPTION_MASK_ISA_POPCNT_SET.
3547 2023-08-01 Andreas Krebbel <krebbel@linux.ibm.com>
3549 * config/s390/s390.cc (s390_encode_section_info): Assume external
3550 symbols without explicit alignment to be unaligned if
3551 -munaligned-symbols has been specified.
3552 * config/s390/s390.opt (-munaligned-symbols): New option.
3554 2023-08-01 Richard Ball <richard.ball@arm.com>
3556 * gimple-fold.cc (fold_ctor_reference):
3557 Add support for poly_int.
3559 2023-08-01 Georg-Johann Lay <avr@gjlay.de>
3562 * config/avr/avr.cc (avr_optimize_casesi): Set JUMP_LABEL and
3563 LABEL_NUSES of new conditional branch instruction.
3565 2023-08-01 Jan Hubicka <jh@suse.cz>
3567 * tree-vect-loop-manip.cc (vect_do_peeling): Fix profile update after
3568 constant prologue peeling.
3570 2023-08-01 Christophe Lyon <christophe.lyon@linaro.org>
3572 * doc/sourcebuild.texi (arm_v8_1m_main_cde_mve_fp): Fix spelling.
3574 2023-08-01 Pan Li <pan2.li@intel.com>
3575 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3577 * config/riscv/riscv.cc (DYNAMIC_FRM_RTL): New macro.
3578 (STATIC_FRM_P): Ditto.
3579 (struct mode_switching_info): New struct for mode switching.
3580 (struct machine_function): Add new field mode switching.
3581 (riscv_emit_frm_mode_set): Add DYN_CALL emit.
3582 (riscv_frm_adjust_mode_after_call): New function for call mode.
3583 (riscv_frm_emit_after_call_in_bb_end): New function for emit
3584 insn when call as the end of bb.
3585 (riscv_frm_mode_needed): New function for frm mode needed.
3586 (frm_unknown_dynamic_p): Remove call check.
3587 (riscv_mode_needed): Extrac function for frm.
3588 (riscv_frm_mode_after): Add DYN_CALL after.
3589 (riscv_mode_entry): Remove backup rtl initialization.
3590 * config/riscv/vector.md (frm_mode): Add dyn_call.
3591 (fsrmsi_restore_exit): Rename to _volatile.
3592 (fsrmsi_restore_volatile): Likewise.
3594 2023-08-01 Pan Li <pan2.li@intel.com>
3596 * config/riscv/riscv-vector-builtins-bases.cc
3597 (class reverse_binop_frm): Add new template for reversed frm.
3598 (vfsub_frm_obj): New obj.
3599 (vfrsub_frm_obj): Likewise.
3600 * config/riscv/riscv-vector-builtins-bases.h:
3601 (vfsub_frm): New declaration.
3602 (vfrsub_frm): Likewise.
3603 * config/riscv/riscv-vector-builtins-functions.def
3604 (vfsub_frm): New function define.
3605 (vfrsub_frm): Likewise.
3607 2023-08-01 Andrew Pinski <apinski@marvell.com>
3609 PR tree-optimization/93044
3610 * match.pd (nested int casts): A truncation (to the same size or smaller)
3611 can always remove the inner cast.
3613 2023-07-31 Hamza Mahfooz <someguy@effective-light.com>
3616 * doc/invoke.texi (-Wmissing-variable-declarations): Document
3619 2023-07-31 Andrew Pinski <apinski@marvell.com>
3621 PR tree-optimization/106164
3622 * match.pd (`a != b & a <= b`, `a != b & a >= b`,
3623 `a == b | a < b`, `a == b | a > b`): Handle these cases
3626 2023-07-31 Andrew Pinski <apinski@marvell.com>
3628 PR tree-optimization/106164
3629 * match.pd: Extend the `(X CMP1 CST1) AND/IOR (X CMP2 CST2)`
3630 patterns to support `(X CMP1 Y) AND/IOR (X CMP2 Y)`.
3632 2023-07-31 Andrew Pinski <apinski@marvell.com>
3634 PR tree-optimization/100864
3635 * generic-match-head.cc (bitwise_inverted_equal_p): New function.
3636 * gimple-match-head.cc (bitwise_inverted_equal_p): New macro.
3637 (gimple_bitwise_inverted_equal_p): New function.
3638 * match.pd ((~x | y) & x): Use bitwise_inverted_equal_p
3639 instead of direct matching bit_not.
3641 2023-07-31 Costas Argyris <costas.argyris@gmail.com>
3644 * gcc-ar.cc (main): Expand argv and use
3645 temporary response file to call ar if any
3646 expansions were made.
3648 2023-07-31 Andrew MacLeod <amacleod@redhat.com>
3650 PR tree-optimization/110582
3651 * gimple-range-fold.cc (fur_list::get_operand): Do not use the
3652 range vector for non-ssa names.
3654 2023-07-31 David Malcolm <dmalcolm@redhat.com>
3657 * diagnostic-client-data-hooks.h (class sarif_object): New forward
3659 (diagnostic_client_data_hooks::add_sarif_invocation_properties):
3661 * diagnostic-format-sarif.cc: Include "diagnostic-format-sarif.h".
3662 (class sarif_invocation): Inherit from sarif_object rather than
3664 (class sarif_result): Likewise.
3665 (class sarif_ice_notification): Likewise.
3666 (sarif_object::get_or_create_properties): New.
3667 (sarif_invocation::prepare_to_flush): Add "context" param. Use it
3668 to call the context's add_sarif_invocation_properties hook.
3669 (sarif_builder::flush_to_file): Pass m_context to
3670 sarif_invocation::prepare_to_flush.
3671 * diagnostic-format-sarif.h: New header.
3672 * doc/invoke.texi (Developer Options): Clarify that -ftime-report
3673 writes to stderr. Document that if SARIF diagnostic output is
3674 requested then any timing information is written in JSON form as
3675 part of the SARIF output, rather than to stderr.
3676 * timevar.cc: Include "json.h".
3677 (timer::named_items::m_hash_map): Split out type into...
3678 (timer::named_items::hash_map_t): ...this new typedef.
3679 (timer::named_items::make_json): New function.
3680 (timevar_diff): New function.
3681 (make_json_for_timevar_time_def): New function.
3682 (timer::timevar_def::make_json): New function.
3683 (timer::make_json): New function.
3684 * timevar.h (class json::value): New forward decl.
3685 (timer::make_json): New decl.
3686 (timer::timevar_def::make_json): New decl.
3687 * tree-diagnostic-client-data-hooks.cc: Include
3688 "diagnostic-format-sarif.h" and "timevar.h".
3689 (compiler_data_hooks::add_sarif_invocation_properties): New vfunc
3692 2023-07-31 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
3694 * combine.cc (simplify_compare_const): Narrow comparison of
3695 memory and constant.
3696 (try_combine): Adapt new function signature.
3697 (simplify_comparison): Adapt new function signature.
3699 2023-07-31 Kito Cheng <kito.cheng@sifive.com>
3701 * config/riscv/riscv-v.cc (expand_vec_series): Drop unused
3703 (expand_vector_init_insert_elems): Ditto.
3705 2023-07-31 Hao Liu <hliu@os.amperecomputing.com>
3708 * config/aarch64/aarch64.cc (count_ops): Only '* count' for
3709 single_defuse_cycle while counting reduction_latency.
3711 2023-07-31 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
3713 * internal-fn.def (DEF_INTERNAL_COND_FN): New macro.
3714 (DEF_INTERNAL_SIGNED_COND_FN): Ditto.
3735 (COND_LEN_ADD): Ditto.
3736 (COND_LEN_SUB): Ditto.
3737 (COND_LEN_MUL): Ditto.
3738 (COND_LEN_DIV): Ditto.
3739 (COND_LEN_MOD): Ditto.
3740 (COND_LEN_RDIV): Ditto.
3741 (COND_LEN_MIN): Ditto.
3742 (COND_LEN_MAX): Ditto.
3743 (COND_LEN_FMIN): Ditto.
3744 (COND_LEN_FMAX): Ditto.
3745 (COND_LEN_AND): Ditto.
3746 (COND_LEN_IOR): Ditto.
3747 (COND_LEN_XOR): Ditto.
3748 (COND_LEN_SHL): Ditto.
3749 (COND_LEN_SHR): Ditto.
3750 (COND_LEN_FMA): Ditto.
3751 (COND_LEN_FMS): Ditto.
3752 (COND_LEN_FNMA): Ditto.
3753 (COND_LEN_FNMS): Ditto.
3754 (COND_LEN_NEG): Ditto.
3755 (ADD): New macro define.
3776 2023-07-31 Roger Sayle <roger@nextmovesoftware.com>
3779 * config/i386/i386-features.cc (compute_convert_gain): Check
3780 TARGET_AVX512VL (not TARGET_AVX512F) when considering V2DImode
3781 and V4SImode rotates in STV.
3782 (general_scalar_chain::convert_rotate): Likewise.
3784 2023-07-31 Kito Cheng <kito.cheng@sifive.com>
3786 * config/riscv/autovec.md (abs<mode>2): Remove `.require ()`.
3787 * config/riscv/riscv-protos.h (get_mask_mode): Update return
3789 * config/riscv/riscv-v.cc (rvv_builder::rvv_builder): Remove
3791 (emit_vlmax_insn): Ditto.
3792 (emit_vlmax_fp_insn): Ditto.
3793 (emit_vlmax_ternary_insn): Ditto.
3794 (emit_vlmax_fp_ternary_insn): Ditto.
3795 (emit_nonvlmax_fp_ternary_tu_insn): Ditto.
3796 (emit_nonvlmax_insn): Ditto.
3797 (emit_vlmax_slide_insn): Ditto.
3798 (emit_nonvlmax_slide_tu_insn): Ditto.
3799 (emit_vlmax_merge_insn): Ditto.
3800 (emit_vlmax_masked_insn): Ditto.
3801 (emit_nonvlmax_masked_insn): Ditto.
3802 (emit_vlmax_masked_store_insn): Ditto.
3803 (emit_nonvlmax_masked_store_insn): Ditto.
3804 (emit_vlmax_masked_mu_insn): Ditto.
3805 (emit_nonvlmax_tu_insn): Ditto.
3806 (emit_nonvlmax_fp_tu_insn): Ditto.
3807 (emit_scalar_move_insn): Ditto.
3808 (emit_vlmax_compress_insn): Ditto.
3809 (emit_vlmax_reduction_insn): Ditto.
3810 (emit_vlmax_fp_reduction_insn): Ditto.
3811 (emit_nonvlmax_fp_reduction_insn): Ditto.
3812 (expand_vec_series): Ditto.
3813 (expand_vector_init_merge_repeating_sequence): Ditto.
3814 (expand_vec_perm): Ditto.
3815 (shuffle_merge_patterns): Ditto.
3816 (shuffle_compress_patterns): Ditto.
3817 (shuffle_decompress_patterns): Ditto.
3818 (expand_reduction): Ditto.
3819 (get_mask_mode): Update return type.
3820 * config/riscv/riscv.cc (riscv_get_mask_mode): Check vector type
3821 is valid, and use new get_mask_mode interface.
3823 2023-07-31 Pan Li <pan2.li@intel.com>
3825 * config/riscv/riscv-vector-builtins-shapes.cc (struct alu_frm_def):
3826 Move rm suffix before mask.
3828 2023-07-31 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3830 * config/riscv/autovec-vls.md (@vec_duplicate<mode>): New pattern.
3831 * config/riscv/riscv-v.cc (autovectorize_vector_modes): Add VLS autovec
3834 2023-07-29 Roger Sayle <roger@nextmovesoftware.com>
3837 * config/i386/i386.md (extv<mode>): Use QImode for offsets.
3838 (extzv<mode>): Likewise.
3839 (insv<mode>): Likewise.
3840 (*testqi_ext_3): Likewise.
3841 (*btr<mode>_2): Likewise.
3842 (define_split): Likewise.
3843 (*btsq_imm): Likewise.
3844 (*btrq_imm): Likewise.
3845 (*btcq_imm): Likewise.
3846 (define_peephole2 x3): Likewise.
3847 (*bt<mode>): Likewise
3848 (*bt<mode>_mask): New define_insn_and_split.
3849 (*jcc_bt<mode>): Use QImode for offsets.
3850 (*jcc_bt<mode>_1): Delete obsolete pattern.
3851 (*jcc_bt<mode>_mask): Use QImode offsets.
3852 (*jcc_bt<mode>_mask_1): Likewise.
3853 (define_split): Likewise.
3854 (*bt<mode>_setcqi): Likewise.
3855 (*bt<mode>_setncqi): Likewise.
3856 (*bt<mode>_setnc<mode>): Likewise.
3857 (*bt<mode>_setncqi_2): Likewise.
3858 (*bt<mode>_setc<mode>_mask): New define_insn_and_split.
3859 (bmi2_bzhi_<mode>3): Use QImode offsets.
3860 (*bmi2_bzhi_<mode>3): Likewise.
3861 (*bmi2_bzhi_<mode>3_1): Likewise.
3862 (*bmi2_bzhi_<mode>3_1_ccz): Likewise.
3863 (@tbm_bextri_<mode>): Likewise.
3865 2023-07-29 Jan Hubicka <jh@suse.cz>
3867 * profile-count.cc (profile_probability::sqrt): New member function.
3868 (profile_probability::pow): Likewise.
3869 * profile-count.h: (profile_probability::sqrt): Declare
3870 (profile_probability::pow): Likewise.
3871 * tree-vect-loop-manip.cc (vect_loop_versioning): Fix profile update.
3873 2023-07-28 Andrew MacLeod <amacleod@redhat.com>
3875 * gimple-range-cache.cc (ssa_cache::merge_range): New.
3876 (ssa_lazy_cache::merge_range): New.
3877 * gimple-range-cache.h (class ssa_cache): Adjust protoypes.
3878 (class ssa_lazy_cache): Ditto.
3879 * gimple-range.cc (assume_query::calculate_op): Use merge_range.
3881 2023-07-28 Andrew MacLeod <amacleod@redhat.com>
3883 * tree-ssa-propagate.cc (substitute_and_fold_engine::value_on_edge):
3884 Move from value-query.cc.
3885 (substitute_and_fold_engine::value_of_stmt): Ditto.
3886 (substitute_and_fold_engine::range_of_expr): New.
3887 * tree-ssa-propagate.h (substitute_and_fold_engine): Inherit from
3888 range_query. New prototypes.
3889 * value-query.cc (value_query::value_on_edge): Relocate.
3890 (value_query::value_of_stmt): Ditto.
3891 * value-query.h (class value_query): Remove.
3892 (class range_query): Remove base class. Adjust prototypes.
3894 2023-07-28 Andrew MacLeod <amacleod@redhat.com>
3896 PR tree-optimization/110205
3897 * gimple-range-cache.h (ranger_cache::m_estimate): Delete.
3898 * range-op-mixed.h (operator_bitwise_xor::op1_op2_relation_effect):
3900 * range-op.cc (operator_lshift): Add missing final overrides.
3901 (operator_rshift): Ditto.
3903 2023-07-28 Jose E. Marchesi <jose.marchesi@oracle.com>
3905 * config/bpf/bpf.cc (bpf_option_override): Disable tail-call
3906 optimizations in BPF target.
3908 2023-07-28 Honza <jh@ryzen4.suse.cz>
3910 * cfgloopmanip.cc (loop_count_in): Break out from ...
3911 (loop_exit_for_scaling): Break out from ...
3912 (update_loop_exit_probability_scale_dom_bbs): Break out from ...;
3913 add more sanity check and debug info.
3914 (scale_loop_profile): ... here.
3915 (create_empty_loop_on_edge): Fix whitespac.
3916 * cfgloopmanip.h (update_loop_exit_probability_scale_dom_bbs): Declare.
3917 * loop-unroll.cc (unroll_loop_constant_iterations): Use
3918 update_loop_exit_probability_scale_dom_bbs.
3919 * tree-ssa-loop-manip.cc (update_exit_probability_after_unrolling): Remove.
3920 (tree_transform_and_unroll_loop): Use
3921 update_loop_exit_probability_scale_dom_bbs.
3922 * tree-ssa-loop-split.cc (split_loop): Use
3923 update_loop_exit_probability_scale_dom_bbs.
3925 2023-07-28 Jan Hubicka <jh@suse.cz>
3928 * tree-ssa-loop-split.cc: Include value-query.h.
3929 (split_at_bb_p): Analyze cases where EQ/NE can be turned
3930 into LT/LE/GT/GE; return updated guard code.
3931 (split_loop): Use guard code.
3933 2023-07-28 Roger Sayle <roger@nextmovesoftware.com>
3934 Richard Biener <rguenther@suse.de>
3937 PR rtl-optimization/110587
3938 * expr.cc (emit_group_load_1): Simplify logic for calling
3939 force_reg on ORIG_SRC, to avoid making a copy if the source
3940 is already in a pseudo register.
3942 2023-07-28 Jan Hubicka <jh@suse.cz>
3944 PR middle-end/106923
3945 * tree-ssa-loop-split.cc (connect_loops): Change probability
3946 of the test preconditioning second loop to very_likely.
3947 (fix_loop_bb_probability): Handle correctly case where
3948 on of the arms of the conditional is empty.
3949 (split_loop): Fold the test guarding first condition to
3950 see if it is constant true; Set correct entry block
3951 probabilities of the split loops; determine correct loop
3954 2023-07-28 xuli <xuli1@eswincomputing.com>
3956 * config/riscv/riscv-vector-builtins-bases.cc: remove rounding mode of
3957 vsadd[u] and vssub[u].
3958 * config/riscv/vector.md: Ditto.
3960 2023-07-28 Jan Hubicka <jh@suse.cz>
3962 * tree-ssa-loop-split.cc (split_loop): Also support NE driven
3963 loops when IV test is not overflowing.
3965 2023-07-28 liuhongt <hongtao.liu@intel.com>
3968 * config/i386/sse.md (avx512cd_maskb_vec_dup<mode>): Add
3970 (avx512cd_maskw_vec_dup<mode>): Ditto.
3972 2023-07-27 David Faust <david.faust@oracle.com>
3976 * config/bpf/bpf.opt (msmov): New option.
3977 * config/bpf/bpf.cc (bpf_option_override): Handle it here.
3978 * config/bpf/bpf.md (*extendsidi2): New.
3984 * doc/invoke.texi (Option Summary): Add -msmov eBPF option.
3985 (eBPF Options): Add -m[no-]smov. Document that -mcpu=v4
3986 also enables -msmov.
3988 2023-07-27 David Faust <david.faust@oracle.com>
3990 * doc/invoke.texi (Option Summary): Remove -mkernel eBPF option.
3991 Add -mbswap and -msdiv eBPF options.
3992 (eBPF Options): Remove -mkernel. Add -mno-{jmpext, jmp32,
3993 alu32, v3-atomics, bswap, sdiv}. Document that -mcpu=v4 also
3996 2023-07-27 David Faust <david.faust@oracle.com>
3998 * config/bpf/bpf.md (add<AM:mode>3): Use %w2 instead of %w1
3999 in pseudo-C dialect output template.
4000 (sub<AM:mode>3): Likewise.
4002 2023-07-27 Jan Hubicka <jh@suse.cz>
4004 * tree-vect-loop.cc (optimize_mask_stores): Make store
4007 2023-07-27 Jan Hubicka <jh@suse.cz>
4009 * cfgloop.h (single_dom_exit): Declare.
4010 * cfgloopmanip.h (update_exit_probability_after_unrolling): Declare.
4011 * cfgrtl.cc (struct cfg_hooks): Fix comment.
4012 * loop-unroll.cc (unroll_loop_constant_iterations): Update exit edge.
4013 * tree-ssa-loop-ivopts.h (single_dom_exit): Do not declare it here.
4014 * tree-ssa-loop-manip.cc (update_exit_probability_after_unrolling):
4016 (tree_transform_and_unroll_loop): ... here;
4018 2023-07-27 Jan Hubicka <jh@suse.cz>
4020 * cfgloopmanip.cc (scale_dominated_blocks_in_loop): Move here from
4021 tree-ssa-loop-manip.cc and avoid recursion.
4022 (scale_loop_profile): Use scale_dominated_blocks_in_loop.
4023 (duplicate_loop_body_to_header_edge): Add DLTHE_FLAG_FLAT_PROFILE
4025 * cfgloopmanip.h (DLTHE_FLAG_FLAT_PROFILE): Define.
4026 (scale_dominated_blocks_in_loop): Declare.
4027 * predict.cc (dump_prediction): Do not ICE on uninitialized probability.
4028 (change_edge_frequency): Remove.
4029 * predict.h (change_edge_frequency): Remove.
4030 * tree-ssa-loop-manip.cc (scale_dominated_blocks_in_loop): Move to
4032 (niter_for_unrolled_loop): Remove.
4033 (tree_transform_and_unroll_loop): Fix profile update.
4035 2023-07-27 Jan Hubicka <jh@suse.cz>
4037 * tree-ssa-loop-im.cc (execute_sm_if_changed): Turn cap probability
4038 to guessed; fix count of new_bb.
4040 2023-07-27 Jan Hubicka <jh@suse.cz>
4042 * profile-count.h (profile_count::apply_probability): Fix
4043 handling of uninitialized probabilities, optimize scaling
4046 2023-07-27 Richard Biener <rguenther@suse.de>
4048 PR tree-optimization/91838
4049 * gimple-match-head.cc: Include attribs.h and asan.h.
4050 * generic-match-head.cc: Likewise.
4051 * match.pd (([rl]shift @0 out-of-bounds) -> zero): New pattern.
4053 2023-07-27 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4055 * config/riscv/riscv-modes.def (VECTOR_BOOL_MODE): Add VLS modes.
4056 (ADJUST_ALIGNMENT): Ditto.
4057 (ADJUST_PRECISION): Ditto.
4059 (VECTOR_MODE_WITH_PREFIX): Ditto.
4060 * config/riscv/riscv-opts.h (TARGET_VECTOR_VLS): New macro.
4061 * config/riscv/riscv-protos.h (riscv_v_ext_vls_mode_p): New function.
4062 * config/riscv/riscv-v.cc (INCLUDE_ALGORITHM): Add include.
4063 (legitimize_move): Enable basic VLS modes support.
4066 (get_vector_mode): Ditto.
4067 * config/riscv/riscv-vector-switch.def (VLS_ENTRY): Add vls modes.
4068 * config/riscv/riscv.cc (riscv_v_ext_vls_mode_p): New function.
4069 (VLS_ENTRY): New macro.
4070 (riscv_v_ext_mode_p): Add vls modes.
4071 (riscv_get_v_regno_alignment): New function.
4072 (riscv_print_operand): Add vls modes.
4073 (riscv_hard_regno_nregs): Ditto.
4074 (riscv_hard_regno_mode_ok): Ditto.
4075 (riscv_regmode_natural_size): Ditto.
4076 (riscv_vectorize_preferred_vector_alignment): Ditto.
4077 * config/riscv/riscv.md: Ditto.
4078 * config/riscv/vector-iterators.md: Ditto.
4079 * config/riscv/vector.md: Ditto.
4080 * config/riscv/autovec-vls.md: New file.
4082 2023-07-27 Pan Li <pan2.li@intel.com>
4084 * config/riscv/riscv_vector.h (enum RVV_CSR): Removed.
4086 (vwrite_csr): Ditto.
4088 2023-07-27 demin.han <demin.han@starfivetech.com>
4090 * config/riscv/autovec.md: Delete which_alternative use in split
4092 2023-07-27 Richard Biener <rguenther@suse.de>
4094 * tree-ssa-sink.cc (sink_code_in_bb): Remove recursion, instead
4096 (pass_sink_code::execute): ... in the caller.
4098 2023-07-27 Kewen Lin <linkw@linux.ibm.com>
4099 Richard Biener <rguenther@suse.de>
4101 PR tree-optimization/110776
4102 * tree-vect-stmts.cc (vectorizable_load): Always cost VMAT_ELEMENTWISE
4105 2023-07-26 Xiao Zeng <zengxiao@eswincomputing.com>
4107 * config/riscv/riscv.md: Include zicond.md
4108 * config/riscv/zicond.md: New file.
4110 2023-07-26 Xiao Zeng <zengxiao@eswincomputing.com>
4112 * common/config/riscv/riscv-common.cc: New extension.
4113 * config/riscv/riscv-opts.h (MASK_ZICOND): New mask.
4114 (TARGET_ZICOND): New target.
4116 2023-07-26 Carl Love <cel@us.ibm.com>
4118 * config/rs6000/rs6000-c.cc (find_instance): Add new parameter that
4119 specifies the number of built-in arguments to check.
4120 (altivec_resolve_overloaded_builtin): Update calls to find_instance
4121 to pass the number of built-in arguments to be checked.
4123 2023-07-26 David Faust <david.faust@oracle.com>
4125 * config/bpf/bpf.opt (mv3-atomics): New option.
4126 * config/bpf/bpf.cc (bpf_option_override): Handle it here.
4127 * config/bpf/bpf.h (enum_reg_class): Add R0 class.
4128 (REG_CLASS_NAMES): Likewise.
4129 (REG_CLASS_CONTENTS): Likewise.
4130 (REGNO_REG_CLASS): Handle R0.
4131 * config/bpf/bpf.md (UNSPEC_XADD): Rename to UNSPEC_AADD.
4132 (UNSPEC_AAND): New unspec.
4133 (UNSPEC_AOR): Likewise.
4134 (UNSPEC_AXOR): Likewise.
4135 (UNSPEC_AFADD): Likewise.
4136 (UNSPEC_AFAND): Likewise.
4137 (UNSPEC_AFOR): Likewise.
4138 (UNSPEC_AFXOR): Likewise.
4139 (UNSPEC_AXCHG): Likewise.
4140 (UNSPEC_ACMPX): Likewise.
4141 (atomic_add<mode>): Use UNSPEC_AADD and atomic type attribute.
4143 * config/bpf/atomic.md: ...Here. New file.
4144 * config/bpf/constraints.md (t): New constraint for R0.
4145 * doc/invoke.texi (eBPF Options): Document -mv3-atomics.
4147 2023-07-26 Matthew Malcomson <matthew.malcomson@arm.com>
4149 * tree-vect-stmts.cc (get_group_load_store_type): Reformat
4152 2023-07-26 Carl Love <cel@us.ibm.com>
4154 * config/rs6000/rs6000-builtins.def: Rename
4155 __builtin_altivec_vreplace_un_uv2di as __builtin_altivec_vreplace_un_udi
4156 __builtin_altivec_vreplace_un_uv4si as __builtin_altivec_vreplace_un_usi
4157 __builtin_altivec_vreplace_un_v2df as __builtin_altivec_vreplace_un_df
4158 __builtin_altivec_vreplace_un_v2di as __builtin_altivec_vreplace_un_di
4159 __builtin_altivec_vreplace_un_v4sf as __builtin_altivec_vreplace_un_sf
4160 __builtin_altivec_vreplace_un_v4si as __builtin_altivec_vreplace_un_si.
4161 Rename VREPLACE_UN_UV2DI as VREPLACE_UN_UDI, VREPLACE_UN_UV4SI as
4162 VREPLACE_UN_USI, VREPLACE_UN_V2DF as VREPLACE_UN_DF,
4163 VREPLACE_UN_V2DI as VREPLACE_UN_DI, VREPLACE_UN_V4SF as
4164 VREPLACE_UN_SF, VREPLACE_UN_V4SI as VREPLACE_UN_SI.
4165 Rename vreplace_un_v2di as vreplace_un_di, vreplace_un_v4si as
4166 vreplace_un_si, vreplace_un_v2df as vreplace_un_df,
4167 vreplace_un_v2di as vreplace_un_di, vreplace_un_v4sf as
4168 vreplace_un_sf, vreplace_un_v4si as vreplace_un_si.
4169 * config/rs6000/rs6000-c.cc (find_instance): Add case
4170 RS6000_OVLD_VEC_REPLACE_UN.
4171 * config/rs6000/rs6000-overload.def (__builtin_vec_replace_un):
4172 Fix first argument type. Rename VREPLACE_UN_UV4SI as
4173 VREPLACE_UN_USI, VREPLACE_UN_V4SI as VREPLACE_UN_SI,
4174 VREPLACE_UN_UV2DI as VREPLACE_UN_UDI, VREPLACE_UN_V2DI as
4175 VREPLACE_UN_DI, VREPLACE_UN_V4SF as VREPLACE_UN_SF,
4176 VREPLACE_UN_V2DF as VREPLACE_UN_DF.
4177 * config/rs6000/vsx.md (REPLACE_ELT): Rename the mode_iterator
4178 REPLACE_ELT_V for vector modes.
4179 (REPLACE_ELT): New scalar mode iterator.
4180 (REPLACE_ELT_char): Add scalar attributes.
4181 (vreplace_un_<mode>): Change iterator and mode attribute.
4183 2023-07-26 David Malcolm <dmalcolm@redhat.com>
4186 * Makefile.in (ANALYZER_OBJS): Add analyzer/symbol.o.
4188 2023-07-26 Richard Biener <rguenther@suse.de>
4190 PR tree-optimization/106081
4191 * tree-vect-slp.cc (vect_optimize_slp_pass::start_choosing_layouts):
4192 Assign layout -1 to splats.
4194 2023-07-26 Aldy Hernandez <aldyh@redhat.com>
4196 * range-op-mixed.h (class operator_cast): Add update_bitmask.
4197 * range-op.cc (operator_cast::update_bitmask): New.
4198 (operator_cast::fold_range): Call update_bitmask.
4200 2023-07-26 Li Xu <xuli1@eswincomputing.com>
4202 * config/riscv/riscv-vector-builtins.def (vfloat16mf4x2_t): Change
4203 scalar type to float16, eliminate warning.
4204 (vfloat16mf4x3_t): Ditto.
4205 (vfloat16mf4x4_t): Ditto.
4206 (vfloat16mf4x5_t): Ditto.
4207 (vfloat16mf4x6_t): Ditto.
4208 (vfloat16mf4x7_t): Ditto.
4209 (vfloat16mf4x8_t): Ditto.
4210 (vfloat16mf2x2_t): Ditto.
4211 (vfloat16mf2x3_t): Ditto.
4212 (vfloat16mf2x4_t): Ditto.
4213 (vfloat16mf2x5_t): Ditto.
4214 (vfloat16mf2x6_t): Ditto.
4215 (vfloat16mf2x7_t): Ditto.
4216 (vfloat16mf2x8_t): Ditto.
4217 (vfloat16m1x2_t): Ditto.
4218 (vfloat16m1x3_t): Ditto.
4219 (vfloat16m1x4_t): Ditto.
4220 (vfloat16m1x5_t): Ditto.
4221 (vfloat16m1x6_t): Ditto.
4222 (vfloat16m1x7_t): Ditto.
4223 (vfloat16m1x8_t): Ditto.
4224 (vfloat16m2x2_t): Ditto.
4225 (vfloat16m2x3_t): Ditto.
4226 (vfloat16m2x4_t): Ditto.
4227 (vfloat16m4x2_t): Ditto.
4228 * config/riscv/vector-iterators.md: add RVVM4x2DF in iterator V4T.
4229 * config/riscv/vector.md: add tuple mode in attr sew.
4231 2023-07-26 Uros Bizjak <ubizjak@gmail.com>
4234 * config/i386/i386.md (plusminusmult): New code iterator.
4235 * config/i386/mmx.md (mmxdoublevecmode): New mode attribute.
4236 (movq_<mode>_to_sse): New expander.
4237 (<plusminusmult:insn>v2sf3): Macroize expander from addv2sf3,
4238 subv2sf3 and mulv2sf3 using plusminusmult code iterator. Rewrite
4239 as a wrapper around V4SFmode operation.
4240 (mmx_addv2sf3): Change operand 1 and operand 2 predicates to
4241 nonimmediate_operand.
4242 (*mmx_addv2sf3): Remove SSE alternatives. Change operand 1 and
4243 operand 2 predicates to nonimmediate_operand.
4244 (mmx_subv2sf3): Change operand 2 predicate to nonimmediate_operand.
4245 (mmx_subrv2sf3): Change operand 1 predicate to nonimmediate_operand.
4246 (*mmx_subv2sf3): Remove SSE alternatives. Change operand 1 and
4247 operand 2 predicates to nonimmediate_operand.
4248 (mmx_mulv2sf3): Change operand 1 and operand 2 predicates to
4249 nonimmediate_operand.
4250 (*mmx_mulv2sf3): Remove SSE alternatives. Change operand 1 and
4251 operand 2 predicates to nonimmediate_operand.
4252 (divv2sf3): Rewrite as a wrapper around V4SFmode operation.
4253 (<smaxmin:code>v2sf3): Ditto.
4254 (mmx_<smaxmin:code>v2sf3): Change operand 1 and operand 2
4255 predicates to nonimmediate_operand.
4256 (*mmx_<smaxmin:code>v2sf3): Remove SSE alternatives. Change
4257 operand 1 and operand 2 predicates to nonimmediate_operand.
4258 (mmx_ieee_<ieee_maxmin>v2sf3): Ditto.
4259 (sqrtv2sf2): Rewrite as a wrapper around V4SFmode operation.
4260 (*mmx_haddv2sf3_low): Ditto.
4261 (*mmx_hsubv2sf3_low): Ditto.
4262 (vec_addsubv2sf3): Ditto.
4263 (*mmx_maskcmpv2sf3_comm): Remove.
4264 (*mmx_maskcmpv2sf3): Remove.
4265 (vec_cmpv2sfv2si): Rewrite as a wrapper around V4SFmode operation.
4266 (vcond<V2FI:mode>v2sf): Ditto.
4271 (fix_truncv2sfv2si2): Ditto.
4272 (fixuns_truncv2sfv2si2): Ditto.
4273 (mmx_fix_truncv2sfv2si2): Remove SSE alternatives.
4274 Change operand 1 predicate to nonimmediate_operand.
4275 (floatv2siv2sf2): Rewrite as a wrapper around V4SFmode operation.
4276 (floatunsv2siv2sf2): Ditto.
4277 (mmx_floatv2siv2sf2): Remove SSE alternatives.
4278 Change operand 1 predicate to nonimmediate_operand.
4279 (nearbyintv2sf2): Rewrite as a wrapper around V4SFmode operation.
4281 (lrintv2sfv2si2): Ditto.
4283 (lceilv2sfv2si2): Ditto.
4284 (floorv2sf2): Ditto.
4285 (lfloorv2sfv2si2): Ditto.
4286 (btruncv2sf2): Ditto.
4287 (roundv2sf2): Ditto.
4288 (lroundv2sfv2si2): Ditto.
4289 (*mmx_roundv2sf2): Remove.
4291 2023-07-26 Jose E. Marchesi <jose.marchesi@oracle.com>
4293 * config/bpf/bpf.md: Fix neg{SI,DI}2 insn.
4295 2023-07-26 Richard Biener <rguenther@suse.de>
4297 PR tree-optimization/110799
4298 * tree-ssa-pre.cc (compute_avail): More thoroughly match
4299 up TBAA behavior of redundant loads.
4301 2023-07-26 Jakub Jelinek <jakub@redhat.com>
4303 PR tree-optimization/110755
4304 * range-op-float.cc (frange_arithmetic): Change +0 result to -0
4305 for PLUS_EXPR or MINUS_EXPR if -frounding-math, inf is negative and
4306 it is exact op1 + (-op1) or op1 - op1.
4308 2023-07-26 Kewen Lin <linkw@linux.ibm.com>
4311 * config/rs6000/vsx.md (define_insn xxeval): Correct vsx
4312 operands output with "x".
4314 2023-07-26 Aldy Hernandez <aldyh@redhat.com>
4316 * range-op.cc (class operator_absu): Add update_bitmask.
4317 (operator_absu::update_bitmask): New.
4319 2023-07-26 Aldy Hernandez <aldyh@redhat.com>
4321 * range-op-mixed.h (class operator_abs): Add update_bitmask.
4322 * range-op.cc (operator_abs::update_bitmask): New.
4324 2023-07-26 Aldy Hernandez <aldyh@redhat.com>
4326 * range-op-mixed.h (class operator_bitwise_not): Add update_bitmask.
4327 * range-op.cc (operator_bitwise_not::update_bitmask): New.
4329 2023-07-26 Aldy Hernandez <aldyh@redhat.com>
4331 * range-op.cc (update_known_bitmask): Handle unary operators.
4333 2023-07-26 Aldy Hernandez <aldyh@redhat.com>
4335 * tree-ssa-ccp.cc (bit_value_unop): Initialize val when appropriate.
4337 2023-07-26 Jin Ma <jinma@linux.alibaba.com>
4339 * config/riscv/riscv.md: Likewise.
4341 2023-07-26 Jan Hubicka <jh@suse.cz>
4343 * profile-count.cc (profile_count::to_sreal_scale): Value is not know
4344 if we divide by zero.
4346 2023-07-25 David Faust <david.faust@oracle.com>
4348 * config/bpf/bpf.cc (bpf_print_operand_address): Don't print
4349 enclosing parentheses for pseudo-C dialect.
4350 * config/bpf/bpf.md (zero_exdendhidi2): Add parentheses around
4351 operands of pseudo-C dialect output templates where needed.
4352 (zero_extendqidi2): Likewise.
4353 (zero_extendsidi2): Likewise.
4354 (*mov<MM:mode>): Likewise.
4356 2023-07-25 Aldy Hernandez <aldyh@redhat.com>
4358 * tree-ssa-ccp.cc (value_mask_to_min_max): Make static.
4359 (bit_value_mult_const): Same.
4360 (get_individual_bits): Same.
4362 2023-07-25 Haochen Gui <guihaoc@gcc.gnu.org>
4365 * config/rs6000/rs6000-builtin.cc (rs6000_gimple_fold_builtin): Gimple
4366 fold RS6000_BIF_XSMINDP and RS6000_BIF_XSMAXDP when fast-math is set.
4367 * config/rs6000/rs6000.md (FMINMAX): New int iterator.
4368 (minmax_op): New int attribute.
4369 (UNSPEC_FMAX, UNSPEC_FMIN): New unspecs.
4370 (f<minmax_op><mode>3): New pattern by UNSPEC_FMAX and UNSPEC_FMIN.
4371 * config/rs6000/rs6000-builtins.def (__builtin_vsx_xsmaxdp): Set
4373 (__builtin_vsx_xsmindp): Set pattern to fmindf3.
4375 2023-07-24 David Faust <david.faust@oracle.com>
4377 * config/bpf/bpf.md (nop): Add pseudo-c asm dialect template.
4379 2023-07-24 Drew Ross <drross@redhat.com>
4380 Jakub Jelinek <jakub@redhat.com>
4382 PR middle-end/109986
4383 * generic-match-head.cc (bitwise_equal_p): New macro.
4384 * gimple-match-head.cc (bitwise_equal_p): New macro.
4385 (gimple_nop_convert): Declare.
4386 (gimple_bitwise_equal_p): Helper for bitwise_equal_p.
4387 * match.pd ((~X | Y) ^ X -> ~(X & Y)): New simplification.
4389 2023-07-24 Jeff Law <jlaw@ventanamicro.com>
4391 * common/config/riscv/riscv-common.cc (riscv_subset_list::add): Use
4392 single quote rather than backquote in diagnostic.
4394 2023-07-24 Jose E. Marchesi <jose.marchesi@oracle.com>
4397 * config/bpf/bpf.opt: New command-line option -msdiv.
4398 * config/bpf/bpf.md: Conditionalize sdiv/smod on bpf_has_sdiv.
4399 * config/bpf/bpf.cc (bpf_option_override): Initialize
4401 * doc/invoke.texi (eBPF Options): Document -msdiv.
4403 2023-07-24 Jeff Law <jlaw@ventanamicro.com>
4405 * config/riscv/riscv.cc (riscv_option_override): Spell out
4406 greater than and use cannot in diagnostic string.
4408 2023-07-24 Richard Biener <rguenther@suse.de>
4410 * tree-vectorizer.h (_slp_tree::push_vec_def): Add.
4411 (_slp_tree::vec_stmts): Remove.
4412 (SLP_TREE_VEC_STMTS): Remove.
4413 * tree-vect-slp.cc (_slp_tree::push_vec_def): Define.
4414 (_slp_tree::_slp_tree): Adjust.
4415 (_slp_tree::~_slp_tree): Likewise.
4416 (vect_get_slp_vect_def): Simplify.
4417 (vect_get_slp_defs): Likewise.
4418 (vect_transform_slp_perm_load_1): Adjust.
4419 (vect_add_slp_permutation): Likewise.
4420 (vect_schedule_slp_node): Likewise.
4421 (vectorize_slp_instance_root_stmt): Likewise.
4422 (vect_schedule_scc): Likewise.
4423 * tree-vect-stmts.cc (vectorizable_bswap): Use push_vec_def.
4424 (vectorizable_call): Likewise.
4425 (vectorizable_call): Likewise.
4426 (vect_create_vectorized_demotion_stmts): Likewise.
4427 (vectorizable_conversion): Likewise.
4428 (vectorizable_assignment): Likewise.
4429 (vectorizable_shift): Likewise.
4430 (vectorizable_operation): Likewise.
4431 (vectorizable_load): Likewise.
4432 (vectorizable_condition): Likewise.
4433 (vectorizable_comparison): Likewise.
4434 * tree-vect-loop.cc (vect_create_epilog_for_reduction): Adjust.
4435 (vectorize_fold_left_reduction): Use push_vec_def.
4436 (vect_transform_reduction): Likewise.
4437 (vect_transform_cycle_phi): Likewise.
4438 (vectorizable_lc_phi): Likewise.
4439 (vectorizable_phi): Likewise.
4440 (vectorizable_recurr): Likewise.
4441 (vectorizable_induction): Likewise.
4442 (vectorizable_live_operation): Likewise.
4444 2023-07-24 Richard Biener <rguenther@suse.de>
4446 * tree-ssa-loop.cc: Remove unused tree-vectorizer.h include.
4448 2023-07-24 Richard Biener <rguenther@suse.de>
4450 * config/i386/i386-builtins.cc: Remove tree-vectorizer.h include.
4451 * config/i386/i386-expand.cc: Likewise.
4452 * config/i386/i386-features.cc: Likewise.
4453 * config/i386/i386-options.cc: Likewise.
4455 2023-07-24 Robin Dapp <rdapp@ventanamicro.com>
4457 * tree-vect-stmts.cc (vectorizable_conversion): Handle
4458 more demotion/promotion for modifier == NONE.
4460 2023-07-24 Roger Sayle <roger@nextmovesoftware.com>
4465 * config/i386/i386.md (extv<mode>): Use QImode for offsets.
4466 (extzv<mode>): Likewise.
4467 (insv<mode>): Likewise.
4468 (*testqi_ext_3): Likewise.
4469 (*btr<mode>_2): Likewise.
4470 (define_split): Likewise.
4471 (*btsq_imm): Likewise.
4472 (*btrq_imm): Likewise.
4473 (*btcq_imm): Likewise.
4474 (define_peephole2 x3): Likewise.
4475 (*bt<mode>): Likewise
4476 (*bt<mode>_mask): New define_insn_and_split.
4477 (*jcc_bt<mode>): Use QImode for offsets.
4478 (*jcc_bt<mode>_1): Delete obsolete pattern.
4479 (*jcc_bt<mode>_mask): Use QImode offsets.
4480 (*jcc_bt<mode>_mask_1): Likewise.
4481 (define_split): Likewise.
4482 (*bt<mode>_setcqi): Likewise.
4483 (*bt<mode>_setncqi): Likewise.
4484 (*bt<mode>_setnc<mode>): Likewise.
4485 (*bt<mode>_setncqi_2): Likewise.
4486 (*bt<mode>_setc<mode>_mask): New define_insn_and_split.
4487 (bmi2_bzhi_<mode>3): Use QImode offsets.
4488 (*bmi2_bzhi_<mode>3): Likewise.
4489 (*bmi2_bzhi_<mode>3_1): Likewise.
4490 (*bmi2_bzhi_<mode>3_1_ccz): Likewise.
4491 (@tbm_bextri_<mode>): Likewise.
4493 2023-07-24 Jose E. Marchesi <jose.marchesi@oracle.com>
4495 * config/bpf/bpf-opts.h (enum bpf_kernel_version): Remove enum.
4496 * config/bpf/bpf.opt (mkernel): Remove option.
4497 * config/bpf/bpf.cc (bpf_target_macros): Do not define
4498 BPF_KERNEL_VERSION_CODE.
4500 2023-07-24 Jose E. Marchesi <jose.marchesi@oracle.com>
4503 * config/bpf/bpf.opt (mcpu): Add ISA_V4 and make it the default.
4504 (mbswap): New option.
4505 * config/bpf/bpf-opts.h (enum bpf_isa_version): New value ISA_V4.
4506 * config/bpf/bpf.cc (bpf_option_override): Set bpf_has_bswap.
4507 * config/bpf/bpf.md: Use bswap instructions if available for
4508 bswap* insn, and fix constraint.
4509 * doc/invoke.texi (eBPF Options): Document -mcpu=v4 and -mbswap.
4511 2023-07-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4513 * config/riscv/autovec.md (fold_left_plus_<mode>): New pattern.
4514 (mask_len_fold_left_plus_<mode>): Ditto.
4515 * config/riscv/riscv-protos.h (enum insn_type): New enum.
4516 (enum reduction_type): Ditto.
4517 (expand_reduction): Add in-order reduction.
4518 * config/riscv/riscv-v.cc (emit_nonvlmax_fp_reduction_insn): New function.
4519 (expand_reduction): Add in-order reduction.
4521 2023-07-24 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
4523 * tree-vect-loop.cc (get_masked_reduction_fn): Add mask_len_fold_left_plus.
4524 (vectorize_fold_left_reduction): Ditto.
4525 (vectorizable_reduction): Ditto.
4526 (vect_transform_reduction): Ditto.
4528 2023-07-24 Richard Biener <rguenther@suse.de>
4530 PR tree-optimization/110777
4531 * tree-ssa-sccvn.cc (eliminate_dom_walker::eliminate_avail):
4532 Avoid propagating abnormals.
4534 2023-07-24 Richard Biener <rguenther@suse.de>
4536 PR tree-optimization/110766
4537 * tree-scalar-evolution.cc
4538 (analyze_and_compute_bitwise_induction_effect): Check the PHI
4539 is defined in the loop header.
4541 2023-07-24 Kewen Lin <linkw@linux.ibm.com>
4543 PR tree-optimization/110740
4544 * tree-vect-loop.cc (vect_analyze_loop_costing): Do not vectorize a
4545 loop with a single scalar iteration.
4547 2023-07-24 Pan Li <pan2.li@intel.com>
4549 * config/riscv/riscv-vector-builtins-shapes.cc
4550 (struct alu_frm_def): Take range check.
4552 2023-07-22 Vineet Gupta <vineetg@rivosinc.com>
4555 * config/riscv/predicates.md (const_0_operand): Add back
4558 2023-07-22 Roger Sayle <roger@nextmovesoftware.com>
4560 * config/i386/i386-expand.cc (ix86_expand_move): Disable the
4561 64-bit insertions into TImode optimizations with -O0, unless
4562 the function has the "naked" attribute (for PR target/110533).
4564 2023-07-22 Andrew Pinski <apinski@marvell.com>
4567 * rtl.h (extended_count): Change last argument type
4570 2023-07-22 Roger Sayle <roger@nextmovesoftware.com>
4572 * config/i386/i386.md (extv<mode>): Use QImode for offsets.
4573 (extzv<mode>): Likewise.
4574 (insv<mode>): Likewise.
4575 (*testqi_ext_3): Likewise.
4576 (*btr<mode>_2): Likewise.
4577 (define_split): Likewise.
4578 (*btsq_imm): Likewise.
4579 (*btrq_imm): Likewise.
4580 (*btcq_imm): Likewise.
4581 (define_peephole2 x3): Likewise.
4582 (*bt<mode>): Likewise
4583 (*bt<mode>_mask): New define_insn_and_split.
4584 (*jcc_bt<mode>): Use QImode for offsets.
4585 (*jcc_bt<mode>_1): Delete obsolete pattern.
4586 (*jcc_bt<mode>_mask): Use QImode offsets.
4587 (*jcc_bt<mode>_mask_1): Likewise.
4588 (define_split): Likewise.
4589 (*bt<mode>_setcqi): Likewise.
4590 (*bt<mode>_setncqi): Likewise.
4591 (*bt<mode>_setnc<mode>): Likewise.
4592 (*bt<mode>_setncqi_2): Likewise.
4593 (*bt<mode>_setc<mode>_mask): New define_insn_and_split.
4594 (bmi2_bzhi_<mode>3): Use QImode offsets.
4595 (*bmi2_bzhi_<mode>3): Likewise.
4596 (*bmi2_bzhi_<mode>3_1): Likewise.
4597 (*bmi2_bzhi_<mode>3_1_ccz): Likewise.
4598 (@tbm_bextri_<mode>): Likewise.
4600 2023-07-22 Jeff Law <jlaw@ventanamicro.com>
4602 * config/bfin/bfin.md (ones): Fix length computation.
4604 2023-07-22 Vladimir N. Makarov <vmakarov@redhat.com>
4606 * lra-eliminations.cc (update_reg_eliminate): Fix the assert.
4607 (lra_update_fp2sp_elimination): Use HARD_FRAME_POINTER_REGNUM
4608 instead of FRAME_POINTER_REGNUM to spill pseudos.
4610 2023-07-21 Roger Sayle <roger@nextmovesoftware.com>
4611 Richard Biener <rguenther@suse.de>
4614 * gimplify.cc (gimplify_compound_lval): If the array's type
4615 is error_mark_node then return GS_ERROR.
4617 2023-07-21 Cupertino Miranda <cupertino.miranda@oracle.com>
4620 * config/bpf/bpf.opt: Added option -masm=<dialect>.
4621 * config/bpf/bpf-opts.h (enum bpf_asm_dialect): New type.
4622 * config/bpf/bpf.cc (bpf_print_register): New function.
4623 (bpf_print_register): Support pseudo-c syntax for registers.
4624 (bpf_print_operand_address): Likewise.
4625 * config/bpf/bpf.h (ASM_SPEC): handle -msasm.
4626 (ASSEMBLER_DIALECT): Define.
4627 * config/bpf/bpf.md: Added pseudo-c templates.
4628 * doc/invoke.texi (-masm=): New eBPF option item.
4630 2023-07-21 Cupertino Miranda <cupertino.miranda@oracle.com>
4632 * config/bpf/bpf.md: fixed template for neg instruction.
4634 2023-07-21 Jan Hubicka <jh@suse.cz>
4637 * tree-vect-loop.cc (scale_profile_for_vect_loop): Avoid scaling flat
4638 profiles by vectorization factor.
4639 (vect_transform_loop): Check for flat profiles.
4641 2023-07-21 Jan Hubicka <jh@suse.cz>
4643 * cfgloop.h (maybe_flat_loop_profile): Declare
4644 * cfgloopanal.cc (maybe_flat_loop_profile): New function.
4645 * tree-cfg.cc (print_loop_info): Print info about flat profiles.
4647 2023-07-21 Jan Hubicka <jh@suse.cz>
4649 * cfgloop.cc (get_estimated_loop_iterations): Use sreal::to_nearest_int
4650 * cfgloopanal.cc (expected_loop_iterations_unbounded): Likewise.
4651 * predict.cc (estimate_bb_frequencies): Likewise.
4652 * profile.cc (branch_prob): Likewise.
4653 * tree-ssa-loop-niter.cc (estimate_numbers_of_iterations): Likewise
4655 2023-07-21 Iain Sandoe <iain@sandoe.co.uk>
4657 * config.in: Regenerate.
4658 * config/darwin.h (DARWIN_LD_DEMANGLE): New.
4659 (LINK_COMMAND_SPEC_A): Add demangle handling.
4660 * configure: Regenerate.
4661 * configure.ac: Detect linker support for '-demangle'.
4663 2023-07-21 Jan Hubicka <jh@suse.cz>
4665 * sreal.cc (sreal::to_nearest_int): New.
4666 (sreal_verify_basics): Verify also to_nearest_int.
4667 (verify_aritmetics): Likewise.
4668 (sreal_verify_conversions): New.
4669 (sreal_cc_tests): Call sreal_verify_conversions.
4670 * sreal.h: (sreal::to_nearest_int): Declare
4672 2023-07-21 Jan Hubicka <jh@suse.cz>
4674 * tree-ssa-loop-ch.cc (enum ch_decision): New enum.
4675 (should_duplicate_loop_header_p): Return info on profitability.
4676 (do_while_loop_p): Watch for constant conditionals.
4677 (update_profile_after_ch): Do not sanity check that all
4678 static exits are taken.
4679 (ch_base::copy_headers): Run on all loops.
4680 (pass_ch::process_loop_p): Improve heuristics by handling also
4681 do_while loop and duplicating shortest sequence containing all
4684 2023-07-21 Jan Hubicka <jh@suse.cz>
4686 * tree-ssa-loop-niter.cc (finite_loop_p): Reorder to do cheap
4687 tests first; update finite_p flag.
4689 2023-07-21 Jan Hubicka <jh@suse.cz>
4691 * cfgloop.cc (flow_loop_dump): Use print_loop_info.
4692 * cfgloop.h (print_loop_info): Declare.
4693 * tree-cfg.cc (print_loop_info): Break out from ...; add
4694 printing of missing fields and profile
4695 (print_loop): ... here.
4697 2023-07-21 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4699 * config/riscv/riscv-v.cc (expand_gather_scatter): Remove redundant variables.
4701 2023-07-21 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4703 * tree-vect-stmts.cc (check_load_store_for_partial_vectors): Change condition order.
4704 (vectorizable_operation): Ditto.
4706 2023-07-21 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4708 * config/riscv/autovec.md: Align order of mask and len.
4709 * config/riscv/riscv-v.cc (expand_load_store): Ditto.
4710 (expand_gather_scatter): Ditto.
4711 * doc/md.texi: Ditto.
4712 * internal-fn.cc (add_len_and_mask_args): Ditto.
4713 (add_mask_and_len_args): Ditto.
4714 (expand_partial_load_optab_fn): Ditto.
4715 (expand_partial_store_optab_fn): Ditto.
4716 (expand_scatter_store_optab_fn): Ditto.
4717 (expand_gather_load_optab_fn): Ditto.
4718 (internal_fn_len_index): Ditto.
4719 (internal_fn_mask_index): Ditto.
4720 (internal_len_load_store_bias): Ditto.
4721 * tree-vect-stmts.cc (vectorizable_store): Ditto.
4722 (vectorizable_load): Ditto.
4724 2023-07-21 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4726 * config/riscv/autovec.md (len_maskload<mode><vm>): Change LEN_MASK into MASK_LEN.
4727 (mask_len_load<mode><vm>): Ditto.
4728 (len_maskstore<mode><vm>): Ditto.
4729 (mask_len_store<mode><vm>): Ditto.
4730 (len_mask_gather_load<RATIO64:mode><RATIO64I:mode>): Ditto.
4731 (mask_len_gather_load<RATIO64:mode><RATIO64I:mode>): Ditto.
4732 (len_mask_gather_load<RATIO32:mode><RATIO32I:mode>): Ditto.
4733 (mask_len_gather_load<RATIO32:mode><RATIO32I:mode>): Ditto.
4734 (len_mask_gather_load<RATIO16:mode><RATIO16I:mode>): Ditto.
4735 (mask_len_gather_load<RATIO16:mode><RATIO16I:mode>): Ditto.
4736 (len_mask_gather_load<RATIO8:mode><RATIO8I:mode>): Ditto.
4737 (mask_len_gather_load<RATIO8:mode><RATIO8I:mode>): Ditto.
4738 (len_mask_gather_load<RATIO4:mode><RATIO4I:mode>): Ditto.
4739 (mask_len_gather_load<RATIO4:mode><RATIO4I:mode>): Ditto.
4740 (len_mask_gather_load<RATIO2:mode><RATIO2I:mode>): Ditto.
4741 (mask_len_gather_load<RATIO2:mode><RATIO2I:mode>): Ditto.
4742 (len_mask_gather_load<RATIO1:mode><RATIO1:mode>): Ditto.
4743 (mask_len_gather_load<RATIO1:mode><RATIO1:mode>): Ditto.
4744 (len_mask_scatter_store<RATIO64:mode><RATIO64I:mode>): Ditto.
4745 (mask_len_scatter_store<RATIO64:mode><RATIO64I:mode>): Ditto.
4746 (len_mask_scatter_store<RATIO32:mode><RATIO32I:mode>): Ditto.
4747 (mask_len_scatter_store<RATIO32:mode><RATIO32I:mode>): Ditto.
4748 (len_mask_scatter_store<RATIO16:mode><RATIO16I:mode>): Ditto.
4749 (mask_len_scatter_store<RATIO16:mode><RATIO16I:mode>): Ditto.
4750 (len_mask_scatter_store<RATIO8:mode><RATIO8I:mode>): Ditto.
4751 (mask_len_scatter_store<RATIO8:mode><RATIO8I:mode>): Ditto.
4752 (len_mask_scatter_store<RATIO4:mode><RATIO4I:mode>): Ditto.
4753 (mask_len_scatter_store<RATIO4:mode><RATIO4I:mode>): Ditto.
4754 (len_mask_scatter_store<RATIO2:mode><RATIO2I:mode>): Ditto.
4755 (mask_len_scatter_store<RATIO2:mode><RATIO2I:mode>): Ditto.
4756 (len_mask_scatter_store<RATIO1:mode><RATIO1:mode>): Ditto.
4757 (mask_len_scatter_store<RATIO1:mode><RATIO1:mode>): Ditto.
4758 * doc/md.texi: Ditto.
4759 * genopinit.cc (main): Ditto.
4760 (CMP_NAME): Ditto. Ditto.
4761 * gimple-fold.cc (arith_overflowed_p): Ditto.
4762 (gimple_fold_partial_load_store_mem_ref): Ditto.
4763 (gimple_fold_call): Ditto.
4764 * internal-fn.cc (len_maskload_direct): Ditto.
4765 (mask_len_load_direct): Ditto.
4766 (len_maskstore_direct): Ditto.
4767 (mask_len_store_direct): Ditto.
4768 (expand_call_mem_ref): Ditto.
4769 (expand_len_maskload_optab_fn): Ditto.
4770 (expand_mask_len_load_optab_fn): Ditto.
4771 (expand_len_maskstore_optab_fn): Ditto.
4772 (expand_mask_len_store_optab_fn): Ditto.
4773 (direct_len_maskload_optab_supported_p): Ditto.
4774 (direct_mask_len_load_optab_supported_p): Ditto.
4775 (direct_len_maskstore_optab_supported_p): Ditto.
4776 (direct_mask_len_store_optab_supported_p): Ditto.
4777 (internal_load_fn_p): Ditto.
4778 (internal_store_fn_p): Ditto.
4779 (internal_gather_scatter_fn_p): Ditto.
4780 (internal_fn_len_index): Ditto.
4781 (internal_fn_mask_index): Ditto.
4782 (internal_fn_stored_value_index): Ditto.
4783 (internal_len_load_store_bias): Ditto.
4784 * internal-fn.def (LEN_MASK_GATHER_LOAD): Ditto.
4785 (MASK_LEN_GATHER_LOAD): Ditto.
4786 (LEN_MASK_LOAD): Ditto.
4787 (MASK_LEN_LOAD): Ditto.
4788 (LEN_MASK_SCATTER_STORE): Ditto.
4789 (MASK_LEN_SCATTER_STORE): Ditto.
4790 (LEN_MASK_STORE): Ditto.
4791 (MASK_LEN_STORE): Ditto.
4792 * optabs-query.cc (supports_vec_gather_load_p): Ditto.
4793 (supports_vec_scatter_store_p): Ditto.
4794 * optabs-tree.cc (target_supports_mask_load_store_p): Ditto.
4795 (target_supports_len_load_store_p): Ditto.
4796 * optabs.def (OPTAB_CD): Ditto.
4797 * tree-ssa-alias.cc (ref_maybe_used_by_call_p_1): Ditto.
4798 (call_may_clobber_ref_p_1): Ditto.
4799 * tree-ssa-dse.cc (initialize_ao_ref_for_dse): Ditto.
4800 (dse_optimize_stmt): Ditto.
4801 * tree-ssa-loop-ivopts.cc (get_mem_type_for_internal_fn): Ditto.
4802 (get_alias_ptr_type_for_ptr_address): Ditto.
4803 * tree-vect-data-refs.cc (vect_gather_scatter_fn_p): Ditto.
4804 * tree-vect-patterns.cc (vect_recog_gather_scatter_pattern): Ditto.
4805 * tree-vect-stmts.cc (check_load_store_for_partial_vectors): Ditto.
4806 (vect_get_strided_load_store_ops): Ditto.
4807 (vectorizable_store): Ditto.
4808 (vectorizable_load): Ditto.
4810 2023-07-21 Haochen Jiang <haochen.jiang@intel.com>
4812 * config/i386/i386.opt: Fix a typo.
4814 2023-07-21 Richard Biener <rguenther@suse.de>
4816 PR tree-optimization/88540
4817 * tree-ssa-phiopt.cc (minmax_replacement): Do not give up
4818 with NaNs but handle the simple case by if-converting to a
4821 2023-07-21 Andrew Pinski <apinski@marvell.com>
4823 * match.pd (minmax<minmax<a,b>,a>->minmax<a,b>): New
4826 2023-07-21 Richard Biener <rguenther@suse.de>
4828 PR tree-optimization/110742
4829 * tree-vect-slp.cc (vect_optimize_slp_pass::get_result_with_layout):
4830 Do not materialize an edge permutation in an external node with
4832 (vect_slp_analyze_node_operations_1): Guard purely internal
4835 2023-07-21 Jan Hubicka <jh@suse.cz>
4837 * cfgloop.cc: Include sreal.h.
4838 (flow_loop_dump): Dump sreal iteration exsitmate.
4839 (get_estimated_loop_iterations): Update.
4840 * cfgloop.h (expected_loop_iterations_by_profile): Declare.
4841 * cfgloopanal.cc (expected_loop_iterations_by_profile): New function.
4842 (expected_loop_iterations_unbounded): Use new API.
4843 * cfgloopmanip.cc (scale_loop_profile): Use
4844 expected_loop_iterations_by_profile
4845 * predict.cc (pass_profile::execute): Likewise.
4846 * profile.cc (branch_prob): Likewise.
4847 * tree-ssa-loop-niter.cc: Include sreal.h.
4848 (estimate_numbers_of_iterations): Likewise
4850 2023-07-21 Kewen Lin <linkw@linux.ibm.com>
4852 PR tree-optimization/110744
4853 * tree-ssa-sccvn.cc (vn_reference_lookup_3): Correct the index of bias
4854 operand for ifn IFN_LEN_STORE.
4856 2023-07-21 liuhongt <hongtao.liu@intel.com>
4859 * common.opt: (fcf-protection=): Add EnumSet attribute to
4860 support combination of params.
4862 2023-07-21 David Malcolm <dmalcolm@redhat.com>
4864 PR middle-end/110612
4865 * text-art/table.cc (table_geometry::table_geometry): Drop m_table
4867 (table_geometry::table_x_to_canvas_x): Add cast to comparison.
4868 (table_geometry::table_y_to_canvas_y): Likewise.
4869 * text-art/table.h (table_geometry::m_table): Drop unused field.
4870 * text-art/widget.h (wrapper_widget::update_child_alloc_rects):
4873 2023-07-20 Uros Bizjak <ubizjak@gmail.com>
4876 * config/i386/i386-features.cc
4877 (general_scalar_chain::compute_convert_gain): Calculate gain
4878 for extend higpart case.
4879 (general_scalar_chain::convert_op): Handle
4880 ASHIFTRT/ASHIFT combined RTX.
4881 (general_scalar_to_vector_candidate_p): Enable ASHIFTRT for
4882 SImode for SSE2 targets. Handle ASHIFTRT/ASHIFT combined RTX.
4883 * config/i386/i386.md (*extend<dwi>2_doubleword_highpart):
4884 New define_insn_and_split pattern.
4885 (*extendv2di2_highpart_stv): Ditto.
4887 2023-07-20 Vladimir N. Makarov <vmakarov@redhat.com>
4889 * lra-constraints.cc (simplify_operand_subreg): Check frame pointer
4892 2023-07-20 Andrew Pinski <apinski@marvell.com>
4894 * combine.cc (dump_combine_stats): Remove.
4895 (dump_combine_total_stats): Remove.
4896 (total_attempts, total_merges, total_extras,
4897 total_successes): Remove.
4898 (combine_instructions): Don't increment total stats
4899 instead use statistics_counter_event.
4900 * dumpfile.cc (print_combine_total_stats): Remove.
4901 * dumpfile.h (print_combine_total_stats): Remove.
4902 (dump_combine_total_stats): Remove.
4903 * passes.cc (finish_optimization_passes):
4904 Don't call print_combine_total_stats.
4905 * rtl.h (dump_combine_total_stats): Remove.
4906 (dump_combine_stats): Remove.
4908 2023-07-20 Jan Hubicka <jh@suse.cz>
4910 * tree-ssa-loop-ch.cc (should_duplicate_loop_header_p): Use BIT instead of TRUTH
4913 2023-07-20 Martin Jambor <mjambor@suse.cz>
4915 * doc/invoke.texi (analyzer-text-art-string-ellipsis-threshold): New.
4916 (analyzer-text-art-ideal-canvas-width): Likewise.
4917 (analyzer-text-art-string-ellipsis-head-len): Likewise.
4918 (analyzer-text-art-string-ellipsis-tail-len): Likewise.
4920 2023-07-20 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
4922 * tree-vect-stmts.cc (check_load_store_for_partial_vectors):
4923 Refine code structure.
4925 2023-07-20 Jan Hubicka <jh@suse.cz>
4927 * tree-ssa-loop-ch.cc (edge_range_query): Rename to ...
4928 (get_range_query): ... this one; do
4929 (static_loop_exit): Add query parametr, turn ranger to reference.
4930 (loop_static_stmt_p): New function.
4931 (loop_static_op_p): New function.
4932 (loop_iv_derived_p): Remove.
4933 (loop_combined_static_and_iv_p): New function.
4934 (should_duplicate_loop_header_p): Discover combined onditionals;
4935 do not track iv derived; improve dumps.
4936 (pass_ch::execute): Fix whitespace.
4938 2023-07-20 Richard Biener <rguenther@suse.de>
4940 PR tree-optimization/110204
4941 * tree-ssa-sccvn.cc (eliminate_dom_walker::eliminate_avail):
4942 Look through copies generated by PRE.
4944 2023-07-20 Matthew Malcomson <matthew.malcomson@arm.com>
4946 * tree-vect-stmts.cc (get_group_load_store_type): Account for
4947 `gap` when checking if need to peel twice.
4949 2023-07-20 Francois-Xavier Coudert <fxcoudert@gcc.gnu.org>
4952 * doc/extend.texi: Document iseqsig builtin.
4953 * builtins.cc (fold_builtin_iseqsig): New function.
4954 (fold_builtin_2): Handle BUILT_IN_ISEQSIG.
4955 (is_inexpensive_builtin): Handle BUILT_IN_ISEQSIG.
4956 * builtins.def (BUILT_IN_ISEQSIG): New built-in.
4958 2023-07-20 Pan Li <pan2.li@intel.com>
4960 * config/riscv/vector.md: Fix incorrect match_operand.
4962 2023-07-20 Roger Sayle <roger@nextmovesoftware.com>
4964 * config/i386/i386-expand.cc (ix86_expand_move): Don't call
4965 force_reg, to use SUBREG rather than create a new pseudo when
4966 inserting DFmode fields into TImode with insvti_{high,low}part.
4967 * config/i386/i386.md (*concat<mode><dwi>3_3): Split into two
4968 define_insn_and_split...
4969 (*concatditi3_3): 64-bit implementation. Provide alternative
4970 that allows register allocation to use SSE registers that is
4971 split into vec_concatv2di after reload.
4972 (*concatsidi3_3): 32-bit implementation.
4974 2023-07-20 Richard Biener <rguenther@suse.de>
4977 * internal-fn.cc (expand_vec_cond_optab_fn): When the
4978 value operands are equal to the original comparison operands
4979 preserve that equality by re-using the comparison expansion.
4980 * optabs.cc (emit_conditional_move): When the value operands
4981 are equal to the comparison operands and would be forced to
4982 a register by prepare_cmp_insn do so earlier, preserving the
4985 2023-07-20 Pan Li <pan2.li@intel.com>
4987 * config/riscv/vector.md: Align pattern format.
4989 2023-07-20 Haochen Jiang <haochen.jiang@intel.com>
4991 * doc/invoke.texi: Remove AVX512VP2INTERSECT in
4992 Granite Rapids{, D} from documentation.
4994 2023-07-20 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4996 * config/riscv/autovec.md
4997 (len_mask_gather_load<VNX16_QHSD:mode><VNX16_QHSDI:mode>):
4998 Refactor RVV machine modes.
4999 (len_mask_gather_load<VNX16_QHS:mode><VNX16_QHSI:mode>): Ditto.
5000 (len_mask_gather_load<VNX32_QHS:mode><VNX32_QHSI:mode>): Ditto.
5001 (len_mask_gather_load<VNX32_QH:mode><VNX32_QHI:mode>): Ditto.
5002 (len_mask_gather_load<VNX64_QH:mode><VNX64_QHI:mode>): Ditto.
5003 (len_mask_gather_load<mode><mode>): Ditto.
5004 (len_mask_gather_load<VNX64_Q:mode><VNX64_Q:mode>): Ditto.
5005 (len_mask_scatter_store<VNX16_QHSD:mode><VNX16_QHSDI:mode>): Ditto.
5006 (len_mask_scatter_store<VNX32_QHS:mode><VNX32_QHSI:mode>): Ditto.
5007 (len_mask_scatter_store<VNX16_QHS:mode><VNX16_QHSI:mode>): Ditto.
5008 (len_mask_scatter_store<VNX64_QH:mode><VNX64_QHI:mode>): Ditto.
5009 (len_mask_scatter_store<VNX32_QH:mode><VNX32_QHI:mode>): Ditto.
5010 (len_mask_scatter_store<mode><mode>): Ditto.
5011 (len_mask_scatter_store<VNX64_Q:mode><VNX64_Q:mode>): Ditto.
5012 * config/riscv/riscv-modes.def (VECTOR_BOOL_MODE): Ditto.
5013 (ADJUST_NUNITS): Ditto.
5014 (ADJUST_ALIGNMENT): Ditto.
5015 (ADJUST_BYTESIZE): Ditto.
5016 (ADJUST_PRECISION): Ditto.
5018 (RVV_WHOLE_MODES): Ditto.
5019 (RVV_FRACT_MODE): Ditto.
5020 (RVV_NF8_MODES): Ditto.
5021 (RVV_NF4_MODES): Ditto.
5022 (VECTOR_MODES_WITH_PREFIX): Ditto.
5023 (VECTOR_MODE_WITH_PREFIX): Ditto.
5024 (RVV_TUPLE_MODES): Ditto.
5025 (RVV_NF2_MODES): Ditto.
5026 (RVV_TUPLE_PARTIAL_MODES): Ditto.
5027 * config/riscv/riscv-v.cc (struct mode_vtype_group): Ditto.
5029 (TUPLE_ENTRY): Ditto.
5033 (preferred_simd_mode): Ditto.
5034 (autovectorize_vector_modes): Ditto.
5035 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_TYPE): Ditto.
5036 * config/riscv/riscv-vector-builtins.def (DEF_RVV_TYPE): Ditto.
5044 (vint8mf8_t): Ditto.
5045 (vuint8mf8_t): Ditto.
5046 (vint8mf4_t): Ditto.
5047 (vuint8mf4_t): Ditto.
5048 (vint8mf2_t): Ditto.
5049 (vuint8mf2_t): Ditto.
5051 (vuint8m1_t): Ditto.
5053 (vuint8m2_t): Ditto.
5055 (vuint8m4_t): Ditto.
5057 (vuint8m8_t): Ditto.
5058 (vint16mf4_t): Ditto.
5059 (vuint16mf4_t): Ditto.
5060 (vint16mf2_t): Ditto.
5061 (vuint16mf2_t): Ditto.
5062 (vint16m1_t): Ditto.
5063 (vuint16m1_t): Ditto.
5064 (vint16m2_t): Ditto.
5065 (vuint16m2_t): Ditto.
5066 (vint16m4_t): Ditto.
5067 (vuint16m4_t): Ditto.
5068 (vint16m8_t): Ditto.
5069 (vuint16m8_t): Ditto.
5070 (vint32mf2_t): Ditto.
5071 (vuint32mf2_t): Ditto.
5072 (vint32m1_t): Ditto.
5073 (vuint32m1_t): Ditto.
5074 (vint32m2_t): Ditto.
5075 (vuint32m2_t): Ditto.
5076 (vint32m4_t): Ditto.
5077 (vuint32m4_t): Ditto.
5078 (vint32m8_t): Ditto.
5079 (vuint32m8_t): Ditto.
5080 (vint64m1_t): Ditto.
5081 (vuint64m1_t): Ditto.
5082 (vint64m2_t): Ditto.
5083 (vuint64m2_t): Ditto.
5084 (vint64m4_t): Ditto.
5085 (vuint64m4_t): Ditto.
5086 (vint64m8_t): Ditto.
5087 (vuint64m8_t): Ditto.
5088 (vfloat16mf4_t): Ditto.
5089 (vfloat16mf2_t): Ditto.
5090 (vfloat16m1_t): Ditto.
5091 (vfloat16m2_t): Ditto.
5092 (vfloat16m4_t): Ditto.
5093 (vfloat16m8_t): Ditto.
5094 (vfloat32mf2_t): Ditto.
5095 (vfloat32m1_t): Ditto.
5096 (vfloat32m2_t): Ditto.
5097 (vfloat32m4_t): Ditto.
5098 (vfloat32m8_t): Ditto.
5099 (vfloat64m1_t): Ditto.
5100 (vfloat64m2_t): Ditto.
5101 (vfloat64m4_t): Ditto.
5102 (vfloat64m8_t): Ditto.
5103 * config/riscv/riscv-vector-switch.def (ENTRY): Ditto.
5104 (TUPLE_ENTRY): Ditto.
5105 * config/riscv/riscv-vsetvl.cc (change_insn): Ditto.
5106 * config/riscv/riscv.cc (riscv_valid_lo_sum_p): Ditto.
5107 (riscv_v_adjust_nunits): Ditto.
5108 (riscv_v_adjust_bytesize): Ditto.
5109 (riscv_v_adjust_precision): Ditto.
5110 (riscv_convert_vector_bits): Ditto.
5111 * config/riscv/riscv.h (riscv_v_adjust_nunits): Ditto.
5112 * config/riscv/riscv.md: Ditto.
5113 * config/riscv/vector-iterators.md: Ditto.
5114 * config/riscv/vector.md
5115 (@pred_indexed_<order>store<VNX16_QHSD:mode><VNX16_QHSDI:mode>): Ditto.
5116 (@pred_indexed_<order>store<VNX16_QHS:mode><VNX16_QHSI:mode>): Ditto.
5117 (@pred_indexed_<order>store<VNX32_QHS:mode><VNX32_QHSI:mode>): Ditto.
5118 (@pred_indexed_<order>store<VNX32_QH:mode><VNX32_QHI:mode>): Ditto.
5119 (@pred_indexed_<order>store<VNX64_QH:mode><VNX64_QHI:mode>): Ditto.
5120 (@pred_indexed_<order>store<VNX64_Q:mode><VNX64_Q:mode>): Ditto.
5121 (@pred_indexed_<order>store<VNX128_Q:mode><VNX128_Q:mode>): Ditto.
5122 (@pred_indexed_<order>load<V1T:mode><V1I:mode>): Ditto.
5123 (@pred_indexed_<order>load<V1T:mode><VNX1_QHSDI:mode>): Ditto.
5124 (@pred_indexed_<order>load<V2T:mode><V2I:mode>): Ditto.
5125 (@pred_indexed_<order>load<V2T:mode><VNX2_QHSDI:mode>): Ditto.
5126 (@pred_indexed_<order>load<V4T:mode><V4I:mode>): Ditto.
5127 (@pred_indexed_<order>load<V4T:mode><VNX4_QHSDI:mode>): Ditto.
5128 (@pred_indexed_<order>load<V8T:mode><V8I:mode>): Ditto.
5129 (@pred_indexed_<order>load<V8T:mode><VNX8_QHSDI:mode>): Ditto.
5130 (@pred_indexed_<order>load<V16T:mode><V16I:mode>): Ditto.
5131 (@pred_indexed_<order>load<V16T:mode><VNX16_QHSI:mode>): Ditto.
5132 (@pred_indexed_<order>load<V32T:mode><V32I:mode>): Ditto.
5133 (@pred_indexed_<order>load<V32T:mode><VNX32_QHI:mode>): Ditto.
5134 (@pred_indexed_<order>load<V64T:mode><V64I:mode>): Ditto.
5135 (@pred_indexed_<order>store<V1T:mode><V1I:mode>): Ditto.
5136 (@pred_indexed_<order>store<V1T:mode><VNX1_QHSDI:mode>): Ditto.
5137 (@pred_indexed_<order>store<V2T:mode><V2I:mode>): Ditto.
5138 (@pred_indexed_<order>store<V2T:mode><VNX2_QHSDI:mode>): Ditto.
5139 (@pred_indexed_<order>store<V4T:mode><V4I:mode>): Ditto.
5140 (@pred_indexed_<order>store<V4T:mode><VNX4_QHSDI:mode>): Ditto.
5141 (@pred_indexed_<order>store<V8T:mode><V8I:mode>): Ditto.
5142 (@pred_indexed_<order>store<V8T:mode><VNX8_QHSDI:mode>): Ditto.
5143 (@pred_indexed_<order>store<V16T:mode><V16I:mode>): Ditto.
5144 (@pred_indexed_<order>store<V16T:mode><VNX16_QHSI:mode>): Ditto.
5145 (@pred_indexed_<order>store<V32T:mode><V32I:mode>): Ditto.
5146 (@pred_indexed_<order>store<V32T:mode><VNX32_QHI:mode>): Ditto.
5147 (@pred_indexed_<order>store<V64T:mode><V64I:mode>): Ditto.
5149 2023-07-19 Vladimir N. Makarov <vmakarov@redhat.com>
5151 * lra-int.h (lra_update_fp2sp_elimination): New prototype.
5152 (lra_asm_insn_error): New prototype.
5153 * lra-spills.cc (remove_pseudos): Add check for pseudo slot memory
5155 (lra_spill): Call lra_update_fp2sp_elimination.
5156 * lra-eliminations.cc: Remove trailing spaces.
5157 (elimination_fp2sp_occured_p): New static flag.
5158 (lra_eliminate_regs_1): Set the flag up.
5159 (update_reg_eliminate): Modify the assert for stack to frame
5160 pointer elimination.
5161 (lra_update_fp2sp_elimination): New function.
5162 (lra_eliminate): Clear flag elimination_fp2sp_occured_p.
5164 2023-07-19 Andrew Carlotti <andrew.carlotti@arm.com>
5166 * config/aarch64/aarch64.h (TARGET_MEMTAG): Remove armv8.5
5168 * config/aarch64/arm_acle.h: Remove unnecessary armv8.x
5169 dependencies from target pragmas.
5170 * config/aarch64/arm_fp16.h (target): Likewise.
5171 * config/aarch64/arm_neon.h (target): Likewise.
5173 2023-07-19 Andrew Pinski <apinski@marvell.com>
5175 PR tree-optimization/110252
5176 * tree-ssa-phiopt.cc (class auto_flow_sensitive): New class.
5177 (auto_flow_sensitive::auto_flow_sensitive): New constructor.
5178 (auto_flow_sensitive::~auto_flow_sensitive): New deconstructor.
5179 (match_simplify_replacement): Temporarily
5180 remove the flow sensitive info on the two statements that might
5183 2023-07-19 Andrew Pinski <apinski@marvell.com>
5185 * gimple-fold.cc (fosa_unwind): Replace `vrange_storage *`
5186 with flow_sensitive_info_storage.
5187 (follow_outer_ssa_edges): Update how to save off the flow
5189 (maybe_fold_comparisons_from_match_pd): Update restoring
5190 of flow sensitive info.
5191 * tree-ssanames.cc (flow_sensitive_info_storage::save): New method.
5192 (flow_sensitive_info_storage::restore): New method.
5193 (flow_sensitive_info_storage::save_and_clear): New method.
5194 (flow_sensitive_info_storage::clear_storage): New method.
5195 * tree-ssanames.h (class flow_sensitive_info_storage): New class.
5197 2023-07-19 Andrew Pinski <apinski@marvell.com>
5199 PR tree-optimization/110726
5200 * match.pd ((a|b)&(a==b),a|(a==b),(a&b)|(a==b)):
5201 Add checks to make sure the type was one bit precision
5204 2023-07-19 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
5206 * doc/md.texi: Add mask_len_fold_left_plus.
5207 * internal-fn.cc (mask_len_fold_left_direct): Ditto.
5208 (expand_mask_len_fold_left_optab_fn): Ditto.
5209 (direct_mask_len_fold_left_optab_supported_p): Ditto.
5210 * internal-fn.def (MASK_LEN_FOLD_LEFT_PLUS): Ditto.
5211 * optabs.def (OPTAB_D): Ditto.
5213 2023-07-19 Jakub Jelinek <jakub@redhat.com>
5215 * tree-switch-conversion.h (class bit_test_cluster): Fix comment typo.
5217 2023-07-19 Jakub Jelinek <jakub@redhat.com>
5219 PR tree-optimization/110731
5220 * wide-int.cc (wi::divmod_internal): Always unpack dividend and
5221 divisor as UNSIGNED regardless of sgn.
5223 2023-07-19 Lehua Ding <lehua.ding@rivai.ai>
5225 * common/config/riscv/riscv-common.cc (riscv_supported_std_ext): Init.
5226 (standard_extensions_p): Add check.
5227 (riscv_subset_list::add): Just return NULL if it failed before.
5228 (riscv_subset_list::parse_std_ext): Continue parse when find a error
5229 (riscv_subset_list::parse): Just return NULL if it failed before.
5230 * config/riscv/riscv-subset.h (class riscv_subset_list): Add field.
5232 2023-07-19 Jan Beulich <jbeulich@suse.com>
5234 * config/i386/i386-expand.cc (ix86_expand_vector_init_duplicate):
5236 (ix86_expand_vector_extract): Use gen_vec_extract_lo /
5238 (expand_vec_perm_broadcast_1): Use gen_vec_interleave_high /
5239 gen_vec_interleave_low. Rename local variable.
5241 2023-07-19 Jan Beulich <jbeulich@suse.com>
5243 * config/i386/sse.md (vec_dupv2df<mask_name>): Add new AVX512F
5244 alternative. Move AVX512VL part of condition to new "enabled"
5247 2023-07-19 liuhongt <hongtao.liu@intel.com>
5250 * config/i386/i386-builtins.cc
5251 (ix86_register_float16_builtin_type): Remove TARGET_SSE2.
5252 (ix86_register_bf16_builtin_type): Ditto.
5253 * config/i386/i386-c.cc (ix86_target_macros): When TARGET_SSE2
5254 isn't available, undef the macros which are used to check the
5255 backend support of the _Float16/__bf16 types when building
5256 libstdc++ and libgcc.
5257 * config/i386/i386.cc (construct_container): Issue errors for
5258 HFmode/BFmode when TARGET_SSE2 is not available.
5259 (function_value_32): Ditto.
5260 (ix86_scalar_mode_supported_p): Remove TARGET_SSE2 for HFmode/BFmode.
5261 (ix86_libgcc_floating_mode_supported_p): Ditto.
5262 (ix86_emit_support_tinfos): Adjust codes.
5263 (ix86_invalid_conversion): Return diagnostic message string
5264 when there's conversion from/to BF/HFmode w/o TARGET_SSE2.
5265 (ix86_invalid_unary_op): New function.
5266 (ix86_invalid_binary_op): Ditto.
5267 (TARGET_INVALID_UNARY_OP): Define.
5268 (TARGET_INVALID_BINARY_OP): Define.
5269 * config/i386/immintrin.h [__SSE2__]: Remove for fp16/bf16
5270 related instrinsics header files.
5271 * config/i386/i386.h (VALID_SSE2_TYPE_MODE): New macro.
5273 2023-07-18 Uros Bizjak <ubizjak@gmail.com>
5275 * dwarf2asm.cc: Change FALSE to false.
5276 * dwarf2cfi.cc (execute_dwarf2_frame): Change return type to void.
5277 * dwarf2out.cc (matches_main_base): Change return type from
5278 int to bool. Change "last_match" variable to bool.
5279 (dump_struct_debug): Change return type from int to bool.
5280 Change "matches" and "result" function arguments to bool.
5281 (is_pseudo_reg): Change return type from int to bool.
5282 (is_tagged_type): Ditto.
5283 (same_loc_p): Ditto.
5284 (same_dw_val_p): Change return type from int to bool and adjust
5285 function body accordingly.
5286 (same_attr_p): Ditto.
5287 (same_die_p): Ditto.
5288 (is_type_die): Ditto.
5289 (is_declaration_die): Ditto.
5290 (should_move_die_to_comdat): Ditto.
5291 (is_base_type): Ditto.
5292 (is_based_loc): Ditto.
5293 (local_scope_p): Ditto.
5294 (class_scope_p): Ditto.
5295 (class_or_namespace_scope_p): Ditto.
5296 (is_tagged_type): Ditto.
5297 (is_rust): Use void argument.
5298 (is_nested_in_subprogram): Change return type from int to bool.
5299 (contains_subprogram_definition): Ditto.
5300 (gen_struct_or_union_type_die): Change "nested", "complete"
5301 and "ns_decl" variables to bool.
5302 (is_naming_typedef_decl): Change FALSE to false.
5304 2023-07-18 Jan Hubicka <jh@suse.cz>
5306 * tree-ssa-loop-ch.cc (edge_range_query): Take loop argument; be ready
5307 for queries not in headers.
5308 (static_loop_exit): Add basic blck parameter; update use of
5310 (should_duplicate_loop_header_p): Add ranger and static_exits
5311 parameter. Do not account statements that will be optimized
5312 out after duplicaiton in overall size. Add ranger query to
5314 (update_profile_after_ch): Take static_exits has set instead of
5315 single eliminated_edge.
5316 (ch_base::copy_headers): Do all analysis in the first pass;
5317 remember invariant_exits and static_exits.
5319 2023-07-18 Jason Merrill <jason@redhat.com>
5321 * fold-const.cc (native_interpret_aggregate): Skip empty fields.
5323 2023-07-18 Gaius Mulley <gaiusmod2@gmail.com>
5325 * doc/gm2.texi (Semantic checking): Change example testwithptr
5328 2023-07-18 Richard Biener <rguenther@suse.de>
5330 PR middle-end/105715
5331 * gimple-isel.cc (gimple_expand_vec_exprs): Merge into...
5332 (pass_gimple_isel::execute): ... this. Duplicate
5333 comparison defs of COND_EXPRs.
5335 2023-07-18 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5337 * config/riscv/riscv-selftests.cc (run_poly_int_selftests): Add more selftests.
5338 * config/riscv/riscv.cc (riscv_legitimize_poly_move): Dynamic adjust size of VLA vectors.
5339 (riscv_convert_vector_bits): Ditto.
5341 2023-07-18 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5343 * config/riscv/autovec.md (vec_shl_insert_<mode>): New patterns.
5344 * config/riscv/riscv-v.cc (shuffle_compress_patterns): Fix bugs.
5346 2023-07-18 Juergen Christ <jchrist@linux.ibm.com>
5348 * config/s390/vx-builtins.md: New vsel pattern.
5350 2023-07-18 liuhongt <hongtao.liu@intel.com>
5353 * config/i386/sse.md (<mask_codefor>one_cmpl<mode>2<mask_name>):
5354 Remove # from assemble output.
5356 2023-07-18 liuhongt <hongtao.liu@intel.com>
5359 * config/i386/sync.md (cmpccxadd_<mode>): Adjust the pattern
5360 to explicitly set FLAGS_REG like *cmp<mode>_1, also add extra
5361 3 define_peephole2 after the pattern.
5363 2023-07-18 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
5365 * rtl-ssa/internals.inl: Fix when mode1 and mode2 are not ordred.
5367 2023-07-18 Pan Li <pan2.li@intel.com>
5368 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5370 * config/riscv/riscv.cc (struct machine_function): Add new field.
5371 (riscv_static_frm_mode_p): New function.
5372 (riscv_emit_frm_mode_set): New function for emit FRM.
5373 (riscv_emit_mode_set): Extract function for FRM.
5374 (riscv_mode_needed): Fix the TODO.
5375 (riscv_mode_entry): Initial dynamic frm RTL.
5376 (riscv_mode_exit): Return DYN_EXIT.
5377 * config/riscv/riscv.md: Add rdfrm.
5378 * config/riscv/vector-iterators.md (unspecv): Add DYN_EXIT unspecv.
5379 * config/riscv/vector.md (frm_modee): Add new mode dyn_exit.
5381 (fsrmsi_backup): New pattern for swap.
5382 (fsrmsi_restore): New pattern for restore.
5383 (fsrmsi_restore_exit): New pattern for restore exit.
5384 (frrmsi): New pattern for backup.
5386 2023-07-17 Arsen Arsenović <arsen@aarsen.me>
5388 * doc/extend.texi: Add @cindex on __auto_type.
5390 2023-07-17 Uros Bizjak <ubizjak@gmail.com>
5392 * combine-stack-adj.cc (stack_memref_p): Change return type from
5393 int to bool and adjust function body accordingly.
5394 (rest_of_handle_stack_adjustments): Change return type to void.
5396 2023-07-17 Uros Bizjak <ubizjak@gmail.com>
5398 * combine.cc (struct reg_stat_type): Change last_set_invalid to bool.
5399 (cant_combine_insn_p): Change return type from int to bool and adjust
5400 function body accordingly.
5401 (can_combine_p): Ditto.
5402 (combinable_i3pat): Ditto. Change "i1_not_in_src" and "i0_not_in_src"
5403 function arguments from int to bool.
5404 (contains_muldiv): Change return type from int to bool and adjust
5405 function body accordingly.
5406 (try_combine): Ditto. Change "new_direct_jump" pointer function
5407 argument from int to bool. Change "substed_i2", "substed_i1",
5408 "substed_i0", "added_sets_0", "added_sets_1", "added_sets_2",
5409 "i2dest_in_i2src", "i1dest_in_i1src", "i2dest_in_i1src",
5410 "i0dest_in_i0src", "i1dest_in_i0src", "i2dest_in_i0src",
5411 "i2dest_killed", "i1dest_killed", "i0dest_killed", "i1_feeds_i2_n",
5412 "i0_feeds_i2_n", "i0_feeds_i1_n", "i3_subst_into_i2", "have_mult",
5413 "swap_i2i3", "split_i2i3" and "changed_i3_dest" variables
5415 (subst): Change "in_dest", "in_cond" and "unique_copy" function
5416 arguments from int to bool.
5417 (combine_simplify_rtx): Change "in_dest" and "in_cond" function
5418 arguments from int to bool.
5419 (make_extraction): Change "unsignedp", "in_dest" and "in_compare"
5420 function argument from int to bool.
5421 (force_int_to_mode): Change "just_select" function argument
5422 from int to bool. Change "next_select" variable to bool.
5423 (rtx_equal_for_field_assignment_p): Change return type from
5424 int to bool and adjust function body accordingly.
5425 (merge_outer_ops): Ditto. Change "pcomp_p" pointer function
5426 argument from int to bool.
5427 (get_last_value_validate): Change return type from int to bool
5428 and adjust function body accordingly.
5429 (reg_dead_at_p): Ditto.
5430 (reg_bitfield_target_p): Ditto.
5431 (combine_instructions): Ditto. Change "new_direct_jump"
5433 (can_combine_p): Change return type from int to bool
5434 and adjust function body accordingly.
5435 (likely_spilled_retval_p): Ditto.
5436 (can_change_dest_mode): Change "added_sets" function argument
5438 (find_split_point): Change "unsignedp" variable to bool.
5439 (simplify_if_then_else): Change "comparison_p" and "swapped"
5441 (simplify_set): Change "other_changed" variable to bool.
5442 (expand_compound_operation): Change "unsignedp" variable to bool.
5443 (force_to_mode): Change "just_select" function argument
5444 from int to bool. Change "next_select" variable to bool.
5445 (extended_count): Change "unsignedp" function argument to bool.
5446 (simplify_shift_const_1): Change "complement_p" variable to bool.
5447 (simplify_comparison): Change "changed" variable to bool.
5448 (rest_of_handle_combine): Change return type to void.
5450 2023-07-17 Andre Vieira <andre.simoesdiasvieira@arm.com>
5453 * Makefile.in (INTERNAL_FN_H): Add insn-opinit.h.
5455 2023-07-17 Senthil Kumar Selvaraj <saaadhu@gcc.gnu.org>
5457 * ira.cc (setup_reg_class_relations): Continue
5458 if regclass cl3 is hard_reg_set_empty_p.
5460 2023-07-17 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5462 * config/riscv/riscv.cc (riscv_option_override): Add sorry check.
5464 2023-07-17 Martin Jambor <mjambor@suse.cz>
5466 * tree-ssa-loop-ivcanon.cc (try_peel_loop): Remove unused variable
5469 2023-07-17 Aldy Hernandez <aldyh@redhat.com>
5471 * tree-ssa-ccp.cc (ccp_finalize): Export value/mask known bits.
5473 2023-07-17 Lehua Ding <lehua.ding@rivai.ai>
5476 * common/config/riscv/riscv-common.cc (riscv_subset_list::handle_implied_ext):
5477 recur add all implied extensions.
5478 (riscv_subset_list::check_implied_ext): Add new method.
5479 (riscv_subset_list::parse): Call checker check_implied_ext.
5480 * config/riscv/riscv-subset.h: Add new method.
5482 2023-07-17 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5484 * config/riscv/autovec.md (reduc_plus_scal_<mode>): New pattern.
5485 (reduc_smax_scal_<mode>): Ditto.
5486 (reduc_umax_scal_<mode>): Ditto.
5487 (reduc_smin_scal_<mode>): Ditto.
5488 (reduc_umin_scal_<mode>): Ditto.
5489 (reduc_and_scal_<mode>): Ditto.
5490 (reduc_ior_scal_<mode>): Ditto.
5491 (reduc_xor_scal_<mode>): Ditto.
5492 * config/riscv/riscv-protos.h (enum insn_type): Add reduction.
5493 (expand_reduction): New function.
5494 * config/riscv/riscv-v.cc (emit_vlmax_reduction_insn): Ditto.
5495 (emit_vlmax_fp_reduction_insn): Ditto.
5496 (get_m1_mode): Ditto.
5497 (expand_cond_len_binop): Fix name.
5498 (expand_reduction): New function
5499 * config/riscv/riscv-vsetvl.cc (gen_vsetvl_pat): Fix VSETVL BUG.
5500 (validate_change_or_fail): New function.
5501 (change_insn): Fix VSETVL BUG.
5502 (change_vsetvl_insn): Ditto.
5503 (pass_vsetvl::backward_demand_fusion): Ditto.
5504 (pass_vsetvl::df_post_optimization): Ditto.
5506 2023-07-17 Aldy Hernandez <aldyh@redhat.com>
5508 * ipa-prop.cc (ipcp_update_bits): Export value/mask known bits.
5510 2023-07-17 Christoph Müllner <christoph.muellner@vrull.eu>
5512 * config/riscv/riscv.cc (riscv_regno_ok_for_index_p):
5513 Remove parameter name from declaration of unused parameter.
5515 2023-07-17 Kewen Lin <linkw@linux.ibm.com>
5517 PR tree-optimization/110652
5518 * tree-vect-stmts.cc (vectorizable_load): Initialize new_temp as
5521 2023-07-17 Richard Biener <rguenther@suse.de>
5523 PR tree-optimization/110669
5524 * tree-scalar-evolution.cc (analyze_and_compute_bitop_with_inv_effect):
5525 Check we matched a header PHI.
5527 2023-07-17 Aldy Hernandez <aldyh@redhat.com>
5529 * tree-ssanames.cc (set_bitmask): New.
5530 * tree-ssanames.h (set_bitmask): New.
5532 2023-07-17 Aldy Hernandez <aldyh@redhat.com>
5534 * value-range.cc (irange_bitmask::verify_mask): Mask need not be
5536 * value-range.h (irange_bitmask::union_): Normalize beforehand.
5537 (irange_bitmask::intersect): Same.
5539 2023-07-17 Andrew Pinski <apinski@marvell.com>
5541 PR tree-optimization/95923
5542 * match.pd ((a|b)&(a==b),a|(a==b),(a&b)|(a==b)): New transformation.
5544 2023-07-17 Roger Sayle <roger@nextmovesoftware.com>
5546 * tree-if-conv.cc (predicate_scalar_phi): Make the arguments
5547 to the std::sort comparison lambda function const.
5549 2023-07-17 Andrew Pinski <apinski@marvell.com>
5551 PR tree-optimization/110666
5552 * match.pd (A NEEQ (A NEEQ CST)): Fix Outer EQ case.
5554 2023-07-17 Mo, Zewei <zewei.mo@intel.com>
5556 * common/config/i386/cpuinfo.h (get_intel_cpu): Handle Lunar Lake,
5557 Arrow Lake and Arrow Lake S.
5558 * common/config/i386/i386-common.cc:
5559 (processor_name): Add arrowlake.
5560 (processor_alias_table): Add arrow lake, arrow lake s and lunar
5562 * common/config/i386/i386-cpuinfo.h (enum processor_subtypes):
5563 Add INTEL_COREI7_ARROWLAKE and INTEL_COREI7_ARROWLAKE_S.
5564 * config.gcc: Add -march=arrowlake and -march=arrowlake-s.
5565 * config/i386/driver-i386.cc (host_detect_local_cpu): Handle
5567 * config/i386/i386-c.cc (ix86_target_macros_internal): Add
5569 * config/i386/i386-options.cc (m_ARROWLAKE): New.
5570 (processor_cost_table): Add arrowlake.
5571 * config/i386/i386.h (enum processor_type):
5572 Add PROCESSOR_ARROWLAKE.
5573 * config/i386/x86-tune.def: Add m_ARROWLAKE.
5574 * doc/extend.texi: Add arrowlake and arrowlake-s.
5575 * doc/invoke.texi: Ditto.
5577 2023-07-17 Haochen Jiang <haochen.jiang@intel.com>
5579 * config/i386/sse.md (VI2_AVX2): Delete V32HI since we actually
5580 have the same iterator. Also renaming all the occurence to
5582 (usdot_prod<mode>): New define_expand.
5583 (udot_prod<mode>): Ditto.
5585 2023-07-17 Haochen Jiang <haochen.jiang@intel.com>
5587 * common/config/i386/cpuinfo.h (get_available_features):
5589 * common/config/i386/i386-common.cc (OPTION_MASK_ISA2_SM4_SET,
5590 OPTION_MASK_ISA2_SM4_UNSET): New.
5591 (OPTION_MASK_ISA2_AVX_UNSET): Add SM4.
5592 (ix86_handle_option): Handle -msm4.
5593 * common/config/i386/i386-cpuinfo.h (enum processor_features):
5595 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
5597 * config.gcc: Add sm4intrin.h.
5598 * config/i386/cpuid.h (bit_SM4): New.
5599 * config/i386/i386-builtin.def (BDESC): Add new builtins.
5600 * config/i386/i386-c.cc (ix86_target_macros_internal): Define
5602 * config/i386/i386-isa.def (SM4): Add DEF_PTA(SM4).
5603 * config/i386/i386-options.cc (isa2_opts): Add -msm4.
5604 (ix86_valid_target_attribute_inner_p): Handle sm4.
5605 * config/i386/i386.opt: Add option -msm4.
5606 * config/i386/immintrin.h: Include sm4intrin.h
5607 * config/i386/sse.md (vsm4key4_<mode>): New define insn.
5608 (vsm4rnds4_<mode>): Ditto.
5609 * doc/extend.texi: Document sm4.
5610 * doc/invoke.texi: Document -msm4.
5611 * doc/sourcebuild.texi: Document target sm4.
5612 * config/i386/sm4intrin.h: New file.
5614 2023-07-17 Haochen Jiang <haochen.jiang@intel.com>
5616 * common/config/i386/cpuinfo.h (get_available_features):
5618 * common/config/i386/i386-common.cc (OPTION_MASK_ISA2_SHA512_SET,
5619 OPTION_MASK_ISA2_SHA512_UNSET): New.
5620 (OPTION_MASK_ISA2_AVX_UNSET): Add SHA512.
5621 (ix86_handle_option): Handle -msha512.
5622 * common/config/i386/i386-cpuinfo.h (enum processor_features):
5624 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
5626 * config.gcc: Add sha512intrin.h.
5627 * config/i386/cpuid.h (bit_SHA512): New.
5628 * config/i386/i386-builtin-types.def:
5629 Add DEF_FUNCTION_TYPE (V4DI, V4DI, V4DI, V2DI).
5630 * config/i386/i386-builtin.def (BDESC): Add new builtins.
5631 * config/i386/i386-c.cc (ix86_target_macros_internal): Define
5633 * config/i386/i386-expand.cc (ix86_expand_args_builtin): Handle
5634 V4DI_FTYPE_V4DI_V4DI_V2DI and V4DI_FTYPE_V4DI_V2DI.
5635 * config/i386/i386-isa.def (SHA512): Add DEF_PTA(SHA512).
5636 * config/i386/i386-options.cc (isa2_opts): Add -msha512.
5637 (ix86_valid_target_attribute_inner_p): Handle sha512.
5638 * config/i386/i386.opt: Add option -msha512.
5639 * config/i386/immintrin.h: Include sha512intrin.h.
5640 * config/i386/sse.md (vsha512msg1): New define insn.
5641 (vsha512msg2): Ditto.
5642 (vsha512rnds2): Ditto.
5643 * doc/extend.texi: Document sha512.
5644 * doc/invoke.texi: Document -msha512.
5645 * doc/sourcebuild.texi: Document target sha512.
5646 * config/i386/sha512intrin.h: New file.
5648 2023-07-17 Haochen Jiang <haochen.jiang@intel.com>
5650 * common/config/i386/cpuinfo.h (get_available_features):
5652 * common/config/i386/i386-common.cc (OPTION_MASK_ISA2_SM3_SET,
5653 OPTION_MASK_ISA2_SM3_UNSET): New.
5654 (OPTION_MASK_ISA2_AVX_UNSET): Add SM3.
5655 (ix86_handle_option): Handle -msm3.
5656 * common/config/i386/i386-cpuinfo.h (enum processor_features):
5658 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
5660 * config.gcc: Add sm3intrin.h
5661 * config/i386/cpuid.h (bit_SM3): New.
5662 * config/i386/i386-builtin-types.def:
5663 Add DEF_FUNCTION_TYPE (V4SI, V4SI, V4SI, V4SI, INT).
5664 * config/i386/i386-builtin.def (BDESC): Add new builtins.
5665 * config/i386/i386-c.cc (ix86_target_macros_internal): Define
5667 * config/i386/i386-expand.cc (ix86_expand_args_builtin): Handle
5668 V4SI_FTYPE_V4SI_V4SI_V4SI_INT.
5669 * config/i386/i386-isa.def (SM3): Add DEF_PTA(SM3).
5670 * config/i386/i386-options.cc (isa2_opts): Add -msm3.
5671 (ix86_valid_target_attribute_inner_p): Handle sm3.
5672 * config/i386/i386.opt: Add option -msm3.
5673 * config/i386/immintrin.h: Include sm3intrin.h.
5674 * config/i386/sse.md (vsm3msg1): New define insn.
5677 * doc/extend.texi: Document sm3.
5678 * doc/invoke.texi: Document -msm3.
5679 * doc/sourcebuild.texi: Document target sm3.
5680 * config/i386/sm3intrin.h: New file.
5682 2023-07-17 Kong Lingling <lingling.kong@intel.com>
5683 Haochen Jiang <haochen.jiang@intel.com>
5685 * common/config/i386/cpuinfo.h (get_available_features): Detect
5687 * common/config/i386/i386-common.cc
5688 (OPTION_MASK_ISA2_AVXVNNIINT16_SET): New.
5689 (OPTION_MASK_ISA2_AVXVNNIINT16_UNSET): Ditto.
5690 (ix86_handle_option): Handle -mavxvnniint16.
5691 * common/config/i386/i386-cpuinfo.h (enum processor_features):
5692 Add FEATURE_AVXVNNIINT16.
5693 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
5695 * config.gcc: Add avxvnniint16.h.
5696 * config/i386/avxvnniint16intrin.h: New file.
5697 * config/i386/cpuid.h (bit_AVXVNNIINT16): New.
5698 * config/i386/i386-builtin.def: Add new builtins.
5699 * config/i386/i386-c.cc (ix86_target_macros_internal): Define
5701 * config/i386/i386-options.cc (isa2_opts): Add -mavxvnniint16.
5702 (ix86_valid_target_attribute_inner_p): Handle avxvnniint16intrin.h.
5703 * config/i386/i386-isa.def: Add DEF_PTA(AVXVNNIINT16).
5704 * config/i386/i386.opt: Add option -mavxvnniint16.
5705 * config/i386/immintrin.h: Include avxvnniint16.h.
5706 * config/i386/sse.md
5707 (vpdp<vpdpwprodtype>_<mode>): New define_insn.
5708 * doc/extend.texi: Document avxvnniint16.
5709 * doc/invoke.texi: Document -mavxvnniint16.
5710 * doc/sourcebuild.texi: Document target avxvnniint16.
5712 2023-07-16 Jan Hubicka <jh@suse.cz>
5714 PR middle-end/110649
5715 * tree-vect-loop.cc (scale_profile_for_vect_loop): Rewrite.
5716 (vect_transform_loop): Move scale_profile_for_vect_loop after
5717 upper bound updates.
5719 2023-07-16 Jan Hubicka <jh@suse.cz>
5721 PR tree-optimization/110649
5722 * tree-vect-loop.cc (optimize_mask_stores): Set correctly
5723 probability of the if-then-else construct.
5725 2023-07-16 Jan Hubicka <jh@suse.cz>
5727 PR middle-end/110649
5728 * tree-ssa-loop-ivcanon.cc (try_peel_loop): Avoid double profile update.
5730 2023-07-15 Andrew Pinski <apinski@marvell.com>
5732 * doc/contrib.texi: Update my entry.
5734 2023-07-15 John David Anglin <danglin@gcc.gnu.org>
5736 * config/pa/pa.md: Define constants R1_REGNUM, R19_REGNUM and
5738 (tgd_load): Restrict to !TARGET_64BIT. Use register constants.
5739 (tld_load): Likewise.
5740 (tgd_load_pic): Change to expander.
5741 (tld_load_pic, tld_offset_load, tp_load): Likewise.
5742 (tie_load_pic, tle_load): Likewise.
5743 (tgd_load_picsi, tgd_load_picdi): New.
5744 (tld_load_picsi, tld_load_picdi): New.
5745 (tld_offset_load<P:mode>): New.
5746 (tp_load<P:mode>): New.
5747 (tie_load_picsi, tie_load_picdi): New.
5748 (tle_load<P:mode>): New.
5750 2023-07-14 Christophe Lyon <christophe.lyon@linaro.org>
5752 * config/arm/arm-mve-builtins-base.cc (vcmlaq, vcmlaq_rot90)
5753 (vcmlaq_rot180, vcmlaq_rot270): New.
5754 * config/arm/arm-mve-builtins-base.def (vcmlaq, vcmlaq_rot90)
5755 (vcmlaq_rot180, vcmlaq_rot270): New.
5756 * config/arm/arm-mve-builtins-base.h: (vcmlaq, vcmlaq_rot90)
5757 (vcmlaq_rot180, vcmlaq_rot270): New.
5758 * config/arm/arm-mve-builtins.cc
5759 (function_instance::has_inactive_argument): Handle vcmlaq,
5760 vcmlaq_rot90, vcmlaq_rot180, vcmlaq_rot270.
5761 * config/arm/arm_mve.h (vcmlaq): Delete.
5762 (vcmlaq_rot180): Delete.
5763 (vcmlaq_rot270): Delete.
5764 (vcmlaq_rot90): Delete.
5766 (vcmlaq_rot180_m): Delete.
5767 (vcmlaq_rot270_m): Delete.
5768 (vcmlaq_rot90_m): Delete.
5769 (vcmlaq_f16): Delete.
5770 (vcmlaq_rot180_f16): Delete.
5771 (vcmlaq_rot270_f16): Delete.
5772 (vcmlaq_rot90_f16): Delete.
5773 (vcmlaq_f32): Delete.
5774 (vcmlaq_rot180_f32): Delete.
5775 (vcmlaq_rot270_f32): Delete.
5776 (vcmlaq_rot90_f32): Delete.
5777 (vcmlaq_m_f32): Delete.
5778 (vcmlaq_m_f16): Delete.
5779 (vcmlaq_rot180_m_f32): Delete.
5780 (vcmlaq_rot180_m_f16): Delete.
5781 (vcmlaq_rot270_m_f32): Delete.
5782 (vcmlaq_rot270_m_f16): Delete.
5783 (vcmlaq_rot90_m_f32): Delete.
5784 (vcmlaq_rot90_m_f16): Delete.
5785 (__arm_vcmlaq_f16): Delete.
5786 (__arm_vcmlaq_rot180_f16): Delete.
5787 (__arm_vcmlaq_rot270_f16): Delete.
5788 (__arm_vcmlaq_rot90_f16): Delete.
5789 (__arm_vcmlaq_f32): Delete.
5790 (__arm_vcmlaq_rot180_f32): Delete.
5791 (__arm_vcmlaq_rot270_f32): Delete.
5792 (__arm_vcmlaq_rot90_f32): Delete.
5793 (__arm_vcmlaq_m_f32): Delete.
5794 (__arm_vcmlaq_m_f16): Delete.
5795 (__arm_vcmlaq_rot180_m_f32): Delete.
5796 (__arm_vcmlaq_rot180_m_f16): Delete.
5797 (__arm_vcmlaq_rot270_m_f32): Delete.
5798 (__arm_vcmlaq_rot270_m_f16): Delete.
5799 (__arm_vcmlaq_rot90_m_f32): Delete.
5800 (__arm_vcmlaq_rot90_m_f16): Delete.
5801 (__arm_vcmlaq): Delete.
5802 (__arm_vcmlaq_rot180): Delete.
5803 (__arm_vcmlaq_rot270): Delete.
5804 (__arm_vcmlaq_rot90): Delete.
5805 (__arm_vcmlaq_m): Delete.
5806 (__arm_vcmlaq_rot180_m): Delete.
5807 (__arm_vcmlaq_rot270_m): Delete.
5808 (__arm_vcmlaq_rot90_m): Delete.
5810 2023-07-14 Christophe Lyon <christophe.lyon@linaro.org>
5812 * config/arm/arm_mve_builtins.def (vcmlaq_rot90_f)
5813 (vcmlaq_rot270_f, vcmlaq_rot180_f, vcmlaq_f): Add "_f" suffix.
5814 * config/arm/iterators.md (MVE_VCMLAQ_M): New.
5815 (mve_insn): Add vcmla.
5816 (rot): Add VCMLAQ_M_F, VCMLAQ_ROT90_M_F, VCMLAQ_ROT180_M_F,
5818 (mve_rot): Add VCMLAQ_M_F, VCMLAQ_ROT90_M_F, VCMLAQ_ROT180_M_F,
5820 * config/arm/mve.md (mve_vcmlaq<mve_rot><mode>): Rename into ...
5821 (@mve_<mve_insn>q<mve_rot>_f<mode>): ... this.
5822 (mve_vcmlaq_m_f<mode>, mve_vcmlaq_rot180_m_f<mode>)
5823 (mve_vcmlaq_rot270_m_f<mode>, mve_vcmlaq_rot90_m_f<mode>): Merge
5825 (@mve_<mve_insn>q<mve_rot>_m_f<mode>): ... this.
5827 2023-07-14 Christophe Lyon <christophe.lyon@linaro.org>
5829 * config/arm/arm-mve-builtins-base.cc (vcmulq, vcmulq_rot90)
5830 (vcmulq_rot180, vcmulq_rot270): New.
5831 * config/arm/arm-mve-builtins-base.def (vcmulq, vcmulq_rot90)
5832 (vcmulq_rot180, vcmulq_rot270): New.
5833 * config/arm/arm-mve-builtins-base.h: (vcmulq, vcmulq_rot90)
5834 (vcmulq_rot180, vcmulq_rot270): New.
5835 * config/arm/arm_mve.h (vcmulq_rot90): Delete.
5836 (vcmulq_rot270): Delete.
5837 (vcmulq_rot180): Delete.
5840 (vcmulq_rot180_m): Delete.
5841 (vcmulq_rot270_m): Delete.
5842 (vcmulq_rot90_m): Delete.
5844 (vcmulq_rot90_x): Delete.
5845 (vcmulq_rot180_x): Delete.
5846 (vcmulq_rot270_x): Delete.
5847 (vcmulq_rot90_f16): Delete.
5848 (vcmulq_rot270_f16): Delete.
5849 (vcmulq_rot180_f16): Delete.
5850 (vcmulq_f16): Delete.
5851 (vcmulq_rot90_f32): Delete.
5852 (vcmulq_rot270_f32): Delete.
5853 (vcmulq_rot180_f32): Delete.
5854 (vcmulq_f32): Delete.
5855 (vcmulq_m_f32): Delete.
5856 (vcmulq_m_f16): Delete.
5857 (vcmulq_rot180_m_f32): Delete.
5858 (vcmulq_rot180_m_f16): Delete.
5859 (vcmulq_rot270_m_f32): Delete.
5860 (vcmulq_rot270_m_f16): Delete.
5861 (vcmulq_rot90_m_f32): Delete.
5862 (vcmulq_rot90_m_f16): Delete.
5863 (vcmulq_x_f16): Delete.
5864 (vcmulq_x_f32): Delete.
5865 (vcmulq_rot90_x_f16): Delete.
5866 (vcmulq_rot90_x_f32): Delete.
5867 (vcmulq_rot180_x_f16): Delete.
5868 (vcmulq_rot180_x_f32): Delete.
5869 (vcmulq_rot270_x_f16): Delete.
5870 (vcmulq_rot270_x_f32): Delete.
5871 (__arm_vcmulq_rot90_f16): Delete.
5872 (__arm_vcmulq_rot270_f16): Delete.
5873 (__arm_vcmulq_rot180_f16): Delete.
5874 (__arm_vcmulq_f16): Delete.
5875 (__arm_vcmulq_rot90_f32): Delete.
5876 (__arm_vcmulq_rot270_f32): Delete.
5877 (__arm_vcmulq_rot180_f32): Delete.
5878 (__arm_vcmulq_f32): Delete.
5879 (__arm_vcmulq_m_f32): Delete.
5880 (__arm_vcmulq_m_f16): Delete.
5881 (__arm_vcmulq_rot180_m_f32): Delete.
5882 (__arm_vcmulq_rot180_m_f16): Delete.
5883 (__arm_vcmulq_rot270_m_f32): Delete.
5884 (__arm_vcmulq_rot270_m_f16): Delete.
5885 (__arm_vcmulq_rot90_m_f32): Delete.
5886 (__arm_vcmulq_rot90_m_f16): Delete.
5887 (__arm_vcmulq_x_f16): Delete.
5888 (__arm_vcmulq_x_f32): Delete.
5889 (__arm_vcmulq_rot90_x_f16): Delete.
5890 (__arm_vcmulq_rot90_x_f32): Delete.
5891 (__arm_vcmulq_rot180_x_f16): Delete.
5892 (__arm_vcmulq_rot180_x_f32): Delete.
5893 (__arm_vcmulq_rot270_x_f16): Delete.
5894 (__arm_vcmulq_rot270_x_f32): Delete.
5895 (__arm_vcmulq_rot90): Delete.
5896 (__arm_vcmulq_rot270): Delete.
5897 (__arm_vcmulq_rot180): Delete.
5898 (__arm_vcmulq): Delete.
5899 (__arm_vcmulq_m): Delete.
5900 (__arm_vcmulq_rot180_m): Delete.
5901 (__arm_vcmulq_rot270_m): Delete.
5902 (__arm_vcmulq_rot90_m): Delete.
5903 (__arm_vcmulq_x): Delete.
5904 (__arm_vcmulq_rot90_x): Delete.
5905 (__arm_vcmulq_rot180_x): Delete.
5906 (__arm_vcmulq_rot270_x): Delete.
5908 2023-07-14 Christophe Lyon <christophe.lyon@linaro.org>
5910 * config/arm/arm_mve_builtins.def (vcmulq_rot90_f)
5911 (vcmulq_rot270_f, vcmulq_rot180_f, vcmulq_f): Add "_f" suffix.
5912 * config/arm/iterators.md (MVE_VCADDQ_VCMULQ)
5913 (MVE_VCADDQ_VCMULQ_M): New.
5914 (mve_insn): Add vcmul.
5915 (rot): Add VCMULQ_M_F, VCMULQ_ROT90_M_F, VCMULQ_ROT180_M_F,
5918 (mve_rot): Add VCMULQ_M_F, VCMULQ_ROT90_M_F, VCMULQ_ROT180_M_F,
5920 * config/arm/mve.md (mve_vcmulq<mve_rot><mode>): Merge into
5921 @mve_<mve_insn>q<mve_rot>_f<mode>.
5922 (mve_vcmulq_m_f<mode>, mve_vcmulq_rot180_m_f<mode>)
5923 (mve_vcmulq_rot270_m_f<mode>, mve_vcmulq_rot90_m_f<mode>): Merge
5924 into @mve_<mve_insn>q<mve_rot>_m_f<mode>.
5926 2023-07-14 Christophe Lyon <christophe.lyon@linaro.org>
5928 * config/arm/arm-mve-builtins-base.cc (vcaddq_rot90)
5929 (vcaddq_rot270, vhcaddq_rot90, vhcaddq_rot270): New.
5930 * config/arm/arm-mve-builtins-base.def (vcaddq_rot90)
5931 (vcaddq_rot270, vhcaddq_rot90, vhcaddq_rot270): New.
5932 * config/arm/arm-mve-builtins-base.h: (vcaddq_rot90)
5933 (vcaddq_rot270, vhcaddq_rot90, vhcaddq_rot270): New.
5934 * config/arm/arm-mve-builtins-functions.h (class
5935 unspec_mve_function_exact_insn_rot): New.
5936 * config/arm/arm_mve.h (vcaddq_rot90): Delete.
5937 (vcaddq_rot270): Delete.
5938 (vhcaddq_rot90): Delete.
5939 (vhcaddq_rot270): Delete.
5940 (vcaddq_rot270_m): Delete.
5941 (vcaddq_rot90_m): Delete.
5942 (vhcaddq_rot270_m): Delete.
5943 (vhcaddq_rot90_m): Delete.
5944 (vcaddq_rot90_x): Delete.
5945 (vcaddq_rot270_x): Delete.
5946 (vhcaddq_rot90_x): Delete.
5947 (vhcaddq_rot270_x): Delete.
5948 (vcaddq_rot90_u8): Delete.
5949 (vcaddq_rot270_u8): Delete.
5950 (vhcaddq_rot90_s8): Delete.
5951 (vhcaddq_rot270_s8): Delete.
5952 (vcaddq_rot90_s8): Delete.
5953 (vcaddq_rot270_s8): Delete.
5954 (vcaddq_rot90_u16): Delete.
5955 (vcaddq_rot270_u16): Delete.
5956 (vhcaddq_rot90_s16): Delete.
5957 (vhcaddq_rot270_s16): Delete.
5958 (vcaddq_rot90_s16): Delete.
5959 (vcaddq_rot270_s16): Delete.
5960 (vcaddq_rot90_u32): Delete.
5961 (vcaddq_rot270_u32): Delete.
5962 (vhcaddq_rot90_s32): Delete.
5963 (vhcaddq_rot270_s32): Delete.
5964 (vcaddq_rot90_s32): Delete.
5965 (vcaddq_rot270_s32): Delete.
5966 (vcaddq_rot90_f16): Delete.
5967 (vcaddq_rot270_f16): Delete.
5968 (vcaddq_rot90_f32): Delete.
5969 (vcaddq_rot270_f32): Delete.
5970 (vcaddq_rot270_m_s8): Delete.
5971 (vcaddq_rot270_m_s32): Delete.
5972 (vcaddq_rot270_m_s16): Delete.
5973 (vcaddq_rot270_m_u8): Delete.
5974 (vcaddq_rot270_m_u32): Delete.
5975 (vcaddq_rot270_m_u16): Delete.
5976 (vcaddq_rot90_m_s8): Delete.
5977 (vcaddq_rot90_m_s32): Delete.
5978 (vcaddq_rot90_m_s16): Delete.
5979 (vcaddq_rot90_m_u8): Delete.
5980 (vcaddq_rot90_m_u32): Delete.
5981 (vcaddq_rot90_m_u16): Delete.
5982 (vhcaddq_rot270_m_s8): Delete.
5983 (vhcaddq_rot270_m_s32): Delete.
5984 (vhcaddq_rot270_m_s16): Delete.
5985 (vhcaddq_rot90_m_s8): Delete.
5986 (vhcaddq_rot90_m_s32): Delete.
5987 (vhcaddq_rot90_m_s16): Delete.
5988 (vcaddq_rot270_m_f32): Delete.
5989 (vcaddq_rot270_m_f16): Delete.
5990 (vcaddq_rot90_m_f32): Delete.
5991 (vcaddq_rot90_m_f16): Delete.
5992 (vcaddq_rot90_x_s8): Delete.
5993 (vcaddq_rot90_x_s16): Delete.
5994 (vcaddq_rot90_x_s32): Delete.
5995 (vcaddq_rot90_x_u8): Delete.
5996 (vcaddq_rot90_x_u16): Delete.
5997 (vcaddq_rot90_x_u32): Delete.
5998 (vcaddq_rot270_x_s8): Delete.
5999 (vcaddq_rot270_x_s16): Delete.
6000 (vcaddq_rot270_x_s32): Delete.
6001 (vcaddq_rot270_x_u8): Delete.
6002 (vcaddq_rot270_x_u16): Delete.
6003 (vcaddq_rot270_x_u32): Delete.
6004 (vhcaddq_rot90_x_s8): Delete.
6005 (vhcaddq_rot90_x_s16): Delete.
6006 (vhcaddq_rot90_x_s32): Delete.
6007 (vhcaddq_rot270_x_s8): Delete.
6008 (vhcaddq_rot270_x_s16): Delete.
6009 (vhcaddq_rot270_x_s32): Delete.
6010 (vcaddq_rot90_x_f16): Delete.
6011 (vcaddq_rot90_x_f32): Delete.
6012 (vcaddq_rot270_x_f16): Delete.
6013 (vcaddq_rot270_x_f32): Delete.
6014 (__arm_vcaddq_rot90_u8): Delete.
6015 (__arm_vcaddq_rot270_u8): Delete.
6016 (__arm_vhcaddq_rot90_s8): Delete.
6017 (__arm_vhcaddq_rot270_s8): Delete.
6018 (__arm_vcaddq_rot90_s8): Delete.
6019 (__arm_vcaddq_rot270_s8): Delete.
6020 (__arm_vcaddq_rot90_u16): Delete.
6021 (__arm_vcaddq_rot270_u16): Delete.
6022 (__arm_vhcaddq_rot90_s16): Delete.
6023 (__arm_vhcaddq_rot270_s16): Delete.
6024 (__arm_vcaddq_rot90_s16): Delete.
6025 (__arm_vcaddq_rot270_s16): Delete.
6026 (__arm_vcaddq_rot90_u32): Delete.
6027 (__arm_vcaddq_rot270_u32): Delete.
6028 (__arm_vhcaddq_rot90_s32): Delete.
6029 (__arm_vhcaddq_rot270_s32): Delete.
6030 (__arm_vcaddq_rot90_s32): Delete.
6031 (__arm_vcaddq_rot270_s32): Delete.
6032 (__arm_vcaddq_rot270_m_s8): Delete.
6033 (__arm_vcaddq_rot270_m_s32): Delete.
6034 (__arm_vcaddq_rot270_m_s16): Delete.
6035 (__arm_vcaddq_rot270_m_u8): Delete.
6036 (__arm_vcaddq_rot270_m_u32): Delete.
6037 (__arm_vcaddq_rot270_m_u16): Delete.
6038 (__arm_vcaddq_rot90_m_s8): Delete.
6039 (__arm_vcaddq_rot90_m_s32): Delete.
6040 (__arm_vcaddq_rot90_m_s16): Delete.
6041 (__arm_vcaddq_rot90_m_u8): Delete.
6042 (__arm_vcaddq_rot90_m_u32): Delete.
6043 (__arm_vcaddq_rot90_m_u16): Delete.
6044 (__arm_vhcaddq_rot270_m_s8): Delete.
6045 (__arm_vhcaddq_rot270_m_s32): Delete.
6046 (__arm_vhcaddq_rot270_m_s16): Delete.
6047 (__arm_vhcaddq_rot90_m_s8): Delete.
6048 (__arm_vhcaddq_rot90_m_s32): Delete.
6049 (__arm_vhcaddq_rot90_m_s16): Delete.
6050 (__arm_vcaddq_rot90_x_s8): Delete.
6051 (__arm_vcaddq_rot90_x_s16): Delete.
6052 (__arm_vcaddq_rot90_x_s32): Delete.
6053 (__arm_vcaddq_rot90_x_u8): Delete.
6054 (__arm_vcaddq_rot90_x_u16): Delete.
6055 (__arm_vcaddq_rot90_x_u32): Delete.
6056 (__arm_vcaddq_rot270_x_s8): Delete.
6057 (__arm_vcaddq_rot270_x_s16): Delete.
6058 (__arm_vcaddq_rot270_x_s32): Delete.
6059 (__arm_vcaddq_rot270_x_u8): Delete.
6060 (__arm_vcaddq_rot270_x_u16): Delete.
6061 (__arm_vcaddq_rot270_x_u32): Delete.
6062 (__arm_vhcaddq_rot90_x_s8): Delete.
6063 (__arm_vhcaddq_rot90_x_s16): Delete.
6064 (__arm_vhcaddq_rot90_x_s32): Delete.
6065 (__arm_vhcaddq_rot270_x_s8): Delete.
6066 (__arm_vhcaddq_rot270_x_s16): Delete.
6067 (__arm_vhcaddq_rot270_x_s32): Delete.
6068 (__arm_vcaddq_rot90_f16): Delete.
6069 (__arm_vcaddq_rot270_f16): Delete.
6070 (__arm_vcaddq_rot90_f32): Delete.
6071 (__arm_vcaddq_rot270_f32): Delete.
6072 (__arm_vcaddq_rot270_m_f32): Delete.
6073 (__arm_vcaddq_rot270_m_f16): Delete.
6074 (__arm_vcaddq_rot90_m_f32): Delete.
6075 (__arm_vcaddq_rot90_m_f16): Delete.
6076 (__arm_vcaddq_rot90_x_f16): Delete.
6077 (__arm_vcaddq_rot90_x_f32): Delete.
6078 (__arm_vcaddq_rot270_x_f16): Delete.
6079 (__arm_vcaddq_rot270_x_f32): Delete.
6080 (__arm_vcaddq_rot90): Delete.
6081 (__arm_vcaddq_rot270): Delete.
6082 (__arm_vhcaddq_rot90): Delete.
6083 (__arm_vhcaddq_rot270): Delete.
6084 (__arm_vcaddq_rot270_m): Delete.
6085 (__arm_vcaddq_rot90_m): Delete.
6086 (__arm_vhcaddq_rot270_m): Delete.
6087 (__arm_vhcaddq_rot90_m): Delete.
6088 (__arm_vcaddq_rot90_x): Delete.
6089 (__arm_vcaddq_rot270_x): Delete.
6090 (__arm_vhcaddq_rot90_x): Delete.
6091 (__arm_vhcaddq_rot270_x): Delete.
6093 2023-07-14 Christophe Lyon <christophe.lyon@linaro.org>
6095 * config/arm/arm_mve_builtins.def (vcaddq_rot90_, vcaddq_rot270_)
6096 (vcaddq_rot90_f, vcaddq_rot90_f): Add "_" or "_f" suffix.
6097 * config/arm/iterators.md (mve_insn): Add vcadd, vhcadd.
6098 (isu): Add UNSPEC_VCADD90, UNSPEC_VCADD270, VCADDQ_ROT270_M_U,
6099 VCADDQ_ROT270_M_S, VCADDQ_ROT90_M_U, VCADDQ_ROT90_M_S,
6100 VHCADDQ_ROT90_M_S, VHCADDQ_ROT270_M_S, VHCADDQ_ROT90_S,
6102 (rot): Add VCADDQ_ROT90_M_F, VCADDQ_ROT90_M_S, VCADDQ_ROT90_M_U,
6103 VCADDQ_ROT270_M_F, VCADDQ_ROT270_M_S, VCADDQ_ROT270_M_U,
6104 VHCADDQ_ROT90_S, VHCADDQ_ROT270_S, VHCADDQ_ROT90_M_S,
6106 (mve_rot): Add VCADDQ_ROT90_M_F, VCADDQ_ROT90_M_S,
6107 VCADDQ_ROT90_M_U, VCADDQ_ROT270_M_F, VCADDQ_ROT270_M_S,
6108 VCADDQ_ROT270_M_U, VHCADDQ_ROT90_S, VHCADDQ_ROT270_S,
6109 VHCADDQ_ROT90_M_S, VHCADDQ_ROT270_M_S.
6110 (supf): Add VHCADDQ_ROT90_M_S, VHCADDQ_ROT270_M_S,
6111 VHCADDQ_ROT90_S, VHCADDQ_ROT270_S, UNSPEC_VCADD90,
6113 (VCADDQ_ROT270_M): Delete.
6114 (VCADDQ_M_F VxCADDQ VxCADDQ_M): New.
6115 (VCADDQ_ROT90_M): Delete.
6116 * config/arm/mve.md (mve_vcaddq<mve_rot><mode>)
6117 (mve_vhcaddq_rot270_s<mode>, mve_vhcaddq_rot90_s<mode>): Merge
6119 (@mve_<mve_insn>q<mve_rot>_<supf><mode>): ... this.
6120 (mve_vcaddq<mve_rot><mode>): Rename into ...
6121 (@mve_<mve_insn>q<mve_rot>_f<mode>): ... this
6122 (mve_vcaddq_rot270_m_<supf><mode>)
6123 (mve_vcaddq_rot90_m_<supf><mode>, mve_vhcaddq_rot270_m_s<mode>)
6124 (mve_vhcaddq_rot90_m_s<mode>): Merge into ...
6125 (@mve_<mve_insn>q<mve_rot>_m_<supf><mode>): ... this.
6126 (mve_vcaddq_rot270_m_f<mode>, mve_vcaddq_rot90_m_f<mode>): Merge
6128 (@mve_<mve_insn>q<mve_rot>_m_f<mode>): ... this.
6130 2023-07-14 Roger Sayle <roger@nextmovesoftware.com>
6133 * config/i386/i386.md (*bt<mode>_setcqi): Prefer string form
6134 preparation statement over braces for a single statement.
6135 (*bt<mode>_setncqi): Likewise.
6136 (*bt<mode>_setncqi_2): New define_insn_and_split.
6138 2023-07-14 Roger Sayle <roger@nextmovesoftware.com>
6140 * config/i386/i386-expand.cc (ix86_expand_move): Generalize special
6141 case inserting of 64-bit values into a TImode register, to handle
6142 both DImode and DFmode using either *insvti_lowpart_1
6143 or *isnvti_highpart_1.
6145 2023-07-14 Uros Bizjak <ubizjak@gmail.com>
6148 * fwprop.cc (contains_paradoxical_subreg_p): Move to ...
6149 * rtlanal.cc (contains_paradoxical_subreg_p): ... here.
6150 * rtlanal.h (contains_paradoxical_subreg_p): Add prototype.
6151 * cprop.cc (try_replace_reg): Do not set REG_EQUAL note
6152 when the original source contains a paradoxical subreg.
6154 2023-07-14 Jan Hubicka <jh@suse.cz>
6156 * passes.cc (execute_function_todo): Remove
6157 TODO_rebuild_frequencies
6158 * passes.def: Add rebuild_frequencies pass.
6159 * predict.cc (estimate_bb_frequencies): Drop
6161 (tree_estimate_probability): Update call of
6162 estimate_bb_frequencies.
6163 (rebuild_frequencies): Turn into a pass; verify CFG profile consistency
6164 first and do not rebuild if not necessary.
6165 (class pass_rebuild_frequencies): New.
6166 (make_pass_rebuild_frequencies): New.
6167 * profile-count.h: Add profile_count::very_large_p.
6168 * tree-inline.cc (optimize_inline_calls): Do not return
6169 TODO_rebuild_frequencies
6170 * tree-pass.h (TODO_rebuild_frequencies): Remove.
6171 (make_pass_rebuild_frequencies): Declare.
6173 2023-07-14 Juzhe-Zhong <juzhe.zhong@rivai.ai>
6175 * config/riscv/autovec.md (cond_len_fma<mode>): New pattern.
6176 * config/riscv/riscv-protos.h (enum insn_type): New enum.
6177 (expand_cond_len_ternop): New function.
6178 * config/riscv/riscv-v.cc (emit_nonvlmax_fp_ternary_tu_insn): Ditto.
6179 (expand_cond_len_ternop): Ditto.
6181 2023-07-14 Jose E. Marchesi <jose.marchesi@oracle.com>
6184 * config/bpf/bpf.md: Enable instruction scheduling.
6186 2023-07-14 Tamar Christina <tamar.christina@arm.com>
6188 PR tree-optimization/109154
6189 * tree-if-conv.cc (INCLUDE_ALGORITHM): Include.
6190 (struct bb_predicate): Add no_predicate_stmts.
6191 (set_bb_predicate): Increase predicate count.
6192 (set_bb_predicate_gimplified_stmts): Conditionally initialize
6194 (get_bb_num_predicate_stmts): New.
6195 (init_bb_predicate): Initialzie no_predicate_stmts.
6196 (release_bb_predicate): Cleanup no_predicate_stmts.
6197 (insert_gimplified_predicates): Preserve no_predicate_stmts.
6199 2023-07-14 Tamar Christina <tamar.christina@arm.com>
6201 PR tree-optimization/109154
6202 * tree-if-conv.cc (gen_simplified_condition,
6203 gen_phi_nest_statement): New.
6204 (gen_phi_arg_condition, predicate_scalar_phi): Use it.
6206 2023-07-14 Richard Biener <rguenther@suse.de>
6208 * gimple.h (gimple_phi_arg): New const overload.
6209 (gimple_phi_arg_def): Make gimple arg const.
6210 (gimple_phi_arg_def_from_edge): New inline function.
6211 * tree-phinodes.h (gimple_phi_arg_imm_use_ptr_from_edge):
6213 * tree-ssa-operands.h (PHI_ARG_DEF_FROM_EDGE): Direct to
6214 new inline function.
6215 (PHI_ARG_DEF_PTR_FROM_EDGE): Likewise.
6217 2023-07-14 Monk Chiang <monk.chiang@sifive.com>
6219 * common/config/riscv/riscv-common.cc:
6220 (riscv_implied_info): Add zihintntl item.
6221 (riscv_ext_version_table): Ditto.
6222 (riscv_ext_flag_table): Ditto.
6223 * config/riscv/riscv-opts.h (MASK_ZIHINTNTL): New macro.
6224 (TARGET_ZIHINTNTL): Ditto.
6226 2023-07-14 Die Li <lidie@eswincomputing.com>
6228 * config/riscv/riscv.md: Remove redundant portion in and<mode>3.
6230 2023-07-14 Oleg Endo <olegendo@gcc.gnu.org>
6233 * config/sh/sh.md (peephole2): Handle case where eliminated reg is also
6234 used by the address of the following memory operand.
6236 2023-07-13 Mikael Pettersson <mikpelinux@gmail.com>
6239 * config/pdp11/pdp11.cc (pdp11_expand_epilogue): Also
6240 deallocate alloca-only frame.
6242 2023-07-13 Iain Sandoe <iain@sandoe.co.uk>
6245 * config/darwin.h (DARWIN_PLATFORM_ID): New.
6246 (LINK_COMMAND_A): Use DARWIN_PLATFORM_ID to pass OS, OS version
6247 and SDK data to the static linker.
6249 2023-07-13 Carl Love <cel@us.ibm.com>
6251 * config/rs6000/rs6000-builtins.def (__builtin_set_fpscr_rn): Update
6252 built-in definition return type.
6253 * config/rs6000/rs6000-c.cc (rs6000_target_modify_macros): Add check,
6254 define __SET_FPSCR_RN_RETURNS_FPSCR__ macro.
6255 * config/rs6000/rs6000.md (rs6000_set_fpscr_rn): Add return
6256 argument to return FPSCR fields.
6257 * doc/extend.texi (__builtin_set_fpscr_rn): Update description for
6258 the return value. Add description for
6259 __SET_FPSCR_RN_RETURNS_FPSCR__ macro.
6261 2023-07-13 Uros Bizjak <ubizjak@gmail.com>
6264 * config/alpha/alpha.cc (alpha_emit_set_long_const):
6265 Always use DImode when constructing long const.
6267 2023-07-13 Uros Bizjak <ubizjak@gmail.com>
6269 * haifa-sched.cc: Change TRUE/FALSE to true/false.
6271 * lra-assigns.cc: Ditto.
6272 * lra-constraints.cc: Ditto.
6273 * sel-sched.cc: Ditto.
6275 2023-07-13 Andrew Pinski <apinski@marvell.com>
6277 PR tree-optimization/110293
6278 PR tree-optimization/110539
6279 * match.pd: Expand the `x != (typeof x)(x == 0)`
6280 pattern to handle where the inner and outer comparsions
6281 are either `!=` or `==` and handle other constants
6284 2023-07-13 Vladimir N. Makarov <vmakarov@redhat.com>
6286 PR middle-end/109520
6287 * lra-int.h (lra_insn_recog_data): Add member asm_reloads_num.
6288 (lra_asm_insn_error): New prototype.
6289 * lra.cc: Include rtl_error.h.
6290 (lra_set_insn_recog_data): Initialize asm_reloads_num.
6291 (lra_asm_insn_error): New func whose code is taken from ...
6292 * lra-assigns.cc (lra_split_hard_reg_for): ... here. Use lra_asm_insn_error.
6293 * lra-constraints.cc (curr_insn_transform): Check reloads nummber for asm.
6295 2023-07-13 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
6297 * genmatch.cc (commutative_op): Add COND_LEN_*
6298 * internal-fn.cc (first_commutative_argument): Ditto.
6300 (get_unconditional_internal_fn): Ditto.
6301 (can_interpret_as_conditional_op_p): Ditto.
6302 (internal_fn_len_index): Ditto.
6303 * internal-fn.h (can_interpret_as_conditional_op_p): Ditt.
6304 * tree-ssa-math-opts.cc (convert_mult_to_fma_1): Ditto.
6305 (convert_mult_to_fma): Ditto.
6306 (math_opts_dom_walker::after_dom_children): Ditto.
6308 2023-07-13 Pan Li <pan2.li@intel.com>
6310 * config/riscv/riscv.cc (vxrm_rtx): New static var.
6312 (global_state_unknown_p): Removed.
6313 (riscv_entity_mode_after): Removed.
6314 (asm_insn_p): New function.
6315 (vxrm_unknown_p): New function for fixed-point.
6316 (riscv_vxrm_mode_after): Ditto.
6317 (frm_unknown_dynamic_p): New function for floating-point.
6318 (riscv_frm_mode_after): Ditto.
6319 (riscv_mode_after): Leverage new functions.
6321 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
6323 * tree-vect-stmts.cc (vect_model_load_cost): Remove.
6324 (vectorizable_load): Adjust the cost handling on VMAT_CONTIGUOUS without
6325 calling vect_model_load_cost.
6327 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
6329 * tree-vect-stmts.cc (vect_model_load_cost): Assert this function only
6330 handle memory_access_type VMAT_CONTIGUOUS, remove some
6331 VMAT_CONTIGUOUS_PERMUTE related handlings.
6332 (vectorizable_load): Adjust the cost handling on VMAT_CONTIGUOUS_PERMUTE
6333 without calling vect_model_load_cost.
6335 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
6337 * tree-vect-stmts.cc (vect_model_load_cost): Assert it won't get
6338 VMAT_CONTIGUOUS_REVERSE any more.
6339 (vectorizable_load): Adjust the costing handling on
6340 VMAT_CONTIGUOUS_REVERSE without calling vect_model_load_cost.
6342 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
6344 * tree-vect-stmts.cc (vectorizable_load): Adjust the cost handling on
6345 VMAT_LOAD_STORE_LANES without calling vect_model_load_cost.
6346 (vectorizable_load): Remove VMAT_LOAD_STORE_LANES related handling and
6347 assert it will never get VMAT_LOAD_STORE_LANES.
6349 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
6351 * tree-vect-stmts.cc (vectorizable_load): Adjust the cost handling on
6352 VMAT_GATHER_SCATTER without calling vect_model_load_cost.
6353 (vect_model_load_cost): Adjut the assertion on VMAT_GATHER_SCATTER,
6354 remove VMAT_GATHER_SCATTER related handlings and the related parameter
6357 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
6359 * tree-vect-stmts.cc (vectorizable_load): Adjust the cost handling
6360 on VMAT_ELEMENTWISE and VMAT_STRIDED_SLP without calling
6361 vect_model_load_cost.
6362 (vect_model_load_cost): Assert it won't get VMAT_ELEMENTWISE and
6363 VMAT_STRIDED_SLP any more, and remove their related handlings.
6365 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
6367 * tree-vect-stmts.cc (hoist_defs_of_uses): Add one argument HOIST_P.
6368 (vectorizable_load): Adjust the handling on VMAT_INVARIANT to respect
6369 hoisting decision and without calling vect_model_load_cost.
6370 (vect_model_load_cost): Assert it won't get VMAT_INVARIANT any more
6371 and remove VMAT_INVARIANT related handlings.
6373 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
6375 * tree-vect-stmts.cc (vect_build_gather_load_calls): Add the handlings
6376 on costing with one extra argument cost_vec.
6377 (vectorizable_load): Adjust the call to vect_build_gather_load_calls.
6378 (vect_model_load_cost): Assert it won't get VMAT_GATHER_SCATTER with
6379 gs_info.decl set any more.
6381 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
6383 * tree-vect-stmts.cc (vectorizable_load): Move and duplicate the call
6384 to vect_model_load_cost down to some different transform paths
6385 according to the handlings of different vect_memory_access_types.
6387 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
6389 * tree.h (wi::from_mpz): Hide from GENERATOR_FILE.
6391 2023-07-13 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
6393 * config/riscv/autovec.md
6394 (len_mask_gather_load<VNX1_QHSD:mode><VNX1_QHSDI:mode>): New pattern.
6395 (len_mask_gather_load<VNX2_QHSD:mode><VNX2_QHSDI:mode>): Ditto.
6396 (len_mask_gather_load<VNX4_QHSD:mode><VNX4_QHSDI:mode>): Ditto.
6397 (len_mask_gather_load<VNX8_QHSD:mode><VNX8_QHSDI:mode>): Ditto.
6398 (len_mask_gather_load<VNX16_QHSD:mode><VNX16_QHSDI:mode>): Ditto.
6399 (len_mask_gather_load<VNX32_QHS:mode><VNX32_QHSI:mode>): Ditto.
6400 (len_mask_gather_load<VNX64_QH:mode><VNX64_QHI:mode>): Ditto.
6401 (len_mask_gather_load<mode><mode>): Ditto.
6402 (len_mask_scatter_store<VNX1_QHSD:mode><VNX1_QHSDI:mode>): Ditto.
6403 (len_mask_scatter_store<VNX2_QHSD:mode><VNX2_QHSDI:mode>): Ditto.
6404 (len_mask_scatter_store<VNX4_QHSD:mode><VNX4_QHSDI:mode>): Ditto.
6405 (len_mask_scatter_store<VNX8_QHSD:mode><VNX8_QHSDI:mode>): Ditto.
6406 (len_mask_scatter_store<VNX16_QHSD:mode><VNX16_QHSDI:mode>): Ditto.
6407 (len_mask_scatter_store<VNX32_QHS:mode><VNX32_QHSI:mode>): Ditto.
6408 (len_mask_scatter_store<VNX64_QH:mode><VNX64_QHI:mode>): Ditto.
6409 (len_mask_scatter_store<mode><mode>): Ditto.
6410 * config/riscv/predicates.md (const_1_operand): New predicate.
6411 (vector_gs_scale_operand_16): Ditto.
6412 (vector_gs_scale_operand_32): Ditto.
6413 (vector_gs_scale_operand_64): Ditto.
6414 (vector_gs_extension_operand): Ditto.
6415 (vector_gs_scale_operand_16_rv32): Ditto.
6416 (vector_gs_scale_operand_32_rv32): Ditto.
6417 * config/riscv/riscv-protos.h (enum insn_type): Add gather/scatter.
6418 (expand_gather_scatter): New function.
6419 * config/riscv/riscv-v.cc (gen_const_vector_dup): Add gather/scatter.
6420 (emit_vlmax_masked_store_insn): New function.
6421 (emit_nonvlmax_masked_store_insn): Ditto.
6422 (modulo_sel_indices): Ditto.
6423 (expand_vec_perm): Fix SLP for gather/scatter.
6424 (prepare_gather_scatter): New function.
6425 (expand_gather_scatter): Ditto.
6426 * config/riscv/riscv.cc (riscv_legitimize_move): Fix bug of
6427 (subreg:SI (DI CONST_POLY_INT)).
6428 * config/riscv/vector-iterators.md: Add gather/scatter.
6429 * config/riscv/vector.md (vec_duplicate<mode>): Use "@" instead.
6430 (@vec_duplicate<mode>): Ditto.
6431 (@pred_indexed_<order>store<VNX16_QHS:mode><VNX16_QHSDI:mode>):
6433 (@pred_indexed_<order>store<VNX16_QHSD:mode><VNX16_QHSDI:mode>): Ditto.
6435 2023-07-12 Juzhe-Zhong <juzhe.zhong@rivai.ai>
6437 * config/riscv/autovec.md (cond_len_<optab><mode>): New pattern.
6438 * config/riscv/riscv-protos.h (enum insn_type): New enum.
6439 (expand_cond_len_binop): New function.
6440 * config/riscv/riscv-v.cc (emit_nonvlmax_tu_insn): Ditto.
6441 (emit_nonvlmax_fp_tu_insn): Ditto.
6442 (need_fp_rounding_p): Ditto.
6443 (expand_cond_len_binop): Ditto.
6444 * config/riscv/riscv.cc (riscv_preferred_else_value): Ditto.
6445 (TARGET_PREFERRED_ELSE_VALUE): New target hook.
6447 2023-07-12 Jan Hubicka <jh@suse.cz>
6449 * tree-cfg.cc (gimple_duplicate_sese_region): Rename to ...
6450 (gimple_duplicate_seme_region): ... this; break out profile updating
6452 * tree-ssa-loop-ch.cc (update_profile_after_ch): ... here.
6453 (ch_base::copy_headers): Update.
6454 * tree-cfg.h (gimple_duplicate_sese_region): Rename to ...
6455 (gimple_duplicate_seme_region): ... this.
6457 2023-07-12 Aldy Hernandez <aldyh@redhat.com>
6459 PR tree-optimization/107043
6460 * range-op.cc (operator_bitwise_and::op1_range): Update bitmask.
6462 2023-07-12 Aldy Hernandez <aldyh@redhat.com>
6464 PR tree-optimization/107053
6465 * gimple-range-op.cc (cfn_popcount): Use known set bits.
6467 2023-07-12 Uros Bizjak <ubizjak@gmail.com>
6469 * ira.cc (equiv_init_varies_p): Change return type from int to bool
6470 and adjust function body accordingly.
6471 (equiv_init_movable_p): Ditto.
6472 (memref_used_between_p): Ditto.
6473 * lra-constraints.cc (valid_address_p): Ditto.
6475 2023-07-12 Aldy Hernandez <aldyh@redhat.com>
6477 * range-op.cc (irange_to_masked_value): Remove.
6478 (update_known_bitmask): Update irange value/mask pair instead of
6479 only updating nonzero bits.
6481 2023-07-12 Jan Hubicka <jh@suse.cz>
6483 * tree-cfg.cc (gimple_duplicate_sese_region): Add ORIG_ELIMINATED_EDGES
6484 parameter and rewrite profile updating code to handle edges elimination.
6485 * tree-cfg.h (gimple_duplicate_sese_region): Update prototpe.
6486 * tree-ssa-loop-ch.cc (loop_invariant_op_p): New function.
6487 (loop_iv_derived_p): New function.
6488 (should_duplicate_loop_header_p): Track invariant exit edges; fix handling
6489 of PHIs and propagation of IV derived variables.
6490 (ch_base::copy_headers): Pass around the invariant edges hash set.
6492 2023-07-12 Uros Bizjak <ubizjak@gmail.com>
6494 * ifcvt.cc (cond_exec_changed_p): Change variable to bool.
6495 (last_active_insn): Change "skip_use_p" function argument to bool.
6496 (noce_operand_ok): Change return type from int to bool.
6497 (find_cond_trap): Ditto.
6498 (block_jumps_and_fallthru_p): Change "fallthru_p" and
6499 "jump_p" variables to bool.
6500 (noce_find_if_block): Change return type from int to bool.
6501 (cond_exec_find_if_block): Ditto.
6502 (find_if_case_1): Ditto.
6503 (find_if_case_2): Ditto.
6504 (dead_or_predicable): Ditto. Change "reversep" function arg to bool.
6505 (block_jumps_and_fallthru): Rename from block_jumps_and_fallthru_p.
6506 (cond_exec_process_insns): Change return type from int to bool.
6507 Change "mod_ok" function arg to bool.
6508 (cond_exec_process_if_block): Change return type from int to bool.
6509 Change "do_multiple_p" function arg to bool. Change "then_mod_ok"
6511 (noce_emit_store_flag): Change return type from int to bool.
6512 Change "reversep" function arg to bool. Change "cond_complex"
6514 (noce_try_move): Change return type from int to bool.
6515 (noce_try_ifelse_collapse): Ditto.
6516 (noce_try_store_flag): Ditto. Change "reversep" variable to bool.
6517 (noce_try_addcc): Change return type from int to bool. Change
6518 "subtract" variable to bool.
6519 (noce_try_store_flag_constants): Change return type from int to bool.
6520 (noce_try_store_flag_mask): Ditto. Change "reversep" variable to bool.
6521 (noce_try_cmove): Change return type from int to bool.
6522 (noce_try_cmove_arith): Ditto. Change "is_mem" variable to bool.
6523 (noce_try_minmax): Change return type from int to bool. Change
6524 "unsignedp" variable to bool.
6525 (noce_try_abs): Change return type from int to bool. Change
6526 "negate" variable to bool.
6527 (noce_try_sign_mask): Change return type from int to bool.
6528 (noce_try_move): Ditto.
6529 (noce_try_store_flag_constants): Ditto.
6530 (noce_try_cmove): Ditto.
6531 (noce_try_cmove_arith): Ditto.
6532 (noce_try_minmax): Ditto. Change "unsignedp" variable to bool.
6533 (noce_try_bitop): Change return type from int to bool.
6534 (noce_operand_ok): Ditto.
6535 (noce_convert_multiple_sets): Ditto.
6536 (noce_convert_multiple_sets_1): Ditto.
6537 (noce_process_if_block): Ditto.
6538 (check_cond_move_block): Ditto.
6539 (cond_move_process_if_block): Ditto. Change "success_p"
6541 (rest_of_handle_if_conversion): Change return type to void.
6543 2023-07-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
6545 * internal-fn.cc (FOR_EACH_CODE_MAPPING): Adapt for COND_LEN_* support.
6547 (get_conditional_len_internal_fn): New function.
6548 * internal-fn.h (get_conditional_len_internal_fn): Ditto.
6549 * tree-vect-stmts.cc (vectorizable_operation): Adapt for COND_LEN_*
6552 2023-07-12 Roger Sayle <roger@nextmovesoftware.com>
6555 * config/i386/i386.md (*add<dwi>3_doubleword_concat_zext): Typo.
6557 2023-07-12 Roger Sayle <roger@nextmovesoftware.com>
6560 * config/i386/i386.md (*add<dwi>3_doubleword_concat_zext): New
6561 define_insn_and_split derived from *add<dwi>3_doubleword_concat
6562 and *add<dwi>3_doubleword_zext.
6564 2023-07-12 Roger Sayle <roger@nextmovesoftware.com>
6567 * config/i386/i386.md (peephole2): Check !reg_mentioned_p when
6568 optimizing rega = 0; rega op= regb for op in [XOR,IOR,PLUS].
6569 (peephole2): Simplify rega = 0; rega op= rega cases.
6571 2023-07-12 Roger Sayle <roger@nextmovesoftware.com>
6573 * config/i386/i386-expand.cc (ix86_expand_int_compare): If
6574 testing a TImode SUBREG of a 128-bit vector register against
6575 zero, use a PTEST instruction instead of first moving it to
6576 a pair of scalar registers.
6578 2023-07-12 Robin Dapp <rdapp@ventanamicro.com>
6580 * genopinit.cc (main): Adjust maximal number of optabs and
6582 * gensupport.cc (find_optab): Shift optab by 20 and mode by
6584 * optabs-query.h (optab_handler): Ditto.
6585 (convert_optab_handler): Ditto.
6587 2023-07-12 Richard Biener <rguenther@suse.de>
6589 PR tree-optimization/110630
6590 * tree-vect-slp.cc (vect_add_slp_permutation): New
6591 offset parameter, honor that for the extract code generation.
6592 (vectorizable_slp_permutation_1): Handle offsetted identities.
6594 2023-07-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
6596 * config/riscv/autovec.md (smul<mode>3_highpart): New pattern.
6597 (umul<mode>3_highpart): Ditto.
6599 2023-07-12 Jan Beulich <jbeulich@suse.com>
6601 * config/i386/i386.md (extendbfsf2_1): Add new AVX512F
6602 alternative. Adjust original last alternative's "prefix"
6603 attribute to maybe_evex.
6605 2023-07-12 Jan Beulich <jbeulich@suse.com>
6607 * config/i386/sse.md (vec_dupv4sf): Make first alternative use
6608 vbroadcastss for AVX2. New AVX512F alternative.
6609 (*vec_dupv4si): New AVX2 and AVX512F alternatives using
6610 vpbroadcastd. Replace sselog1 by sseshuf1 in "type" attribute.
6612 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
6614 * config/riscv/peephole.md: Remove XThead* peephole passes.
6615 * config/riscv/thead.md: Include thead-peephole.md.
6616 * config/riscv/thead-peephole.md: New file.
6618 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
6620 * config/riscv/riscv-protos.h (riscv_regno_ok_for_index_p):
6622 (riscv_index_reg_class): Likewise.
6623 * config/riscv/riscv.cc (riscv_regno_ok_for_index_p): New function.
6624 (riscv_index_reg_class): New function.
6625 * config/riscv/riscv.h (INDEX_REG_CLASS): Call new function
6626 riscv_index_reg_class().
6627 (REGNO_OK_FOR_INDEX_P): Call new function
6628 riscv_regno_ok_for_index_p().
6630 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
6632 * config/riscv/riscv-protos.h (enum riscv_address_type):
6633 New location of type definition.
6634 (struct riscv_address_info): Likewise.
6635 * config/riscv/riscv.cc (enum riscv_address_type):
6636 Old location of type definition.
6637 (struct riscv_address_info): Likewise.
6639 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
6641 * config/riscv/riscv.h (Xmode): New macro.
6643 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
6645 * config/riscv/riscv.cc (riscv_print_operand_address): Use
6646 output_addr_const rather than riscv_print_operand.
6648 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
6650 * config/riscv/thead.md: Adjust constraints of th_addsl.
6652 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
6654 * config/riscv/thead.cc (th_mempair_operands_p):
6655 Fix documentation of th_mempair_order_operands().
6657 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
6659 * config/riscv/thead.cc (th_mempair_save_regs):
6660 Emit REG_FRAME_RELATED_EXPR notes in prologue.
6662 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
6664 * config/riscv/riscv.md: No base-ISA extension splitter for XThead*.
6665 * config/riscv/thead.md (*extend<SHORT:mode><SUPERQI:mode>2_th_ext):
6666 New XThead extension INSN.
6667 (*zero_extendsidi2_th_extu): New XThead extension INSN.
6668 (*zero_extendhi<GPR:mode>2_th_extu): New XThead extension INSN.
6670 2023-07-12 liuhongt <hongtao.liu@intel.com>
6674 * config/i386/predicates.md
6675 (int_float_vector_all_ones_operand): New predicate.
6676 * config/i386/sse.md (*vmov<mode>_constm1_pternlog_false_dep): New
6678 (*<avx512>_cvtmask2<ssemodesuffix><mode>_pternlog_false_dep):
6680 (*<avx512>_cvtmask2<ssemodesuffix><mode>_pternlog_false_dep):
6682 (*<avx512>_cvtmask2<ssemodesuffix><mode>): Adjust to
6683 define_insn_and_split to avoid false dependence.
6684 (*<avx512>_cvtmask2<ssemodesuffix><mode>): Ditto.
6685 (<mask_codefor>one_cmpl<mode>2<mask_name>): Adjust constraint
6686 of operands 1 to '0' to avoid false dependence.
6687 (*andnot<mode>3): Ditto.
6688 (iornot<mode>3): Ditto.
6689 (*<nlogic><mode>3): Ditto.
6691 2023-07-12 Mo, Zewei <zewei.mo@intel.com>
6693 * common/config/i386/cpuinfo.h
6694 (get_intel_cpu): Handle Granite Rapids D.
6695 * common/config/i386/i386-common.cc:
6696 (processor_alias_table): Add graniterapids-d.
6697 * common/config/i386/i386-cpuinfo.h
6698 (enum processor_subtypes): Add INTEL_COREI7_GRANITERAPIDS_D.
6699 * config.gcc: Add -march=graniterapids-d.
6700 * config/i386/driver-i386.cc (host_detect_local_cpu):
6701 Handle graniterapids-d.
6702 * config/i386/i386.h: (PTA_GRANITERAPIDS_D): New.
6703 * doc/extend.texi: Add graniterapids-d.
6704 * doc/invoke.texi: Ditto.
6706 2023-07-12 Haochen Jiang <haochen.jiang@intel.com>
6708 * config/i386/i386-builtins.cc (ix86_init_mmx_sse_builtins):
6709 Add OPTION_MASK_ISA_AVX512VL.
6710 * config/i386/i386-expand.cc (ix86_check_builtin_isa_match):
6713 2023-07-11 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
6715 * config/riscv/riscv-protos.h (enum insn_type): Add vcompress optimization.
6716 * config/riscv/riscv-v.cc (emit_vlmax_compress_insn): Ditto.
6717 (shuffle_compress_patterns): Ditto.
6718 (expand_vec_perm_const_1): Ditto.
6720 2023-07-11 Uros Bizjak <ubizjak@gmail.com>
6722 * cfghooks.cc (verify_flow_info): Change "err" variable to bool.
6723 * cfghooks.h (struct cfg_hooks): Change return type of
6724 verify_flow_info from integer to bool.
6725 * cfgrtl.cc (can_delete_note_p): Change return type from int to bool.
6726 (can_delete_label_p): Ditto.
6727 (rtl_verify_flow_info): Change return type from int to bool
6728 and adjust function body accordingly. Change "err" variable to bool.
6729 (rtl_verify_flow_info_1): Ditto.
6730 (free_bb_for_insn): Change return type to void.
6731 (rtl_merge_blocks): Change "b_empty" variable to bool.
6732 (try_redirect_by_replacing_jump): Change "fallthru" variable to bool.
6733 (verify_hot_cold_block_grouping): Change return type from int to bool.
6734 Change "err" variable to bool.
6735 (rtl_verify_edges): Ditto.
6736 (rtl_verify_bb_insns): Ditto.
6737 (rtl_verify_bb_pointers): Ditto.
6738 (rtl_verify_bb_insn_chain): Ditto.
6739 (rtl_verify_fallthru): Ditto.
6740 (rtl_verify_bb_layout): Ditto.
6741 (purge_all_dead_edges): Change "purged" variable to bool.
6742 * cfgrtl.h (free_bb_for_insn): Change return type from int to void.
6743 * postreload-gcse.cc (expr_hasher::equal): Change "equiv_p" to bool.
6744 (load_killed_in_block_p): Change return type from int to bool
6745 and adjust function body accordingly.
6746 (oprs_unchanged_p): Return true/false.
6747 (rest_of_handle_gcse2): Change return type to void.
6748 * tree-cfg.cc (gimple_verify_flow_info): Change return type from
6749 int to bool. Change "err" variable to bool.
6751 2023-07-11 Gaius Mulley <gaiusmod2@gmail.com>
6753 * doc/gm2.texi (-Wuninit-variable-checking=) New item.
6755 2023-07-11 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
6757 * doc/md.texi: Add COND_LEN_* operations for loop control with length.
6758 * internal-fn.cc (cond_len_unary_direct): Ditto.
6759 (cond_len_binary_direct): Ditto.
6760 (cond_len_ternary_direct): Ditto.
6761 (expand_cond_len_unary_optab_fn): Ditto.
6762 (expand_cond_len_binary_optab_fn): Ditto.
6763 (expand_cond_len_ternary_optab_fn): Ditto.
6764 (direct_cond_len_unary_optab_supported_p): Ditto.
6765 (direct_cond_len_binary_optab_supported_p): Ditto.
6766 (direct_cond_len_ternary_optab_supported_p): Ditto.
6767 * internal-fn.def (COND_LEN_ADD): Ditto.
6768 (COND_LEN_SUB): Ditto.
6769 (COND_LEN_MUL): Ditto.
6770 (COND_LEN_DIV): Ditto.
6771 (COND_LEN_MOD): Ditto.
6772 (COND_LEN_RDIV): Ditto.
6773 (COND_LEN_MIN): Ditto.
6774 (COND_LEN_MAX): Ditto.
6775 (COND_LEN_FMIN): Ditto.
6776 (COND_LEN_FMAX): Ditto.
6777 (COND_LEN_AND): Ditto.
6778 (COND_LEN_IOR): Ditto.
6779 (COND_LEN_XOR): Ditto.
6780 (COND_LEN_SHL): Ditto.
6781 (COND_LEN_SHR): Ditto.
6782 (COND_LEN_FMA): Ditto.
6783 (COND_LEN_FMS): Ditto.
6784 (COND_LEN_FNMA): Ditto.
6785 (COND_LEN_FNMS): Ditto.
6786 (COND_LEN_NEG): Ditto.
6787 * optabs.def (OPTAB_D): Ditto.
6789 2023-07-11 Richard Biener <rguenther@suse.de>
6791 PR tree-optimization/110614
6792 * tree-vect-data-refs.cc (vect_supportable_dr_alignment):
6793 SLP splats are not suitable for re-align ops.
6795 2023-07-10 Peter Bergner <bergner@linux.ibm.com>
6797 * config/rs6000/predicates.md (quad_memory_operand): Remove redundant
6799 (vsx_quad_dform_memory_operand): Likewise.
6801 2023-07-10 Uros Bizjak <ubizjak@gmail.com>
6803 * reorg.cc (stop_search_p): Change return type from int to bool
6804 and adjust function body accordingly.
6805 (resource_conflicts_p): Ditto.
6806 (insn_references_resource_p): Change return type from int to bool.
6807 (insn_sets_resource_p): Ditto.
6808 (redirect_with_delay_slots_safe_p): Ditto.
6809 (condition_dominates_p): Change return type from int to bool
6810 and adjust function body accordingly.
6811 (redirect_with_delay_list_safe_p): Ditto.
6812 (check_annul_list_true_false): Ditto. Change "annul_true_p"
6813 function argument to bool.
6814 (steal_delay_list_from_target): Change "pannul_p" function
6815 argument to bool pointer. Change "must_annul" and "used_annul"
6816 variables from int to bool.
6817 (steal_delay_list_from_fallthrough): Ditto.
6818 (own_thread_p): Change return type from int to bool and adjust
6819 function body accordingly. Change "allow_fallthrough" function
6821 (reorg_redirect_jump): Change return type from int to bool.
6822 (fill_simple_delay_slots): Change "non_jumps_p" function
6823 argument from int to bool. Change "maybe_never" varible to bool.
6824 (fill_slots_from_thread): Change "likely", "thread_if_true" and
6825 "own_thread" function arguments to bool. Change "lose" and
6826 "must_annul" variables to bool.
6827 (delete_from_delay_slot): Change "had_barrier" variable to bool.
6828 (try_merge_delay_insns): Change "annul_p" variable to bool.
6829 (fill_eager_delay_slots): Change "own_target" and "own_fallthrouhg"
6831 (rest_of_handle_delay_slots): Change return type from int to void
6832 and adjust function body accordingly.
6834 2023-07-10 Kito Cheng <kito.cheng@sifive.com>
6836 * doc/extend.texi (RISC-V Operand Modifiers): New.
6838 2023-07-10 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
6840 * config/riscv/riscv-vsetvl.cc (add_label_notes): Remove it.
6841 (insert_insn_end_basic_block): Ditto.
6842 (pass_vsetvl::commit_vsetvls): Adapt for new helper function.
6843 * gcse.cc (insert_insn_end_basic_block): Export as global function.
6844 * gcse.h (insert_insn_end_basic_block): Ditto.
6846 2023-07-10 Christophe Lyon <christophe.lyon@linaro.org>
6849 * config/arm/arm-builtins.cc (arm_init_mve_builtins): Handle LTO.
6850 (arm_builtin_decl): Hahndle MVE builtins.
6851 * config/arm/arm-mve-builtins.cc (builtin_decl): New function.
6852 (add_unique_function): Fix handling of
6853 __ARM_MVE_PRESERVE_USER_NAMESPACE.
6854 (add_overloaded_function): Likewise.
6855 * config/arm/arm-protos.h (builtin_decl): New declaration.
6857 2023-07-10 Christophe Lyon <christophe.lyon@linaro.org>
6859 * doc/sourcebuild.texi (arm_v8_1m_main_cde_mve_fp): Document.
6861 2023-07-10 Xi Ruoyao <xry111@xry111.site>
6863 PR tree-optimization/110557
6864 * tree-vect-patterns.cc (vect_recog_bitfield_ref_pattern):
6865 Ensure the output sign-extended if necessary.
6867 2023-07-10 Roger Sayle <roger@nextmovesoftware.com>
6869 * config/i386/i386.md (peephole2): Transform xchg insn with a
6870 REG_UNUSED note to a (simple) move.
6871 (*insvti_lowpart_1): New define_insn_and_split.
6872 (*insvdi_lowpart_1): Likewise.
6874 2023-07-10 Roger Sayle <roger@nextmovesoftware.com>
6876 * config/i386/i386-features.cc (compute_convert_gain): Tweak
6877 gains/costs for ROTATE/ROTATERT by integer constant on AVX512VL.
6878 (general_scalar_chain::convert_rotate): On TARGET_AVX512F generate
6879 avx512vl_rolv2di or avx412vl_rolv4si when appropriate.
6881 2023-07-10 liuhongt <hongtao.liu@intel.com>
6884 * config/i386/i386.md (*ieee_max<mode>3_1): New pre_reload
6885 splitter to detect fp max pattern.
6886 (*ieee_min<mode>3_1): Ditto, but for fp min pattern.
6888 2023-07-09 Jan Hubicka <jh@suse.cz>
6890 * cfg.cc (check_bb_profile): Dump counts with relative frequency.
6891 (dump_edge_info): Likewise.
6892 (dump_bb_info): Likewise.
6893 * profile-count.cc (profile_count::dump): Add comma between quality and
6896 2023-07-08 Jan Hubicka <jh@suse.cz>
6898 PR tree-optimization/110600
6899 * cfgloopmanip.cc (scale_loop_profile): Add mising profile_dump check.
6901 2023-07-08 Jan Hubicka <jh@suse.cz>
6903 PR middle-end/110590
6904 * cfgloopmanip.cc (scale_loop_profile): Avoid scaling exits within
6905 inner loops and be more careful about inconsistent profiles.
6906 (duplicate_loop_body_to_header_edge): Fix profile update when eliminated
6907 exit is followed by other exit.
6909 2023-07-08 Uros Bizjak <ubizjak@gmail.com>
6911 * cprop.cc (reg_available_p): Change return type from int to bool.
6912 (reg_not_set_p): Ditto.
6913 (try_replace_reg): Ditto. Change "success" variable to bool.
6914 (cprop_jump): Change return type from int to void
6915 and adjust function body accordingly.
6916 (constprop_register): Ditto.
6917 (cprop_insn): Ditto. Change "changed" variable to bool.
6918 (local_cprop_pass): Change return type from int to void
6919 and adjust function body accordingly.
6920 (bypass_block): Ditto. Change "change", "may_be_loop_header"
6921 and "removed_p" variables to bool.
6922 (bypass_conditional_jumps): Change return type from int to void
6923 and adjust function body accordingly. Change "changed"
6925 (one_cprop_pass): Ditto.
6927 2023-07-08 Uros Bizjak <ubizjak@gmail.com>
6929 * gcse.cc (expr_equiv_p): Change return type from int to bool.
6930 (oprs_unchanged_p): Change return type from int to void
6931 and adjust function body accordingly.
6932 (oprs_anticipatable_p): Ditto.
6933 (oprs_available_p): Ditto.
6934 (insert_expr_in_table): Ditto. Change "antic_p" and "avail_p"
6935 arguments to bool. Change "found" variable to bool.
6936 (load_killed_in_block_p): Change return type from int to void and
6937 adjust function body accordingly. Change "avail_p" argument to bool.
6938 (pre_expr_reaches_here_p): Change return type from int to void
6939 and adjust function body accordingly.
6940 (pre_delete): Ditto. Change "changed" variable to bool.
6941 (pre_gcse): Change return type from int to void
6942 and adjust function body accordingly. Change "did_insert" and
6943 "changed" variables to bool.
6944 (one_pre_gcse_pass): Change return type from int to void
6945 and adjust function body accordingly. Change "changed" variable
6947 (should_hoist_expr_to_dom): Change return type from int to void
6948 and adjust function body accordingly. Change
6949 "visited_allocated_locally" variable to bool.
6950 (hoist_code): Change return type from int to void and adjust
6951 function body accordingly. Change "changed" variable to bool.
6952 (one_code_hoisting_pass): Ditto.
6953 (pre_edge_insert): Change return type from int to void and adjust
6954 function body accordingly. Change "did_insert" variable to bool.
6955 (pre_expr_reaches_here_p_work): Change return type from int to void
6956 and adjust function body accordingly.
6957 (simple_mem): Ditto.
6958 (want_to_gcse_p): Change return type from int to void
6959 and adjust function body accordingly.
6960 (can_assign_to_reg_without_clobbers_p): Update function body
6961 for bool return type.
6962 (hash_scan_set): Change "antic_p" and "avail_p" variables to bool.
6963 (pre_insert_copies): Change "added_copy" variable to bool.
6965 2023-07-08 Jonathan Wakely <jwakely@redhat.com>
6969 * doc/invoke.texi (Warning Options): Fix typos.
6971 2023-07-07 Jan Hubicka <jh@suse.cz>
6973 * profile-count.cc (profile_count::dump): Add FUN
6974 parameter; print relative frequency.
6975 (profile_count::debug): Update.
6976 * profile-count.h (profile_count::dump): Update
6979 2023-07-07 Roger Sayle <roger@nextmovesoftware.com>
6983 * config/i386/i386-expand.cc (ix86_expand_move): Convert SETs of
6984 TImode destinations from paradoxical SUBREGs (setting the lowpart)
6985 into explicit zero extensions. Use *insvti_highpart_1 instruction
6986 to set the highpart of a TImode destination.
6988 2023-07-07 Jan Hubicka <jh@suse.cz>
6990 * predict.cc (force_edge_cold): Use
6991 set_edge_probability_and_rescale_others; improve dumps.
6993 2023-07-07 Jan Hubicka <jh@suse.cz>
6995 * cfgloopmanip.cc (scale_loop_profile): Fix computation of count_in and scaling blocks
6997 * tree-vect-loop-manip.cc (vect_do_peeling): Scale loop profile of the epilogue if bound
7000 2023-07-07 Juergen Christ <jchrist@linux.ibm.com>
7002 * config/s390/s390.cc (vec_init): Fix default case
7004 2023-07-07 Vladimir N. Makarov <vmakarov@redhat.com>
7006 * lra-assigns.cc (assign_by_spills): Add reload insns involving
7007 reload pseudos with non-refined class to be processed on the next
7009 * lra-constraints.cc (enough_allocatable_hard_regs_p): New func.
7010 (in_class_p): Use it.
7011 (print_curr_insn_alt): New func.
7012 (process_alt_operands): Use it. Improve debug info.
7013 (curr_insn_transform): Use print_curr_insn_alt. Refine reload
7014 pseudo class if it is not refined yet.
7016 2023-07-07 Aldy Hernandez <aldyh@redhat.com>
7018 * value-range.cc (irange::get_bitmask_from_range): Return all the
7019 known bits for a singleton.
7020 (irange::set_range_from_bitmask): Set a range of a singleton when
7023 2023-07-07 Aldy Hernandez <aldyh@redhat.com>
7025 * value-range.cc (irange::intersect): Leave normalization to
7028 2023-07-07 Aldy Hernandez <aldyh@redhat.com>
7030 * data-streamer-in.cc (streamer_read_value_range): Adjust for
7032 * data-streamer-out.cc (streamer_write_vrange): Same.
7033 * range-op.cc (operator_cast::fold_range): Same.
7034 * value-range-pretty-print.cc
7035 (vrange_printer::print_irange_bitmasks): Same.
7036 * value-range-storage.cc (irange_storage::write_lengths_address):
7038 (irange_storage::set_irange): Same.
7039 (irange_storage::get_irange): Same.
7040 (irange_storage::size): Same.
7041 (irange_storage::dump): Same.
7042 * value-range-storage.h: Same.
7043 * value-range.cc (debug): New.
7044 (irange_bitmask::dump): New.
7045 (add_vrange): Adjust for value/mask.
7046 (irange::operator=): Same.
7047 (irange::set): Same.
7048 (irange::verify_range): Same.
7049 (irange::operator==): Same.
7050 (irange::contains_p): Same.
7051 (irange::irange_single_pair_union): Same.
7052 (irange::union_): Same.
7053 (irange::intersect): Same.
7054 (irange::invert): Same.
7055 (irange::get_nonzero_bits_from_range): Rename to...
7056 (irange::get_bitmask_from_range): ...this.
7057 (irange::set_range_from_nonzero_bits): Rename to...
7058 (irange::set_range_from_bitmask): ...this.
7059 (irange::set_nonzero_bits): Rename to...
7060 (irange::update_bitmask): ...this.
7061 (irange::get_nonzero_bits): Rename to...
7062 (irange::get_bitmask): ...this.
7063 (irange::intersect_nonzero_bits): Rename to...
7064 (irange::intersect_bitmask): ...this.
7065 (irange::union_nonzero_bits): Rename to...
7066 (irange::union_bitmask): ...this.
7067 (irange_bitmask::verify_mask): New.
7068 * value-range.h (class irange_bitmask): New.
7069 (irange_bitmask::set_unknown): New.
7070 (irange_bitmask::unknown_p): New.
7071 (irange_bitmask::irange_bitmask): New.
7072 (irange_bitmask::get_precision): New.
7073 (irange_bitmask::get_nonzero_bits): New.
7074 (irange_bitmask::set_nonzero_bits): New.
7075 (irange_bitmask::operator==): New.
7076 (irange_bitmask::union_): New.
7077 (irange_bitmask::intersect): New.
7078 (class irange): Friend vrange_printer.
7079 (irange::varying_compatible_p): Adjust for bitmask.
7080 (irange::set_varying): Same.
7081 (irange::set_nonzero): Same.
7083 2023-07-07 Jan Beulich <jbeulich@suse.com>
7085 * config/i386/sse.md (*vec_extractv2ti): Drop g modifiers.
7087 2023-07-07 Jan Beulich <jbeulich@suse.com>
7089 * config/i386/sse.md (@vec_extract_hi_<mode>): Drop last
7090 alternative. Switch new last alternative's "isa" attribute to
7092 (vec_extract_hi_v32qi): Likewise.
7094 2023-07-07 Pan Li <pan2.li@intel.com>
7095 Robin Dapp <rdapp@ventanamicro.com>
7097 * config/riscv/riscv.cc (riscv_emit_mode_set): Avoid emit insn
7099 (riscv_mode_entry): Take FRM_MODE_DYN as entry mode.
7100 (riscv_mode_exit): Likewise for exit mode.
7101 (riscv_mode_needed): Likewise for needed mode.
7102 (riscv_mode_after): Likewise for after mode.
7104 2023-07-07 Pan Li <pan2.li@intel.com>
7106 * config/riscv/vector.md: Fix typo.
7108 2023-07-06 Jan Hubicka <jh@suse.cz>
7111 * tree-ssa-loop-ch.cc (ch_base::copy_headers): Scale loop frequency to maximal number
7112 of iterations determined.
7113 * tree-ssa-loop-ivcanon.cc (try_unroll_loop_completely): Likewise.
7115 2023-07-06 Jan Hubicka <jh@suse.cz>
7117 * cfgloopmanip.cc (scale_loop_profile): Rewrite exit edge
7118 probability update to be safe on loops with subloops.
7119 Make bound parameter to be iteration bound.
7120 * tree-ssa-loop-ivcanon.cc (try_peel_loop): Update call
7121 of scale_loop_profile.
7122 * tree-vect-loop-manip.cc (vect_do_peeling): Likewise.
7124 2023-07-06 Hao Liu OS <hliu@os.amperecomputing.com>
7126 PR tree-optimization/110449
7127 * tree-vect-loop.cc (vectorizable_induction): use vec_n to replace
7128 vec_loop for the unrolled loop.
7130 2023-07-06 Jan Hubicka <jh@suse.cz>
7132 * cfg.cc (set_edge_probability_and_rescale_others): New function.
7133 (update_bb_profile_for_threading): Use it; simplify the rest.
7134 * cfg.h (set_edge_probability_and_rescale_others): Declare.
7135 * profile-count.h (profile_probability::apply_scale): New.
7137 2023-07-06 Claudiu Zissulescu <claziss@gmail.com>
7139 * doc/extend.texi (ARC Built-in Functions): Update documentation
7140 with missing builtins.
7142 2023-07-06 Richard Biener <rguenther@suse.de>
7144 PR tree-optimization/110556
7145 * tree-ssa-tail-merge.cc (gimple_equal_p): Check
7146 assign code and all operands of non-stores.
7148 2023-07-06 Richard Biener <rguenther@suse.de>
7150 PR tree-optimization/110563
7151 * tree-vectorizer.h (vect_determine_partial_vectors_and_peeling):
7152 Remove second argument.
7153 * tree-vect-loop.cc (vect_determine_partial_vectors_and_peeling):
7154 Remove for_epilogue_p argument. Merge assert ...
7155 (vect_analyze_loop_2): ... with check done before determining
7156 partial vectors by moving it after.
7157 * tree-vect-loop-manip.cc (vect_do_peeling): Adjust.
7159 2023-07-06 Thomas Schwinge <thomas@codesourcery.com>
7161 * ggc-common.cc (gt_pch_note_reorder, gt_pch_save): Tighten up a
7162 few things re 'reorder' option and strings.
7163 * stringpool.cc (gt_pch_p_S): This is now 'gcc_unreachable'.
7165 2023-07-06 Thomas Schwinge <thomas@codesourcery.com>
7167 * gengtype-parse.cc: Clean up obsolete parametrized structs
7169 * gengtype.cc: Likewise.
7170 * gengtype.h: Likewise.
7172 2023-07-06 Thomas Schwinge <thomas@codesourcery.com>
7174 * gengtype.cc (struct walk_type_data): Remove 'needs_cast_p'.
7177 2023-07-06 Thomas Schwinge <thomas@codesourcery.com>
7179 * gengtype-parse.cc (token_names): Add '"user"'.
7180 * gengtype.h (gty_token): Add 'UNUSED_PARAM_IS' for use with
7181 'FIRST_TOKEN_WITH_VALUE'.
7183 2023-07-06 Thomas Schwinge <thomas@codesourcery.com>
7185 * doc/gty.texi (GTY Options) <string_length>: Enhance.
7187 2023-07-06 Thomas Schwinge <thomas@codesourcery.com>
7189 * gengtype.cc (write_root, write_roots): Explicitly reject
7190 'string_length' option.
7191 * doc/gty.texi (GTY Options) <string_length>: Document.
7193 2023-07-06 Thomas Schwinge <thomas@codesourcery.com>
7195 * ggc-internal.h (ggc_pch_count_object, ggc_pch_alloc_object)
7196 (ggc_pch_write_object): Remove 'bool is_string' argument.
7197 * ggc-common.cc: Adjust.
7198 * ggc-page.cc: Likewise.
7200 2023-07-06 Roger Sayle <roger@nextmovesoftware.com>
7202 * dwarf2out.cc (mem_loc_descriptor): Handle COPYSIGN.
7204 2023-07-06 Hongyu Wang <hongyu.wang@intel.com>
7206 * doc/extend.texi: Move x86 inlining rule to a new subsubsection
7207 and add description for inling of function with arch and tune
7210 2023-07-06 Richard Biener <rguenther@suse.de>
7212 PR tree-optimization/110515
7213 * tree-ssa-pre.cc (compute_avail): Make code dealing
7214 with hoisting loads with different alias-sets more
7217 2023-07-06 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
7219 * tree-vect-stmts.cc (vect_get_strided_load_store_ops): Fix ICE.
7221 2023-07-06 Hongyu Wang <hongyu.wang@intel.com>
7223 * config/i386/i386.cc (ix86_can_inline_p): If callee has
7224 default arch=x86-64 and tune=generic, do not block the
7225 inlining to its caller. Also allow callee with different
7226 arch= to be inlined if it has always_inline attribute and
7227 it's ISA is subset of caller's.
7229 2023-07-06 liuhongt <hongtao.liu@intel.com>
7231 * config/i386/i386.cc (ix86_rtx_costs): Adjust rtx_cost for
7232 DF/SFmode AND/IOR/XOR/ANDN operations.
7234 2023-07-06 Andrew Pinski <apinski@marvell.com>
7236 PR middle-end/110554
7237 * tree-vect-generic.cc (expand_vector_condition): For comparisons,
7238 just build using boolean_type_node instead of the cond_type.
7239 For non-comparisons/non-scalar-bitmask, build a ` != 0` gimple
7240 that will feed into the COND_EXPR.
7242 2023-07-06 liuhongt <hongtao.liu@intel.com>
7245 * config/i386/i386.md (movdf_internal): Disparage slightly for
7246 2 alternatives (r,v) and (v,r) by adding constraint modifier
7249 2023-07-06 Jeevitha Palanisamy <jeevitha@linux.ibm.com>
7252 * config/rs6000/rs6000.cc (rs6000_expand_vector_extract): Remove redundant
7253 initialization of new_addr.
7255 2023-07-06 Hao Liu <hliu@os.amperecomputing.com>
7257 PR tree-optimization/110474
7258 * tree-vect-loop.cc (vect_analyze_loop_2): unscale the VF by suggested
7259 unroll factor while selecting the epilog vect loop VF.
7261 2023-07-05 Andrew MacLeod <amacleod@redhat.com>
7263 * gimple-range-gori.cc (compute_operand_range): Convert to a tail
7266 2023-07-05 Andrew MacLeod <amacleod@redhat.com>
7268 * gimple-range-gori.cc (compute_operand_range): After calling
7269 compute_operand2_range, recursively call self if needed.
7270 (compute_operand2_range): Turn into a leaf function.
7271 (gori_compute::compute_operand1_and_operand2_range): Finish
7272 operand2 calculation.
7273 * gimple-range-gori.h (compute_operand2_range): Remove name param.
7275 2023-07-05 Andrew MacLeod <amacleod@redhat.com>
7277 * gimple-range-gori.cc (compute_operand_range): After calling
7278 compute_operand1_range, recursively call self if needed.
7279 (compute_operand1_range): Turn into a leaf function.
7280 (gori_compute::compute_operand1_and_operand2_range): Finish
7281 operand1 calculation.
7282 * gimple-range-gori.h (compute_operand1_range): Remove name param.
7284 2023-07-05 Andrew MacLeod <amacleod@redhat.com>
7286 * gimple-range-gori.cc (compute_operand_range): Check for
7287 operand interdependence when both op1 and op2 are computed.
7288 (compute_operand1_and_operand2_range): No checks required now.
7290 2023-07-05 Andrew MacLeod <amacleod@redhat.com>
7292 * gimple-range-gori.cc (compute_operand_range): Check for
7293 a relation between op1 and op2 and use that instead.
7294 (compute_operand1_range): Don't look for a relation override.
7295 (compute_operand2_range): Ditto.
7297 2023-07-05 Jonathan Wakely <jwakely@redhat.com>
7299 * doc/contrib.texi (Contributors): Update my entry.
7301 2023-07-05 Filip Kastl <filip.kastl@gmail.com>
7303 * value-prof.cc (gimple_mod_subtract_transform): Correct edge
7306 2023-07-05 Uros Bizjak <ubizjak@gmail.com>
7308 * sched-int.h (struct haifa_sched_info): Change can_schedule_ready_p,
7309 scehdule_more_p and contributes_to_priority indirect frunction
7310 type from int to bool.
7311 (no_real_insns_p): Change return type from int to bool.
7312 (contributes_to_priority): Ditto.
7313 * haifa-sched.cc (no_real_insns_p): Change return type from
7314 int to bool and adjust function body accordingly.
7315 * modulo-sched.cc (try_scheduling_node_in_cycle): Change "success"
7316 variable type from int to bool.
7317 (ps_insn_advance_column): Change return type from int to bool.
7318 (ps_has_conflicts): Ditto. Change "has_conflicts"
7319 variable type from int to bool.
7320 * sched-deps.cc (deps_may_trap_p): Change return type from int to bool.
7321 (conditions_mutex_p): Ditto.
7322 * sched-ebb.cc (schedule_more_p): Ditto.
7323 (ebb_contributes_to_priority): Change return type from
7324 int to bool and adjust function body accordingly.
7325 * sched-rgn.cc (is_cfg_nonregular): Ditto.
7326 (check_live_1): Ditto.
7328 (find_conditional_protection): Ditto.
7329 (is_conditionally_protected): Ditto.
7331 (is_exception_free): Ditto.
7332 (haifa_find_rgns): Change "unreachable" and "too_large_failure"
7333 variables from int to bool.
7334 (extend_rgns): Change "rescan" variable from int to bool.
7335 (check_live): Change return type from
7336 int to bool and adjust function body accordingly.
7337 (can_schedule_ready_p): Ditto.
7338 (schedule_more_p): Ditto.
7339 (contributes_to_priority): Ditto.
7341 2023-07-05 Robin Dapp <rdapp@ventanamicro.com>
7343 * doc/md.texi: Document that vec_set and vec_extract must not
7345 * gimple-isel.cc (gimple_expand_vec_set_expr): Rename this...
7346 (gimple_expand_vec_set_extract_expr): ...to this.
7347 (gimple_expand_vec_exprs): Call renamed function.
7348 * internal-fn.cc (vec_extract_direct): Add.
7349 (expand_vec_extract_optab_fn): New function to expand
7351 (direct_vec_extract_optab_supported_p): Add.
7352 * internal-fn.def (VEC_EXTRACT): Add.
7353 * optabs.cc (can_vec_extract_var_idx_p): New function.
7354 * optabs.h (can_vec_extract_var_idx_p): Declare.
7356 2023-07-05 Robin Dapp <rdapp@ventanamicro.com>
7358 * config/riscv/autovec.md: Add gen_lowpart.
7360 2023-07-05 Robin Dapp <rdapp@ventanamicro.com>
7362 * config/riscv/autovec.md: Allow register index operand.
7364 2023-07-05 Pan Li <pan2.li@intel.com>
7366 * config/riscv/riscv-vector-builtins.cc
7367 (function_expander::use_exact_insn): Use FRM_DYN instead of const0.
7369 2023-07-05 Robin Dapp <rdapp@ventanamicro.com>
7371 * config/riscv/autovec.md: Use float_truncate.
7373 2023-07-05 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
7375 * internal-fn.cc (internal_fn_len_index): Apply
7376 LEN_MASK_GATHER_LOAD/SCATTER_STORE into vectorizer.
7377 (internal_fn_mask_index): Ditto.
7378 * optabs-query.cc (supports_vec_gather_load_p): Ditto.
7379 (supports_vec_scatter_store_p): Ditto.
7380 * tree-vect-data-refs.cc (vect_gather_scatter_fn_p): Ditto.
7381 * tree-vect-patterns.cc (vect_recog_gather_scatter_pattern): Ditto.
7382 * tree-vect-stmts.cc (check_load_store_for_partial_vectors): Ditto.
7383 (vect_get_strided_load_store_ops): Ditto.
7384 (vectorizable_store): Ditto.
7385 (vectorizable_load): Ditto.
7387 2023-07-05 Robin Dapp <rdapp@ventanamicro.com>
7388 Juzhe-Zhong <juzhe.zhong@rivai.ai>
7390 * simplify-rtx.cc (native_encode_rtx): Ditto.
7391 (native_decode_vector_rtx): Ditto.
7392 (simplify_const_vector_byte_offset): Ditto.
7393 (simplify_const_vector_subreg): Ditto.
7394 * tree.cc (build_truth_vector_type_for_mode): Ditto.
7395 * varasm.cc (output_constant_pool_2): Ditto.
7397 2023-07-05 YunQiang Su <yunqiang.su@cipunited.com>
7399 * config/mips/mips.cc (mips_expand_block_move): don't expand for
7400 r6 with -mno-unaligned-access option if one or both of src and
7401 dest are unaligned. restruct: return directly if length is not const.
7402 (mips_block_move_straight): emit_move if ISA_HAS_UNALIGNED_ACCESS.
7404 2023-07-05 Jan Beulich <jbeulich@suse.com>
7407 * config/i386/sse.md: New splitters to simplify
7408 not;vec_duplicate as a singular vpternlog.
7409 (one_cmpl<mode>2): Allow broadcast for operand 1.
7410 (<mask_codefor>one_cmpl<mode>2<mask_name>): Likewise.
7412 2023-07-05 Jan Beulich <jbeulich@suse.com>
7415 * config/i386/sse.md: New splitters to simplify
7416 not;vec_duplicate;{ior,xor} as vec_duplicate;{iornot,xnor}.
7418 2023-07-05 Jan Beulich <jbeulich@suse.com>
7421 * config/i386/sse.md: Permit non-immediate operand 1 in AVX2
7422 form of splitter for PR target/100711.
7424 2023-07-05 Richard Biener <rguenther@suse.de>
7426 PR middle-end/110541
7427 * tree.def (VEC_PERM_EXPR): Adjust documentation to reflect
7430 2023-07-05 Jan Beulich <jbeulich@suse.com>
7433 * config/i386/sse.md (*andnot<mode>3): Add new alternatives
7434 for memory form operand 1.
7436 2023-07-05 Jan Beulich <jbeulich@suse.com>
7439 * config/i386/i386.cc (ix86_rtx_costs): Further special-case
7440 bitwise vector operations.
7441 * config/i386/sse.md (*iornot<mode>3): New insn.
7442 (*xnor<mode>3): Likewise.
7443 (*<nlogic><mode>3): Likewise.
7444 (andor): New code iterator.
7445 (nlogic): New code attribute.
7446 (ternlog_nlogic): Likewise.
7448 2023-07-05 Richard Biener <rguenther@suse.de>
7450 * tree-vect-stmts.cc (vect_mark_relevant): Fix typo.
7452 2023-07-05 yulong <shiyulong@iscas.ac.cn>
7454 * config/riscv/vector.md: Add float16 attr at sew、vlmul and ratio.
7456 2023-07-05 yulong <shiyulong@iscas.ac.cn>
7458 * config/riscv/genrvv-type-indexer.cc (valid_type): Enable FP16 tuple.
7459 * config/riscv/riscv-modes.def (RVV_TUPLE_MODES): New macro.
7460 (ADJUST_ALIGNMENT): Ditto.
7461 (RVV_TUPLE_PARTIAL_MODES): Ditto.
7462 (ADJUST_NUNITS): Ditto.
7463 * config/riscv/riscv-vector-builtins-types.def (vfloat16mf4x2_t):
7465 (vfloat16mf4x3_t): Ditto.
7466 (vfloat16mf4x4_t): Ditto.
7467 (vfloat16mf4x5_t): Ditto.
7468 (vfloat16mf4x6_t): Ditto.
7469 (vfloat16mf4x7_t): Ditto.
7470 (vfloat16mf4x8_t): Ditto.
7471 (vfloat16mf2x2_t): Ditto.
7472 (vfloat16mf2x3_t): Ditto.
7473 (vfloat16mf2x4_t): Ditto.
7474 (vfloat16mf2x5_t): Ditto.
7475 (vfloat16mf2x6_t): Ditto.
7476 (vfloat16mf2x7_t): Ditto.
7477 (vfloat16mf2x8_t): Ditto.
7478 (vfloat16m1x2_t): Ditto.
7479 (vfloat16m1x3_t): Ditto.
7480 (vfloat16m1x4_t): Ditto.
7481 (vfloat16m1x5_t): Ditto.
7482 (vfloat16m1x6_t): Ditto.
7483 (vfloat16m1x7_t): Ditto.
7484 (vfloat16m1x8_t): Ditto.
7485 (vfloat16m2x2_t): Ditto.
7486 (vfloat16m2x3_t): Ditto.
7487 (vfloat16m2x4_t): Ditto.
7488 (vfloat16m4x2_t): Ditto.
7489 * config/riscv/riscv-vector-builtins.def (vfloat16mf4x2_t): New macro.
7490 (vfloat16mf4x3_t): Ditto.
7491 (vfloat16mf4x4_t): Ditto.
7492 (vfloat16mf4x5_t): Ditto.
7493 (vfloat16mf4x6_t): Ditto.
7494 (vfloat16mf4x7_t): Ditto.
7495 (vfloat16mf4x8_t): Ditto.
7496 (vfloat16mf2x2_t): Ditto.
7497 (vfloat16mf2x3_t): Ditto.
7498 (vfloat16mf2x4_t): Ditto.
7499 (vfloat16mf2x5_t): Ditto.
7500 (vfloat16mf2x6_t): Ditto.
7501 (vfloat16mf2x7_t): Ditto.
7502 (vfloat16mf2x8_t): Ditto.
7503 (vfloat16m1x2_t): Ditto.
7504 (vfloat16m1x3_t): Ditto.
7505 (vfloat16m1x4_t): Ditto.
7506 (vfloat16m1x5_t): Ditto.
7507 (vfloat16m1x6_t): Ditto.
7508 (vfloat16m1x7_t): Ditto.
7509 (vfloat16m1x8_t): Ditto.
7510 (vfloat16m2x2_t): Ditto.
7511 (vfloat16m2x3_t): Ditto.
7512 (vfloat16m2x4_t): Ditto.
7513 (vfloat16m4x2_t): Ditto.
7514 * config/riscv/riscv-vector-switch.def (TUPLE_ENTRY): New.
7515 * config/riscv/riscv.md: New.
7516 * config/riscv/vector-iterators.md: New.
7518 2023-07-04 Andrew Pinski <apinski@marvell.com>
7520 PR tree-optimization/110487
7521 * match.pd (a !=/== CST1 ? CST2 : CST3): Always
7522 build a nonstandard integer and use that.
7524 2023-07-04 Andrew Pinski <apinski@marvell.com>
7526 * match.pd (a?-1:0): Cast type an integer type
7527 rather the type before the negative.
7530 2023-07-04 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
7532 * config/xtensa/xtensa.cc (machine_function, xtensa_expand_prologue):
7533 Change to use HARD_REG_BIT and its macros.
7534 * config/xtensa/xtensa.md
7535 (peephole2: regmove elimination during DFmode input reload):
7538 2023-07-04 Richard Biener <rguenther@suse.de>
7540 PR tree-optimization/110491
7541 * tree-ssa-phiopt.cc (match_simplify_replacement): Check
7542 whether the PHI args are possibly undefined before folding
7545 2023-07-04 Pan Li <pan2.li@intel.com>
7546 Thomas Schwinge <thomas@codesourcery.com>
7548 * lto-streamer-in.cc (lto_input_mode_table): Stream in the mode
7549 bits for machine mode table.
7550 * lto-streamer-out.cc (lto_write_mode_table): Stream out the
7551 HOST machine mode bits.
7552 * lto-streamer.h (struct lto_file_decl_data): New fields mode_bits.
7553 * tree-streamer.cc (streamer_mode_table): Take MAX_MACHINE_MODE
7555 * tree-streamer.h (streamer_mode_table): Ditto.
7556 (bp_pack_machine_mode): Take 1 << ceil_log2 (MAX_MACHINE_MODE)
7557 as the packing limit.
7558 (bp_unpack_machine_mode): Ditto with 'file_data->mode_bits'.
7560 2023-07-04 Thomas Schwinge <thomas@codesourcery.com>
7562 * lto-streamer.h (class lto_input_block): Capture
7563 'lto_file_decl_data *file_data' instead of just
7564 'unsigned char *mode_table'.
7565 * ipa-devirt.cc (ipa_odr_read_section): Adjust.
7566 * ipa-fnsummary.cc (inline_read_section): Likewise.
7567 * ipa-icf.cc (sem_item_optimizer::read_section): Likewise.
7568 * ipa-modref.cc (read_section): Likewise.
7569 * ipa-prop.cc (ipa_prop_read_section, read_replacements_section):
7571 * ipa-sra.cc (isra_read_summary_section): Likewise.
7572 * lto-cgraph.cc (input_cgraph_opt_section): Likewise.
7573 * lto-section-in.cc (lto_create_simple_input_block): Likewise.
7574 * lto-streamer-in.cc (lto_read_body_or_constructor)
7575 (lto_input_toplevel_asms): Likewise.
7576 * tree-streamer.h (bp_unpack_machine_mode): Likewise.
7578 2023-07-04 Richard Biener <rguenther@suse.de>
7580 * tree-ssa-phiopt.cc (pass_phiopt::execute): Mark SSA undefs.
7581 (empty_bb_or_one_feeding_into_p): Check for them.
7582 * tree-ssa.h (gimple_uses_undefined_value_p): Remove.
7583 * tree-ssa.cc (gimple_uses_undefined_value_p): Likewise.
7585 2023-07-04 Richard Biener <rguenther@suse.de>
7587 * tree-vect-loop.cc (vect_analyze_loop_costing): Remove
7588 check guarding scalar_niter underflow.
7590 2023-07-04 Hao Liu <hliu@os.amperecomputing.com>
7592 PR tree-optimization/110531
7593 * tree-vect-loop.cc (vect_analyze_loop_1): initialize
7594 slp_done_for_suggested_uf to false.
7596 2023-07-04 Richard Biener <rguenther@suse.de>
7598 PR tree-optimization/110228
7599 * tree-ssa-ifcombine.cc (pass_tree_ifcombine::execute):
7600 Mark SSA may-undefs.
7601 (bb_no_side_effects_p): Check stmt uses for undefs.
7603 2023-07-04 Richard Biener <rguenther@suse.de>
7605 PR tree-optimization/110436
7606 * tree-vect-stmts.cc (vect_mark_relevant): Expand dumping,
7607 force live but not relevant pattern stmts relevant.
7609 2023-07-04 Lili Cui <lili.cui@intel.com>
7611 * config/i386/i386.h: Add PTA_ENQCMD and PTA_UINTR to PTA_SIERRAFOREST.
7612 * doc/invoke.texi: Update new isa to march=sierraforest and grandridge.
7614 2023-07-04 Richard Biener <rguenther@suse.de>
7616 PR middle-end/110495
7617 * tree.h (TREE_OVERFLOW): Do not mention VECTOR_CSTs
7618 since we do not set TREE_OVERFLOW on those since the
7619 introduction of VL vectors.
7620 * match.pd (x +- CST +- CST): For VECTOR_CST do not look
7621 at TREE_OVERFLOW to determine validity of association.
7623 2023-07-04 Richard Biener <rguenther@suse.de>
7625 PR tree-optimization/110310
7626 * tree-vect-loop.cc (vect_determine_partial_vectors_and_peeling):
7627 Move costing part ...
7628 (vect_analyze_loop_costing): ... here. Integrate better
7629 estimate for epilogues from ...
7630 (vect_analyze_loop_2): Call vect_determine_partial_vectors_and_peeling
7631 with actual epilogue status.
7632 * tree-vect-loop-manip.cc (vect_do_peeling): ... here and
7633 avoid cancelling epilogue vectorization.
7634 (vect_update_epilogue_niters): Remove. No longer update
7635 epilogue LOOP_VINFO_NITERS.
7637 2023-07-04 Pan Li <pan2.li@intel.com>
7640 2023-07-03 Pan Li <pan2.li@intel.com>
7642 * config/riscv/vector.md: Fix typo.
7644 2023-07-04 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
7646 * doc/md.texi: Add len_mask_gather_load/len_mask_scatter_store.
7647 * internal-fn.cc (expand_scatter_store_optab_fn): Ditto.
7648 (expand_gather_load_optab_fn): Ditto.
7649 (internal_load_fn_p): Ditto.
7650 (internal_store_fn_p): Ditto.
7651 (internal_gather_scatter_fn_p): Ditto.
7652 (internal_fn_len_index): Ditto.
7653 (internal_fn_mask_index): Ditto.
7654 (internal_fn_stored_value_index): Ditto.
7655 * internal-fn.def (LEN_MASK_GATHER_LOAD): Ditto.
7656 (LEN_MASK_SCATTER_STORE): Ditto.
7657 * optabs.def (OPTAB_CD): Ditto.
7659 2023-07-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
7661 * config/riscv/riscv-vsetvl.cc
7662 (vector_insn_info::parse_insn): Add early break.
7664 2023-07-04 Hans-Peter Nilsson <hp@axis.com>
7666 * config/cris/cris.md (CRIS_UNSPEC_SWAP_BITS): Remove.
7667 ("cris_swap_bits", "ctzsi2"): Use bitreverse instead.
7669 2023-07-04 Hans-Peter Nilsson <hp@axis.com>
7671 * dwarf2out.cc (mem_loc_descriptor): Handle BITREVERSE.
7673 2023-07-03 Christoph Müllner <christoph.muellner@vrull.eu>
7675 * common/config/riscv/riscv-common.cc: Add support for zvbb,
7676 zvbc, zvkg, zvkned, zvknha, zvknhb, zvksed, zvksh, zvkn,
7677 zvknc, zvkng, zvks, zvksc, zvksg, zvkt and the implied subsets.
7678 * config/riscv/arch-canonicalize: Add canonicalization info for
7679 zvkn, zvknc, zvkng, zvks, zvksc, zvksg.
7680 * config/riscv/riscv-opts.h (MASK_ZVBB): New macro.
7681 (MASK_ZVBC): Likewise.
7682 (TARGET_ZVBB): Likewise.
7683 (TARGET_ZVBC): Likewise.
7684 (MASK_ZVKG): Likewise.
7685 (MASK_ZVKNED): Likewise.
7686 (MASK_ZVKNHA): Likewise.
7687 (MASK_ZVKNHB): Likewise.
7688 (MASK_ZVKSED): Likewise.
7689 (MASK_ZVKSH): Likewise.
7690 (MASK_ZVKN): Likewise.
7691 (MASK_ZVKNC): Likewise.
7692 (MASK_ZVKNG): Likewise.
7693 (MASK_ZVKS): Likewise.
7694 (MASK_ZVKSC): Likewise.
7695 (MASK_ZVKSG): Likewise.
7696 (MASK_ZVKT): Likewise.
7697 (TARGET_ZVKG): Likewise.
7698 (TARGET_ZVKNED): Likewise.
7699 (TARGET_ZVKNHA): Likewise.
7700 (TARGET_ZVKNHB): Likewise.
7701 (TARGET_ZVKSED): Likewise.
7702 (TARGET_ZVKSH): Likewise.
7703 (TARGET_ZVKN): Likewise.
7704 (TARGET_ZVKNC): Likewise.
7705 (TARGET_ZVKNG): Likewise.
7706 (TARGET_ZVKS): Likewise.
7707 (TARGET_ZVKSC): Likewise.
7708 (TARGET_ZVKSG): Likewise.
7709 (TARGET_ZVKT): Likewise.
7710 * config/riscv/riscv.opt: Introduction of riscv_zv{b,k}_subext.
7712 2023-07-03 Andrew Pinski <apinski@marvell.com>
7714 PR middle-end/110510
7715 * except.h (struct eh_landing_pad_d): Add chain_next GTY.
7717 2023-07-03 Iain Sandoe <iain@sandoe.co.uk>
7719 * config/darwin.h: Avoid duplicate multiply_defined specs on
7720 earlier Darwin versions with shared libgcc.
7722 2023-07-03 Uros Bizjak <ubizjak@gmail.com>
7724 * tree.h (tree_int_cst_equal): Change return type from int to bool.
7725 (operand_equal_for_phi_arg_p): Ditto.
7726 (tree_map_base_marked_p): Ditto.
7727 * tree.cc (contains_placeholder_p): Update function body
7728 for bool return type.
7729 (type_cache_hasher::equal): Ditto.
7730 (tree_map_base_hash): Change return type
7731 from int to void and adjust function body accordingly.
7732 (tree_int_cst_equal): Ditto.
7733 (operand_equal_for_phi_arg_p): Ditto.
7734 (get_narrower): Change "first" variable to bool.
7735 (cl_option_hasher::equal): Update function body for bool return type.
7736 * ggc.h (ggc_set_mark): Change return type from int to bool.
7737 (ggc_marked_p): Ditto.
7738 * ggc-page.cc (gt_ggc_mx): Change return type
7739 from int to void and adjust function body accordingly.
7740 (ggc_set_mark): Ditto.
7742 2023-07-03 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
7744 * config/riscv/autovec.md: Change order of
7745 LEN_MASK_LOAD/LEN_MASK_STORE/LEN_LOAD/LEN_STORE arguments.
7746 * config/riscv/riscv-v.cc (expand_load_store): Ditto.
7747 * doc/md.texi: Ditto.
7748 * gimple-fold.cc (gimple_fold_partial_load_store_mem_ref): Ditto.
7749 * internal-fn.cc (len_maskload_direct): Ditto.
7750 (len_maskstore_direct): Ditto.
7751 (add_len_and_mask_args): New function.
7752 (expand_partial_load_optab_fn): Change order of
7753 LEN_MASK_LOAD/LEN_MASK_STORE/LEN_LOAD/LEN_STORE arguments.
7754 (expand_partial_store_optab_fn): Ditto.
7755 (internal_fn_len_index): New function.
7756 (internal_fn_mask_index): Change order of
7757 LEN_MASK_LOAD/LEN_MASK_STORE/LEN_LOAD/LEN_STORE arguments.
7758 (internal_fn_stored_value_index): Ditto.
7759 (internal_len_load_store_bias): Ditto.
7760 * internal-fn.h (internal_fn_len_index): New function.
7761 * tree-ssa-dse.cc (initialize_ao_ref_for_dse): Change order of
7762 LEN_MASK_LOAD/LEN_MASK_STORE/LEN_LOAD/LEN_STORE arguments.
7763 * tree-vect-stmts.cc (vectorizable_store): Ditto.
7764 (vectorizable_load): Ditto.
7766 2023-07-03 Gaius Mulley <gaiusmod2@gmail.com>
7769 * doc/gm2.texi (Semantic checking): Include examples using
7770 -Wuninit-variable-checking.
7772 2023-07-03 Juzhe-Zhong <juzhe.zhong@rivai.ai>
7774 * config/riscv/autovec-opt.md (*double_widen_fnma<mode>): New pattern.
7775 (*single_widen_fnma<mode>): Ditto.
7776 (*double_widen_fms<mode>): Ditto.
7777 (*single_widen_fms<mode>): Ditto.
7778 (*double_widen_fnms<mode>): Ditto.
7779 (*single_widen_fnms<mode>): Ditto.
7781 2023-07-03 Juzhe-Zhong <juzhe.zhong@rivai.ai>
7783 * config/riscv/autovec-opt.md (@pred_single_widen_mul<any_extend:su><mode>): Change "@"
7784 into "*" in pattern name which simplifies build files.
7785 (*pred_single_widen_mul<any_extend:su><mode>): Ditto.
7786 (*pred_single_widen_mul<mode>): New pattern.
7788 2023-07-03 Richard Sandiford <richard.sandiford@arm.com>
7790 * config/aarch64/aarch64-simd.md (vec_extract<mode><Vhalf>): Expect
7791 the index to be 0 or 1.
7793 2023-07-03 Lehua Ding <lehua.ding@rivai.ai>
7796 2023-07-03 Juzhe-Zhong <juzhe.zhong@rivai.ai>
7798 * config/riscv/autovec-opt.md (*double_widen_fnma<mode>): New pattern.
7799 (*single_widen_fnma<mode>): Ditto.
7800 (*double_widen_fms<mode>): Ditto.
7801 (*single_widen_fms<mode>): Ditto.
7802 (*double_widen_fnms<mode>): Ditto.
7803 (*single_widen_fnms<mode>): Ditto.
7805 2023-07-03 Juzhe-Zhong <juzhe.zhong@rivai.ai>
7807 * config/riscv/autovec-opt.md (*double_widen_fnma<mode>): New pattern.
7808 (*single_widen_fnma<mode>): Ditto.
7809 (*double_widen_fms<mode>): Ditto.
7810 (*single_widen_fms<mode>): Ditto.
7811 (*double_widen_fnms<mode>): Ditto.
7812 (*single_widen_fnms<mode>): Ditto.
7814 2023-07-03 Pan Li <pan2.li@intel.com>
7816 * config/riscv/vector.md: Fix typo.
7818 2023-07-03 Richard Biener <rguenther@suse.de>
7820 PR tree-optimization/110506
7821 * tree-vect-patterns.cc (vect_recog_rotate_pattern): Re-order
7822 TYPE_PRECISION access with INTEGRAL_TYPE_P check.
7824 2023-07-03 Richard Biener <rguenther@suse.de>
7826 PR tree-optimization/110506
7827 * tree-ssa-ccp.cc (get_value_for_expr): Check for integral
7828 type before relying on TYPE_PRECISION to produce a nonzero mask.
7830 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
7832 * config/mips/mips.md(*and<mode>3_mips16): Generates
7833 ZEB/ZEH instructions.
7835 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
7837 * config/mips/mips.cc(mips_9bit_offset_address_p): Restrict the
7838 address register to M16_REGS for MIPS16.
7839 (BUILTIN_AVAIL_MIPS16E2): Defined a new macro.
7840 (AVAIL_MIPS16E2_OR_NON_MIPS16): Same as above.
7841 (AVAIL_NON_MIPS16 (cache..)): Update to
7842 AVAIL_MIPS16E2_OR_NON_MIPS16.
7843 * config/mips/mips.h (ISA_HAS_CACHE): Add clause for ISA_HAS_MIPS16E2.
7844 * config/mips/mips.md (mips_cache): Mark as extended MIPS16.
7846 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
7848 * config/mips/mips.h(ISA_HAS_9BIT_DISPLACEMENT): Add clause
7849 for ISA_HAS_MIPS16E2.
7850 (ISA_HAS_SYNC): Same as above.
7851 (ISA_HAS_LL_SC): Same as above.
7853 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
7855 * config/mips/mips.cc(mips_expand_ins_as_unaligned_store):
7856 Add logics for generating instruction.
7857 * config/mips/mips.h(ISA_HAS_LWL_LWR): Add clause for ISA_HAS_MIPS16E2.
7858 * config/mips/mips.md(mov_<load>l): Generates instructions.
7859 (mov_<load>r): Same as above.
7860 (mov_<store>l): Adjusted for the conditions above.
7861 (mov_<store>r): Same as above.
7862 (mov_<store>l_mips16e2): Add machine description for `define_insn mov_<store>l_mips16e2`.
7863 (mov_<store>r_mips16e2): Add machine description for `define_insn mov_<store>r_mips16e2`.
7865 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
7867 * config/mips/mips.cc(mips_symbol_insns_1): Generates LUI instruction.
7868 (mips_const_insns): Same as above.
7869 (mips_output_move): Same as above.
7870 (mips_output_function_prologue): Same as above.
7871 * config/mips/mips.md: Same as above
7873 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
7875 * config/mips/constraints.md(Yz): New constraints for mips16e2.
7876 * config/mips/mips-protos.h(mips_bit_clear_p): Declared new function.
7877 (mips_bit_clear_info): Same as above.
7878 * config/mips/mips.cc(mips_bit_clear_info): New function for
7879 generating instructions.
7880 (mips_bit_clear_p): Same as above.
7881 * config/mips/mips.h(ISA_HAS_EXT_INS): Add clause for ISA_HAS_MIPS16E2.
7882 * config/mips/mips.md(extended_mips16): Generates EXT and INS instructions.
7883 (*and<mode>3): Generates INS instruction.
7884 (*and<mode>3_mips16): Generates EXT, INS and ANDI instructions.
7885 (ior<mode>3): Add logics for ORI instruction.
7886 (*ior<mode>3_mips16_asmacro): Generates ORI instrucion.
7887 (*ior<mode>3_mips16): Add logics for XORI instruction.
7888 (*xor<mode>3_mips16): Generates XORI instrucion.
7889 (*extzv<mode>): Add logics for EXT instruction.
7890 (*insv<mode>): Add logics for INS instruction.
7891 * config/mips/predicates.md(bit_clear_operand): New predicate for
7892 generating bitwise instructions.
7893 (and_reg_operand): Add logics for generating bitwise instructions.
7895 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
7897 * config/mips/mips.cc(mips_regno_mode_ok_for_base_p): Generate instructions
7898 that uses global pointer register.
7899 (mips16_unextended_reference_p): Same as above.
7900 (mips_pic_base_register): Same as above.
7901 (mips_init_relocs): Same as above.
7902 * config/mips/mips.h(MIPS16_GP_LOADS): Defined a new macro.
7903 (GLOBAL_POINTER_REGNUM): Moved to machine description `mips.md`.
7904 * config/mips/mips.md(GLOBAL_POINTER_REGNUM): Moved to here from above.
7905 (*lowsi_mips16_gp):New `define_insn *low<mode>_mips16`.
7907 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
7909 * config/mips/mips.h(ISA_HAS_CONDMOVE): Add condition for ISA_HAS_MIPS16E2.
7910 * config/mips/mips.md(*mov<GPR:mode>_on_<MOVECC:mode>): Add logics for MOVx insts.
7911 (*mov<GPR:mode>_on_<MOVECC:mode>_mips16e2): Generate MOVx instruction.
7912 (*mov<GPR:mode>_on_<GPR2:mode>_ne): Add logics for MOVx insts.
7913 (*mov<GPR:mode>_on_<GPR2:mode>_ne_mips16e2): Generate MOVx instruction.
7914 * config/mips/predicates.md(reg_or_0_operand_mips16e2): New predicate for MOVx insts.
7916 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
7918 * config/mips/mips.cc(mips_file_start): Add mips16e2 info
7920 * config/mips/mips.h(__mips_mips16e2): Defined a new
7922 (ISA_HAS_MIPS16E2): Defined a new macro.
7923 (ASM_SPEC): Pass mmips16e2 to the assembler.
7924 * config/mips/mips.opt: Add -m(no-)mips16e2 option.
7925 * config/mips/predicates.md: Add clause for TARGET_MIPS16E2.
7926 * doc/invoke.texi: Add -m(no-)mips16e2 option..
7928 2023-07-02 Jakub Jelinek <jakub@redhat.com>
7930 PR tree-optimization/110508
7931 * tree-ssa-math-opts.cc (match_uaddc_usubc): Only replace re2 with
7932 REALPART_EXPR opf nlhs if re2 is non-NULL.
7934 2023-07-02 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
7936 * config/xtensa/xtensa.cc (xtensa_match_CLAMPS_imms_p):
7938 * config/xtensa/xtensa.md (*xtensa_clamps):
7939 Add TARGET_MINMAX to the condition.
7941 2023-07-02 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
7943 * config/xtensa/xtensa.md (*eqne_INT_MIN):
7944 Add missing ":SI" to the match_operator.
7946 2023-07-02 Iain Sandoe <iain@sandoe.co.uk>
7949 * config/darwin.opt: Add fconstant-cfstrings alias to
7950 mconstant-cfstrings.
7951 * doc/invoke.texi: Amend invocation descriptions to reflect
7952 that the fconstant-cfstrings is a target-option alias and to
7953 add the missing mconstant-cfstrings option description to the
7956 2023-07-01 Jan Hubicka <jh@suse.cz>
7958 * tree-cfg.cc (gimple_duplicate_sese_region): Add elliminated_edge
7959 parmaeter; update profile.
7960 * tree-cfg.h (gimple_duplicate_sese_region): Update prototype.
7961 * tree-ssa-loop-ch.cc (entry_loop_condition_is_static): Rename to ...
7962 (static_loop_exit): ... this; return the edge to be elliminated.
7963 (ch_base::copy_headers): Handle profile updating for eliminated exits.
7965 2023-07-01 Roger Sayle <roger@nextmovesoftware.com>
7967 * config/i386/i386-features.cc (compute_convert_gain): Provide
7968 gains/costs for ROTATE and ROTATERT (by an integer constant).
7969 (general_scalar_chain::convert_rotate): New helper function to
7970 convert a DImode or SImode rotation by an integer constant into
7972 (general_scalar_chain::convert_insn): Call the new convert_rotate
7973 for ROTATE and ROTATERT.
7974 (general_scalar_to_vector_candidate_p): Consider ROTATE and
7975 ROTATERT to be candidates if the second operand is an integer
7976 constant, valid for a rotation (or shift) in the given mode.
7977 * config/i386/i386-features.h (general_scalar_chain): Add new
7978 helper method convert_rotate.
7980 2023-07-01 Jan Hubicka <jh@suse.cz>
7982 PR tree-optimization/103680
7983 * cfg.cc (update_bb_profile_for_threading): Fix profile update;
7984 make message clearer.
7986 2023-06-30 Qing Zhao <qing.zhao@oracle.com>
7988 PR tree-optimization/101832
7989 * tree-object-size.cc (addr_object_size): Handle structure/union type
7990 when it has flexible size.
7992 2023-06-30 Eric Botcazou <ebotcazou@adacore.com>
7994 * gimple-fold.cc (fold_array_ctor_reference): Fix head comment.
7995 (fold_nonarray_ctor_reference): Likewise. Specifically deal
7996 with integral bit-fields.
7997 (fold_ctor_reference): Make sure that the constructor uses the
7998 native storage order.
8000 2023-06-30 Jan Hubicka <jh@suse.cz>
8002 PR middle-end/109849
8003 * predict.cc (estimate_bb_frequencies): Turn to static function.
8004 (expr_expected_value_1): Fix handling of binary expressions with
8006 * predict.def (PRED_MALLOC_NONNULL): Move later in the priority queue.
8007 (PRED_BUILTIN_EXPECT_WITH_PROBABILITY): Move to almost top of the priority
8009 * predict.h (estimate_bb_frequencies): No longer declare it.
8011 2023-06-30 Uros Bizjak <ubizjak@gmail.com>
8013 * fold-const.h (multiple_of_p): Change return type from int to bool.
8014 * fold-const.cc (split_tree): Change negl_p, neg_litp_p,
8015 neg_conp_p and neg_var_p variables to bool.
8016 (const_binop): Change sat_p variable to bool.
8017 (merge_ranges): Change no_overlap variable to bool.
8018 (extract_muldiv_1): Change same_p variable to bool.
8019 (tree_swap_operands_p): Update function body for bool return type.
8020 (fold_truth_andor): Change commutative variable to bool.
8021 (multiple_of_p): Change return type
8022 from int to void and adjust function body accordingly.
8023 * optabs.h (expand_twoval_unop): Change return type from int to bool.
8024 (expand_twoval_binop): Ditto.
8025 (can_compare_p): Ditto.
8026 (have_add2_insn): Ditto.
8027 (have_addptr3_insn): Ditto.
8028 (have_sub2_insn): Ditto.
8029 (have_insn_for): Ditto.
8030 * optabs.cc (add_equal_note): Ditto.
8031 (widen_operand): Change no_extend argument from int to bool.
8032 (expand_binop): Ditto.
8033 (expand_twoval_unop): Change return type
8034 from int to void and adjust function body accordingly.
8035 (expand_twoval_binop): Ditto.
8036 (can_compare_p): Ditto.
8037 (have_add2_insn): Ditto.
8038 (have_addptr3_insn): Ditto.
8039 (have_sub2_insn): Ditto.
8040 (have_insn_for): Ditto.
8042 2023-06-30 Oluwatamilore Adebayo <oluwatamilore.adebayo@arm.com>
8044 * config/aarch64/aarch64-simd.md
8045 (vec_widen_<su>abdl_lo_<mode>, vec_widen_<su>abdl_hi_<mode>):
8046 Expansions for abd vec widen optabs.
8047 (aarch64_<su>abdl<mode>_insn): VQW based abdl RTL.
8048 * config/aarch64/iterators.md (USMAX_EXT): Code attributes
8049 that give the appropriate extend RTL for the max RTL.
8051 2023-06-30 Oluwatamilore Adebayo <oluwatamilore.adebayo@arm.com>
8053 * internal-fn.def (VEC_WIDEN_ABD): New internal hilo optab.
8054 * optabs.def (vec_widen_sabd_optab,
8055 vec_widen_sabd_hi_optab, vec_widen_sabd_lo_optab,
8056 vec_widen_sabd_odd_even, vec_widen_sabd_even_optab,
8057 vec_widen_uabd_optab,
8058 vec_widen_uabd_hi_optab, vec_widen_uabd_lo_optab,
8059 vec_widen_uabd_odd_even, vec_widen_uabd_even_optab):
8061 * doc/md.texi: Document them.
8062 * tree-vect-patterns.cc (vect_recog_abd_pattern): Update to
8063 to build a VEC_WIDEN_ABD call if the input precision is smaller
8064 than the precision of the output.
8065 (vect_recog_widen_abd_pattern): Should an ABD expression be
8066 found preceeding an extension, replace the two with a
8069 2023-06-30 Pan Li <pan2.li@intel.com>
8071 * config/riscv/vector.md: Refactor the common condition.
8073 2023-06-30 Richard Biener <rguenther@suse.de>
8075 PR tree-optimization/110496
8076 * gimple-ssa-store-merging.cc (find_bswap_or_nop_1): Re-order
8077 verifying and TYPE_PRECISION query for the BIT_FIELD_REF case.
8079 2023-06-30 Richard Biener <rguenther@suse.de>
8081 PR middle-end/110489
8082 * statistics.cc (curr_statistics_hash): Add argument
8083 indicating whether we should allocate the hash.
8084 (statistics_fini_pass): If the hash isn't allocated
8085 only print the summary header.
8087 2023-06-30 Segher Boessenkool <segher@kernel.crashing.org>
8088 Thomas Schwinge <thomas@codesourcery.com>
8090 * config/nvptx/nvptx.cc (TARGET_LRA_P): Remove.
8092 2023-06-30 Jovan Dmitrović <jovan.dmitrovic@syrmia.com>
8095 * config/mips/mips.cc (mips_function_arg_alignment): Returns
8096 the alignment of function argument. In case of typedef type,
8097 it returns the aligment of the aliased type.
8098 (mips_function_arg_boundary): Relocated calculation of the
8099 aligment of function arguments.
8101 2023-06-29 Jan Hubicka <jh@suse.cz>
8103 PR tree-optimization/109849
8104 * ipa-fnsummary.cc (decompose_param_expr): Skip
8105 functions returning its parameter.
8106 (set_cond_stmt_execution_predicate): Return early
8107 if predicate was constructed.
8109 2023-06-29 Qing Zhao <qing.zhao@oracle.com>
8112 * doc/extend.texi: Document GCC extension on a structure containing
8113 a flexible array member to be a member of another structure.
8115 2023-06-29 Qing Zhao <qing.zhao@oracle.com>
8117 * print-tree.cc (print_node): Print new bit type_include_flexarray.
8118 * tree-core.h (struct tree_type_common): Use bit no_named_args_stdarg_p
8119 as type_include_flexarray for RECORD_TYPE or UNION_TYPE.
8120 * tree-streamer-in.cc (unpack_ts_type_common_value_fields): Stream
8121 in bit no_named_args_stdarg_p properly for its corresponding type.
8122 * tree-streamer-out.cc (pack_ts_type_common_value_fields): Stream
8123 out bit no_named_args_stdarg_p properly for its corresponding type.
8124 * tree.h (TYPE_INCLUDES_FLEXARRAY): New macro TYPE_INCLUDES_FLEXARRAY.
8126 2023-06-29 Aldy Hernandez <aldyh@redhat.com>
8128 * tree-vrp.cc (maybe_set_nonzero_bits): Move from here...
8129 * tree-ssa-dom.cc (maybe_set_nonzero_bits): ...to here.
8130 * tree-vrp.h (maybe_set_nonzero_bits): Remove.
8132 2023-06-29 Aldy Hernandez <aldyh@redhat.com>
8134 * value-range.cc (frange::set): Do not call verify_range.
8135 (frange::normalize_kind): Verify range.
8136 (frange::union_nans): Do not call verify_range.
8137 (frange::union_): Same.
8138 (frange::intersect): Same.
8139 (irange::irange_single_pair_union): Call normalize_kind if
8141 (irange::union_): Same.
8142 (irange::intersect): Same.
8143 (irange::set_range_from_nonzero_bits): Verify range.
8144 (irange::set_nonzero_bits): Call normalize_kind if necessary.
8145 (irange::get_nonzero_bits): Tweak comment.
8146 (irange::intersect_nonzero_bits): Call normalize_kind if
8148 (irange::union_nonzero_bits): Same.
8149 * value-range.h (irange::normalize_kind): Verify range.
8151 2023-06-29 Uros Bizjak <ubizjak@gmail.com>
8153 * cselib.h (rtx_equal_for_cselib_1):
8154 Change return type from int to bool.
8155 (references_value_p): Ditto.
8156 (rtx_equal_for_cselib_p): Ditto.
8157 * expr.h (can_store_by_pieces): Ditto.
8158 (try_casesi): Ditto.
8159 (try_tablejump): Ditto.
8160 (safe_from_p): Ditto.
8161 * sbitmap.h (bitmap_equal_p): Ditto.
8162 * cselib.cc (references_value_p): Change return type
8163 from int to void and adjust function body accordingly.
8164 (rtx_equal_for_cselib_1): Ditto.
8165 * expr.cc (is_aligning_offset): Ditto.
8166 (can_store_by_pieces): Ditto.
8167 (mostly_zeros_p): Ditto.
8168 (all_zeros_p): Ditto.
8169 (safe_from_p): Ditto.
8170 (is_aligning_offset): Ditto.
8171 (try_casesi): Ditto.
8172 (try_tablejump): Ditto.
8173 (store_constructor): Change "need_to_clear" and
8174 "const_bounds_p" variables to bool.
8175 * sbitmap.cc (bitmap_equal_p): Change return type from int to bool.
8177 2023-06-29 Robin Dapp <rdapp@ventanamicro.com>
8179 * tree-ssa-math-opts.cc (divmod_candidate_p): Use
8182 2023-06-29 Richard Biener <rguenther@suse.de>
8184 PR tree-optimization/110460
8185 * tree-vect-stmts.cc (get_related_vectype_for_scalar_type):
8186 Only allow integral, pointer and scalar float type scalar_type.
8188 2023-06-29 Lili Cui <lili.cui@intel.com>
8190 PR tree-optimization/110148
8191 * tree-ssa-reassoc.cc (rewrite_expr_tree_parallel): Handle loop-carried
8192 ops in this function.
8194 2023-06-29 Richard Biener <rguenther@suse.de>
8196 PR middle-end/110452
8197 * expr.cc (store_constructor): Handle uniform boolean
8198 vectors with integer mode specially.
8200 2023-06-29 Richard Biener <rguenther@suse.de>
8202 PR middle-end/110461
8203 * match.pd (bitop (convert@2 @0) (convert?@3 @1)): Disable
8206 2023-06-29 Richard Sandiford <richard.sandiford@arm.com>
8208 * vec.h (gt_pch_nx): Add overloads for va_gc_atomic.
8209 (array_slice): Relax va_gc constructor to handle all vectors
8210 with a vl_embed layout.
8212 2023-06-29 Pan Li <pan2.li@intel.com>
8214 * config/riscv/riscv.cc (riscv_emit_mode_set): Add emit for FRM.
8215 (riscv_mode_needed): Likewise.
8216 (riscv_entity_mode_after): Likewise.
8217 (riscv_mode_after): Likewise.
8218 (riscv_mode_entry): Likewise.
8219 (riscv_mode_exit): Likewise.
8220 * config/riscv/riscv.h (NUM_MODES_FOR_MODE_SWITCHING): Add number
8222 * config/riscv/riscv.md: Add FRM register.
8223 * config/riscv/vector-iterators.md: Add FRM type.
8224 * config/riscv/vector.md (frm_mode): Define new attr for FRM mode.
8225 (fsrm): Define new insn for fsrm instruction.
8227 2023-06-29 Pan Li <pan2.li@intel.com>
8229 * config/riscv/riscv-protos.h (enum floating_point_rounding_mode):
8230 Add macro for static frm min and max.
8231 * config/riscv/riscv-vector-builtins-bases.cc
8232 (class binop_frm): New class for floating-point with frm.
8233 (BASE): Add vfadd for frm.
8234 * config/riscv/riscv-vector-builtins-bases.h: Likewise.
8235 * config/riscv/riscv-vector-builtins-functions.def
8236 (vfadd_frm): Likewise.
8237 * config/riscv/riscv-vector-builtins-shapes.cc
8238 (struct alu_frm_def): New struct for alu with frm.
8239 (SHAPE): Add alu with frm.
8240 * config/riscv/riscv-vector-builtins-shapes.h: Likewise.
8241 * config/riscv/riscv-vector-builtins.cc
8242 (function_checker::report_out_of_range_and_not): New function
8243 for report out of range and not val.
8244 (function_checker::require_immediate_range_or): New function
8245 for checking in range or one val.
8246 * config/riscv/riscv-vector-builtins.h: Add function decl.
8248 2023-06-29 Cui, Lili <lili.cui@intel.com>
8250 * common/config/i386/cpuinfo.h (get_intel_cpu): Remove model value 0xa8
8251 from Rocketlake, move model value 0xbf from Alderlake to Raptorlake.
8253 2023-06-28 Hans-Peter Nilsson <hp@axis.com>
8256 * config/cris/cris.cc (cris_postdbr_cmpelim): Don't apply PATTERN
8257 to insn before validating it.
8259 2023-06-28 Jan Hubicka <jh@suse.cz>
8261 PR middle-end/110334
8262 * ipa-fnsummary.h (ipa_fn_summary): Add
8263 safe_to_inline_to_always_inline.
8264 * ipa-inline.cc (can_early_inline_edge_p): ICE
8265 if SSA is not built; do cycle checking for
8266 always_inline functions.
8267 (inline_always_inline_functions): Be recrusive;
8268 watch for cycles; do not updat overall summary.
8269 (early_inliner): Do not give up on always_inlines.
8270 * ipa-utils.cc (ipa_reverse_postorder): Do not skip
8273 2023-06-28 Uros Bizjak <ubizjak@gmail.com>
8275 * output.h (leaf_function_p): Change return type from int to bool.
8276 (final_forward_branch_p): Ditto.
8277 (only_leaf_regs_used): Ditto.
8278 (maybe_assemble_visibility): Ditto.
8279 * varasm.h (supports_one_only): Ditto.
8280 * rtl.h (compute_alignments): Change return type from int to void.
8281 * final.cc (app_on): Change return type from int to bool.
8282 (compute_alignments): Change return type from int to void
8283 and adjust function body accordingly.
8284 (shorten_branches): Change "something_changed" variable
8285 type from int to bool.
8286 (leaf_function_p): Change return type from int to bool
8287 and adjust function body accordingly.
8288 (final_forward_branch_p): Ditto.
8289 (only_leaf_regs_used): Ditto.
8290 * varasm.cc (contains_pointers_p): Change return type from
8291 int to bool and adjust function body accordingly.
8292 (compare_constant): Ditto.
8293 (maybe_assemble_visibility): Ditto.
8294 (supports_one_only): Ditto.
8296 2023-06-28 Manolis Tsamis <manolis.tsamis@vrull.eu>
8299 * regcprop.cc (maybe_mode_change): Check stack_pointer_rtx mode.
8300 (maybe_copy_reg_attrs): New function.
8301 (find_oldest_value_reg): Use maybe_copy_reg_attrs.
8302 (copyprop_hardreg_forward_1): Ditto.
8304 2023-06-28 Richard Biener <rguenther@suse.de>
8306 PR tree-optimization/110434
8307 * tree-nrv.cc (pass_nrv::execute): Remove CLOBBERs of
8308 VAR we replace with <retval>.
8310 2023-06-28 Richard Biener <rguenther@suse.de>
8312 PR tree-optimization/110451
8313 * tree-ssa-loop-im.cc (stmt_cost): [VEC_]COND_EXPR and
8314 tcc_comparison are expensive.
8316 2023-06-28 Roger Sayle <roger@nextmovesoftware.com>
8318 * config/i386/i386-expand.cc (ix86_expand_branch): Also use ptest
8319 for TImode comparisons on 32-bit architectures.
8320 * config/i386/i386.md (cbranch<mode>4): Change from SDWIM to
8321 SWIM1248x to exclude/avoid TImode being conditional on -m64.
8322 (cbranchti4): New define_expand for TImode on both TARGET_64BIT
8323 and/or with TARGET_SSE4_1.
8324 * config/i386/predicates.md (ix86_timode_comparison_operator):
8325 New predicate that depends upon TARGET_64BIT.
8326 (ix86_timode_comparison_operand): Likewise.
8328 2023-06-28 Roger Sayle <roger@nextmovesoftware.com>
8331 * config/i386/i386-features.cc (compute_convert_gain): Provide
8332 more accurate gains for conversion of scalar comparisons to
8335 2023-06-28 Richard Biener <rguenther@suse.de>
8337 PR tree-optimization/110443
8338 * tree-vect-slp.cc (vect_build_slp_tree_1): Reject non-grouped
8341 2023-06-28 Haochen Gui <guihaoc@gcc.gnu.org>
8343 * config/rs6000/rs6000.md (peephole2 for compare_and_move): New.
8344 (peephole2 for move_and_compare): New.
8345 (mode_iterator WORD): New. Set the mode to SI/DImode by
8347 (*mov<mode>_internal2): Change the mode iterator from P to WORD.
8348 (split pattern for compare_and_move): Likewise.
8350 2023-06-28 Juzhe-Zhong <juzhe.zhong@rivai.ai>
8352 * config/riscv/autovec-opt.md (*double_widen_fma<mode>): New pattern.
8353 (*single_widen_fma<mode>): Ditto.
8355 2023-06-28 Haochen Gui <guihaoc@gcc.gnu.org>
8358 * config/rs6000/altivec.md (*altivec_vupkhs<VU_char>_direct): Rename
8360 (altivec_vupkhs<VU_char>_direct): ...this.
8361 * config/rs6000/predicates.md (vspltisw_vupkhsw_constant_split): New
8362 predicate to test if a constant can be loaded with vspltisw and
8364 (easy_vector_constant): Call vspltisw_vupkhsw_constant_p to Check if
8365 a vector constant can be synthesized with a vspltisw and a vupkhsw.
8366 * config/rs6000/rs6000-protos.h (vspltisw_vupkhsw_constant_p):
8368 * config/rs6000/rs6000.cc (vspltisw_vupkhsw_constant_p): New
8369 function to return true if OP mode is V2DI and can be synthesized
8370 with vupkhsw and vspltisw.
8371 * config/rs6000/vsx.md (*vspltisw_v2di_split): New insn to load up
8372 constants with vspltisw and vupkhsw.
8374 2023-06-28 Jan Hubicka <jh@suse.cz>
8376 PR tree-optimization/110377
8377 * ipa-prop.cc (ipa_compute_jump_functions_for_edge): Pass statement to
8379 (ipa_analyze_node): Enable ranger.
8381 2023-06-28 Richard Biener <rguenther@suse.de>
8383 * tree.h (TYPE_PRECISION): Check for non-VECTOR_TYPE.
8384 (TYPE_PRECISION_RAW): Provide raw access to the precision
8386 * tree.cc (verify_type_variant): Compare TYPE_PRECISION_RAW.
8387 (gimple_canonical_types_compatible_p): Likewise.
8388 * tree-streamer-out.cc (pack_ts_type_common_value_fields):
8389 Stream TYPE_PRECISION_RAW.
8390 * tree-streamer-in.cc (unpack_ts_type_common_value_fields):
8392 * lto-streamer-out.cc (hash_tree): Hash TYPE_PRECISION_RAW.
8394 2023-06-28 Alexandre Oliva <oliva@adacore.com>
8396 * doc/extend.texi (zero-call-used-regs): Document leafy and
8398 * flag-types.h (zero_regs_flags): Add LEAFY_MODE, as well as
8400 * function.cc (gen_call_ued_regs_seq): Set only_used for leaf
8401 functions in leafy mode.
8402 * opts.cc (zero_call_used_regs_opts): Add leafy and variants.
8404 2023-06-28 Juzhe-Zhong <juzhe.zhong@rivai.ai>
8406 * config/riscv/riscv-vector-builtins-bases.cc: Adapt expand.
8407 * config/riscv/vector.md (@pred_single_widen_<plus_minus:optab><mode>):
8409 (@pred_single_widen_add<mode>): New pattern.
8410 (@pred_single_widen_sub<mode>): New pattern.
8412 2023-06-28 liuhongt <hongtao.liu@intel.com>
8414 * config/i386/i386.cc (ix86_invalid_conversion): New function.
8415 (TARGET_INVALID_CONVERSION): Define as
8416 ix86_invalid_conversion.
8418 2023-06-27 Robin Dapp <rdapp@ventanamicro.com>
8420 * config/riscv/autovec.md (<optab><vnconvert><mode>2): New
8422 (<float_cvt><vnconvert><mode>2): Ditto.
8423 (<optab><mode><vnconvert>2): Ditto.
8424 (<float_cvt><mode><vnconvert>2): Ditto.
8425 * config/riscv/vector-iterators.md: Add vnconvert.
8427 2023-06-27 Robin Dapp <rdapp@ventanamicro.com>
8429 * config/riscv/autovec.md (extend<v_double_trunc><mode>2): New
8431 (extend<v_quad_trunc><mode>2): Ditto.
8432 (trunc<mode><v_double_trunc>2): Ditto.
8433 (trunc<mode><v_quad_trunc>2): Ditto.
8434 * config/riscv/vector-iterators.md: Add VQEXTF and HF to
8435 V_QUAD_TRUNC and v_quad_trunc.
8437 2023-06-27 Robin Dapp <rdapp@ventanamicro.com>
8439 * config/riscv/autovec.md (<float_cvt><vconvert><mode>2): New
8442 2023-06-27 Robin Dapp <rdapp@ventanamicro.com>
8444 * config/riscv/autovec.md (copysign<mode>3): Add expander.
8445 (xorsign<mode>3): Ditto.
8446 * config/riscv/riscv-vector-builtins-bases.cc (class vfsgnjn):
8448 * config/riscv/vector-iterators.md (copysign): Remove ncopysign.
8452 * config/riscv/vector.md (@pred_ncopysign<mode>): Split off.
8453 (@pred_ncopysign<mode>_scalar): Ditto.
8455 2023-06-27 Robin Dapp <rdapp@ventanamicro.com>
8457 * config/riscv/autovec.md: VF_AUTO -> VF.
8458 * config/riscv/vector-iterators.md: Introduce VF_ZVFHMIN,
8459 VWEXTF_ZVFHMIN and use TARGET_ZVFH in VWCONVERTI, VHF and
8461 * config/riscv/vector.md: Use new iterators.
8463 2023-06-27 Robin Dapp <rdapp@ventanamicro.com>
8465 * match.pd: Use element_mode and check if target supports
8466 operation with new type.
8468 2023-06-27 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
8470 * config/aarch64/aarch64-sve-builtins-base.cc
8471 (svdupq_impl::fold_nonconst_dupq): New method.
8472 (svdupq_impl::fold): Call fold_nonconst_dupq.
8474 2023-06-27 Andrew Pinski <apinski@marvell.com>
8476 PR middle-end/110420
8477 PR middle-end/103979
8479 * gimplify.cc (gimplify_asm_expr): Mark asm with labels as volatile.
8481 2023-06-27 Aldy Hernandez <aldyh@redhat.com>
8483 * ipa-cp.cc (decide_whether_version_node): Adjust comment.
8484 * ipa-fnsummary.cc (evaluate_conditions_for_known_args): Adjust
8486 (set_switch_stmt_execution_predicate): Same.
8487 * ipa-prop.cc (ipa_compute_jump_functions_for_edge): Same.
8489 2023-06-27 Aldy Hernandez <aldyh@redhat.com>
8491 * ipa-prop.cc (struct ipa_vr_ggc_hash_traits): Adjust for use with
8492 ipa_vr instead of value_range.
8495 (ipa_get_value_range): Same.
8496 * value-range.cc (gt_pch_nx): Move to ipa-prop.cc and adjust for
8500 2023-06-27 Aldy Hernandez <aldyh@redhat.com>
8502 * ipa-cp.cc (ipa_vr_operation_and_type_effects): New.
8503 * ipa-prop.cc (ipa_get_value_range): Adjust for ipa_vr.
8504 (ipa_set_jfunc_vr): Take a range.
8505 (ipa_compute_jump_functions_for_edge): Pass range to
8507 (ipa_write_jump_function): Call streamer write helper.
8508 (ipa_read_jump_function): Call streamer read helper.
8509 * ipa-prop.h (class ipa_vr): Change m_vr to an ipa_vr.
8511 2023-06-27 Richard Sandiford <richard.sandiford@arm.com>
8513 * gengtype-parse.cc (consume_until_comma_or_eos): Parse "= { ... }"
8514 as a probable initializer rather than a probable complete statement.
8516 2023-06-27 Richard Biener <rguenther@suse.de>
8518 PR tree-optimization/96208
8519 * tree-vect-slp.cc (vect_build_slp_tree_1): Allow
8520 a non-grouped load if it is the same for all lanes.
8521 (vect_build_slp_tree_2): Handle not grouped loads.
8522 (vect_optimize_slp_pass::remove_redundant_permutations):
8524 (vect_transform_slp_perm_load_1): Likewise.
8525 * tree-vect-stmts.cc (vect_model_load_cost): Likewise.
8526 (get_group_load_store_type): Likewise. Handle
8528 (vectorizable_load): Likewise.
8530 2023-06-27 liuhongt <hongtao.liu@intel.com>
8532 PR rtl-optimization/110237
8533 * config/i386/sse.md (<avx512>_store<mode>_mask): Refine with
8535 (maskstore<mode><avx512fmaskmodelower): Ditto.
8536 (*<avx512>_store<mode>_mask): New define_insn, it's renamed
8537 from original <avx512>_store<mode>_mask.
8539 2023-06-27 liuhongt <hongtao.liu@intel.com>
8541 * config/i386/i386-features.cc (pass_insert_vzeroupper:gate):
8542 Move flag_expensive_optimizations && !optimize_size to ..
8543 * config/i386/i386-options.cc (ix86_option_override_internal):
8544 .. this, it makes -mvzeroupper independent of optimization
8545 level, but still keeps the behavior of architecture
8546 tuning(emit_vzeroupper) unchanged.
8548 2023-06-27 liuhongt <hongtao.liu@intel.com>
8551 * config/i386/i386.cc (ix86_avx_u127_mode_needed): Don't emit
8552 vzeroupper for vzeroupper call_insn.
8554 2023-06-27 Andrew Pinski <apinski@marvell.com>
8556 * doc/extend.texi (__builtin_alloca_with_align_and_max): Fix
8559 2023-06-27 Juzhe-Zhong <juzhe.zhong@rivai.ai>
8561 * config/riscv/riscv-v.cc (expand_const_vector): Fix stepped vector
8564 2023-06-26 Andrew Pinski <apinski@marvell.com>
8566 * doc/extend.texi (access attribute): Add
8568 (interrupt/interrupt_handler attribute):
8571 2023-06-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8573 * config/aarch64/aarch64-simd.md (aarch64_sqrshrun_n<mode>_insn):
8574 Use <DWI> instead of <V2XWIDE>.
8575 (aarch64_sqrshrun_n<mode>): Likewise.
8577 2023-06-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8579 * config/aarch64/aarch64-protos.h (aarch64_const_vec_rsra_rnd_imm_p):
8581 (aarch64_rnd_imm_p): ... This.
8582 * config/aarch64/predicates.md (aarch64_simd_rsra_rnd_imm_vec):
8584 (aarch64_int_rnd_operand): ... This.
8585 (aarch64_simd_rshrn_imm_vec): Delete.
8586 * config/aarch64/aarch64-simd.md (aarch64_<sra_op>rsra_n<mode>_insn):
8587 Adjust for the above.
8588 (aarch64_<sra_op>rshr_n<mode><vczle><vczbe>_insn): Likewise.
8589 (*aarch64_<shrn_op>rshrn_n<mode>_insn): Likewise.
8590 (*aarch64_sqrshrun_n<mode>_insn<vczle><vczbe>): Likewise.
8591 (aarch64_sqrshrun_n<mode>_insn): Likewise.
8592 (aarch64_<shrn_op>rshrn2_n<mode>_insn_le): Likewise.
8593 (aarch64_<shrn_op>rshrn2_n<mode>_insn_be): Likewise.
8594 (aarch64_sqrshrun2_n<mode>_insn_le): Likewise.
8595 (aarch64_sqrshrun2_n<mode>_insn_be): Likewise.
8596 * config/aarch64/aarch64.cc (aarch64_const_vec_rsra_rnd_imm_p):
8598 (aarch64_rnd_imm_p): ... This.
8600 2023-06-26 Andreas Krebbel <krebbel@linux.ibm.com>
8602 * config/s390/s390.cc (s390_encode_section_info): Set
8603 SYMBOL_FLAG_SET_NOTALIGN2 only if the symbol has explicitely been
8606 2023-06-26 Jan Hubicka <jh@suse.cz>
8608 PR tree-optimization/109849
8609 * tree-ssa-dce.cc (make_forwarders_with_degenerate_phis): Fix profile
8610 count of newly constructed forwarder block.
8612 2023-06-26 Andrew Carlotti <andrew.carlotti@arm.com>
8614 * doc/optinfo.texi: Fix "steam" -> "stream".
8616 2023-06-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
8618 * tree-ssa-dse.cc (initialize_ao_ref_for_dse): Add LEN_MASK_STORE and
8620 (dse_optimize_stmt): Add LEN_MASK_STORE.
8622 2023-06-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
8624 * gimple-fold.cc (gimple_fold_partial_load_store_mem_ref): Fix gimple
8625 fold of LOAD/STORE with length.
8627 2023-06-26 Andrew MacLeod <amacleod@redhat.com>
8629 * gimple-range-gori.cc (compute_operand1_and_operand2_range):
8630 Check for interdependence between operands 1 and 2.
8632 2023-06-26 Richard Sandiford <richard.sandiford@arm.com>
8634 * tree-vect-stmts.cc (vectorizable_conversion): Take multi_step_cvt
8635 into account when costing non-widening/truncating conversions.
8637 2023-06-26 Richard Biener <rguenther@suse.de>
8639 PR tree-optimization/110381
8640 * tree-vect-slp.cc (vect_optimize_slp_pass::start_choosing_layouts):
8641 Materialize permutes before fold-left reductions.
8643 2023-06-26 Pan Li <pan2.li@intel.com>
8645 * config/riscv/riscv-vector-builtins-bases.h: Remove duplicated decl.
8647 2023-06-26 Richard Biener <rguenther@suse.de>
8649 * varasm.cc (initializer_constant_valid_p_1): Also
8650 constrain the type of value to be scalar integral
8651 before dispatching to narrowing_initializer_constant_valid_p.
8653 2023-06-26 Richard Biener <rguenther@suse.de>
8655 * tree-ssa-scopedtables.cc (hashable_expr_equal_p):
8656 Use element_precision.
8658 2023-06-26 Juzhe-Zhong <juzhe.zhong@rivai.ai>
8660 * config/riscv/autovec.md (vcond<V:mode><VI:mode>): Remove redundant
8662 (vcondu<V:mode><VI:mode>): Ditto.
8663 * config/riscv/riscv-protos.h (expand_vcond): Ditto.
8664 * config/riscv/riscv-v.cc (expand_vcond): Ditto.
8666 2023-06-26 Richard Biener <rguenther@suse.de>
8668 PR tree-optimization/110392
8669 * gimple-predicate-analysis.cc (uninit_analysis::is_use_guarded):
8670 Do early exits on true/false predicate only after normalization.
8672 2023-06-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
8674 * tree-ssa-sccvn.cc (vn_reference_lookup_3): Change name "len" into
8677 2023-06-26 Roger Sayle <roger@nextmovesoftware.com>
8679 * config/i386/i386.md (peephole2): Simplify zeroing a register
8680 followed by an IOR, XOR or PLUS operation on it, into a move.
8681 (*ashl<dwi>3_doubleword_highpart): New define_insn_and_split to
8682 eliminate (and hide from reload) unnecessary word to doubleword
8683 extensions that are followed by left shifts by sufficiently large,
8684 but valid, bit counts.
8686 2023-06-26 liuhongt <hongtao.liu@intel.com>
8688 PR tree-optimization/110371
8689 PR tree-optimization/110018
8690 * tree-vect-stmts.cc (vectorizable_conversion): Use cvt_op to
8691 save intermediate type operand instead of "subtle" vec_dest
8694 2023-06-26 liuhongt <hongtao.liu@intel.com>
8696 PR tree-optimization/110371
8697 PR tree-optimization/110018
8698 * tree-vect-stmts.cc (vectorizable_conversion): Don't use
8699 intermiediate type for FIX_TRUNC_EXPR when ftrapping-math.
8701 2023-06-26 Hongyu Wang <hongyu.wang@intel.com>
8703 * config/i386/i386-options.cc (ix86_valid_target_attribute_tree):
8704 Override tune_string with arch_string if tune_string is not
8705 explicitly specified.
8707 2023-06-25 Juzhe-Zhong <juzhe.zhong@rivai.ai>
8709 * config/riscv/riscv-vsetvl.cc (vector_insn_info::parse_insn): Ehance
8711 * config/riscv/riscv-vsetvl.h: New function.
8713 2023-06-25 Li Xu <xuli1@eswincomputing.com>
8715 * config/riscv/riscv-vector-builtins-bases.cc: change emit_insn to
8718 2023-06-25 Juzhe-Zhong <juzhe.zhong@rivai.ai>
8720 * config/riscv/autovec.md (len_load_<mode>): Remove.
8721 (len_maskload<mode><vm>): Remove.
8722 (len_store_<mode>): New pattern.
8723 (len_maskstore<mode><vm>): New pattern.
8724 * config/riscv/predicates.md (autovec_length_operand): New predicate.
8725 * config/riscv/riscv-protos.h (enum insn_type): New enum.
8726 (expand_load_store): New function.
8727 * config/riscv/riscv-v.cc (emit_vlmax_masked_insn): Ditto.
8728 (emit_nonvlmax_masked_insn): Ditto.
8729 (expand_load_store): Ditto.
8730 * config/riscv/riscv-vector-builtins.cc
8731 (function_expander::use_contiguous_store_insn): Add avl_type operand
8733 * config/riscv/vector.md: Ditto.
8735 2023-06-25 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
8737 * internal-fn.cc (expand_partial_store_optab_fn): Fix bug of BIAS
8740 2023-06-25 Pan Li <pan2.li@intel.com>
8742 * config/riscv/vector.md: Revert.
8744 2023-06-25 Pan Li <pan2.li@intel.com>
8746 * config/riscv/genrvv-type-indexer.cc (valid_type): Revert changes.
8747 * config/riscv/riscv-modes.def (RVV_TUPLE_MODES): Ditto.
8748 (ADJUST_ALIGNMENT): Ditto.
8749 (RVV_TUPLE_PARTIAL_MODES): Ditto.
8750 (ADJUST_NUNITS): Ditto.
8751 * config/riscv/riscv-vector-builtins-types.def (vfloat16mf4x2_t): Ditto.
8752 (vfloat16mf4x3_t): Ditto.
8753 (vfloat16mf4x4_t): Ditto.
8754 (vfloat16mf4x5_t): Ditto.
8755 (vfloat16mf4x6_t): Ditto.
8756 (vfloat16mf4x7_t): Ditto.
8757 (vfloat16mf4x8_t): Ditto.
8758 (vfloat16mf2x2_t): Ditto.
8759 (vfloat16mf2x3_t): Ditto.
8760 (vfloat16mf2x4_t): Ditto.
8761 (vfloat16mf2x5_t): Ditto.
8762 (vfloat16mf2x6_t): Ditto.
8763 (vfloat16mf2x7_t): Ditto.
8764 (vfloat16mf2x8_t): Ditto.
8765 (vfloat16m1x2_t): Ditto.
8766 (vfloat16m1x3_t): Ditto.
8767 (vfloat16m1x4_t): Ditto.
8768 (vfloat16m1x5_t): Ditto.
8769 (vfloat16m1x6_t): Ditto.
8770 (vfloat16m1x7_t): Ditto.
8771 (vfloat16m1x8_t): Ditto.
8772 (vfloat16m2x2_t): Ditto.
8773 (vfloat16m2x3_t): Diito.
8774 (vfloat16m2x4_t): Diito.
8775 (vfloat16m4x2_t): Diito.
8776 * config/riscv/riscv-vector-builtins.def (vfloat16mf4x2_t): Ditto.
8777 (vfloat16mf4x3_t): Ditto.
8778 (vfloat16mf4x4_t): Ditto.
8779 (vfloat16mf4x5_t): Ditto.
8780 (vfloat16mf4x6_t): Ditto.
8781 (vfloat16mf4x7_t): Ditto.
8782 (vfloat16mf4x8_t): Ditto.
8783 (vfloat16mf2x2_t): Ditto.
8784 (vfloat16mf2x3_t): Ditto.
8785 (vfloat16mf2x4_t): Ditto.
8786 (vfloat16mf2x5_t): Ditto.
8787 (vfloat16mf2x6_t): Ditto.
8788 (vfloat16mf2x7_t): Ditto.
8789 (vfloat16mf2x8_t): Ditto.
8790 (vfloat16m1x2_t): Ditto.
8791 (vfloat16m1x3_t): Ditto.
8792 (vfloat16m1x4_t): Ditto.
8793 (vfloat16m1x5_t): Ditto.
8794 (vfloat16m1x6_t): Ditto.
8795 (vfloat16m1x7_t): Ditto.
8796 (vfloat16m1x8_t): Ditto.
8797 (vfloat16m2x2_t): Ditto.
8798 (vfloat16m2x3_t): Ditto.
8799 (vfloat16m2x4_t): Ditto.
8800 (vfloat16m4x2_t): Ditto.
8801 * config/riscv/riscv-vector-switch.def (TUPLE_ENTRY): Ditto.
8802 * config/riscv/riscv.md: Ditto.
8803 * config/riscv/vector-iterators.md: Ditto.
8805 2023-06-25 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
8807 * gimple-fold.cc (arith_overflowed_p): Apply LEN_MASK_{LOAD,STORE}.
8808 (gimple_fold_partial_load_store_mem_ref): Ditto.
8809 (gimple_fold_partial_store): Ditto.
8810 (gimple_fold_call): Ditto.
8812 2023-06-25 liuhongt <hongtao.liu@intel.com>
8815 * config/i386/sse.md (maskload<mode><avx512fmaskmodelower>):
8816 Refine pattern with UNSPEC_MASKLOAD.
8817 (maskload<mode><avx512fmaskmodelower>): Ditto.
8818 (*<avx512>_load<mode>_mask): Extend mode iterator to
8820 (*<avx512>_load<mode>): Ditto.
8822 2023-06-25 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
8824 * tree-ssa-alias.cc (call_may_clobber_ref_p_1): Add LEN_MASK_STORE.
8826 2023-06-25 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
8828 * tree-ssa-alias.cc (ref_maybe_used_by_call_p_1): Apply
8829 LEN_MASK_{LOAD,STORE}
8831 2023-06-25 yulong <shiyulong@iscas.ac.cn>
8833 * config/riscv/vector.md: Add float16 attr at sew、vlmul and ratio.
8835 2023-06-24 Roger Sayle <roger@nextmovesoftware.com>
8837 * config/i386/i386.md (*<code>qi_ext<mode>_3): New define_insn.
8839 2023-06-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
8841 * config/riscv/autovec.md (*fma<mode>): set clobber to Pmode in expand stage.
8842 (*fma<VI:mode><P:mode>): Ditto.
8843 (*fnma<mode>): Ditto.
8844 (*fnma<VI:mode><P:mode>): Ditto.
8846 2023-06-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
8848 * config/riscv/autovec.md (fma<mode>4): New pattern.
8849 (*fma<mode>): Ditto.
8850 (fnma<mode>4): Ditto.
8851 (*fnma<mode>): Ditto.
8852 (fms<mode>4): Ditto.
8853 (*fms<mode>): Ditto.
8854 (fnms<mode>4): Ditto.
8855 (*fnms<mode>): Ditto.
8856 * config/riscv/riscv-protos.h (emit_vlmax_fp_ternary_insn):
8858 * config/riscv/riscv-v.cc (emit_vlmax_fp_ternary_insn): Ditto.
8859 * config/riscv/vector.md: Fix attribute bug.
8861 2023-06-24 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
8863 * tree-ssa-loop-ivopts.cc (get_mem_type_for_internal_fn):
8864 Apply LEN_MASK_{LOAD,STORE}.
8866 2023-06-24 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
8868 * tree-ssa-loop-ivopts.cc (get_alias_ptr_type_for_ptr_address):
8869 Add LEN_MASK_{LOAD,STORE}.
8871 2023-06-24 David Malcolm <dmalcolm@redhat.com>
8873 * diagnostic-format-sarif.cc: Add #define INCLUDE_VECTOR.
8874 * diagnostic.cc: Likewise.
8875 * text-art/box-drawing.cc: Likewise.
8876 * text-art/canvas.cc: Likewise.
8877 * text-art/ruler.cc: Likewise.
8878 * text-art/selftests.cc: Likewise.
8879 * text-art/selftests.h (text_art::canvas): New forward decl.
8880 * text-art/style.cc: Add #define INCLUDE_VECTOR.
8881 * text-art/styled-string.cc: Likewise.
8882 * text-art/table.cc: Likewise.
8883 * text-art/table.h: Remove #include <vector>.
8884 * text-art/theme.cc: Add #define INCLUDE_VECTOR.
8885 * text-art/types.h: Check that INCLUDE_VECTOR is defined.
8886 Remove #include of <vector> and <string>.
8887 * text-art/widget.cc: Add #define INCLUDE_VECTOR.
8888 * text-art/widget.h: Remove #include <vector>.
8890 2023-06-24 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
8892 * internal-fn.cc (expand_partial_store_optab_fn): Adapt for LEN_MASK_STORE.
8893 (internal_load_fn_p): Add LEN_MASK_LOAD.
8894 (internal_store_fn_p): Add LEN_MASK_STORE.
8895 (internal_fn_mask_index): Add LEN_MASK_{LOAD,STORE}.
8896 (internal_fn_stored_value_index): Add LEN_MASK_STORE.
8897 (internal_len_load_store_bias): Add LEN_MASK_{LOAD,STORE}.
8898 * optabs-tree.cc (can_vec_mask_load_store_p): Adapt for LEN_MASK_{LOAD,STORE}.
8899 (get_len_load_store_mode): Ditto.
8900 * optabs-tree.h (can_vec_mask_load_store_p): Ditto.
8901 (get_len_load_store_mode): Ditto.
8902 * tree-vect-stmts.cc (check_load_store_for_partial_vectors): Ditto.
8903 (get_all_ones_mask): New function.
8904 (vectorizable_store): Apply LEN_MASK_{LOAD,STORE} into vectorizer.
8905 (vectorizable_load): Ditto.
8907 2023-06-23 Marek Polacek <polacek@redhat.com>
8909 * doc/cpp.texi (__cplusplus): Document value for -std=c++26 and
8910 -std=gnu++26. Document that for C++23, its value is 202302L.
8911 * doc/invoke.texi: Document -std=c++26 and -std=gnu++26.
8912 * dwarf2out.cc (highest_c_language): Handle GNU C++26.
8913 (gen_compile_unit_die): Likewise.
8915 2023-06-23 Jan Hubicka <jh@suse.cz>
8917 * tree-ssa-phiprop.cc (propagate_with_phi): Compute post dominators on
8919 (pass_phiprop::execute): Do not compute it here; return
8920 update_ssa_only_virtuals if something changed.
8921 (pass_data_phiprop): Remove TODO_update_ssa from todos.
8923 2023-06-23 Michael Meissner <meissner@linux.ibm.com>
8924 Aaron Sawdey <acsawdey@linux.ibm.com>
8927 * config/rs6000/genfusion.pl (gen_ld_cmpi_p10_one): Fix problems that
8928 allowed prefixed lwa to be generated.
8929 * config/rs6000/fusion.md: Regenerate.
8930 * config/rs6000/predicates.md (ds_form_mem_operand): Delete.
8931 * config/rs6000/rs6000.md (prefixed attribute): Add support for load
8932 plus compare immediate fused insns.
8933 (maybe_prefixed): Likewise.
8935 2023-06-23 Roger Sayle <roger@nextmovesoftware.com>
8937 * simplify-rtx.cc (simplify_subreg): Optimize lowpart SUBREGs
8938 of ASHIFT to const0_rtx with sufficiently large shift count.
8939 Optimize highpart SUBREGs of ASHIFT as the shift operand when
8940 the shift count is the correct offset. Optimize SUBREGs of
8941 multi-word logic operations if the SUBREGs of both operands
8944 2023-06-23 Richard Biener <rguenther@suse.de>
8946 * varasm.cc (initializer_constant_valid_p_1): Only
8947 allow conversions between scalar floating point types.
8949 2023-06-23 Richard Biener <rguenther@suse.de>
8951 * tree-vect-stmts.cc (vectorizable_assignment):
8952 Properly handle non-integral operands when analyzing
8955 2023-06-23 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
8957 PR tree-optimization/110280
8958 * match.pd (vec_perm_expr(v, v, mask) -> v): Explicitly build vector
8959 using build_vector_from_val with the element of input operand, and
8960 mask's type if operand and mask's types don't match.
8962 2023-06-23 Richard Biener <rguenther@suse.de>
8964 * fold-const.cc (tree_simple_nonnegative_warnv_p): Guard
8965 the truth_value_p case with !VECTOR_TYPE_P.
8967 2023-06-23 Richard Biener <rguenther@suse.de>
8969 * tree-vect-patterns.cc (vect_look_through_possible_promotion):
8970 Exit early when the type isn't scalar integral.
8972 2023-06-23 Richard Biener <rguenther@suse.de>
8974 * match.pd ((outertype)((innertype0)a+(innertype1)b)
8975 -> ((newtype)a+(newtype)b)): Use element_precision
8978 2023-06-23 Richard Biener <rguenther@suse.de>
8980 * fold-const.cc (fold_binary_loc): Use element_precision
8981 when trying (double)float1 CMP (double)float2 to
8982 float1 CMP float2 simplification.
8983 * match.pd: Likewise.
8985 2023-06-23 Richard Biener <rguenther@suse.de>
8987 * tree-vect-stmts.cc (vectorizable_load): Avoid useless
8988 copies of VMAT_INVARIANT vectorized stmts, fix SLP support.
8990 2023-06-23 Richard Biener <rguenther@suse.de>
8992 * tree-vect-stmts.cc (vector_vector_composition_type):
8993 Handle composition of a vector from a number of elements that
8994 happens to match its number of lanes.
8996 2023-06-22 Marek Polacek <polacek@redhat.com>
8998 * configure.ac (--enable-host-bind-now): New check. Add
8999 -Wl,-z,now to LD_PICFLAG if --enable-host-bind-now.
9000 * configure: Regenerate.
9001 * doc/install.texi: Document --enable-host-bind-now.
9003 2023-06-22 Di Zhao OS <dizhao@os.amperecomputing.com>
9005 * config/aarch64/aarch64.cc: Change fma_reassoc_width for ampere1.
9007 2023-06-22 Richard Biener <rguenther@suse.de>
9009 PR tree-optimization/110332
9010 * tree-ssa-phiprop.cc (propagate_with_phi): Always
9011 check aliasing with edge inserted loads.
9013 2023-06-22 Roger Sayle <roger@nextmovesoftware.com>
9014 Uros Bizjak <ubizjak@gmail.com>
9016 * config/i386/i386-expand.cc (ix86_expand_sse_ptest): Recognize
9017 expansion of ptestc with equal operands as producing const1_rtx.
9018 * config/i386/i386.cc (ix86_rtx_costs): Provide accurate cost
9019 estimates of UNSPEC_PTEST, where the ptest performs the PAND
9020 or PAND of its operands.
9021 * config/i386/sse.md (define_split): Transform CCCmode UNSPEC_PTEST
9022 of reg_equal_p operands into an x86_stc instruction.
9023 (define_split): Split pandn/ptestz/set{n?}e into ptestc/set{n?}c.
9024 (define_split): Similar to above for strict_low_part destinations.
9025 (define_split): Split pandn/ptestz/j{n?}e into ptestc/j{n?}c.
9027 2023-06-22 David Malcolm <dmalcolm@redhat.com>
9030 * Makefile.in (ANALYZER_OBJS): Add analyzer/access-diagram.o.
9031 * doc/invoke.texi (Wanalyzer-out-of-bounds): Add description of
9033 (fanalyzer-debug-text-art): New.
9035 2023-06-22 David Malcolm <dmalcolm@redhat.com>
9037 * Makefile.in (OBJS-libcommon): Add text-art/box-drawing.o,
9038 text-art/canvas.o, text-art/ruler.o, text-art/selftests.o,
9039 text-art/style.o, text-art/styled-string.o, text-art/table.o,
9040 text-art/theme.o, and text-art/widget.o.
9041 * color-macros.h (COLOR_FG_BRIGHT_BLACK): New.
9042 (COLOR_FG_BRIGHT_RED): New.
9043 (COLOR_FG_BRIGHT_GREEN): New.
9044 (COLOR_FG_BRIGHT_YELLOW): New.
9045 (COLOR_FG_BRIGHT_BLUE): New.
9046 (COLOR_FG_BRIGHT_MAGENTA): New.
9047 (COLOR_FG_BRIGHT_CYAN): New.
9048 (COLOR_FG_BRIGHT_WHITE): New.
9049 (COLOR_BG_BRIGHT_BLACK): New.
9050 (COLOR_BG_BRIGHT_RED): New.
9051 (COLOR_BG_BRIGHT_GREEN): New.
9052 (COLOR_BG_BRIGHT_YELLOW): New.
9053 (COLOR_BG_BRIGHT_BLUE): New.
9054 (COLOR_BG_BRIGHT_MAGENTA): New.
9055 (COLOR_BG_BRIGHT_CYAN): New.
9056 (COLOR_BG_BRIGHT_WHITE): New.
9057 * common.opt (fdiagnostics-text-art-charset=): New option.
9058 (diagnostic-text-art.h): New SourceInclude.
9059 (diagnostic_text_art_charset) New Enum and EnumValues.
9060 * configure: Regenerate.
9061 * configure.ac (gccdepdir): Add text-art to loop.
9062 * diagnostic-diagram.h: New file.
9063 * diagnostic-format-json.cc (json_emit_diagram): New.
9064 (diagnostic_output_format_init_json): Wire it up to
9065 context->m_diagrams.m_emission_cb.
9066 * diagnostic-format-sarif.cc: Include "diagnostic-diagram.h" and
9067 "text-art/canvas.h".
9068 (sarif_result::on_nested_diagnostic): Move code to...
9069 (sarif_result::add_related_location): ...this new function.
9070 (sarif_result::on_diagram): New.
9071 (sarif_builder::emit_diagram): New.
9072 (sarif_builder::make_message_object_for_diagram): New.
9073 (sarif_emit_diagram): New.
9074 (diagnostic_output_format_init_sarif): Set
9075 context->m_diagrams.m_emission_cb to sarif_emit_diagram.
9076 * diagnostic-text-art.h: New file.
9077 * diagnostic.cc: Include "diagnostic-text-art.h",
9078 "diagnostic-diagram.h", and "text-art/theme.h".
9079 (diagnostic_initialize): Initialize context->m_diagrams and
9080 call diagnostics_text_art_charset_init.
9081 (diagnostic_finish): Clean up context->m_diagrams.m_theme.
9082 (diagnostic_emit_diagram): New.
9083 (diagnostics_text_art_charset_init): New.
9084 * diagnostic.h (text_art::theme): New forward decl.
9085 (class diagnostic_diagram): Likewise.
9086 (diagnostic_context::m_diagrams): New field.
9087 (diagnostic_emit_diagram): New decl.
9088 * doc/invoke.texi (Diagnostic Message Formatting Options): Add
9089 -fdiagnostics-text-art-charset=.
9090 (-fdiagnostics-plain-output): Add
9091 -fdiagnostics-text-art-charset=none.
9092 * gcc.cc: Include "diagnostic-text-art.h".
9093 (driver_handle_option): Handle OPT_fdiagnostics_text_art_charset_.
9094 * opts-common.cc (decode_cmdline_options_to_array): Add
9095 "-fdiagnostics-text-art-charset=none" to expanded_args for
9096 -fdiagnostics-plain-output.
9097 * opts.cc: Include "diagnostic-text-art.h".
9098 (common_handle_option): Handle OPT_fdiagnostics_text_art_charset_.
9099 * pretty-print.cc (pp_unicode_character): New.
9100 * pretty-print.h (pp_unicode_character): New decl.
9101 * selftest-run-tests.cc: Include "text-art/selftests.h".
9102 (selftest::run_tests): Call text_art_tests.
9103 * text-art/box-drawing-chars.inc: New file, generated by
9104 contrib/unicode/gen-box-drawing-chars.py.
9105 * text-art/box-drawing.cc: New file.
9106 * text-art/box-drawing.h: New file.
9107 * text-art/canvas.cc: New file.
9108 * text-art/canvas.h: New file.
9109 * text-art/ruler.cc: New file.
9110 * text-art/ruler.h: New file.
9111 * text-art/selftests.cc: New file.
9112 * text-art/selftests.h: New file.
9113 * text-art/style.cc: New file.
9114 * text-art/styled-string.cc: New file.
9115 * text-art/table.cc: New file.
9116 * text-art/table.h: New file.
9117 * text-art/theme.cc: New file.
9118 * text-art/theme.h: New file.
9119 * text-art/types.h: New file.
9120 * text-art/widget.cc: New file.
9121 * text-art/widget.h: New file.
9123 2023-06-21 Uros Bizjak <ubizjak@gmail.com>
9125 * function.h (emit_initial_value_sets):
9126 Change return type from int to void.
9127 (aggregate_value_p): Change return type from int to bool.
9128 (prologue_contains): Ditto.
9129 (epilogue_contains): Ditto.
9130 (prologue_epilogue_contains): Ditto.
9131 * function.cc (temp_slot): Make "in_use" variable bool.
9132 (make_slot_available): Update for changed "in_use" variable.
9133 (assign_stack_temp_for_type): Ditto.
9134 (emit_initial_value_sets): Change return type from int to void
9135 and update function body accordingly.
9136 (instantiate_virtual_regs): Ditto.
9137 (rest_of_handle_thread_prologue_and_epilogue): Ditto.
9138 (safe_insn_predicate): Change return type from int to bool.
9139 (aggregate_value_p): Change return type from int to bool
9140 and update function body accordingly.
9141 (prologue_contains): Change return type from int to bool.
9142 (prologue_epilogue_contains): Ditto.
9144 2023-06-21 Alexander Monakov <amonakov@ispras.ru>
9146 * common.opt (fp_contract_mode) [on]: Remove fallback.
9147 * config/sh/sh.md (*fmasf4): Correct flag_fp_contract_mode test.
9148 * doc/invoke.texi (-ffp-contract): Update.
9149 * trans-mem.cc (diagnose_tm_1): Skip internal function calls.
9151 2023-06-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9153 * config/aarch64/aarch64-sve.md (mask_gather_load<mode><v_int_container>):
9154 Add alternatives to prefer to avoid same input and output Z register.
9155 (mask_gather_load<mode><v_int_container>): Likewise.
9156 (*mask_gather_load<mode><v_int_container>_<su>xtw_unpacked): Likewise.
9157 (*mask_gather_load<mode><v_int_container>_sxtw): Likewise.
9158 (*mask_gather_load<mode><v_int_container>_uxtw): Likewise.
9159 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_4HSI:mode><SVE_4BHI:mode>):
9161 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode><SVE_2BHSI:mode>):
9163 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
9164 <SVE_2BHSI:mode>_<ANY_EXTEND2:su>xtw_unpacked): Likewise.
9165 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
9166 <SVE_2BHSI:mode>_sxtw): Likewise.
9167 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
9168 <SVE_2BHSI:mode>_uxtw): Likewise.
9169 (@aarch64_ldff1_gather<mode>): Likewise.
9170 (@aarch64_ldff1_gather<mode>): Likewise.
9171 (*aarch64_ldff1_gather<mode>_sxtw): Likewise.
9172 (*aarch64_ldff1_gather<mode>_uxtw): Likewise.
9173 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx4_WIDE:mode>
9174 <VNx4_NARROW:mode>): Likewise.
9175 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
9176 <VNx2_NARROW:mode>): Likewise.
9177 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
9178 <VNx2_NARROW:mode>_sxtw): Likewise.
9179 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
9180 <VNx2_NARROW:mode>_uxtw): Likewise.
9181 * config/aarch64/aarch64-sve2.md (@aarch64_gather_ldnt<mode>): Likewise.
9182 (@aarch64_gather_ldnt_<ANY_EXTEND:optab><SVE_FULL_SDI:mode>
9183 <SVE_PARTIAL_I:mode>): Likewise.
9185 2023-06-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9187 * config/aarch64/aarch64-sve.md (mask_gather_load<mode><v_int_container>):
9188 Convert to compact alternatives syntax.
9189 (mask_gather_load<mode><v_int_container>): Likewise.
9190 (*mask_gather_load<mode><v_int_container>_<su>xtw_unpacked): Likewise.
9191 (*mask_gather_load<mode><v_int_container>_sxtw): Likewise.
9192 (*mask_gather_load<mode><v_int_container>_uxtw): Likewise.
9193 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_4HSI:mode><SVE_4BHI:mode>):
9195 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode><SVE_2BHSI:mode>):
9197 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
9198 <SVE_2BHSI:mode>_<ANY_EXTEND2:su>xtw_unpacked): Likewise.
9199 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
9200 <SVE_2BHSI:mode>_sxtw): Likewise.
9201 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
9202 <SVE_2BHSI:mode>_uxtw): Likewise.
9203 (@aarch64_ldff1_gather<mode>): Likewise.
9204 (@aarch64_ldff1_gather<mode>): Likewise.
9205 (*aarch64_ldff1_gather<mode>_sxtw): Likewise.
9206 (*aarch64_ldff1_gather<mode>_uxtw): Likewise.
9207 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx4_WIDE:mode>
9208 <VNx4_NARROW:mode>): Likewise.
9209 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
9210 <VNx2_NARROW:mode>): Likewise.
9211 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
9212 <VNx2_NARROW:mode>_sxtw): Likewise.
9213 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
9214 <VNx2_NARROW:mode>_uxtw): Likewise.
9215 * config/aarch64/aarch64-sve2.md (@aarch64_gather_ldnt<mode>): Likewise.
9216 (@aarch64_gather_ldnt_<ANY_EXTEND:optab><SVE_FULL_SDI:mode>
9217 <SVE_PARTIAL_I:mode>): Likewise.
9219 2023-06-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9222 2023-06-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9224 * config/aarch64/aarch64-sve.md (mask_gather_load<mode><v_int_container>):
9225 Convert to compact alternatives syntax.
9226 (mask_gather_load<mode><v_int_container>): Likewise.
9227 (*mask_gather_load<mode><v_int_container>_<su>xtw_unpacked): Likewise.
9228 (*mask_gather_load<mode><v_int_container>_sxtw): Likewise.
9229 (*mask_gather_load<mode><v_int_container>_uxtw): Likewise.
9230 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_4HSI:mode><SVE_4BHI:mode>):
9232 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode><SVE_2BHSI:mode>):
9234 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
9235 <SVE_2BHSI:mode>_<ANY_EXTEND2:su>xtw_unpacked): Likewise.
9236 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
9237 <SVE_2BHSI:mode>_sxtw): Likewise.
9238 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
9239 <SVE_2BHSI:mode>_uxtw): Likewise.
9240 (@aarch64_ldff1_gather<mode>): Likewise.
9241 (@aarch64_ldff1_gather<mode>): Likewise.
9242 (*aarch64_ldff1_gather<mode>_sxtw): Likewise.
9243 (*aarch64_ldff1_gather<mode>_uxtw): Likewise.
9244 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx4_WIDE:mode>
9245 <VNx4_NARROW:mode>): Likewise.
9246 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
9247 <VNx2_NARROW:mode>): Likewise.
9248 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
9249 <VNx2_NARROW:mode>_sxtw): Likewise.
9250 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
9251 <VNx2_NARROW:mode>_uxtw): Likewise.
9252 * config/aarch64/aarch64-sve2.md (@aarch64_gather_ldnt<mode>): Likewise.
9253 (@aarch64_gather_ldnt_<ANY_EXTEND:optab><SVE_FULL_SDI:mode>
9254 <SVE_PARTIAL_I:mode>): Likewise.
9256 2023-06-21 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
9258 * optabs-query.cc (can_vec_mask_load_store_p): Move to optabs-tree.cc.
9259 (get_len_load_store_mode): Ditto.
9260 * optabs-query.h (can_vec_mask_load_store_p): Move to optabs-tree.h.
9261 (get_len_load_store_mode): Ditto.
9262 * optabs-tree.cc (can_vec_mask_load_store_p): New function.
9263 (get_len_load_store_mode): Ditto.
9264 * optabs-tree.h (can_vec_mask_load_store_p): Ditto.
9265 (get_len_load_store_mode): Ditto.
9266 * tree-if-conv.cc: include optabs-tree instead of optabs-query
9268 2023-06-21 Richard Biener <rguenther@suse.de>
9270 * tree-ssa-loop-ivopts.cc (add_iv_candidate_for_use): Use
9271 split_constant_offset for the POINTER_PLUS_EXPR case.
9273 2023-06-21 Richard Biener <rguenther@suse.de>
9275 * tree-ssa-loop-ivopts.cc (record_group_use): Use
9276 split_constant_offset.
9278 2023-06-21 Richard Biener <rguenther@suse.de>
9280 * tree-loop-distribution.cc (classify_builtin_st): Use
9281 split_constant_offset.
9282 * tree-ssa-loop-ivopts.h (strip_offset): Remove.
9283 * tree-ssa-loop-ivopts.cc (strip_offset): Make static.
9285 2023-06-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9287 * config/aarch64/aarch64-sve.md (mask_gather_load<mode><v_int_container>):
9288 Convert to compact alternatives syntax.
9289 (mask_gather_load<mode><v_int_container>): Likewise.
9290 (*mask_gather_load<mode><v_int_container>_<su>xtw_unpacked): Likewise.
9291 (*mask_gather_load<mode><v_int_container>_sxtw): Likewise.
9292 (*mask_gather_load<mode><v_int_container>_uxtw): Likewise.
9293 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_4HSI:mode><SVE_4BHI:mode>):
9295 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode><SVE_2BHSI:mode>):
9297 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
9298 <SVE_2BHSI:mode>_<ANY_EXTEND2:su>xtw_unpacked): Likewise.
9299 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
9300 <SVE_2BHSI:mode>_sxtw): Likewise.
9301 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
9302 <SVE_2BHSI:mode>_uxtw): Likewise.
9303 (@aarch64_ldff1_gather<mode>): Likewise.
9304 (@aarch64_ldff1_gather<mode>): Likewise.
9305 (*aarch64_ldff1_gather<mode>_sxtw): Likewise.
9306 (*aarch64_ldff1_gather<mode>_uxtw): Likewise.
9307 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx4_WIDE:mode>
9308 <VNx4_NARROW:mode>): Likewise.
9309 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
9310 <VNx2_NARROW:mode>): Likewise.
9311 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
9312 <VNx2_NARROW:mode>_sxtw): Likewise.
9313 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
9314 <VNx2_NARROW:mode>_uxtw): Likewise.
9315 * config/aarch64/aarch64-sve2.md (@aarch64_gather_ldnt<mode>): Likewise.
9316 (@aarch64_gather_ldnt_<ANY_EXTEND:optab><SVE_FULL_SDI:mode>
9317 <SVE_PARTIAL_I:mode>): Likewise.
9319 2023-06-21 Tamar Christina <tamar.christina@arm.com>
9322 * doc/md.texi: Replace backslashchar.
9324 2023-06-21 Richard Biener <rguenther@suse.de>
9326 * config/i386/i386.cc (ix86_vector_costs::finish_cost):
9327 Overload. For masked main loops make sure the vectorization
9328 factor isn't more than double the number of iterations.
9330 2023-06-21 Jan Beulich <jbeulich@suse.com>
9332 * config/i386/i386-expand.cc (ix86_expand_copysign): Request
9333 value duplication by ix86_build_signbit_mask() when AVX512F and
9335 * config/i386/sse.md (*<avx512>_vternlog<mode>_all): Convert to
9336 2-alternative form. Adjust "mode" attribute. Add "enabled"
9338 (*<avx512>_vpternlog<mode>_1): Also permit when TARGET_AVX512F
9339 && !TARGET_PREFER_AVX256.
9340 (*<avx512>_vpternlog<mode>_2): Likewise.
9341 (*<avx512>_vpternlog<mode>_3): Likewise.
9343 2023-06-21 liuhongt <hongtao.liu@intel.com>
9346 * tree-vect-stmts.cc (vectorizable_conversion): Use
9347 intermiediate integer type for float_expr/fix_trunc_expr when
9348 direct optab is not existed.
9350 2023-06-20 Tamar Christina <tamar.christina@arm.com>
9353 * gensupport.cc (convert_syntax): Explicitly check for RTX code.
9355 2023-06-20 Richard Sandiford <richard.sandiford@arm.com>
9357 * config/aarch64/aarch64.md (stack_tie): Hard-code the first
9358 register operand to the stack pointer. Require the second register
9359 operand to have the number specified in a separate const_int operand.
9360 * config/aarch64/aarch64.cc (aarch64_emit_stack_tie): New function.
9361 (aarch64_allocate_and_probe_stack_space): Use it.
9362 (aarch64_expand_prologue, aarch64_expand_epilogue): Likewise.
9363 (aarch64_expand_epilogue): Likewise.
9365 2023-06-20 Jakub Jelinek <jakub@redhat.com>
9368 * tree-ssa-math-opts.cc (match_uaddc_usubc): Remember lhs of
9369 IMAGPART_EXPR of arg2/arg3 and use that as arg3 if it has the right
9372 2023-06-20 Uros Bizjak <ubizjak@gmail.com>
9374 * calls.h (setjmp_call_p): Change return type from int to bool.
9375 * calls.cc (struct arg_data): Change "pass_on_stack" to bool.
9376 (store_one_arg): Change return type from int to bool
9377 and adjust function body accordingly. Change "sibcall_failure"
9379 (finalize_must_preallocate): Ditto. Change *must_preallocate pointer
9380 argument to bool. Change "partial_seen" variable to bool.
9381 (load_register_parameters): Change *sibcall_failure
9382 pointer argument to bool.
9383 (check_sibcall_argument_overlap_1): Change return type from int to bool
9384 and adjust function body accordingly.
9385 (check_sibcall_argument_overlap): Ditto. Change
9386 "mark_stored_args_map" argument to bool.
9387 (emit_call_1): Change "already_popped" variable to bool.
9388 (setjmp_call_p): Change return type from int to bool
9389 and adjust function body accordingly.
9390 (initialize_argument_information): Change *must_preallocate
9391 pointer argument to bool.
9392 (expand_call): Change "pcc_struct_value", "must_preallocate"
9393 and "sibcall_failure" variables to bool.
9394 (emit_library_call_value_1): Change "pcc_struct_value"
9397 2023-06-20 Martin Jambor <mjambor@suse.cz>
9400 * ipa-sra.cc (struct caller_issues): New field there_is_one.
9401 (check_for_caller_issues): Set it.
9402 (check_all_callers_for_issues): Check it.
9404 2023-06-20 Martin Jambor <mjambor@suse.cz>
9406 * ipa-prop.h (ipa_uid_to_idx_map_elt): New type.
9407 (struct ipcp_transformation): Rearrange members according to
9408 C++ class coding convention, add m_uid_to_idx,
9409 get_param_index and maybe_create_parm_idx_map.
9410 * ipa-cp.cc (ipcp_transformation::get_param_index): New function.
9411 (compare_uids): Likewise.
9412 (ipcp_transformation::maype_create_parm_idx_map): Likewise.
9413 * ipa-prop.cc (ipcp_get_parm_bits): Use get_param_index.
9414 (ipcp_update_bits): Accept TS as a parameter, assume it is not NULL.
9415 (ipcp_update_vr): Likewise.
9416 (ipcp_transform_function): Call, maybe_create_parm_idx_map of TS, bail
9417 out quickly if empty, pass it to ipcp_update_bits and ipcp_update_vr.
9419 2023-06-20 Carl Love <cel@us.ibm.com>
9421 * config/rs6000/rs6000-builtin.cc (rs6000_expand_builtin):
9422 Rename CODE_FOR_xsxsigqp_tf to CODE_FOR_xsxsigqp_tf_ti.
9423 Rename CODE_FOR_xsxsigqp_kf to CODE_FOR_xsxsigqp_kf_ti.
9424 Rename CCDE_FOR_xsxexpqp_tf to CODE_FOR_xsxexpqp_tf_di.
9425 Rename CODE_FOR_xsxexpqp_kf to CODE_FOR_xsxexpqp_kf_di.
9426 (CODE_FOR_xsxexpqp_kf_v2di, CODE_FOR_xsxsigqp_kf_v1ti,
9427 CODE_FOR_xsiexpqp_kf_v2di): Add case statements.
9428 * config/rs6000/rs6000-builtins.def
9429 (__builtin_vsx_scalar_extract_exp_to_vec,
9430 __builtin_vsx_scalar_extract_sig_to_vec,
9431 __builtin_vsx_scalar_insert_exp_vqp): Add new builtin definitions.
9432 Rename xsxexpqp_kf, xsxsigqp_kf, xsiexpqp_kf to xsexpqp_kf_di,
9433 xsxsigqp_kf_ti, xsiexpqp_kf_di respectively.
9434 * config/rs6000/rs6000-c.cc (altivec_resolve_overloaded_builtin):
9435 Update case RS6000_OVLD_VEC_VSIE to handle MODE_VECTOR_INT for new
9436 overloaded instance. Update comments.
9437 * config/rs6000/rs6000-overload.def
9438 (__builtin_vec_scalar_insert_exp): Add new overload definition with
9440 (scalar_extract_exp_to_vec, scalar_extract_sig_to_vec): New
9441 overloaded definitions.
9442 * config/rs6000/vsx.md (V2DI_DI): New mode iterator.
9443 (DI_to_TI): New mode attribute.
9444 Rename xsxexpqp_<mode> to sxexpqp_<IEEE128:mode>_<V2DI_DI:mode>.
9445 Rename xsxsigqp_<mode> to xsxsigqp_<IEEE128:mode>_<VEC_TI:mode>.
9446 Rename xsiexpqp_<mode> to xsiexpqp_<IEEE128:mode>_<V2DI_DI:mode>.
9447 * doc/extend.texi (scalar_extract_exp_to_vec,
9448 scalar_extract_sig_to_vec): Add documentation for new builtins.
9449 (scalar_insert_exp): Add new overloaded builtin definition.
9451 2023-06-20 Li Xu <xuli1@eswincomputing.com>
9453 * config/riscv/riscv.cc (riscv_regmode_natural_size): set the natural
9454 size of vector mask mode to one rvv register.
9456 2023-06-20 Juzhe-Zhong <juzhe.zhong@rivai.ai>
9458 * config/riscv/riscv-v.cc (expand_const_vector): Optimize codegen.
9460 2023-06-20 Lehua Ding <lehua.ding@rivai.ai>
9462 * config/riscv/riscv.cc (riscv_arg_has_vector): Add default
9465 2023-06-20 Richard Biener <rguenther@suse.de>
9467 * tree-ssa-dse.cc (dse_classify_store): When we found
9468 no defs and the basic-block with the original definition
9469 ends in __builtin_unreachable[_trap] the store is dead.
9471 2023-06-20 Richard Biener <rguenther@suse.de>
9473 * tree-ssa-phiprop.cc (phiprop_insert_phi): For simple loads
9474 keep the virtual SSA form up-to-date.
9476 2023-06-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9478 * config/aarch64/aarch64-simd.md (*aarch64_addp_same_reg<mode>):
9479 New define_insn_and_split.
9481 2023-06-20 Tamar Christina <tamar.christina@arm.com>
9483 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Drop test comment.
9485 2023-06-20 Jan Beulich <jbeulich@suse.com>
9487 * config/i386/sse.md (vec_dupv2di): Correct %vmovddup input
9488 constraint. Add new AVX512F alternative.
9490 2023-06-20 Richard Biener <rguenther@suse.de>
9493 * dwarf2out.cc (process_scope_var): Continue processing
9494 the decl after setting a parent in case the existing DIE
9497 2023-06-20 Lehua Ding <lehua.ding@rivai.ai>
9499 * config/riscv/riscv.cc (riscv_scalable_vector_type_p): Delete.
9500 (riscv_arg_has_vector): Simplify.
9501 (riscv_pass_in_vector_p): Adjust warning message.
9503 2023-06-19 Jin Ma <jinma@linux.alibaba.com>
9505 * config/riscv/riscv.cc (riscv_compute_frame_info): Allocate frame for FCSR.
9506 (riscv_for_each_saved_reg): Save and restore FCSR in interrupt functions.
9507 * config/riscv/riscv.md (riscv_frcsr): New patterns.
9508 (riscv_fscsr): Likewise.
9510 2023-06-19 Toru Kisuki <tkisuki@tachyum.com>
9512 PR rtl-optimization/110305
9513 * simplify-rtx.cc (simplify_context::simplify_binary_operation_1):
9514 Handle HONOR_SNANS for x + 0.0.
9516 2023-06-19 Jan Hubicka <jh@suse.cz>
9518 PR tree-optimization/109811
9519 PR tree-optimization/109849
9520 * passes.def: Add phiprop to early optimization passes.
9521 * tree-ssa-phiprop.cc: Allow clonning.
9523 2023-06-19 Tamar Christina <tamar.christina@arm.com>
9525 * config/aarch64/aarch64.md (arches): Add nosimd.
9526 (*mov<mode>_aarch64, *movsi_aarch64, *movdi_aarch64): Rewrite to
9529 2023-06-19 Tamar Christina <tamar.christina@arm.com>
9530 Omar Tahir <Omar.Tahir2@arm.com>
9532 * gensupport.cc (class conlist, add_constraints, add_attributes,
9533 skip_spaces, expect_char, preprocess_compact_syntax,
9534 parse_section_layout, parse_section, convert_syntax): New.
9535 (process_rtx): Check for conversion.
9536 * genoutput.cc (process_template): Check for unresolved iterators.
9537 (class data): Add compact_syntax_p.
9539 * gensupport.h (compact_syntax): New.
9540 (hash-set.h): Include.
9541 * doc/md.texi: Document it.
9543 2023-06-19 Uros Bizjak <ubizjak@gmail.com>
9545 * recog.h (check_asm_operands): Change return type from int to bool.
9546 (insn_invalid_p): Ditto.
9547 (verify_changes): Ditto.
9548 (apply_change_group): Ditto.
9549 (constrain_operands): Ditto.
9550 (constrain_operands_cached): Ditto.
9551 (validate_replace_rtx_subexp): Ditto.
9552 (validate_replace_rtx): Ditto.
9553 (validate_replace_rtx_part): Ditto.
9554 (validate_replace_rtx_part_nosimplify): Ditto.
9555 (added_clobbers_hard_reg_p): Ditto.
9556 (peep2_regno_dead_p): Ditto.
9557 (peep2_reg_dead_p): Ditto.
9558 (store_data_bypass_p): Ditto.
9559 (if_test_bypass_p): Ditto.
9560 * rtl.h (split_all_insns_noflow): Change
9561 return type from unsigned int to void.
9562 * genemit.cc (output_added_clobbers_hard_reg_p): Change return type
9563 of generated added_clobbers_hard_reg_p from int to bool and adjust
9564 function body accordingly. Change "used" variable type from
9566 * recog.cc (check_asm_operands): Change return type
9567 from int to bool and adjust function body accordingly.
9568 (insn_invalid_p): Ditto. Change "is_asm" variable to bool.
9569 (verify_changes): Change return type from int to bool.
9570 (apply_change_group): Change return type from int to bool
9571 and adjust function body accordingly.
9572 (validate_replace_rtx_subexp): Change return type from int to bool.
9573 (validate_replace_rtx): Ditto.
9574 (validate_replace_rtx_part): Ditto.
9575 (validate_replace_rtx_part_nosimplify): Ditto.
9576 (constrain_operands_cached): Ditto.
9577 (constrain_operands): Ditto. Change "lose" and "win"
9578 variables type from int to bool.
9579 (split_all_insns_noflow): Change return type from unsigned int
9580 to void and adjust function body accordingly.
9581 (peep2_regno_dead_p): Change return type from int to bool.
9582 (peep2_reg_dead_p): Ditto.
9583 (peep2_find_free_register): Change "success"
9584 variable type from int to bool
9585 (store_data_bypass_p_1): Change return type from int to bool.
9586 (store_data_bypass_p): Ditto.
9588 2023-06-19 Li Xu <xuli1@eswincomputing.com>
9590 * config/riscv/vector-iterators.md: zvfh/zvfhmin depends on the
9593 2023-06-19 Pan Li <pan2.li@intel.com>
9596 * config/riscv/riscv-vector-builtins-bases.cc: Adjust expand for
9598 * config/riscv/vector-iterators.md: Remove VWLMUL1, VWLMUL1_ZVE64,
9599 VWLMUL1_ZVE32, VI_ZVE64, VI_ZVE32, VWI, VWI_ZVE64, VWI_ZVE32,
9600 VF_ZVE63 and VF_ZVE32.
9601 * config/riscv/vector.md
9602 (@pred_widen_reduc_plus<v_su><mode><vwlmul1>): Removed.
9603 (@pred_widen_reduc_plus<v_su><mode><vwlmul1_zve64>): Ditto.
9604 (@pred_widen_reduc_plus<v_su><mode><vwlmul1_zve32>): Ditto.
9605 (@pred_widen_reduc_plus<order><mode><vwlmul1>): Ditto.
9606 (@pred_widen_reduc_plus<order><mode><vwlmul1_zve64>): Ditto.
9607 (@pred_widen_reduc_plus<v_su><VQI:mode><VHI_LMUL1:mode>): New pattern.
9608 (@pred_widen_reduc_plus<v_su><VHI:mode><VSI_LMUL1:mode>): Ditto.
9609 (@pred_widen_reduc_plus<v_su><VSI:mode><VDI_LMUL1:mode>): Ditto.
9610 (@pred_widen_reduc_plus<order><VHF:mode><VSF_LMUL1:mode>): Ditto.
9611 (@pred_widen_reduc_plus<order><VSF:mode><VDF_LMUL1:mode>): Ditto.
9613 2023-06-19 Pan Li <pan2.li@intel.com>
9616 * config/riscv/riscv-vector-builtins-bases.cc: Adjust expand for
9618 * config/riscv/vector-iterators.md: Add VHF, VSF, VDF,
9619 VHF_LMUL1, VSF_LMUL1, VDF_LMUL1, and remove unused attr.
9620 * config/riscv/vector.md (@pred_reduc_<reduc><mode><vlmul1>): Removed.
9621 (@pred_reduc_<reduc><mode><vlmul1_zve64>): Ditto.
9622 (@pred_reduc_<reduc><mode><vlmul1_zve32>): Ditto.
9623 (@pred_reduc_plus<order><mode><vlmul1>): Ditto.
9624 (@pred_reduc_plus<order><mode><vlmul1_zve32>): Ditto.
9625 (@pred_reduc_plus<order><mode><vlmul1_zve64>): Ditto.
9626 (@pred_reduc_<reduc><VHF:mode><VHF_LMUL1:mode>): New pattern.
9627 (@pred_reduc_<reduc><VSF:mode><VSF_LMUL1:mode>): Ditto.
9628 (@pred_reduc_<reduc><VDF:mode><VDF_LMUL1:mode>): Ditto.
9629 (@pred_reduc_plus<order><VHF:mode><VHF_LMUL1:mode>): Ditto.
9630 (@pred_reduc_plus<order><VSF:mode><VSF_LMUL1:mode>): Ditto.
9631 (@pred_reduc_plus<order><VDF:mode><VDF_LMUL1:mode>): Ditto.
9633 2023-06-19 Andrew Stubbs <ams@codesourcery.com>
9635 * config/gcn/gcn.cc (gcn_expand_divmod_libfunc): New function.
9636 (gcn_init_libfuncs): Add div and mod functions for all modes.
9637 Add placeholders for divmod functions.
9638 (TARGET_EXPAND_DIVMOD_LIBFUNC): Define.
9640 2023-06-19 Andrew Stubbs <ams@codesourcery.com>
9642 * tree-vect-generic.cc: Include optabs-libfuncs.h.
9643 (get_compute_type): Check optab_libfunc.
9644 * tree-vect-stmts.cc: Include optabs-libfuncs.h.
9645 (vectorizable_operation): Check optab_libfunc.
9647 2023-06-19 Andrew Stubbs <ams@codesourcery.com>
9649 * config/gcn/gcn-protos.h (vgpr_4reg_mode_p): New function.
9650 * config/gcn/gcn-valu.md (V_4REG, V_4REG_ALT): New iterators.
9651 (V_MOV, V_MOV_ALT): Likewise.
9652 (scalar_mode, SCALAR_MODE): Add TImode.
9653 (vnsi, VnSI, vndi, VnDI): Likewise.
9654 (vec_merge, vec_merge_with_clobber, vec_merge_with_vcc): Use V_MOV.
9655 (mov<mode>, mov<mode>_unspec): Use V_MOV.
9656 (*mov<mode>_4reg): New insn.
9657 (mov<mode>_exec): New 4reg variant.
9658 (mov<mode>_sgprbase): Likewise.
9659 (reload_in<mode>, reload_out<mode>): Use V_MOV.
9660 (vec_set<mode>): Likewise.
9661 (vec_duplicate<mode><exec>): New 4reg variant.
9662 (vec_extract<mode><scalar_mode>): Likewise.
9663 (vec_extract<V_ALL:mode><V_ALL_ALT:mode>): Rename to ...
9664 (vec_extract<V_MOV:mode><V_MOV_ALT:mode>): ... this, and use V_MOV.
9665 (vec_extract<V_4REG:mode><V_4REG_ALT:mode>_nop): New 4reg variant.
9666 (fold_extract_last_<mode>): Use V_MOV.
9667 (vec_init<V_ALL:mode><V_ALL_ALT:mode>): Rename to ...
9668 (vec_init<V_MOV:mode><V_MOV_ALT:mode>): ... this, and use V_MOV.
9669 (gather_load<mode><vnsi>, gather<mode>_expr<exec>,
9670 gather<mode>_insn_1offset<exec>, gather<mode>_insn_1offset_ds<exec>,
9671 gather<mode>_insn_2offsets<exec>): Use V_MOV.
9672 (scatter_store<mode><vnsi>, scatter<mode>_expr<exec_scatter>,
9673 scatter<mode>_insn_1offset<exec_scatter>,
9674 scatter<mode>_insn_1offset_ds<exec_scatter>,
9675 scatter<mode>_insn_2offsets<exec_scatter>): Likewise.
9676 (maskload<mode>di, maskstore<mode>di, mask_gather_load<mode><vnsi>,
9677 mask_scatter_store<mode><vnsi>): Likewise.
9678 * config/gcn/gcn.cc (gcn_class_max_nregs): Use vgpr_4reg_mode_p.
9679 (gcn_hard_regno_mode_ok): Likewise.
9680 (GEN_VNM): Add TImode support.
9681 (USE_TI): New macro. Separate TImode operations from non-TImode ones.
9682 (gcn_vector_mode_supported_p): Add V64TImode, V32TImode, V16TImode,
9683 V8TImode, and V2TImode.
9684 (print_operand): Add 'J' and 'K' print codes.
9686 2023-06-19 Richard Biener <rguenther@suse.de>
9688 PR tree-optimization/110298
9689 * tree-ssa-loop-ivcanon.cc (tree_unroll_loops_completely):
9690 Clear number of iterations info before cleaning up the CFG.
9692 2023-06-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9694 * simplify-rtx.cc (simplify_context::simplify_binary_operation_1):
9695 Simplify vec_concat of lowpart subreg and high part vec_select.
9697 2023-06-19 Tobias Burnus <tobias@codesourcery.com>
9699 * doc/invoke.texi (-foffload-options): Remove '-O3' from the examples.
9701 2023-06-19 Richard Sandiford <richard.sandiford@arm.com>
9703 * tree-vect-loop-manip.cc (vect_set_loop_condition_partial_vectors):
9704 Handle null niters_skip.
9706 2023-06-19 Richard Biener <rguenther@suse.de>
9708 * config/aarch64/aarch64.cc
9709 (aarch64_vector_costs::analyze_loop_vinfo): Fix reference
9710 to LOOP_VINFO_MASKS.
9712 2023-06-19 Senthil Kumar Selvaraj <saaadhu@gcc.gnu.org>
9715 * common/config/avr/avr-common.cc: Remove setting
9716 of OPT_fdelete_null_pointer_checks.
9717 * config/avr/avr.cc (avr_option_override): Clear
9718 flag_delete_null_pointer_checks if zero_address_valid.
9719 (avr_addr_space_zero_address_valid): New function.
9720 (TARGET_ADDR_SPACE_ZERO_ADDRESS_VALID): Provide target
9723 2023-06-19 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
9724 Robin Dapp <rdapp.gcc@gmail.com>
9726 * doc/md.texi: Add len_mask{load,store}.
9727 * genopinit.cc (main): Ditto.
9729 * internal-fn.cc (len_maskload_direct): Ditto.
9730 (len_maskstore_direct): Ditto.
9731 (expand_call_mem_ref): Ditto.
9732 (expand_partial_load_optab_fn): Ditto.
9733 (expand_len_maskload_optab_fn): Ditto.
9734 (expand_partial_store_optab_fn): Ditto.
9735 (expand_len_maskstore_optab_fn): Ditto.
9736 (direct_len_maskload_optab_supported_p): Ditto.
9737 (direct_len_maskstore_optab_supported_p): Ditto.
9738 * internal-fn.def (LEN_MASK_LOAD): Ditto.
9739 (LEN_MASK_STORE): Ditto.
9740 * optabs.def (OPTAB_CD): Ditto.
9742 2023-06-19 Robin Dapp <rdapp@ventanamicro.com>
9744 * config/riscv/autovec.md (<optab><mode>2): Add unop expanders.
9746 2023-06-19 Robin Dapp <rdapp@ventanamicro.com>
9748 * config/riscv/autovec.md (<optab><mode>3): Implement binop
9750 * config/riscv/riscv-protos.h (emit_vlmax_fp_insn): Declare.
9751 (enum vxrm_field_enum): Rename this...
9752 (enum fixed_point_rounding_mode): ...to this.
9753 (enum frm_field_enum): Rename this...
9754 (enum floating_point_rounding_mode): ...to this.
9755 * config/riscv/riscv-v.cc (emit_vlmax_fp_insn): New function
9756 * config/riscv/riscv.cc (riscv_const_insns): Clarify const
9758 (riscv_libgcc_floating_mode_supported_p): Adjust comment.
9759 (riscv_excess_precision): Do not convert to float for ZVFH.
9760 * config/riscv/vector-iterators.md: Add VF_AUTO iterator.
9762 2023-06-19 Robin Dapp <rdapp@ventanamicro.com>
9764 * config/riscv/vector-iterators.md: Add VI_QH iterator.
9765 * config/riscv/autovec-opt.md
9766 (@pred_extract_first_sextdi<mode>): New vmv.x.s pattern
9767 that includes sign extension.
9768 (@pred_extract_first_sextsi<mode>): Dito for SImode.
9770 2023-06-19 Robin Dapp <rdapp@ventanamicro.com>
9772 * config/riscv/autovec.md (vec_set<mode>): Implement.
9773 (vec_extract<mode><vel>): Implement.
9774 * config/riscv/riscv-protos.h (enum insn_type): Add slide insn.
9775 (emit_vlmax_slide_insn): Declare.
9776 (emit_nonvlmax_slide_tu_insn): Declare.
9777 (emit_scalar_move_insn): Export.
9778 (emit_nonvlmax_integer_move_insn): Export.
9779 * config/riscv/riscv-v.cc (emit_vlmax_slide_insn): New function.
9780 (emit_nonvlmax_slide_tu_insn): New function.
9781 (emit_vlmax_masked_mu_insn): No change.
9782 (emit_vlmax_integer_move_insn): Export.
9784 2023-06-19 Richard Biener <rguenther@suse.de>
9786 * tree-vectorizer.h (enum vect_partial_vector_style): New.
9787 (_loop_vec_info::partial_vector_style): Likewise.
9788 (LOOP_VINFO_PARTIAL_VECTORS_STYLE): Likewise.
9789 (rgroup_controls::compare_type): Add.
9790 (vec_loop_masks): Change from a typedef to auto_vec<>
9792 * tree-vect-loop-manip.cc (vect_set_loop_condition_partial_vectors):
9793 Adjust. Convert niters_skip to compare_type.
9794 (vect_set_loop_condition_partial_vectors_avx512): New function
9795 implementing the AVX512 partial vector codegen.
9796 (vect_set_loop_condition): Dispatch to the correct
9797 vect_set_loop_condition_partial_vectors_* function based on
9798 LOOP_VINFO_PARTIAL_VECTORS_STYLE.
9799 (vect_prepare_for_masked_peels): Compute LOOP_VINFO_MASK_SKIP_NITERS
9800 in the original niter type.
9801 * tree-vect-loop.cc (_loop_vec_info::_loop_vec_info): Initialize
9802 partial_vector_style.
9803 (can_produce_all_loop_masks_p): Adjust.
9804 (vect_verify_full_masking): Produce the rgroup_controls vector
9805 here. Set LOOP_VINFO_PARTIAL_VECTORS_STYLE on success.
9806 (vect_verify_full_masking_avx512): New function implementing
9807 verification of AVX512 style masking.
9808 (vect_verify_loop_lens): Set LOOP_VINFO_PARTIAL_VECTORS_STYLE.
9809 (vect_analyze_loop_2): Also try AVX512 style masking.
9811 (vect_estimate_min_profitable_iters): Implement AVX512 style
9812 mask producing cost.
9813 (vect_record_loop_mask): Do not build the rgroup_controls
9814 vector here but record masks in a hash-set.
9815 (vect_get_loop_mask): Implement AVX512 style mask query,
9816 complementing the existing while_ult style.
9818 2023-06-19 Richard Biener <rguenther@suse.de>
9820 * tree-vectorizer.h (vect_get_loop_mask): Add loop_vec_info
9822 * tree-vect-loop.cc (vect_get_loop_mask): Likewise.
9823 (vectorize_fold_left_reduction): Adjust.
9824 (vect_transform_reduction): Likewise.
9825 (vectorizable_live_operation): Likewise.
9826 * tree-vect-stmts.cc (vectorizable_call): Likewise.
9827 (vectorizable_operation): Likewise.
9828 (vectorizable_store): Likewise.
9829 (vectorizable_load): Likewise.
9830 (vectorizable_condition): Likewise.
9832 2023-06-19 Senthil Kumar Selvaraj <saaadhu@gcc.gnu.org>
9835 * config/avr/avr.opt (mgas-isr-prologues, mmain-is-OS_task):
9836 Add Optimization option property.
9838 2023-06-19 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
9840 * config/xtensa/xtensa.cc (xtensa_constantsynth_2insn):
9841 Add new pattern for the abovementioned case.
9843 2023-06-19 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
9845 * config/xtensa/xtensa.cc
9846 (TARGET_MEMORY_MOVE_COST, xtensa_memory_move_cost): Remove.
9848 2023-06-19 Jiufu Guo <guojiufu@linux.ibm.com>
9850 * config/rs6000/rs6000.cc (TARGET_CONST_ANCHOR): New define.
9852 2023-06-19 Jiufu Guo <guojiufu@linux.ibm.com>
9854 * cse.cc (try_const_anchors): Check SCALAR_INT_MODE.
9856 2023-06-19 liuhongt <hongtao.liu@intel.com>
9859 * config/i386/sse.md (<sse2_avx2>_packsswb<mask_name>):
9861 (sse2_packsswb<mask_name>): .. this, ..
9862 (avx2_packsswb<mask_name>): .. this and ..
9863 (avx512bw_packsswb<mask_name>): .. this.
9864 (<sse2_avx2>_packssdw<mask_name>): Substitute with ..
9865 (sse2_packssdw<mask_name>): .. this, ..
9866 (avx2_packssdw<mask_name>): .. this and ..
9867 (avx512bw_packssdw<mask_name>): .. this.
9869 2023-06-19 liuhongt <hongtao.liu@intel.com>
9872 * config/i386/i386-expand.cc (ix86_split_mmx_pack): Use
9873 UNSPEC_US_TRUNCATE instead of original us_truncate for
9875 * config/i386/mmx.md (mmx_pack<s_trunsuffix>swb): Substitute
9877 (mmx_packsswb): .. this and ..
9878 (mmx_packuswb): .. this.
9879 (mmx_packusdw): Use UNSPEC_US_TRUNCATE instead of original
9881 (s_trunsuffix): Removed code iterator.
9882 (any_s_truncate): Ditto.
9883 * config/i386/sse.md (<sse2_avx2>_packuswb<mask_name>): Use
9884 UNSPEC_US_TRUNCATE instead of original us_truncate.
9885 (<sse4_1_avx2>_packusdw<mask_name>): Ditto.
9886 * config/i386/i386.md (UNSPEC_US_TRUNCATE): New unspec_c_enum.
9888 2023-06-18 Pan Li <pan2.li@intel.com>
9890 * config/riscv/riscv-vector-builtins-bases.cc: Fix one typo.
9892 2023-06-18 Uros Bizjak <ubizjak@gmail.com>
9894 * rtl.h (*rtx_equal_p_callback_function):
9895 Change return type from int to bool.
9896 (rtx_equal_p): Ditto.
9897 (*hash_rtx_callback_function): Ditto.
9898 * rtl.cc (rtx_equal_p): Change return type from int to bool
9899 and adjust function body accordingly.
9900 * early-remat.cc (scratch_equal): Ditto.
9901 * sel-sched-ir.cc (skip_unspecs_callback): Ditto.
9902 (hash_with_unspec_callback): Ditto.
9904 2023-06-18 Jeff Law <jlaw@ventanamicro.com>
9906 * config/arc/arc.md (movqi_insn): Allow certain constants to
9907 be stored into memory in the pattern's condition.
9908 (movsf_insn): Similarly.
9910 2023-06-18 Honza <jh@ryzen3.suse.cz>
9912 PR tree-optimization/109849
9913 * ipa-fnsummary.cc (evaluate_conditions_for_known_args): Add new parameter
9914 ES; handle ipa_predicate::not_sra_candidate.
9915 (evaluate_properties_for_edge): Pass es to
9916 evaluate_conditions_for_known_args.
9917 (ipa_fn_summary_t::duplicate): Handle sra candidates.
9918 (dump_ipa_call_summary): Dump points_to_possible_sra_candidate.
9919 (load_or_store_of_ptr_parameter): New function.
9920 (points_to_possible_sra_candidate_p): New function.
9921 (analyze_function_body): Initialize points_to_possible_sra_candidate;
9922 determine sra predicates.
9923 (estimate_ipcp_clone_size_and_time): Update call of
9924 evaluate_conditions_for_known_args.
9925 (remap_edge_params): Update points_to_possible_sra_candidate.
9926 (read_ipa_call_summary): Stream points_to_possible_sra_candidate
9927 (write_ipa_call_summary): Likewise.
9928 * ipa-predicate.cc (ipa_predicate::add_clause): Handle not_sra_candidate.
9929 (dump_condition): Dump it.
9930 * ipa-predicate.h (struct inline_param_summary): Add
9931 points_to_possible_sra_candidate.
9933 2023-06-18 Roger Sayle <roger@nextmovesoftware.com>
9935 * config/i386/i386-expand.cc (ix86_expand_carry): New helper
9936 function for setting the carry flag.
9937 (ix86_expand_builtin) <handlecarry>: Use it here.
9938 * config/i386/i386-protos.h (ix86_expand_carry): Prototype here.
9939 * config/i386/i386.md (uaddc<mode>5): Use ix86_expand_carry.
9940 (usubc<mode>5): Likewise.
9942 2023-06-18 Roger Sayle <roger@nextmovesoftware.com>
9944 * config/i386/i386.md (*concat<mode><dwi>3_1): Use QImode
9945 for the immediate constant shift count.
9946 (*concat<mode><dwi>3_2): Likewise.
9947 (*concat<mode><dwi>3_3): Likewise.
9948 (*concat<mode><dwi>3_4): Likewise.
9949 (*concat<mode><dwi>3_5): Likewise.
9950 (*concat<mode><dwi>3_6): Likewise.
9952 2023-06-18 Uros Bizjak <ubizjak@gmail.com>
9954 * cse.cc (hash_rtx_cb): Rename to hash_rtx.
9956 * early-remat.cc (remat_candidate_hasher::equal): Update
9957 to call rtx_equal_p with rtx_equal_p_callback_function argument.
9958 * rtl.cc (rtx_equal_p_cb): Rename to rtx_equal_p.
9959 (rtx_equal_p): Remove.
9960 * rtl.h (rtx_equal_p): Add rtx_equal_p_callback_function
9961 argument with NULL default value.
9962 (rtx_equal_p_cb): Remove function declaration.
9963 (hash_rtx_cb): Ditto.
9964 (hash_rtx): Add hash_rtx_callback_function argument
9965 with NULL default value.
9966 * sel-sched-ir.cc (free_nop_pool): Update function comment.
9967 (skip_unspecs_callback): Ditto.
9968 (vinsn_init): Update to call hash_rtx with
9969 hash_rtx_callback_function argument.
9970 (vinsn_equal_p): Ditto.
9972 2023-06-18 yulong <shiyulong@iscas.ac.cn>
9974 * config/riscv/genrvv-type-indexer.cc (valid_type): Enable FP16 tuple.
9975 * config/riscv/riscv-modes.def (RVV_TUPLE_MODES): New macro.
9976 (ADJUST_ALIGNMENT): Ditto.
9977 (RVV_TUPLE_PARTIAL_MODES): Ditto.
9978 (ADJUST_NUNITS): Ditto.
9979 * config/riscv/riscv-vector-builtins-types.def (vfloat16mf4x2_t):
9981 (vfloat16mf4x3_t): Ditto.
9982 (vfloat16mf4x4_t): Ditto.
9983 (vfloat16mf4x5_t): Ditto.
9984 (vfloat16mf4x6_t): Ditto.
9985 (vfloat16mf4x7_t): Ditto.
9986 (vfloat16mf4x8_t): Ditto.
9987 (vfloat16mf2x2_t): Ditto.
9988 (vfloat16mf2x3_t): Ditto.
9989 (vfloat16mf2x4_t): Ditto.
9990 (vfloat16mf2x5_t): Ditto.
9991 (vfloat16mf2x6_t): Ditto.
9992 (vfloat16mf2x7_t): Ditto.
9993 (vfloat16mf2x8_t): Ditto.
9994 (vfloat16m1x2_t): Ditto.
9995 (vfloat16m1x3_t): Ditto.
9996 (vfloat16m1x4_t): Ditto.
9997 (vfloat16m1x5_t): Ditto.
9998 (vfloat16m1x6_t): Ditto.
9999 (vfloat16m1x7_t): Ditto.
10000 (vfloat16m1x8_t): Ditto.
10001 (vfloat16m2x2_t): Ditto.
10002 (vfloat16m2x3_t): Ditto.
10003 (vfloat16m2x4_t): Ditto.
10004 (vfloat16m4x2_t): Ditto.
10005 * config/riscv/riscv-vector-builtins.def (vfloat16mf4x2_t): New macro.
10006 (vfloat16mf4x3_t): Ditto.
10007 (vfloat16mf4x4_t): Ditto.
10008 (vfloat16mf4x5_t): Ditto.
10009 (vfloat16mf4x6_t): Ditto.
10010 (vfloat16mf4x7_t): Ditto.
10011 (vfloat16mf4x8_t): Ditto.
10012 (vfloat16mf2x2_t): Ditto.
10013 (vfloat16mf2x3_t): Ditto.
10014 (vfloat16mf2x4_t): Ditto.
10015 (vfloat16mf2x5_t): Ditto.
10016 (vfloat16mf2x6_t): Ditto.
10017 (vfloat16mf2x7_t): Ditto.
10018 (vfloat16mf2x8_t): Ditto.
10019 (vfloat16m1x2_t): Ditto.
10020 (vfloat16m1x3_t): Ditto.
10021 (vfloat16m1x4_t): Ditto.
10022 (vfloat16m1x5_t): Ditto.
10023 (vfloat16m1x6_t): Ditto.
10024 (vfloat16m1x7_t): Ditto.
10025 (vfloat16m1x8_t): Ditto.
10026 (vfloat16m2x2_t): Ditto.
10027 (vfloat16m2x3_t): Ditto.
10028 (vfloat16m2x4_t): Ditto.
10029 (vfloat16m4x2_t): Ditto.
10030 * config/riscv/riscv-vector-switch.def (TUPLE_ENTRY): New.
10031 * config/riscv/riscv.md: New.
10032 * config/riscv/vector-iterators.md: New.
10034 2023-06-17 Roger Sayle <roger@nextmovesoftware.com>
10036 * config/i386/i386-expand.cc (ix86_expand_move): Check that OP1 is
10037 CONST_WIDE_INT_P before calling ix86_convert_wide_int_to_broadcast.
10038 Generalize special case for converting TImode to V1TImode to handle
10039 all 128-bit vector conversions.
10041 2023-06-17 Costas Argyris <costas.argyris@gmail.com>
10043 * gcc-ar.cc (main): Refactor to slightly reduce code
10044 duplication. Avoid unnecessary elements in nargv.
10046 2023-06-16 Pan Li <pan2.li@intel.com>
10049 * config/riscv/riscv-vector-builtins-bases.cc: Add ret_mode for
10050 integer reduction expand.
10051 * config/riscv/vector-iterators.md: Add VQI, VHI, VSI and VDI,
10052 and the LMUL1 attr respectively.
10053 * config/riscv/vector.md
10054 (@pred_reduc_<reduc><mode><vlmul1>): Removed.
10055 (@pred_reduc_<reduc><mode><vlmul1_zve64>): Likewise.
10056 (@pred_reduc_<reduc><mode><vlmul1_zve32>): Likewise.
10057 (@pred_reduc_<reduc><VQI:mode><VQI_LMUL1:mode>): New pattern.
10058 (@pred_reduc_<reduc><VHI:mode><VHI_LMUL1:mode>): Likewise.
10059 (@pred_reduc_<reduc><VSI:mode><VSI_LMUL1:mode>): Likewise.
10060 (@pred_reduc_<reduc><VDI:mode><VDI_LMUL1:mode>): Likewise.
10062 2023-06-16 Juzhe-Zhong <juzhe.zhong@rivai.ai>
10065 * config/riscv/riscv-vsetvl.cc (insert_vsetvl): Fix bug.
10067 2023-06-16 Jakub Jelinek <jakub@redhat.com>
10069 PR middle-end/79173
10070 * builtin-types.def (BT_FN_UINT_UINT_UINT_UINT_UINTPTR,
10071 BT_FN_ULONG_ULONG_ULONG_ULONG_ULONGPTR,
10072 BT_FN_ULONGLONG_ULONGLONG_ULONGLONG_ULONGLONG_ULONGLONGPTR): New
10074 * builtins.def (BUILT_IN_ADDC, BUILT_IN_ADDCL, BUILT_IN_ADDCLL,
10075 BUILT_IN_SUBC, BUILT_IN_SUBCL, BUILT_IN_SUBCLL): New builtins.
10076 * builtins.cc (fold_builtin_addc_subc): New function.
10077 (fold_builtin_varargs): Handle BUILT_IN_{ADD,SUB}C{,L,LL}.
10078 * doc/extend.texi (__builtin_addc, __builtin_subc): Document.
10080 2023-06-16 Jakub Jelinek <jakub@redhat.com>
10082 PR tree-optimization/110271
10083 * tree-ssa-math-opts.cc (math_opts_dom_walker::after_dom_children)
10084 <case PLUS_EXPR>: Ignore return value from match_arith_overflow,
10085 instead call match_uaddc_usubc only if gsi_stmt (gsi) is still stmt.
10087 2023-06-16 Martin Jambor <mjambor@suse.cz>
10089 * configure: Regenerate.
10091 2023-06-16 Roger Sayle <roger@nextmovesoftware.com>
10092 Uros Bizjak <ubizjak@gmail.com>
10095 * config/i386/i386.md (*add<dwi>3_doubleword_concat): New
10096 define_insn_and_split combine *add<dwi>3_doubleword with
10097 a *concat<mode><dwi>3 for more efficient lowering after reload.
10099 2023-06-16 Vladimir N. Makarov <vmakarov@redhat.com>
10101 * ira-lives.cc: Include except.h.
10102 (process_bb_node_lives): Ignore conflicts from cleanup exceptions
10103 when the pseudo does not live at the exception landing pad.
10105 2023-06-16 Alex Coplan <alex.coplan@arm.com>
10107 * doc/invoke.texi: Document -Welaborated-enum-base.
10109 2023-06-16 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10111 * config/aarch64/aarch64-simd-builtins.def (shrn2_n): Rename builtins to...
10112 (ushrn2_n): ... This.
10113 (sqshrn2_n): Rename builtins to...
10114 (ssqshrn2_n): ... This.
10115 (uqshrn2_n): Rename builtins to...
10116 (uqushrn2_n): ... This.
10117 * config/aarch64/arm_neon.h (vqshrn_high_n_s16): Adjust for the above.
10118 (vqshrn_high_n_s32): Likewise.
10119 (vqshrn_high_n_s64): Likewise.
10120 (vqshrn_high_n_u16): Likewise.
10121 (vqshrn_high_n_u32): Likewise.
10122 (vqshrn_high_n_u64): Likewise.
10123 (vshrn_high_n_s16): Likewise.
10124 (vshrn_high_n_s32): Likewise.
10125 (vshrn_high_n_s64): Likewise.
10126 (vshrn_high_n_u16): Likewise.
10127 (vshrn_high_n_u32): Likewise.
10128 (vshrn_high_n_u64): Likewise.
10129 * config/aarch64/aarch64-simd.md (aarch64_<shrn_op>shrn2_n<mode>_insn_le):
10131 (aarch64_<shrn_op><sra_op>shrn2_n<mode>_insn_le): ... This.
10132 Use SHIFTRT iterator and AARCH64_VALID_SHRN_OP check.
10133 (aarch64_<shrn_op>shrn2_n<mode>_insn_be): Rename to...
10134 (aarch64_<shrn_op><sra_op>shrn2_n<mode>_insn_be): ... This.
10135 Use SHIFTRT iterator and AARCH64_VALID_SHRN_OP check.
10136 (aarch64_<shrn_op>shrn2_n<mode>): Rename to...
10137 (aarch64_<shrn_op><sra_op>shrn2_n<mode>): ... This.
10138 Update expander for the above.
10140 2023-06-16 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10142 * config/aarch64/aarch64-simd-builtins.def (shrn2): Rename builtins to...
10143 (shrn2_n): ... This.
10144 (rshrn2): Rename builtins to...
10145 (rshrn2_n): ... This.
10146 * config/aarch64/arm_neon.h (vrshrn_high_n_s16): Adjust for the above.
10147 (vrshrn_high_n_s32): Likewise.
10148 (vrshrn_high_n_s64): Likewise.
10149 (vrshrn_high_n_u16): Likewise.
10150 (vrshrn_high_n_u32): Likewise.
10151 (vrshrn_high_n_u64): Likewise.
10152 (vshrn_high_n_s16): Likewise.
10153 (vshrn_high_n_s32): Likewise.
10154 (vshrn_high_n_s64): Likewise.
10155 (vshrn_high_n_u16): Likewise.
10156 (vshrn_high_n_u32): Likewise.
10157 (vshrn_high_n_u64): Likewise.
10158 * config/aarch64/aarch64-simd.md (*aarch64_<srn_op>shrn<mode>2_vect_le):
10160 (*aarch64_<srn_op>shrn<mode>2_vect_be): Likewise.
10161 (aarch64_shrn2<mode>_insn_le): Likewise.
10162 (aarch64_shrn2<mode>_insn_be): Likewise.
10163 (aarch64_shrn2<mode>): Likewise.
10164 (aarch64_rshrn2<mode>_insn_le): Likewise.
10165 (aarch64_rshrn2<mode>_insn_be): Likewise.
10166 (aarch64_rshrn2<mode>): Likewise.
10167 (aarch64_<sur>q<r>shr<u>n2_n<mode>_insn_le): Likewise.
10168 (aarch64_<shrn_op>shrn2_n<mode>_insn_le): New define_insn.
10169 (aarch64_<sur>q<r>shr<u>n2_n<mode>_insn_be): Delete.
10170 (aarch64_<shrn_op>shrn2_n<mode>_insn_be): New define_insn.
10171 (aarch64_<sur>q<r>shr<u>n2_n<mode>): Delete.
10172 (aarch64_<shrn_op>shrn2_n<mode>): New define_expand.
10173 (aarch64_<shrn_op>rshrn2_n<mode>_insn_le): New define_insn.
10174 (aarch64_<shrn_op>rshrn2_n<mode>_insn_be): New define_insn.
10175 (aarch64_<shrn_op>rshrn2_n<mode>): New define_expand.
10176 (aarch64_sqshrun2_n<mode>_insn_le): New define_insn.
10177 (aarch64_sqshrun2_n<mode>_insn_be): New define_insn.
10178 (aarch64_sqshrun2_n<mode>): New define_expand.
10179 (aarch64_sqrshrun2_n<mode>_insn_le): New define_insn.
10180 (aarch64_sqrshrun2_n<mode>_insn_be): New define_insn.
10181 (aarch64_sqrshrun2_n<mode>): New define_expand.
10182 * config/aarch64/iterators.md (UNSPEC_SQSHRUN, UNSPEC_SQRSHRUN,
10183 UNSPEC_SQSHRN, UNSPEC_UQSHRN, UNSPEC_SQRSHRN, UNSPEC_UQRSHRN):
10184 Delete unspec values.
10185 (VQSHRN_N): Delete int iterator.
10187 2023-06-16 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10189 * config/aarch64/aarch64.h (AARCH64_VALID_SHRN_OP): Define.
10190 * config/aarch64/aarch64-simd.md
10191 (*aarch64_<shrn_op>shrn_n<mode>_insn<vczle><vczbe>): Rename to...
10192 (*aarch64_<shrn_op><shrn_s>shrn_n<mode>_insn<vczle><vczbe>): ... This.
10193 Use SHIFTRT iterator and add AARCH64_VALID_SHRN_OP to condition.
10194 * config/aarch64/iterators.md (shrn_s): New code attribute.
10196 2023-06-16 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10198 * config/aarch64/aarch64-simd.md (aarch64_<sur>q<r>shr<u>n_n<mode>):
10200 (aarch64_<shrn_op>shrn_n<mode>): ... This. Reimplement with RTL codes.
10201 (*aarch64_<shrn_op>rshrn_n<mode>_insn): New define_insn.
10202 (aarch64_sqrshrun_n<mode>_insn): Likewise.
10203 (aarch64_sqshrun_n<mode>_insn): Likewise.
10204 (aarch64_<shrn_op>rshrn_n<mode>): New define_expand.
10205 (aarch64_sqshrun_n<mode>): Likewise.
10206 (aarch64_sqrshrun_n<mode>): Likewise.
10207 * config/aarch64/iterators.md (V2XWIDE): Add HI and SI modes.
10209 2023-06-16 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10211 * config/aarch64/aarch64-simd-builtins.def (shrn): Rename builtins to...
10212 (shrn_n): ... This.
10213 (rshrn): Rename builtins to...
10214 (rshrn_n): ... This.
10215 * config/aarch64/arm_neon.h (vshrn_n_s16): Adjust for the above.
10216 (vshrn_n_s32): Likewise.
10217 (vshrn_n_s64): Likewise.
10218 (vshrn_n_u16): Likewise.
10219 (vshrn_n_u32): Likewise.
10220 (vshrn_n_u64): Likewise.
10221 (vrshrn_n_s16): Likewise.
10222 (vrshrn_n_s32): Likewise.
10223 (vrshrn_n_s64): Likewise.
10224 (vrshrn_n_u16): Likewise.
10225 (vrshrn_n_u32): Likewise.
10226 (vrshrn_n_u64): Likewise.
10227 * config/aarch64/aarch64-simd.md
10228 (*aarch64_<srn_op>shrn<mode><vczle><vczbe>): Delete.
10229 (aarch64_shrn<mode>): Likewise.
10230 (aarch64_rshrn<mode><vczle><vczbe>_insn): Likewise.
10231 (aarch64_rshrn<mode>): Likewise.
10232 (aarch64_<sur>q<r>shr<u>n_n<mode>_insn<vczle><vczbe>): Likewise.
10233 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
10234 (*aarch64_<shrn_op>shrn_n<mode>_insn<vczle><vczbe>): New define_insn.
10235 (*aarch64_<shrn_op>rshrn_n<mode>_insn<vczle><vczbe>): Likewise.
10236 (*aarch64_sqshrun_n<mode>_insn<vczle><vczbe>): Likewise.
10237 (*aarch64_sqrshrun_n<mode>_insn<vczle><vczbe>): Likewise.
10238 (aarch64_<shrn_op>shrn_n<mode>): New define_expand.
10239 (aarch64_<shrn_op>rshrn_n<mode>): Likewise.
10240 (aarch64_sqshrun_n<mode>): Likewise.
10241 (aarch64_sqrshrun_n<mode>): Likewise.
10242 * config/aarch64/iterators.md (ALL_TRUNC): New code iterator.
10243 (TRUNCEXTEND): New code attribute.
10244 (TRUNC_SHIFT): Likewise.
10245 (shrn_op): Likewise.
10246 * config/aarch64/predicates.md (aarch64_simd_umax_quarter_mode):
10249 2023-06-16 Pan Li <pan2.li@intel.com>
10251 * config/riscv/riscv-vsetvl.cc
10252 (pass_vsetvl::global_eliminate_vsetvl_insn): Initialize var by NULL.
10254 2023-06-16 Richard Biener <rguenther@suse.de>
10256 PR tree-optimization/110278
10257 * match.pd (uns < (typeof uns)(uns != 0) -> false): New.
10258 (x != (typeof x)(x == 0) -> true): Likewise.
10260 2023-06-16 Pali Rohár <pali@kernel.org>
10262 * config/i386/mingw-w64.h (CPP_SPEC): Adjust for -mcrtdll=.
10263 (REAL_LIBGCC_SPEC): New define.
10264 * config/i386/mingw.opt: Add mcrtdll=
10265 * config/i386/mingw32.h (CPP_SPEC): Adjust for -mcrtdll=.
10266 (REAL_LIBGCC_SPEC): Adjust for -mcrtdll=.
10267 (STARTFILE_SPEC): Adjust for -mcrtdll=.
10268 * doc/invoke.texi: Add mcrtdll= documentation.
10270 2023-06-16 Simon Dardis <simon.dardis@imgtec.com>
10272 * config/mips/mips.cc (enum mips_code_readable_setting):New enmu.
10273 (mips_handle_code_readable_attr):New static function.
10274 (mips_get_code_readable_attr):New static enum function.
10275 (mips_set_current_function):Set the code_readable mode.
10276 (mips_option_override):Same as above.
10277 * doc/extend.texi:Document code_readable.
10279 2023-06-16 Richard Biener <rguenther@suse.de>
10281 PR tree-optimization/110269
10282 * fold-const.cc (fold_binary_loc): Merge x != 0 folding
10283 with tree_expr_nonzero_p ...
10284 * match.pd (cmp (convert? addr@0) integer_zerop): With this
10287 2023-06-15 Marek Polacek <polacek@redhat.com>
10289 * Makefile.in: Set LD_PICFLAG. Use it. Set enable_host_pie.
10290 Remove NO_PIE_CFLAGS and NO_PIE_FLAG. Pass LD_PICFLAG to
10291 ALL_LINKERFLAGS. Use the "pic" build of libiberty if --enable-host-pie.
10292 * configure.ac (--enable-host-shared): Don't set PICFLAG here.
10293 (--enable-host-pie): New check. Set PICFLAG and LD_PICFLAG after this
10295 * configure: Regenerate.
10296 * doc/install.texi: Document --enable-host-pie.
10298 2023-06-15 Manolis Tsamis <manolis.tsamis@vrull.eu>
10300 * regcprop.cc (maybe_mode_change): Enable stack pointer
10303 2023-06-15 Andrew MacLeod <amacleod@redhat.com>
10305 PR tree-optimization/110266
10306 * gimple-range-fold.cc (adjust_imagpart_expr): Check for integer
10308 (adjust_realpart_expr): Ditto.
10310 2023-06-15 Jan Beulich <jbeulich@suse.com>
10312 * config/i386/sse.md (<avx512>_vec_dup<mode><mask_name>): Use
10315 2023-06-15 Jan Beulich <jbeulich@suse.com>
10317 * config/i386/constraints.md: Mention k and r for B.
10319 2023-06-15 Lulu Cheng <chenglulu@loongson.cn>
10320 Andrew Pinski <apinski@marvell.com>
10323 * config/loongarch/loongarch.md: Modify the register constraints for template
10324 "jumptable" and "indirect_jump" from "r" to "e".
10326 2023-06-15 Xi Ruoyao <xry111@xry111.site>
10328 * config/loongarch/loongarch-tune.h (loongarch_align): New
10330 * config/loongarch/loongarch-def.h (loongarch_cpu_align): New
10332 * config/loongarch/loongarch-def.c (loongarch_cpu_align): Define
10334 * config/loongarch/loongarch.cc
10335 (loongarch_option_override_internal): Set the value of
10336 -falign-functions= if -falign-functions is enabled but no value
10337 is given. Likewise for -falign-labels=.
10339 2023-06-15 Jakub Jelinek <jakub@redhat.com>
10341 PR middle-end/79173
10342 * internal-fn.def (UADDC, USUBC): New internal functions.
10343 * internal-fn.cc (expand_UADDC, expand_USUBC): New functions.
10344 (commutative_ternary_fn_p): Return true also for IFN_UADDC.
10345 * optabs.def (uaddc5_optab, usubc5_optab): New optabs.
10346 * tree-ssa-math-opts.cc (uaddc_cast, uaddc_ne0, uaddc_is_cplxpart,
10347 match_uaddc_usubc): New functions.
10348 (math_opts_dom_walker::after_dom_children): Call match_uaddc_usubc
10349 for PLUS_EXPR, MINUS_EXPR, BIT_IOR_EXPR and BIT_XOR_EXPR unless
10350 other optimizations have been successful for those.
10351 * gimple-fold.cc (gimple_fold_call): Handle IFN_UADDC and IFN_USUBC.
10352 * fold-const-call.cc (fold_const_call): Likewise.
10353 * gimple-range-fold.cc (adjust_imagpart_expr): Likewise.
10354 * tree-ssa-dce.cc (eliminate_unnecessary_stmts): Likewise.
10355 * doc/md.texi (uaddc<mode>5, usubc<mode>5): Document new named
10357 * config/i386/i386.md (uaddc<mode>5, usubc<mode>5): New
10358 define_expand patterns.
10359 (*setcc_qi_addqi3_cconly_overflow_1_<mode>, *setccc): Split
10360 into NOTE_INSN_DELETED note rather than nop instruction.
10361 (*setcc_qi_negqi_ccc_1_<mode>, *setcc_qi_negqi_ccc_2_<mode>):
10364 2023-06-15 Jakub Jelinek <jakub@redhat.com>
10366 PR middle-end/79173
10367 * config/i386/i386.md (subborrow<mode>): Add alternative with
10368 memory destination and add for it define_peephole2
10369 TARGET_READ_MODIFY_WRITE/-Os patterns to prefer using memory
10370 destination in these patterns.
10372 2023-06-15 Jakub Jelinek <jakub@redhat.com>
10374 PR middle-end/79173
10375 * config/i386/i386.md (*sub<mode>_3, @add<mode>3_carry,
10376 addcarry<mode>, @sub<mode>3_carry, *add<mode>3_cc_overflow_1): Add
10377 define_peephole2 TARGET_READ_MODIFY_WRITE/-Os patterns to prefer
10378 using memory destination in these patterns.
10380 2023-06-15 Jakub Jelinek <jakub@redhat.com>
10382 * gimple-fold.cc (gimple_fold_call): Move handling of arg0
10383 as well as arg1 INTEGER_CSTs for .UBSAN_CHECK_{ADD,SUB,MUL}
10384 and .{ADD,SUB,MUL}_OVERFLOW calls from here...
10385 * fold-const-call.cc (fold_const_call): ... here.
10387 2023-06-15 Oluwatamilore Adebayo <oluwatamilore.adebayo@arm.com>
10389 * config/aarch64/aarch64-simd.md (aarch64_<su>abd<mode>):
10390 Rename to <su>abd<mode>3.
10391 * config/aarch64/aarch64-sve.md (<su>abd<mode>_3): Rename
10394 2023-06-15 Oluwatamilore Adebayo <oluwatamilore.adebayo@arm.com>
10396 * doc/md.texi (sabd, uabd): Document them.
10397 * internal-fn.def (ABD): Use new optab.
10398 * optabs.def (sabd_optab, uabd_optab): New optabs,
10399 * tree-vect-patterns.cc (vect_recog_absolute_difference):
10400 Recognize the following idiom abs (a - b).
10401 (vect_recog_sad_pattern): Refactor to use
10402 vect_recog_absolute_difference.
10403 (vect_recog_abd_pattern): Use patterns found by
10404 vect_recog_absolute_difference to build a new ABD
10407 2023-06-15 chenxiaolong <chenxl04200420@163.com>
10409 * config/loongarch/loongarch.h (LARCH_CALL_RATIO): Modify the value
10410 of macro LARCH_CALL_RATIO on LoongArch to make it perform optimally.
10412 2023-06-15 Juzhe-Zhong <juzhe.zhong@rivai.ai>
10414 * config/riscv/riscv-v.cc (shuffle_merge_patterns): New pattern.
10415 (expand_vec_perm_const_1): Add merge optmization.
10417 2023-06-15 Lehua Ding <lehua.ding@rivai.ai>
10420 * config/riscv/riscv.cc (riscv_get_arg_info): Return NULL_RTX for vector mode
10421 (riscv_pass_by_reference): Return true for vector mode
10423 2023-06-15 Pan Li <pan2.li@intel.com>
10425 * config/riscv/autovec-opt.md: Align the predictor sytle.
10426 * config/riscv/autovec.md: Ditto.
10428 2023-06-15 Pan Li <pan2.li@intel.com>
10430 * config/riscv/riscv-v.cc (rvv_builder::get_merge_scalar_mask):
10431 Take elen instead of scalar BITS_PER_WORD.
10432 (expand_vector_init_merge_repeating_sequence): Use inner_bits_size
10433 instead of scaler BITS_PER_WORD.
10435 2023-06-14 Jivan Hakobyan <jivanhakobyan9@gmail.com>
10437 * config/moxie/uclinux.h (MFWRAP_SPEC): Remove
10439 2023-06-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10441 * config/aarch64/aarch64-sve-builtins-base.cc (svlast_impl::fold):
10442 Fix signed comparison warning in loop from npats to enelts.
10444 2023-06-14 Thomas Schwinge <thomas@codesourcery.com>
10446 * gcc.cc (driver_handle_option): Forward host '-lgfortran', '-lm'
10447 to offloading compilation.
10448 * config/gcn/mkoffload.cc (main): Adjust.
10449 * config/nvptx/mkoffload.cc (main): Likewise.
10450 * doc/invoke.texi (foffload-options): Update example.
10452 2023-06-14 liuhongt <hongtao.liu@intel.com>
10455 * config/i386/sse.md (mov<mode>_internal>): Use x instead of v
10456 for alternative 2 since there's no evex version for vpcmpeqd
10459 2023-06-13 Jeff Law <jlaw@ventanamicro.com>
10461 * gcc.cc (LINK_COMMAND_SPEC): Remove mudflap spec handling.
10463 2023-06-13 Jeff Law <jlaw@ventanamicro.com>
10465 * config/sh/divtab.cc: Remove.
10467 2023-06-13 Jakub Jelinek <jakub@redhat.com>
10469 * config/i386/i386.cc (standard_sse_constant_opcode): Remove
10470 superfluous spaces around \t for vpcmpeqd.
10472 2023-06-13 Roger Sayle <roger@nextmovesoftware.com>
10474 * expr.cc (store_constructor) <case VECTOR_TYPE>: Don't bother
10475 clearing vectors with only a single element. Set CLEARED if the
10476 vector was initialized to zero.
10478 2023-06-13 Lehua Ding <lehua.ding@rivai.ai>
10480 * config/riscv/riscv-v.cc (struct mode_vtype_group): Remove duplicate
10483 (TUPLE_ENTRY): Undef.
10485 2023-06-13 Juzhe-Zhong <juzhe.zhong@rivai.ai>
10487 * config/riscv/riscv-v.cc (rvv_builder::single_step_npatterns_p): Add comment.
10488 (shuffle_generic_patterns): Ditto.
10489 (expand_vec_perm_const_1): Ditto.
10491 2023-06-13 Juzhe-Zhong <juzhe.zhong@rivai.ai>
10493 * config/riscv/riscv-v.cc (emit_vlmax_decompress_insn): Fix bug.
10494 (shuffle_decompress_patterns): Ditto.
10496 2023-06-13 Richard Biener <rguenther@suse.de>
10498 * tree-ssa-loop-ch.cc (ch_base::copy_headers): Free loop BBs.
10500 2023-06-13 Yanzhang Wang <yanzhang.wang@intel.com>
10501 Kito Cheng <kito.cheng@sifive.com>
10503 * config/riscv/riscv-protos.h (riscv_init_cumulative_args): Set
10504 warning flag if func is not builtin
10505 * config/riscv/riscv.cc
10506 (riscv_scalable_vector_type_p): Determine whether the type is scalable vector.
10507 (riscv_arg_has_vector): Determine whether the arg is vector type.
10508 (riscv_pass_in_vector_p): Check the vector type param is passed by value.
10509 (riscv_init_cumulative_args): The same as header.
10510 (riscv_get_arg_info): Add the checking.
10511 (riscv_function_value): Check the func return and set warning flag
10512 * config/riscv/riscv.h (INIT_CUMULATIVE_ARGS): Add a flag to
10513 determine whether warning psabi or not.
10515 2023-06-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10517 * config/arm/arm-opts.h (enum arm_tp_type): Remove TP_CP15.
10518 Add TP_TPIDRURW, TP_TPIDRURO, TP_TPIDRPRW values.
10519 * config/arm/arm-protos.h (arm_output_load_tpidr): Declare prototype.
10520 * config/arm/arm.cc (arm_option_reconfigure_globals): Replace TP_CP15
10522 (arm_output_load_tpidr): Define.
10523 * config/arm/arm.h (TARGET_HARD_TP): Define in terms of TARGET_SOFT_TP.
10524 * config/arm/arm.md (load_tp_hard): Call arm_output_load_tpidr to output
10526 (reload_tp_hard): Likewise.
10527 * config/arm/arm.opt (tpidrurw, tpidruro, tpidrprw): New values for
10529 * doc/invoke.texi (Arm Options, mtp): Document new values.
10531 2023-06-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10534 * config/aarch64/aarch64-opts.h (enum aarch64_tp_reg): Add
10535 AARCH64_TPIDRRO_EL0 value.
10536 * config/aarch64/aarch64.cc (aarch64_output_load_tp): Define.
10537 * config/aarch64/aarch64.opt (tpidr_el0, tpidr_el1, tpidr_el2,
10538 tpidr_el3, tpidrro_el3): New accepted values to -mtp=.
10539 * doc/invoke.texi (AArch64 Options): Document new -mtp= options.
10541 2023-06-13 Alexandre Oliva <oliva@adacore.com>
10543 * range-op-float.cc (frange_nextafter): Drop inline.
10544 (frelop_early_resolve): Add static.
10545 (frange_float): Likewise.
10547 2023-06-13 Richard Biener <rguenther@suse.de>
10549 PR middle-end/110232
10550 * fold-const.cc (native_interpret_vector): Use TYPE_SIZE_UNIT
10551 to check whether the buffer covers the whole vector.
10553 2023-06-13 Richard Biener <rguenther@suse.de>
10555 * tree-ssa-alias.cc (ref_maybe_used_by_call_p_1): For
10556 .MASK_LOAD and friends set the size of the access to unknown.
10558 2023-06-13 Tejas Belagod <tbelagod@arm.com>
10561 * config/aarch64/aarch64-sve-builtins-base.cc (svlast_impl::fold): Fold sve
10562 calls that have a constant input predicate vector.
10563 (svlast_impl::is_lasta): Query to check if intrinsic is svlasta.
10564 (svlast_impl::is_lastb): Query to check if intrinsic is svlastb.
10565 (svlast_impl::vect_all_same): Check if all vector elements are equal.
10567 2023-06-13 Andi Kleen <ak@linux.intel.com>
10569 * config/i386/gcc-auto-profile: Regenerate.
10571 2023-06-13 Juzhe-Zhong <juzhe.zhong@rivai.ai>
10573 * config/riscv/vector-iterators.md: Fix requirement.
10575 2023-06-13 Juzhe-Zhong <juzhe.zhong@rivai.ai>
10577 * config/riscv/riscv-v.cc (emit_vlmax_decompress_insn): New function.
10578 (shuffle_decompress_patterns): New function.
10579 (expand_vec_perm_const_1): Add decompress optimization.
10581 2023-06-12 Jeff Law <jlaw@ventanamicro.com>
10583 PR rtl-optimization/101188
10584 * postreload.cc (reload_cse_move2add_invalidate): New function,
10586 (reload_cse_move2add): Call reload_cse_move2add_invalidate.
10588 2023-06-12 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
10590 * config/aarch64/aarch64.cc (aarch64_expand_vector_init): Tweak condition
10591 if (n_var == n_elts && n_elts <= 16) to allow a single constant,
10592 and if maxv == 1, use constant element for duplicating into register.
10594 2023-06-12 Tobias Burnus <tobias@codesourcery.com>
10596 * gimplify.cc (gimplify_adjust_omp_clauses_1): Use
10597 GOMP_MAP_FORCE_PRESENT for 'present alloc' implicit mapping.
10598 (gimplify_adjust_omp_clauses): Change
10599 GOMP_MAP_PRESENT_{TO,TOFROM,FROM,ALLOC} to the equivalent
10600 GOMP_MAP_FORCE_PRESENT.
10601 * omp-low.cc (lower_omp_target): Remove handling of no-longer valid
10602 GOMP_MAP_PRESENT_{TO,TOFROM,FROM,ALLOC}; update map kinds used for
10603 to/from clauses with present modifier.
10605 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10607 PR tree-optimization/110205
10608 * range-op-float.cc (range_operator::fold_range): Add default FII
10610 * range-op-mixed.h (class operator_gt): Add missing final overrides.
10611 * range-op.cc (range_op_handler::fold_range): Add RO_FII case.
10612 (operator_lshift ::update_bitmask): Add final override.
10613 (operator_rshift ::update_bitmask): Add final override.
10614 * range-op.h (range_operator::fold_range): Add FII prototype.
10616 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10618 * gimple-range-op.cc (gimple_range_op_handler::maybe_non_standard):
10619 Use range_op_handler directly.
10620 * range-op.cc (range_op_handler::range_op_handler): Unsigned
10621 param instead of tree-code.
10622 (ptr_op_widen_plus_signed): Delete.
10623 (ptr_op_widen_plus_unsigned): Delete.
10624 (ptr_op_widen_mult_signed): Delete.
10625 (ptr_op_widen_mult_unsigned): Delete.
10626 (range_op_table::initialize_integral_ops): Add new opcodes.
10627 * range-op.h (range_op_handler): Use unsigned.
10628 (OP_WIDEN_MULT_SIGNED): New.
10629 (OP_WIDEN_MULT_UNSIGNED): New.
10630 (OP_WIDEN_PLUS_SIGNED): New.
10631 (OP_WIDEN_PLUS_UNSIGNED): New.
10632 (RANGE_OP_TABLE_SIZE): New.
10633 (range_op_table::operator []): Use unsigned.
10634 (range_op_table::set): Use unsigned.
10635 (m_range_tree): Make unsigned.
10636 (ptr_op_widen_mult_signed): Remove.
10637 (ptr_op_widen_mult_unsigned): Remove.
10638 (ptr_op_widen_plus_signed): Remove.
10639 (ptr_op_widen_plus_unsigned): Remove.
10641 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10643 * gimple-range-op.cc (gimple_range_op_handler): Set m_operator
10644 manually as there is no access to the default operator.
10645 (cfn_copysign::fold_range): Don't check for validity.
10646 (cfn_ubsan::fold_range): Ditto.
10647 (gimple_range_op_handler::maybe_builtin_call): Don't set to NULL.
10648 * range-op.cc (default_operator): New.
10649 (range_op_handler::range_op_handler): Use default_operator
10651 (range_op_handler::operator bool): Move from header, compare
10652 against default operator.
10653 (range_op_handler::range_op): New.
10654 * range-op.h (range_op_handler::operator bool): Move.
10656 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10658 * range-op.cc (unified_table): Delete.
10659 (range_op_table operator_table): Instantiate.
10660 (range_op_table::range_op_table): Rename from unified_table.
10661 (range_op_handler::range_op_handler): Use range_op_table.
10662 * range-op.h (range_op_table::operator []): Inline.
10663 (range_op_table::set): Inline.
10665 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10667 * gimple-range-gori.cc (gori_compute::condexpr_adjust): Do not
10669 * gimple-range-op.cc (get_code): Rename from get_code_and_type
10671 (gimple_range_op_handler::supported_p): No need for type.
10672 (gimple_range_op_handler::gimple_range_op_handler): Ditto.
10673 (cfn_copysign::fold_range): Ditto.
10674 (cfn_ubsan::fold_range): Ditto.
10675 * ipa-cp.cc (ipa_vr_operation_and_type_effects): Ditto.
10676 * ipa-fnsummary.cc (evaluate_conditions_for_known_args): Ditto.
10677 * range-op-float.cc (operator_plus::op1_range): Ditto.
10678 (operator_mult::op1_range): Ditto.
10679 (range_op_float_tests): Ditto.
10680 * range-op.cc (get_op_handler): Remove.
10681 (range_op_handler::set_op_handler): Remove.
10682 (operator_plus::op1_range): No need for type.
10683 (operator_minus::op1_range): Ditto.
10684 (operator_mult::op1_range): Ditto.
10685 (operator_exact_divide::op1_range): Ditto.
10686 (operator_cast::op1_range): Ditto.
10687 (perator_bitwise_not::fold_range): Ditto.
10688 (operator_negate::fold_range): Ditto.
10689 * range-op.h (range_op_handler::range_op_handler): Remove type param.
10690 (range_cast): No need for type.
10691 (range_op_table::operator[]): Check for enum_code >= 0.
10692 * tree-data-ref.cc (compute_distributive_range): No need for type.
10693 * tree-ssa-loop-unswitch.cc (unswitch_predicate): Ditto.
10694 * value-query.cc (range_query::get_tree_range): Ditto.
10695 * value-relation.cc (relation_oracle::validate_relation): Ditto.
10696 * vr-values.cc (range_of_var_in_loop): Ditto.
10697 (simplify_using_ranges::fold_cond_with_ops): Ditto.
10699 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10701 * range-op-mixed.h (operator_max): Remove final.
10702 * range-op-ptr.cc (pointer_table::pointer_table): Remove MAX_EXPR.
10703 (pointer_table::pointer_table): Remove.
10704 (class hybrid_max_operator): New.
10705 (range_op_table::initialize_pointer_ops): Add hybrid_max_operator.
10706 * range-op.cc (pointer_tree_table): Remove.
10707 (unified_table::unified_table): Comment out MAX_EXPR.
10708 (get_op_handler): Remove check of pointer table.
10709 * range-op.h (class pointer_table): Remove.
10711 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10713 * range-op-mixed.h (operator_min): Remove final.
10714 * range-op-ptr.cc (pointer_table::pointer_table): Remove MIN_EXPR.
10715 (class hybrid_min_operator): New.
10716 (range_op_table::initialize_pointer_ops): Add hybrid_min_operator.
10717 * range-op.cc (unified_table::unified_table): Comment out MIN_EXPR.
10719 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10721 * range-op-mixed.h (operator_bitwise_or): Remove final.
10722 * range-op-ptr.cc (pointer_table::pointer_table): Remove BIT_IOR_EXPR.
10723 (class hybrid_or_operator): New.
10724 (range_op_table::initialize_pointer_ops): Add hybrid_or_operator.
10725 * range-op.cc (unified_table::unified_table): Comment out BIT_IOR_EXPR.
10727 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10729 * range-op-mixed.h (operator_bitwise_and): Remove final.
10730 * range-op-ptr.cc (pointer_table::pointer_table): Remove BIT_AND_EXPR.
10731 (class hybrid_and_operator): New.
10732 (range_op_table::initialize_pointer_ops): Add hybrid_and_operator.
10733 * range-op.cc (unified_table::unified_table): Comment out BIT_AND_EXPR.
10735 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10737 * Makefile.in (OBJS): Add range-op-ptr.o.
10738 * range-op-mixed.h (update_known_bitmask): Move prototype here.
10739 (minus_op1_op2_relation_effect): Move prototype here.
10740 (wi_includes_zero_p): Move function to here.
10741 (wi_zero_p): Ditto.
10742 * range-op.cc (update_known_bitmask): Remove static.
10743 (wi_includes_zero_p): Move to header.
10744 (wi_zero_p): Move to header.
10745 (minus_op1_op2_relation_effect): Remove static.
10746 (operator_pointer_diff): Move class and routines to range-op-ptr.cc.
10747 (pointer_plus_operator): Ditto.
10748 (pointer_min_max_operator): Ditto.
10749 (pointer_and_operator): Ditto.
10750 (pointer_or_operator): Ditto.
10751 (pointer_table): Ditto.
10752 (range_op_table::initialize_pointer_ops): Ditto.
10753 * range-op-ptr.cc: New.
10755 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10757 * range-op-mixed.h (class operator_max): Move from...
10758 * range-op.cc (unified_table::unified_table): Add MAX_EXPR.
10759 (get_op_handler): Remove the integral table.
10760 (class operator_max): Move from here.
10761 (integral_table::integral_table): Delete.
10762 * range-op.h (class integral_table): Delete.
10764 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10766 * range-op-mixed.h (class operator_min): Move from...
10767 * range-op.cc (unified_table::unified_table): Add MIN_EXPR.
10768 (class operator_min): Move from here.
10769 (integral_table::integral_table): Remove MIN_EXPR.
10771 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10773 * range-op-mixed.h (class operator_bitwise_or): Move from...
10774 * range-op.cc (unified_table::unified_table): Add BIT_IOR_EXPR.
10775 (class operator_bitwise_or): Move from here.
10776 (integral_table::integral_table): Remove BIT_IOR_EXPR.
10778 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10780 * range-op-mixed.h (class operator_bitwise_and): Move from...
10781 * range-op.cc (unified_table::unified_table): Add BIT_AND_EXPR.
10782 (get_op_handler): Check for a pointer table entry first.
10783 (class operator_bitwise_and): Move from here.
10784 (integral_table::integral_table): Remove BIT_AND_EXPR.
10786 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10788 * range-op-mixed.h (class operator_bitwise_xor): Move from...
10789 * range-op.cc (unified_table::unified_table): Add BIT_XOR_EXPR.
10790 (class operator_bitwise_xor): Move from here.
10791 (integral_table::integral_table): Remove BIT_XOR_EXPR.
10792 (pointer_table::pointer_table): Remove BIT_XOR_EXPR.
10794 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10796 * range-op-mixed.h (class operator_bitwise_not): Move from...
10797 * range-op.cc (unified_table::unified_table): Add BIT_NOT_EXPR.
10798 (class operator_bitwise_not): Move from here.
10799 (integral_table::integral_table): Remove BIT_NOT_EXPR.
10800 (pointer_table::pointer_table): Remove BIT_NOT_EXPR.
10802 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
10804 * range-op-mixed.h (class operator_addr_expr): Move from...
10805 * range-op.cc (unified_table::unified_table): Add ADDR_EXPR.
10806 (class operator_addr_expr): Move from here.
10807 (integral_table::integral_table): Remove ADDR_EXPR.
10808 (pointer_table::pointer_table): Remove ADDR_EXPR.
10810 2023-06-12 Pan Li <pan2.li@intel.com>
10812 * config/riscv/riscv-vector-builtins-types.def
10813 (vfloat16m1_t): Add type to lmul1 ops.
10814 (vfloat16m2_t): Likewise.
10815 (vfloat16m4_t): Likewise.
10817 2023-06-12 Richard Biener <rguenther@suse.de>
10819 * tree-ssa-alias.cc (call_may_clobber_ref_p_1): For
10820 .MASK_STORE and friend set the size of the access to
10823 2023-06-12 Tamar Christina <tamar.christina@arm.com>
10825 * config.in: Regenerate.
10826 * configure: Regenerate.
10827 * configure.ac: Remove DEFAULT_MATCHPD_PARTITIONS.
10829 2023-06-12 Juzhe-Zhong <juzhe.zhong@rivai.ai>
10831 * config/riscv/autovec-opt.md
10832 (*v<any_shiftrt:optab><any_extend:optab>trunc<mode>): New pattern.
10833 (*<any_shiftrt:optab>trunc<mode>): Ditto.
10834 * config/riscv/autovec.md (<optab><mode>3): Change to
10835 define_insn_and_split.
10836 (v<optab><mode>3): Ditto.
10837 (trunc<mode><v_double_trunc>2): Ditto.
10839 2023-06-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10841 * simplify-rtx.cc (simplify_const_unary_operation):
10842 Handle US_TRUNCATE, SS_TRUNCATE.
10844 2023-06-12 Eric Botcazou <ebotcazou@adacore.com>
10847 * doc/gm2.texi (Standard procedures): Fix Next link.
10849 2023-06-12 Tamar Christina <tamar.christina@arm.com>
10851 * config.in: Regenerate.
10853 2023-06-12 Andre Vieira <andre.simoesdiasvieira@arm.com>
10855 PR middle-end/110142
10856 * tree-vect-patterns.cc (vect_recog_widen_op_pattern): Don't pass
10857 subtype to vect_widened_op_tree and remove subtype parameter, also
10858 remove superfluous overloaded function definition.
10859 (vect_recog_widen_plus_pattern): Remove subtype parameter and dont pass
10860 to call to vect_recog_widen_op_pattern.
10861 (vect_recog_widen_minus_pattern): Likewise.
10863 2023-06-12 liuhongt <hongtao.liu@intel.com>
10865 * config/i386/sse.md (vec_pack<floatprefix>_float_<mode>): New expander.
10866 (vec_unpack_<fixprefix>fix_trunc_lo_<mode>): Ditto.
10867 (vec_unpack_<fixprefix>fix_trunc_hi_<mode>): Ditto.
10868 (vec_unpacks_lo_<mode>): Ditto.
10869 (vec_unpacks_hi_<mode>): Ditto.
10870 (sse_movlhps_<mode>): New define_insn.
10871 (ssse3_palignr<mode>_perm): Extend to V_128H.
10872 (V_128H): New mode iterator.
10873 (ssepackPHmode): New mode attribute.
10874 (vunpck_extract_mode): Ditto.
10875 (vpckfloat_concat_mode): Extend to VxSI/VxSF for _Float16.
10876 (vpckfloat_temp_mode): Ditto.
10877 (vpckfloat_op_mode): Ditto.
10878 (vunpckfixt_mode): Extend to VxHF.
10879 (vunpckfixt_model): Ditto.
10880 (vunpckfixt_extract_mode): Ditto.
10882 2023-06-12 Richard Biener <rguenther@suse.de>
10884 PR middle-end/110200
10885 * genmatch.cc (expr::gen_transform): Put braces around
10886 the if arm for the (convert ...) short-cut.
10888 2023-06-12 Kewen Lin <linkw@linux.ibm.com>
10891 * config/rs6000/rs6000-builtins.def (__builtin_pack_vector_int128,
10892 __builtin_unpack_vector_int128): Move from stanza power7 to vsx.
10894 2023-06-12 Kewen Lin <linkw@linux.ibm.com>
10897 * config/rs6000/rs6000.cc (output_toc): Use the mode of the 128-bit
10898 floating constant itself for real_to_target call.
10900 2023-06-12 Pan Li <pan2.li@intel.com>
10902 * config/riscv/riscv-vector-builtins-types.def
10903 (vfloat16mf4_t): Add type to X2/X4/X8/X16/X32 vlmul ext ops.
10904 (vfloat16mf2_t): Ditto.
10905 (vfloat16m1_t): Ditto.
10906 (vfloat16m2_t): Ditto.
10907 (vfloat16m4_t): Ditto.
10909 2023-06-12 David Edelsohn <dje.gcc@gmail.com>
10911 * config/rs6000/rs6000-logue.cc (rs6000_stack_info):
10912 Do not require a stack frame when debugging is enabled for AIX.
10914 2023-06-11 Georg-Johann Lay <avr@gjlay.de>
10916 * config/avr/avr.md (adjust_len) [insv_notbit_0, insv_notbit_7]:
10917 Remove attribute values.
10918 (insv_notbit): New post-reload insn.
10919 (*insv.not-shiftrt_split, *insv.xor1-bit.0_split)
10920 (*insv.not-bit.0_split, *insv.not-bit.7_split)
10921 (*insv.xor-extract_split): Split to insv_notbit.
10922 (*insv.not-shiftrt, *insv.xor1-bit.0, *insv.not-bit.0, *insv.not-bit.7)
10923 (*insv.xor-extract): Remove post-reload insns.
10924 * config/avr/avr.cc (avr_out_insert_notbit) [bitno]: Remove parameter.
10925 (avr_adjust_insn_length): Adjust call of avr_out_insert_notbit.
10926 [ADJUST_LEN_INSV_NOTBIT_0, ADJUST_LEN_INSV_NOTBIT_7]: Remove cases.
10927 * config/avr/avr-protos.h (avr_out_insert_notbit): Adjust prototype.
10929 2023-06-11 Georg-Johann Lay <avr@gjlay.de>
10932 * config/avr/avr.md (adjust_len) [extr, extr_not]: New elements.
10933 (MSB, SIZE): New mode attributes.
10934 (any_shift): New code iterator.
10935 (*lshr<mode>3_split, *lshr<mode>3, lshr<mode>3)
10936 (*lshr<mode>3_const_split): Add constraint alternative for
10937 the case of shift-offset = MSB. Ditch "length" attribute.
10938 (extzv<mode): New. replaces extzv. Adjust following patterns.
10939 Use avr_out_extr, avr_out_extr_not to print asm.
10940 (*extzv.subreg.<mode>, *extzv.<mode>.subreg, *extzv.xor)
10941 (*extzv<mode>.ge, *neg.ashiftrt<mode>.msb, *extzv.io.lsr7): New.
10942 * config/avr/constraints.md (C15, C23, C31, Yil): New
10943 * config/avr/predicates.md (reg_or_low_io_operand)
10944 (const7_operand, reg_or_low_io_operand)
10945 (const15_operand, const_0_to_15_operand)
10946 (const23_operand, const_0_to_23_operand)
10947 (const31_operand, const_0_to_31_operand): New.
10948 * config/avr/avr-protos.h (avr_out_extr, avr_out_extr_not): New.
10949 * config/avr/avr.cc (avr_out_extr, avr_out_extr_not): New funcs.
10950 (lshrqi3_out, lshrhi3_out, lshrpsi3_out, lshrsi3_out): Adjust
10951 MSB case to new insn constraint "r" for operands[1].
10952 (avr_adjust_insn_length) [ADJUST_LEN_EXTR_NOT, ADJUST_LEN_EXTR]:
10953 Handle these cases.
10954 (avr_rtx_costs_1): Adjust cost for a new pattern.
10956 2023-06-11 Juzhe-Zhong <juzhe.zhong@rivai.ai>
10958 * config/riscv/riscv-vsetvl.cc (available_occurrence_p): Enhance user vsetvl optimization.
10959 (vector_insn_info::parse_insn): Add rtx_insn parse.
10960 (pass_vsetvl::local_eliminate_vsetvl_insn): Enhance user vsetvl optimization.
10961 (get_first_vsetvl): New function.
10962 (pass_vsetvl::global_eliminate_vsetvl_insn): Ditto.
10963 (pass_vsetvl::cleanup_insns): Remove it.
10964 (pass_vsetvl::ssa_post_optimization): New function.
10965 (has_no_uses): Ditto.
10966 (pass_vsetvl::propagate_avl): Remove it.
10967 (pass_vsetvl::df_post_optimization): New function.
10968 (pass_vsetvl::lazy_vsetvl): Rework Phase 5 && Phase 6.
10969 * config/riscv/riscv-vsetvl.h: Adapt declaration.
10971 2023-06-10 Aldy Hernandez <aldyh@redhat.com>
10973 * ipa-cp.cc (ipcp_vr_lattice::init): Take type argument.
10974 (ipcp_vr_lattice::print): Call dump method.
10975 (ipcp_vr_lattice::meet_with): Adjust for m_vr being a
10977 (ipcp_vr_lattice::meet_with_1): Make argument a reference.
10978 (ipcp_vr_lattice::set_to_bottom): Set varying for an unsupported
10980 (initialize_node_lattices): Pass type when appropriate.
10981 (ipa_vr_operation_and_type_effects): Make type agnostic.
10982 (ipa_value_range_from_jfunc): Same.
10983 (propagate_vr_across_jump_function): Same.
10984 * ipa-fnsummary.cc (evaluate_conditions_for_known_args): Same.
10985 (evaluate_properties_for_edge): Same.
10986 * ipa-prop.cc (ipa_vr::get_vrange): Same.
10987 (ipcp_update_vr): Same.
10988 * ipa-prop.h (ipa_value_range_from_jfunc): Same.
10989 (ipa_range_set_and_normalize): Same.
10991 2023-06-10 Georg-Johann Lay <avr@gjlay.de>
10995 * config/avr/avr-passes.def (avr_pass_ifelse): Insert new pass.
10996 * config/avr/avr.cc (avr_pass_ifelse): New RTL pass.
10997 (avr_pass_data_ifelse): New pass_data for it.
10998 (make_avr_pass_ifelse, avr_redundant_compare, avr_cbranch_cost)
10999 (avr_canonicalize_comparison, avr_out_plus_set_ZN)
11000 (avr_out_cmp_ext): New functions.
11001 (compare_condtition): Make sure REG_CC dies in the branch insn.
11002 (avr_rtx_costs_1): Add computation of cbranch costs.
11003 (avr_adjust_insn_length) [ADJUST_LEN_ADD_SET_ZN, ADJUST_LEN_CMP_ZEXT]:
11004 [ADJUST_LEN_CMP_SEXT]Handle them.
11005 (TARGET_CANONICALIZE_COMPARISON): New define.
11006 (avr_simplify_comparison_p, compare_diff_p, avr_compare_pattern)
11007 (avr_reorg_remove_redundant_compare, avr_reorg): Remove functions.
11008 (TARGET_MACHINE_DEPENDENT_REORG): Remove define.
11009 * config/avr/avr-protos.h (avr_simplify_comparison_p): Remove proto.
11010 (make_avr_pass_ifelse, avr_out_plus_set_ZN, cc_reg_rtx)
11011 (avr_out_cmp_zext): New Protos
11012 * config/avr/avr.md (branch, difficult_branch): Don't split insns.
11013 (*cbranchhi.zero-extend.0", *cbranchhi.zero-extend.1")
11014 (*swapped_tst<mode>, *add.for.eqne.<mode>): New insns.
11015 (*cbranch<mode>4): Rename to cbranch<mode>4_insn.
11016 (define_peephole): Add dead_or_set_regno_p(insn,REG_CC) as needed.
11017 (define_deephole2): Add peep2_regno_dead_p(*,REG_CC) as needed.
11018 Add new RTL peepholes for decrement-and-branch and *swapped_tst<mode>.
11019 Rework signtest-and-branch peepholes for *sbrx_branch<mode>.
11020 (adjust_len) [add_set_ZN, cmp_zext]: New.
11021 (QIPSI): New mode iterator.
11022 (ALLs1, ALLs2, ALLs4, ALLs234): New mode iterators.
11023 (gelt): New code iterator.
11024 (gelt_eqne): New code attribute.
11025 (rvbranch, *rvbranch, difficult_rvbranch, *difficult_rvbranch)
11026 (branch_unspec, *negated_tst<mode>, *reversed_tst<mode>)
11027 (*cmpqi_sign_extend): Remove insns.
11028 (define_c_enum "unspec") [UNSPEC_IDENTITY]: Remove.
11029 * config/avr/avr-dimode.md (cbranch<mode>4): Canonicalize comparisons.
11030 * config/avr/predicates.md (scratch_or_d_register_operand): New.
11031 * config/avr/constraints.md (Yxx): New constraint.
11033 2023-06-10 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11035 * config/riscv/autovec.md (select_vl<mode>): New pattern.
11036 * config/riscv/riscv-protos.h (expand_select_vl): New function.
11037 * config/riscv/riscv-v.cc (expand_select_vl): Ditto.
11039 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
11041 * range-op-float.cc (foperator_mult_div_base): Delete.
11042 (foperator_mult_div_base::find_range): Make static local function.
11043 (foperator_mult): Remove. Move prototypes to range-op-mixed.h
11044 (operator_mult::op1_range): Rename from foperator_mult.
11045 (operator_mult::op2_range): Ditto.
11046 (operator_mult::rv_fold): Ditto.
11047 (float_table::float_table): Remove MULT_EXPR.
11048 (class foperator_div): Inherit from range_operator.
11049 (float_table::float_table): Delete.
11050 * range-op-mixed.h (class operator_mult): Combined from integer
11052 * range-op.cc (float_tree_table): Delete.
11053 (op_mult): New object.
11054 (unified_table::unified_table): Add MULT_EXPR.
11055 (get_op_handler): Do not check float table any longer.
11056 (class cross_product_operator): Move to range-op-mixed.h.
11057 (class operator_mult): Move to range-op-mixed.h.
11058 (integral_table::integral_table): Remove MULT_EXPR.
11059 (pointer_table::pointer_table): Remove MULT_EXPR.
11060 * range-op.h (float_table): Remove.
11062 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
11064 * range-op-float.cc (foperator_negate): Remove. Move prototypes
11065 to range-op-mixed.h
11066 (operator_negate::fold_range): Rename from foperator_negate.
11067 (operator_negate::op1_range): Ditto.
11068 (float_table::float_table): Remove NEGATE_EXPR.
11069 * range-op-mixed.h (class operator_negate): Combined from integer
11071 * range-op.cc (op_negate): New object.
11072 (unified_table::unified_table): Add NEGATE_EXPR.
11073 (class operator_negate): Move to range-op-mixed.h.
11074 (integral_table::integral_table): Remove NEGATE_EXPR.
11075 (pointer_table::pointer_table): Remove NEGATE_EXPR.
11077 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
11079 * range-op-float.cc (foperator_minus): Remove. Move prototypes
11080 to range-op-mixed.h
11081 (operator_minus::fold_range): Rename from foperator_minus.
11082 (operator_minus::op1_range): Ditto.
11083 (operator_minus::op2_range): Ditto.
11084 (operator_minus::rv_fold): Ditto.
11085 (float_table::float_table): Remove MINUS_EXPR.
11086 * range-op-mixed.h (class operator_minus): Combined from integer
11088 * range-op.cc (op_minus): New object.
11089 (unified_table::unified_table): Add MINUS_EXPR.
11090 (class operator_minus): Move to range-op-mixed.h.
11091 (integral_table::integral_table): Remove MINUS_EXPR.
11092 (pointer_table::pointer_table): Remove MINUS_EXPR.
11094 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
11096 * range-op-float.cc (foperator_abs): Remove. Move prototypes
11097 to range-op-mixed.h
11098 (operator_abs::fold_range): Rename from foperator_abs.
11099 (operator_abs::op1_range): Ditto.
11100 (float_table::float_table): Remove ABS_EXPR.
11101 * range-op-mixed.h (class operator_abs): Combined from integer
11103 * range-op.cc (op_abs): New object.
11104 (unified_table::unified_table): Add ABS_EXPR.
11105 (class operator_abs): Move to range-op-mixed.h.
11106 (integral_table::integral_table): Remove ABS_EXPR.
11107 (pointer_table::pointer_table): Remove ABS_EXPR.
11109 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
11111 * range-op-float.cc (foperator_plus): Remove. Move prototypes
11112 to range-op-mixed.h
11113 (operator_plus::fold_range): Rename from foperator_plus.
11114 (operator_plus::op1_range): Ditto.
11115 (operator_plus::op2_range): Ditto.
11116 (operator_plus::rv_fold): Ditto.
11117 (float_table::float_table): Remove PLUS_EXPR.
11118 * range-op-mixed.h (class operator_plus): Combined from integer
11120 * range-op.cc (op_plus): New object.
11121 (unified_table::unified_table): Add PLUS_EXPR.
11122 (class operator_plus): Move to range-op-mixed.h.
11123 (integral_table::integral_table): Remove PLUS_EXPR.
11124 (pointer_table::pointer_table): Remove PLUS_EXPR.
11126 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
11128 * range-op-mixed.h (class operator_cast): Combined from integer
11130 * range-op.cc (op_cast): New object.
11131 (unified_table::unified_table): Add op_cast
11132 (class operator_cast): Move to range-op-mixed.h.
11133 (integral_table::integral_table): Remove op_cast
11134 (pointer_table::pointer_table): Remove op_cast.
11136 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
11138 * range-op-float.cc (operator_cst::fold_range): New.
11139 * range-op-mixed.h (class operator_cst): Move from integer file.
11140 * range-op.cc (op_cst): New object.
11141 (unified_table::unified_table): Add op_cst. Also use for REAL_CST.
11142 (class operator_cst): Move to range-op-mixed.h.
11143 (integral_table::integral_table): Remove op_cst.
11144 (pointer_table::pointer_table): Remove op_cst.
11146 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
11148 * range-op-float.cc (foperator_identity): Remove. Move prototypes
11149 to range-op-mixed.h
11150 (operator_identity::fold_range): Rename from foperator_identity.
11151 (operator_identity::op1_range): Ditto.
11152 (float_table::float_table): Remove fop_identity.
11153 * range-op-mixed.h (class operator_identity): Combined from integer
11155 * range-op.cc (op_identity): New object.
11156 (unified_table::unified_table): Add op_identity.
11157 (class operator_identity): Move to range-op-mixed.h.
11158 (integral_table::integral_table): Remove identity.
11159 (pointer_table::pointer_table): Remove identity.
11161 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
11163 * range-op-float.cc (foperator_ge): Remove. Move prototypes
11164 to range-op-mixed.h
11165 (operator_ge::fold_range): Rename from foperator_ge.
11166 (operator_ge::op1_range): Ditto.
11167 (float_table::float_table): Remove GE_EXPR.
11168 * range-op-mixed.h (class operator_ge): Combined from integer
11170 * range-op.cc (op_ge): New object.
11171 (unified_table::unified_table): Add GE_EXPR.
11172 (class operator_ge): Move to range-op-mixed.h.
11173 (ge_op1_op2_relation): Fold into
11174 operator_ge::op1_op2_relation.
11175 (integral_table::integral_table): Remove GE_EXPR.
11176 (pointer_table::pointer_table): Remove GE_EXPR.
11177 * range-op.h (ge_op1_op2_relation): Delete.
11179 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
11181 * range-op-float.cc (foperator_gt): Remove. Move prototypes
11182 to range-op-mixed.h
11183 (operator_gt::fold_range): Rename from foperator_gt.
11184 (operator_gt::op1_range): Ditto.
11185 (float_table::float_table): Remove GT_EXPR.
11186 * range-op-mixed.h (class operator_gt): Combined from integer
11188 * range-op.cc (op_gt): New object.
11189 (unified_table::unified_table): Add GT_EXPR.
11190 (class operator_gt): Move to range-op-mixed.h.
11191 (gt_op1_op2_relation): Fold into
11192 operator_gt::op1_op2_relation.
11193 (integral_table::integral_table): Remove GT_EXPR.
11194 (pointer_table::pointer_table): Remove GT_EXPR.
11195 * range-op.h (gt_op1_op2_relation): Delete.
11197 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
11199 * range-op-float.cc (foperator_le): Remove. Move prototypes
11200 to range-op-mixed.h
11201 (operator_le::fold_range): Rename from foperator_le.
11202 (operator_le::op1_range): Ditto.
11203 (float_table::float_table): Remove LE_EXPR.
11204 * range-op-mixed.h (class operator_le): Combined from integer
11206 * range-op.cc (op_le): New object.
11207 (unified_table::unified_table): Add LE_EXPR.
11208 (class operator_le): Move to range-op-mixed.h.
11209 (le_op1_op2_relation): Fold into
11210 operator_le::op1_op2_relation.
11211 (integral_table::integral_table): Remove LE_EXPR.
11212 (pointer_table::pointer_table): Remove LE_EXPR.
11213 * range-op.h (le_op1_op2_relation): Delete.
11215 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
11217 * range-op-float.cc (foperator_lt): Remove. Move prototypes
11218 to range-op-mixed.h
11219 (operator_lt::fold_range): Rename from foperator_lt.
11220 (operator_lt::op1_range): Ditto.
11221 (float_table::float_table): Remove LT_EXPR.
11222 * range-op-mixed.h (class operator_lt): Combined from integer
11224 * range-op.cc (op_lt): New object.
11225 (unified_table::unified_table): Add LT_EXPR.
11226 (class operator_lt): Move to range-op-mixed.h.
11227 (lt_op1_op2_relation): Fold into
11228 operator_lt::op1_op2_relation.
11229 (integral_table::integral_table): Remove LT_EXPR.
11230 (pointer_table::pointer_table): Remove LT_EXPR.
11231 * range-op.h (lt_op1_op2_relation): Delete.
11233 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
11235 * range-op-float.cc (foperator_not_equal): Remove. Move prototypes
11236 to range-op-mixed.h
11237 (operator_equal::fold_range): Rename from foperator_not_equal.
11238 (operator_equal::op1_range): Ditto.
11239 (float_table::float_table): Remove NE_EXPR.
11240 * range-op-mixed.h (class operator_not_equal): Combined from integer
11242 * range-op.cc (op_equal): New object.
11243 (unified_table::unified_table): Add NE_EXPR.
11244 (class operator_not_equal): Move to range-op-mixed.h.
11245 (not_equal_op1_op2_relation): Fold into
11246 operator_not_equal::op1_op2_relation.
11247 (integral_table::integral_table): Remove NE_EXPR.
11248 (pointer_table::pointer_table): Remove NE_EXPR.
11249 * range-op.h (not_equal_op1_op2_relation): Delete.
11251 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
11253 * range-op-float.cc (foperator_equal): Remove. Move prototypes
11254 to range-op-mixed.h
11255 (operator_equal::fold_range): Rename from foperator_equal.
11256 (operator_equal::op1_range): Ditto.
11257 (float_table::float_table): Remove EQ_EXPR.
11258 * range-op-mixed.h (class operator_equal): Combined from integer
11260 * range-op.cc (op_equal): New object.
11261 (unified_table::unified_table): Add EQ_EXPR.
11262 (class operator_equal): Move to range-op-mixed.h.
11263 (equal_op1_op2_relation): Fold into
11264 operator_equal::op1_op2_relation.
11265 (integral_table::integral_table): Remove EQ_EXPR.
11266 (pointer_table::pointer_table): Remove EQ_EXPR.
11267 * range-op.h (equal_op1_op2_relation): Delete.
11269 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
11271 * range-op-float.cc (class float_table): Move to header.
11272 (float_table::float_table): Move float only operators to...
11273 (range_op_table::initialize_float_ops): Here.
11274 * range-op-mixed.h: New.
11275 * range-op.cc (integral_tree_table, pointer_tree_table): Moved
11277 (float_tree_table): Moved from range-op-float.cc.
11278 (unified_tree_table): New.
11279 (unified_table::unified_table): New. Call initialize routines.
11280 (get_op_handler): Check unified table first.
11281 (range_op_handler::range_op_handler): Handle no type constructor.
11282 (integral_table::integral_table): Move integral only operators to...
11283 (range_op_table::initialize_integral_ops): Here.
11284 (pointer_table::pointer_table): Move pointer only operators to...
11285 (range_op_table::initialize_pointer_ops): Here.
11286 * range-op.h (enum bool_range_state): Move to range-op-mixed.h.
11287 (get_bool_state): Ditto.
11288 (empty_range_varying): Ditto.
11289 (relop_early_resolve): Ditto.
11290 (class range_op_table): Add new init methods for range types.
11291 (class integral_table): Move declaration to here.
11292 (class pointer_table): Move declaration to here.
11293 (class float_table): Move declaration to here.
11295 2023-06-09 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
11296 Richard Sandiford <richard.sandiford@arm.com>
11297 Richard Biener <rguenther@suse.de>
11299 * doc/md.texi: Add SELECT_VL support.
11300 * internal-fn.def (SELECT_VL): Ditto.
11301 * optabs.def (OPTAB_D): Ditto.
11302 * tree-vect-loop-manip.cc (vect_set_loop_controls_directly): Ditto.
11303 * tree-vect-loop.cc (_loop_vec_info::_loop_vec_info): Ditto.
11304 * tree-vect-stmts.cc (get_select_vl_data_ref_ptr): Ditto.
11305 (vectorizable_store): Ditto.
11306 (vectorizable_load): Ditto.
11307 * tree-vectorizer.h (LOOP_VINFO_USING_SELECT_VL_P): Ditto.
11309 2023-06-09 Andrew MacLeod <amacleod@redhat.com>
11312 * ipa-prop.cc (ipa_compute_jump_functions_for_edge): Check param
11315 2023-06-09 Andrew MacLeod <amacleod@redhat.com>
11317 * range-op.cc (range_cast): Move to...
11318 * range-op.h (range_cast): Here and add generic a version.
11320 2023-06-09 Marek Polacek <polacek@redhat.com>
11324 * doc/invoke.texi: Clarify that -Wmissing-field-initializers doesn't
11325 warn about designated initializers in C only.
11327 2023-06-09 Andrew Pinski <apinski@marvell.com>
11329 PR tree-optimization/97711
11330 PR tree-optimization/110155
11331 * match.pd ((zero_one == 0) ? y : z <op> y): Add plus to the op.
11332 ((zero_one != 0) ? z <op> y : y): Likewise.
11334 2023-06-09 Andrew Pinski <apinski@marvell.com>
11336 * match.pd ((zero_one ==/!= 0) ? y : z <op> y): Use
11337 multiply rather than negation/bit_and.
11339 2023-06-09 Andrew Pinski <apinski@marvell.com>
11341 * match.pd (`X & -Y -> X * Y`): Allow for truncation
11342 and the same type for unsigned types.
11344 2023-06-09 Andrew Pinski <apinski@marvell.com>
11346 PR tree-optimization/110165
11347 PR tree-optimization/110166
11348 * match.pd (zero_one_valued_p): Don't accept
11349 signed 1-bit integers.
11351 2023-06-09 Richard Biener <rguenther@suse.de>
11353 * match.pd (two conversions in a row): Use element_precision
11354 to DTRT for VECTOR_TYPE.
11356 2023-06-09 Pan Li <pan2.li@intel.com>
11358 * config/riscv/riscv.md (enabled): Move to another place, and
11359 add fp_vector_disabled to the cond.
11360 (fp_vector_disabled): New attr defined for disabling fp.
11361 * config/riscv/vector-iterators.md: Fix V_WHOLE and V_FRACT.
11363 2023-06-09 Pan Li <pan2.li@intel.com>
11365 * config/riscv/riscv-protos.h (enum frm_field_enum): Adjust
11368 2023-06-09 liuhongt <hongtao.liu@intel.com>
11371 * config/i386/i386.cc (ix86_gimple_fold_builtin): Explicitly
11372 view_convert_expr mask to signed type when folding pblendvb
11375 2023-06-09 liuhongt <hongtao.liu@intel.com>
11378 * config/i386/i386.cc (ix86_gimple_fold_builtin): Fold
11379 _mm{,256,512}_abs_{epi8,epi16,epi32,epi64} into gimple
11380 ABSU_EXPR + VCE, don't fold _mm_abs_{pi8,pi16,pi32} w/o
11382 * config/i386/i386-builtin.def: Replace CODE_FOR_nothing with
11383 real codename for __builtin_ia32_pabs{b,w,d}.
11385 2023-06-08 Andrew MacLeod <amacleod@redhat.com>
11387 * gimple-range-op.cc
11388 (gimple_range_op_handler::gimple_range_op_handler): Adjust.
11389 (gimple_range_op_handler::maybe_builtin_call): Adjust.
11390 * gimple-range-op.h (operand1, operand2): Use m_operator.
11391 * range-op.cc (integral_table, pointer_table): Relocate.
11392 (get_op_handler): Rename from get_handler and handle all types.
11393 (range_op_handler::range_op_handler): Relocate.
11394 (range_op_handler::set_op_handler): Relocate and adjust.
11395 (range_op_handler::range_op_handler): Relocate.
11396 (dispatch_trio): New.
11397 (RO_III, RO_IFI, RO_IFF, RO_FFF, RO_FIF, RO_FII): New consts.
11398 (range_op_handler::dispatch_kind): New.
11399 (range_op_handler::fold_range): Relocate and Use new dispatch value.
11400 (range_op_handler::op1_range): Ditto.
11401 (range_op_handler::op2_range): Ditto.
11402 (range_op_handler::lhs_op1_relation): Ditto.
11403 (range_op_handler::lhs_op2_relation): Ditto.
11404 (range_op_handler::op1_op2_relation): Ditto.
11405 (range_op_handler::set_op_handler): Use m_operator member.
11406 * range-op.h (range_op_handler::operator bool): Use m_operator.
11407 (range_op_handler::dispatch_kind): New.
11408 (range_op_handler::m_valid): Delete.
11409 (range_op_handler::m_int): Delete
11410 (range_op_handler::m_float): Delete
11411 (range_op_handler::m_operator): New.
11412 (range_op_table::operator[]): Relocate from .cc file.
11413 (range_op_table::set): Ditto.
11414 * value-range.h (class vrange): Make range_op_handler a friend.
11416 2023-06-08 Andrew MacLeod <amacleod@redhat.com>
11418 * gimple-range-op.cc (cfn_constant_float_p): Change base class.
11419 (cfn_pass_through_arg1): Adjust using statemenmt.
11420 (cfn_signbit): Change base class, adjust using statement.
11421 (cfn_copysign): Ditto.
11423 (cfn_sincos): Ditto.
11424 * range-op-float.cc (fold_range): Change class to range_operator.
11428 (lhs_op1_relation): Ditto.
11429 (lhs_op2_relation): Ditto.
11430 (op1_op2_relation): Ditto.
11431 (foperator_*): Ditto.
11432 (class float_table): New. Inherit from range_op_table.
11433 (floating_tree_table) Change to range_op_table pointer.
11434 (class floating_op_table): Delete.
11435 * range-op.cc (operator_equal): Adjust using statement.
11436 (operator_not_equal): Ditto.
11437 (operator_lt, operator_le, operator_gt, operator_ge): Ditto.
11438 (operator_minus, operator_cast): Ditto.
11439 (operator_bitwise_and, pointer_plus_operator): Ditto.
11440 (get_float_handle): Change return type.
11441 * range-op.h (range_operator_float): Delete. Relocate all methods
11442 into class range_operator.
11443 (range_op_handler::m_float): Change type to range_operator.
11444 (floating_op_table): Delete.
11445 (floating_tree_table): Change type.
11447 2023-06-08 Andrew MacLeod <amacleod@redhat.com>
11449 * range-op.cc (range_operator::fold_range): Call virtual routine.
11450 (range_operator::update_bitmask): New.
11451 (operator_equal::update_bitmask): New.
11452 (operator_not_equal::update_bitmask): New.
11453 (operator_lt::update_bitmask): New.
11454 (operator_le::update_bitmask): New.
11455 (operator_gt::update_bitmask): New.
11456 (operator_ge::update_bitmask): New.
11457 (operator_ge::update_bitmask): New.
11458 (operator_plus::update_bitmask): New.
11459 (operator_minus::update_bitmask): New.
11460 (operator_pointer_diff::update_bitmask): New.
11461 (operator_min::update_bitmask): New.
11462 (operator_max::update_bitmask): New.
11463 (operator_mult::update_bitmask): New.
11464 (operator_div:operator_div):New.
11465 (operator_div::update_bitmask): New.
11466 (operator_div::m_code): New member.
11467 (operator_exact_divide::operator_exact_divide): New constructor.
11468 (operator_lshift::update_bitmask): New.
11469 (operator_rshift::update_bitmask): New.
11470 (operator_bitwise_and::update_bitmask): New.
11471 (operator_bitwise_or::update_bitmask): New.
11472 (operator_bitwise_xor::update_bitmask): New.
11473 (operator_trunc_mod::update_bitmask): New.
11474 (op_ident, op_unknown, op_ptr_min_max): New.
11475 (op_nop, op_convert): Delete.
11476 (op_ssa, op_paren, op_obj_type): Delete.
11477 (op_realpart, op_imagpart): Delete.
11478 (op_ptr_min, op_ptr_max): Delete.
11479 (pointer_plus_operator:update_bitmask): New.
11480 (range_op_table::set): Do not use m_code.
11481 (integral_table::integral_table): Adjust to single instances.
11482 * range-op.h (range_operator::range_operator): Delete.
11483 (range_operator::m_code): Delete.
11484 (range_operator::update_bitmask): New.
11486 2023-06-08 Andrew MacLeod <amacleod@redhat.com>
11488 * range-op-float.cc (range_operator_float::fold_range): Return
11489 NAN of the result type.
11491 2023-06-08 Jakub Jelinek <jakub@redhat.com>
11493 * optabs.cc (expand_ffs): Add forward declaration.
11494 (expand_doubleword_clz): Rename to ...
11495 (expand_doubleword_clz_ctz_ffs): ... this. Add UNOPTAB argument,
11496 handle also doubleword CTZ and FFS in addition to CLZ.
11497 (expand_unop): Adjust caller. Also call it for doubleword
11498 ctz_optab and ffs_optab.
11500 2023-06-08 Jakub Jelinek <jakub@redhat.com>
11503 * config/i386/i386-expand.cc (ix86_expand_vector_init_general): For
11504 n_words == 2 recurse with mmx_ok as first argument rather than false.
11506 2023-06-07 Roger Sayle <roger@nextmovesoftware.com>
11508 * wide-int.cc (wi::bitreverse_large): Use HOST_WIDE_INT_1U to
11509 avoid sign extension/undefined behaviour when setting each bit.
11511 2023-06-07 Roger Sayle <roger@nextmovesoftware.com>
11512 Uros Bizjak <ubizjak@gmail.com>
11514 * config/i386/i386-expand.cc (ix86_expand_builtin) <handlecarry>:
11515 Use new x86_stc instruction when the carry flag must be set.
11516 * config/i386/i386.cc (ix86_cc_mode): Use CCCmode for *x86_cmc.
11517 (ix86_rtx_costs): Provide accurate rtx_costs for *x86_cmc.
11518 * config/i386/i386.h (TARGET_SLOW_STC): New define.
11519 * config/i386/i386.md (UNSPEC_STC): New UNSPEC for stc.
11520 (x86_stc): New define_insn.
11521 (define_peephole2): Convert x86_stc into alternate implementation
11522 on pentium4 without -Os when a QImode register is available.
11523 (*x86_cmc): New define_insn.
11524 (define_peephole2): Convert *x86_cmc into alternate implementation
11525 on pentium4 without -Os when a QImode register is available.
11526 (*setccc): New define_insn_and_split for a no-op CCCmode move.
11527 (*setcc_qi_negqi_ccc_1_<mode>): New define_insn_and_split to
11528 recognize (and eliminate) the carry flag being copied to itself.
11529 (*setcc_qi_negqi_ccc_2_<mode>): Likewise.
11530 * config/i386/x86-tune.def (X86_TUNE_SLOW_STC): New tuning flag.
11532 2023-06-07 Andrew Pinski <apinski@marvell.com>
11534 * match.pd: Fix comment for the
11535 `(zero_one ==/!= 0) ? y : z <op> y` patterns.
11537 2023-06-07 Jeff Law <jlaw@ventanamicro.com>
11538 Jeff Law <jlaw@ventanamicro.com>
11540 * config/riscv/bitmanip.md (rotrdi3, rotrsi3, rotlsi3): New expanders.
11541 (rotrsi3_sext): Expose generator.
11542 (rotlsi3 pattern): Hide generator.
11543 * config/riscv/riscv-protos.h (riscv_emit_binary): New function
11545 * config/riscv/riscv.cc (riscv_emit_binary): Removed static
11546 * config/riscv/riscv.md (addsi3, subsi3, negsi2): Hide generator.
11547 (mulsi3, <optab>si3): Likewise.
11548 (addsi3, subsi3, negsi2, mulsi3, <optab>si3): New expanders.
11549 (addv<mode>4, subv<mode>4, mulv<mode>4): Use riscv_emit_binary.
11550 (<u>mulsidi3): Likewise.
11551 (addsi3_extended, subsi3_extended, negsi2_extended): Expose generator.
11552 (mulsi3_extended, <optab>si3_extended): Likewise.
11553 (splitter for shadd feeding divison): Update RTL pattern to account
11554 for changes in how 32 bit ops are expanded for TARGET_64BIT.
11555 * loop-iv.cc (get_biv_step_1): Process src of extension when it PLUS.
11557 2023-06-07 Dimitar Dimitrov <dimitar@dinux.eu>
11560 * config/riscv/riscv.cc (riscv_print_operand): Calculate
11561 memmodel only when it is valid.
11563 2023-06-07 Dimitar Dimitrov <dimitar@dinux.eu>
11565 * config/riscv/riscv.cc (riscv_const_insns): Recursively call
11566 for constant element of a vector.
11568 2023-06-07 Jakub Jelinek <jakub@redhat.com>
11570 * match.pd (zero_one_valued_p): Don't handle integer_zerop specially,
11571 instead compare tree_nonzero_bits <= 1U rather than just == 1.
11573 2023-06-07 Alex Coplan <alex.coplan@arm.com>
11576 * config/aarch64/aarch64-builtins.cc (aarch64_general_simulate_builtin):
11578 (aarch64_init_ls64_builtins): ... here. Switch to declaring public ACLE
11579 names for builtins.
11580 (aarch64_general_init_builtins): Ensure we invoke the arm_acle.h
11581 setup if in_lto_p, just like we do for SVE.
11582 * config/aarch64/arm_acle.h: (__arm_ld64b): Delete.
11583 (__arm_st64b): Delete.
11584 (__arm_st64bv): Delete.
11585 (__arm_st64bv0): Delete.
11587 2023-06-07 Alex Coplan <alex.coplan@arm.com>
11590 * config/aarch64/aarch64-builtins.cc (aarch64_expand_builtin_ls64):
11591 Use input operand for the destination address.
11592 * config/aarch64/aarch64.md (st64b): Fix constraint on address
11595 2023-06-07 Alex Coplan <alex.coplan@arm.com>
11598 * config/aarch64/aarch64-builtins.cc (aarch64_init_ls64_builtins_types):
11599 Replace eight consecutive spaces with tabs.
11600 (aarch64_init_ls64_builtins): Likewise.
11601 (aarch64_expand_builtin_ls64): Likewise.
11602 * config/aarch64/aarch64.md (ld64b): Likewise.
11605 (st64bv0): Likewise.
11607 2023-06-07 Vladimir N. Makarov <vmakarov@redhat.com>
11609 * ira-costs.cc: (find_costs_and_classes): Constrain classes of pic
11610 offset table pseudo to a general reg subset.
11612 2023-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11614 * config/aarch64/aarch64-simd.md (aarch64_sqmovun<mode><vczle><vczbe>):
11616 (*aarch64_sqmovun<mode>_insn<vczle><vczbe>): ... This. Reimplement
11618 (aarch64_sqmovun<mode> [SD_HSDI]): Reimplement with RTL codes.
11619 (aarch64_sqxtun2<mode>_le): Likewise.
11620 (aarch64_sqxtun2<mode>_be): Likewise.
11621 (aarch64_sqxtun2<mode>): Adjust for the above.
11622 (aarch64_sqmovun<mode>): New define_expand.
11623 * config/aarch64/iterators.md (UNSPEC_SQXTUN): Delete.
11624 (half_mask): New mode attribute.
11625 * config/aarch64/predicates.md (aarch64_simd_umax_half_mode):
11628 2023-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11630 * config/aarch64/aarch64-simd.md (aarch64_addp<mode><vczle><vczbe>):
11632 (aarch64_addp<mode>_insn): ... This...
11633 (aarch64_addp<mode><vczle><vczbe>_insn): ... And this.
11634 (aarch64_addp<mode>): New define_expand.
11636 2023-06-07 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11638 * config/riscv/riscv-protos.h (expand_vec_perm_const): New function.
11639 * config/riscv/riscv-v.cc
11640 (rvv_builder::can_duplicate_repeating_sequence_p): Support POLY
11642 (rvv_builder::single_step_npatterns_p): New function.
11643 (rvv_builder::npatterns_all_equal_p): Ditto.
11644 (const_vec_all_in_range_p): Support POLY handling.
11645 (gen_const_vector_dup): Ditto.
11646 (emit_vlmax_gather_insn): Add vrgatherei16.
11647 (emit_vlmax_masked_gather_mu_insn): Ditto.
11648 (expand_const_vector): Add VLA SLP const vector support.
11649 (expand_vec_perm): Support POLY.
11650 (struct expand_vec_perm_d): New struct.
11651 (shuffle_generic_patterns): New function.
11652 (expand_vec_perm_const_1): Ditto.
11653 (expand_vec_perm_const): Ditto.
11654 * config/riscv/riscv.cc (riscv_vectorize_vec_perm_const): Ditto.
11655 (TARGET_VECTORIZE_VEC_PERM_CONST): New targethook.
11657 2023-06-07 Andrew Pinski <apinski@marvell.com>
11659 PR middle-end/110117
11660 * expr.cc (expand_single_bit_test): Handle
11661 const_int from expand_expr.
11663 2023-06-07 Andrew Pinski <apinski@marvell.com>
11665 * expr.cc (do_store_flag): Rearrange the
11666 TER code so that it overrides the nonzero bits
11667 info if we had `a & POW2`.
11669 2023-06-07 Andrew Pinski <apinski@marvell.com>
11671 PR tree-optimization/110134
11672 * match.pd (-A CMP -B -> B CMP A): Allow EQ/NE for all integer
11674 (-A CMP CST -> B CMP (-CST)): Likewise.
11676 2023-06-07 Andrew Pinski <apinski@marvell.com>
11678 PR tree-optimization/89263
11679 PR tree-optimization/99069
11680 PR tree-optimization/20083
11681 PR tree-optimization/94898
11682 * match.pd: Add patterns to optimize `a ? onezero : onezero` with
11683 one of the operands are constant.
11685 2023-06-07 Andrew Pinski <apinski@marvell.com>
11687 * match.pd (zero_one_valued_p): Match 0 integer constant
11690 2023-06-07 Pan Li <pan2.li@intel.com>
11692 * config/riscv/riscv-vector-builtins-types.def
11693 (vfloat32mf2_t): Take RVV_REQUIRE_ELEN_FP_16 as requirement.
11694 (vfloat32m1_t): Ditto.
11695 (vfloat32m2_t): Ditto.
11696 (vfloat32m4_t): Ditto.
11697 (vfloat32m8_t): Ditto.
11698 (vint16mf4_t): Ditto.
11699 (vint16mf2_t): Ditto.
11700 (vint16m1_t): Ditto.
11701 (vint16m2_t): Ditto.
11702 (vint16m4_t): Ditto.
11703 (vint16m8_t): Ditto.
11704 (vuint16mf4_t): Ditto.
11705 (vuint16mf2_t): Ditto.
11706 (vuint16m1_t): Ditto.
11707 (vuint16m2_t): Ditto.
11708 (vuint16m4_t): Ditto.
11709 (vuint16m8_t): Ditto.
11710 (vint32mf2_t): Ditto.
11711 (vint32m1_t): Ditto.
11712 (vint32m2_t): Ditto.
11713 (vint32m4_t): Ditto.
11714 (vint32m8_t): Ditto.
11715 (vuint32mf2_t): Ditto.
11716 (vuint32m1_t): Ditto.
11717 (vuint32m2_t): Ditto.
11718 (vuint32m4_t): Ditto.
11719 (vuint32m8_t): Ditto.
11721 2023-06-07 Jason Merrill <jason@redhat.com>
11724 * doc/invoke.texi: Document it.
11726 2023-06-06 Roger Sayle <roger@nextmovesoftware.com>
11728 * doc/rtl.texi (bitreverse, copysign): Document new RTX codes.
11729 * rtl.def (BITREVERSE, COPYSIGN): Define new RTX codes.
11730 * simplify-rtx.cc (simplify_unary_operation_1): Optimize
11731 NOT (BITREVERSE x) as BITREVERSE (NOT x).
11732 Optimize POPCOUNT (BITREVERSE x) as POPCOUNT x.
11733 Optimize PARITY (BITREVERSE x) as PARITY x.
11734 Optimize BITREVERSE (BITREVERSE x) as x.
11735 (simplify_const_unary_operation) <case BITREVERSE>: Evaluate
11736 BITREVERSE of a constant integer at compile-time.
11737 (simplify_binary_operation_1) <case COPYSIGN>: Optimize
11738 COPY_SIGN (x, x) as x. Optimize COPYSIGN (x, C) as ABS x
11739 or NEG (ABS x) for constant C. Optimize COPYSIGN (ABS x, y)
11740 and COPYSIGN (NEG x, y) as COPYSIGN (x, y).
11741 Optimize COPYSIGN (x, ABS y) as ABS x.
11742 Optimize COPYSIGN (COPYSIGN (x, y), z) as COPYSIGN (x, z).
11743 Optimize COPYSIGN (x, COPYSIGN (y, z)) as COPYSIGN (x, z).
11744 (simplify_const_binary_operation): Evaluate COPYSIGN of constant
11745 arguments at compile-time.
11747 2023-06-06 Uros Bizjak <ubizjak@gmail.com>
11749 * rtl.h (function_invariant_p): Change return type from int to bool.
11750 * reload1.cc (function_invariant_p): Change return type from
11751 int to bool and adjust function body accordingly.
11753 2023-06-06 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11755 * config/riscv/autovec-opt.md (*<optab>_fma<mode>): New pattern.
11756 (*single_<optab>mult_plus<mode>): Ditto.
11757 (*double_<optab>mult_plus<mode>): Ditto.
11758 (*sign_zero_extend_fma): Ditto.
11759 (*zero_sign_extend_fma): Ditto.
11760 * config/riscv/riscv-protos.h (enum insn_type): New enum.
11762 2023-06-06 Kwok Cheung Yeung <kcy@codesourcery.com>
11763 Tobias Burnus <tobias@codesourcery.com>
11765 * gimplify.cc (omp_notice_variable): Apply GOVD_MAP_ALLOC_ONLY flag
11766 and defaultmap flags if the defaultmap has GOVD_MAP_FORCE_PRESENT flag
11768 (omp_get_attachment): Handle map clauses with 'present' modifier.
11769 (omp_group_base): Likewise.
11770 (gimplify_scan_omp_clauses): Reorder present maps to come first.
11771 Set GOVD flags for present defaultmaps.
11772 (gimplify_adjust_omp_clauses_1): Set map kind for present defaultmaps.
11773 * omp-low.cc (scan_sharing_clauses): Handle 'always, present' map
11775 (lower_omp_target): Handle map clauses with 'present' modifier.
11776 Handle 'to' and 'from' clauses with 'present'.
11777 * tree-core.h (enum omp_clause_defaultmap_kind): Add
11778 OMP_CLAUSE_DEFAULTMAP_PRESENT defaultmap kind.
11779 * tree-pretty-print.cc (dump_omp_clause): Handle 'map', 'to' and
11780 'from' clauses with 'present' modifier. Handle present defaultmap.
11781 * tree.h (OMP_CLAUSE_MOTION_PRESENT): New #define.
11783 2023-06-06 Segher Boessenkool <segher@kernel.crashing.org>
11785 * config/rs6000/genfusion.pl: Delete some dead code.
11787 2023-06-06 Segher Boessenkool <segher@kernel.crashing.org>
11789 * config/rs6000/genfusion.pl (gen_ld_cmpi_p10_one): New, rewritten and
11791 (gen_ld_cmpi_p10): ... this.
11793 2023-06-06 Jeevitha Palanisamy <jeevitha@linux.ibm.com>
11796 * config/rs6000/rs6000.cc (vec_const_128bit_to_bytes): Remove
11797 duplicate expression.
11799 2023-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11801 * config/aarch64/aarch64-builtins.cc (aarch64_general_gimple_fold_builtin):
11802 Handle unsigned reduc_plus_scal_ builtins.
11803 * config/aarch64/aarch64-simd-builtins.def (addp): Delete DImode instances.
11804 * config/aarch64/aarch64-simd.md (aarch64_addpdi): Delete.
11805 * config/aarch64/arm_neon.h (vpaddd_s64): Reimplement with
11806 __builtin_aarch64_reduc_plus_scal_v2di.
11807 (vpaddd_u64): Reimplement with __builtin_aarch64_reduc_plus_scal_v2di_uu.
11809 2023-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11811 * config/aarch64/aarch64-simd.md (aarch64_<sur>shr_n<mode>): Delete.
11812 (aarch64_<sra_op>rshr_n<mode><vczle><vczbe>_insn): New define_insn.
11813 (aarch64_<sra_op>rshr_n<mode>): New define_expand.
11815 2023-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11817 * config/aarch64/aarch64-simd.md (aarch64_shrn<mode>_insn_le): Delete.
11818 (aarch64_shrn<mode>_insn_be): Delete.
11819 (*aarch64_<srn_op>shrn<mode>_vect): Rename to...
11820 (*aarch64_<srn_op>shrn<mode><vczle><vczbe>): ... This.
11821 (aarch64_shrn<mode>): Remove reference to the above deleted patterns.
11822 (aarch64_rshrn<mode>_insn_le): Delete.
11823 (aarch64_rshrn<mode>_insn_be): Delete.
11824 (aarch64_rshrn<mode><vczle><vczbe>_insn): New define_insn.
11825 (aarch64_rshrn<mode>): Remove references to the above deleted patterns.
11827 2023-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11829 * config/aarch64/aarch64-protos.h (aarch64_parallel_select_half_p):
11831 (aarch64_pars_overlap_p): Likewise.
11832 * config/aarch64/aarch64-simd.md (aarch64_<su>addlv<mode>):
11833 Express in terms of UNSPEC_ADDV.
11834 (*aarch64_<su>addlv<VDQV_L:mode>_ze<GPI:mode>): Likewise.
11835 (*aarch64_<su>addlv<mode>_reduction): Define.
11836 (*aarch64_uaddlv<mode>_reduction_2): Likewise.
11837 * config/aarch64/aarch64.cc (aarch64_parallel_select_half_p): Define.
11838 (aarch64_pars_overlap_p): Likewise.
11839 * config/aarch64/iterators.md (UNSPEC_SADDLV, UNSPEC_UADDLV): Delete.
11840 (VQUADW): New mode attribute.
11841 (VWIDE2X_S): Likewise.
11843 (su): Delete handling of UNSPEC_SADDLV, UNSPEC_UADDLV.
11844 * config/aarch64/predicates.md (vect_par_cnst_select_half): Define.
11846 2023-06-06 Richard Biener <rguenther@suse.de>
11848 PR middle-end/110055
11849 * gimplify.cc (gimplify_target_expr): Do not emit
11850 CLOBBERs for variables which have static storage duration
11851 after gimplifying their initializers.
11853 2023-06-06 Richard Biener <rguenther@suse.de>
11855 PR tree-optimization/109143
11856 * tree-ssa-structalias.cc (solution_set_expand): Avoid
11857 one bitmap iteration and optimize bit range setting.
11859 2023-06-06 Hans-Peter Nilsson <hp@axis.com>
11861 PR bootstrap/110120
11862 * postreload.cc (reload_cse_move2add, move2add_use_add2_insn): Use
11863 XVECEXP, not XEXP, to access first item of a PARALLEL.
11865 2023-06-06 Pan Li <pan2.li@intel.com>
11867 * config/riscv/riscv-vector-builtins-types.def
11868 (vfloat16mf4_t): Add vfloat16mf4_t to WF operations.
11869 (vfloat16mf2_t): Likewise.
11870 (vfloat16m1_t): Likewise.
11871 (vfloat16m2_t): Likewise.
11872 (vfloat16m4_t): Likewise.
11873 (vfloat16m8_t): Likewise.
11874 * config/riscv/vector-iterators.md: Add FP=16 to VWF, VWF_ZVE64,
11875 VWLMUL1, VWLMUL1_ZVE64, vwlmul1 and vwlmul1_zve64.
11877 2023-06-06 Fei Gao <gaofei@eswincomputing.com>
11879 * config/riscv/riscv.cc (riscv_adjust_libcall_cfi_prologue): Use Pmode
11880 for cfi reg/mem machmode
11881 (riscv_adjust_libcall_cfi_epilogue): Use Pmode for cfi reg machmode
11883 2023-06-06 Li Xu <xuli1@eswincomputing.com>
11885 * config/riscv/vector-iterators.md:
11886 Fix 'REQUIREMENT' for machine_mode 'MODE'.
11887 * config/riscv/vector.md (@pred_indexed_<order>store<VNX16_QHS:mode>
11888 <VNX16_QHSI:mode>): change VNX16_QHSI to VNX16_QHSDI.
11889 (@pred_indexed_<order>store<VNX16_QHS:mode><VNX16_QHSDI:mode>): Ditto.
11891 2023-06-06 Pan Li <pan2.li@intel.com>
11893 * config/riscv/vector-iterators.md: Fix typo in mode attr.
11895 2023-06-05 Andre Vieira <andre.simoesdiasvieira@arm.com>
11896 Joel Hutton <joel.hutton@arm.com>
11898 * doc/generic.texi: Remove old tree codes.
11899 * expr.cc (expand_expr_real_2): Remove old tree code cases.
11900 * gimple-pretty-print.cc (dump_binary_rhs): Likewise.
11901 * optabs-tree.cc (optab_for_tree_code): Likewise.
11902 (supportable_half_widening_operation): Likewise.
11903 * tree-cfg.cc (verify_gimple_assign_binary): Likewise.
11904 * tree-inline.cc (estimate_operator_cost): Likewise.
11905 (op_symbol_code): Likewise.
11906 * tree-vect-data-refs.cc (vect_get_smallest_scalar_type): Likewise.
11907 (vect_analyze_data_ref_accesses): Likewise.
11908 * tree-vect-generic.cc (expand_vector_operations_1): Likewise.
11909 * cfgexpand.cc (expand_debug_expr): Likewise.
11910 * tree-vect-stmts.cc (vectorizable_conversion): Likewise.
11911 (supportable_widening_operation): Likewise.
11912 * gimple-range-op.cc (gimple_range_op_handler::maybe_non_standard):
11914 * optabs.def (vec_widen_ssubl_hi_optab, vec_widen_ssubl_lo_optab,
11915 vec_widen_saddl_hi_optab, vec_widen_saddl_lo_optab,
11916 vec_widen_usubl_hi_optab, vec_widen_usubl_lo_optab,
11917 vec_widen_uaddl_hi_optab, vec_widen_uaddl_lo_optab): Remove optabs.
11918 * tree-pretty-print.cc (dump_generic_node): Remove tree code definition.
11919 * tree.def (WIDEN_PLUS_EXPR, WIDEN_MINUS_EXPR, VEC_WIDEN_PLUS_HI_EXPR,
11920 VEC_WIDEN_PLUS_LO_EXPR, VEC_WIDEN_MINUS_HI_EXPR,
11921 VEC_WIDEN_MINUS_LO_EXPR): Likewise.
11923 2023-06-05 Andre Vieira <andre.simoesdiasvieira@arm.com>
11924 Joel Hutton <joel.hutton@arm.com>
11925 Tamar Christina <tamar.christina@arm.com>
11927 * config/aarch64/aarch64-simd.md (vec_widen_<su>addl_lo_<mode>): Rename
11929 (vec_widen_<su>add_lo_<mode>): ... to this.
11930 (vec_widen_<su>addl_hi_<mode>): Rename this ...
11931 (vec_widen_<su>add_hi_<mode>): ... to this.
11932 (vec_widen_<su>subl_lo_<mode>): Rename this ...
11933 (vec_widen_<su>sub_lo_<mode>): ... to this.
11934 (vec_widen_<su>subl_hi_<mode>): Rename this ...
11935 (vec_widen_<su>sub_hi_<mode>): ...to this.
11936 * doc/generic.texi: Document new IFN codes.
11937 * internal-fn.cc (lookup_hilo_internal_fn): Add lookup function.
11938 (commutative_binary_fn_p): Add widen_plus fn's.
11939 (widening_fn_p): New function.
11940 (narrowing_fn_p): New function.
11941 (direct_internal_fn_optab): Change visibility.
11942 * internal-fn.def (DEF_INTERNAL_WIDENING_OPTAB_FN): Macro to define an
11943 internal_fn that expands into multiple internal_fns for widening.
11944 (IFN_VEC_WIDEN_PLUS, IFN_VEC_WIDEN_PLUS_HI, IFN_VEC_WIDEN_PLUS_LO,
11945 IFN_VEC_WIDEN_PLUS_EVEN, IFN_VEC_WIDEN_PLUS_ODD,
11946 IFN_VEC_WIDEN_MINUS, IFN_VEC_WIDEN_MINUS_HI,
11947 IFN_VEC_WIDEN_MINUS_LO, IFN_VEC_WIDEN_MINUS_ODD,
11948 IFN_VEC_WIDEN_MINUS_EVEN): Define widening plus,minus functions.
11949 * internal-fn.h (direct_internal_fn_optab): Declare new prototype.
11950 (lookup_hilo_internal_fn): Likewise.
11951 (widening_fn_p): Likewise.
11952 (Narrowing_fn_p): Likewise.
11953 * optabs.cc (commutative_optab_p): Add widening plus optabs.
11954 * optabs.def (OPTAB_D): Define widen add, sub optabs.
11955 * tree-vect-patterns.cc (vect_recog_widen_op_pattern): Support
11956 patterns with a hi/lo or even/odd split.
11957 (vect_recog_sad_pattern): Refactor to use new IFN codes.
11958 (vect_recog_widen_plus_pattern): Likewise.
11959 (vect_recog_widen_minus_pattern): Likewise.
11960 (vect_recog_average_pattern): Likewise.
11961 * tree-vect-stmts.cc (vectorizable_conversion): Add support for
11963 (supportable_widening_operation): Likewise.
11964 * tree.def (WIDEN_SUM_EXPR): Update example to use new IFNs.
11966 2023-06-05 Andre Vieira <andre.simoesdiasvieira@arm.com>
11967 Joel Hutton <joel.hutton@arm.com>
11969 * tree-vect-patterns.cc: Add include for gimple-iterator.
11970 (vect_recog_widen_op_pattern): Refactor to use code_helper.
11971 (vect_gimple_build): New function.
11972 * tree-vect-stmts.cc (simple_integer_narrowing): Refactor to use
11974 (vectorizable_call): Likewise.
11975 (vect_gen_widened_results_half): Likewise.
11976 (vect_create_vectorized_demotion_stmts): Likewise.
11977 (vect_create_vectorized_promotion_stmts): Likewise.
11978 (vect_create_half_widening_stmts): Likewise.
11979 (vectorizable_conversion): Likewise.
11980 (supportable_widening_operation): Likewise.
11981 (supportable_narrowing_operation): Likewise.
11982 * tree-vectorizer.h (supportable_widening_operation): Change
11983 prototype to use code_helper.
11984 (supportable_narrowing_operation): Likewise.
11985 (vect_gimple_build): New function prototype.
11986 * tree.h (code_helper::safe_as_tree_code): New function.
11987 (code_helper::safe_as_fn_code): New function.
11989 2023-06-05 Roger Sayle <roger@nextmovesoftware.com>
11991 * wide-int.cc (wi::bitreverse_large): New function implementing
11992 bit reversal of an integer.
11993 * wide-int.h (wi::bitreverse): New (template) function prototype.
11994 (bitreverse_large): Prototype helper function/implementation.
11995 (wi::bitreverse): New template wrapper around bitreverse_large.
11997 2023-06-05 Uros Bizjak <ubizjak@gmail.com>
11999 * rtl.h (print_rtl_single): Change return type from int to void.
12000 (print_rtl_single_with_indent): Ditto.
12001 * print-rtl.h (class rtx_writer): Ditto. Change m_sawclose to bool.
12002 * print-rtl.cc (rtx_writer::rtx_writer): Update for m_sawclose change.
12003 (rtx_writer::print_rtx_operand_code_0): Ditto.
12004 (rtx_writer::print_rtx_operand_codes_E_and_V): Ditto.
12005 (rtx_writer::print_rtx_operand_code_i): Ditto.
12006 (rtx_writer::print_rtx_operand_code_u): Ditto.
12007 (rtx_writer::print_rtx_operand): Ditto.
12008 (rtx_writer::print_rtx): Ditto.
12009 (rtx_writer::finish_directive): Ditto.
12010 (print_rtl_single): Change return type from int to void
12011 and adjust function body accordingly.
12012 (rtx_writer::print_rtl_single_with_indent): Ditto.
12014 2023-06-05 Uros Bizjak <ubizjak@gmail.com>
12016 * rtl.h (reg_classes_intersect_p): Change return type from int to bool.
12017 (reg_class_subset_p): Ditto.
12018 * reginfo.cc (reg_classes_intersect_p): Ditto.
12019 (reg_class_subset_p): Ditto.
12021 2023-06-05 Pan Li <pan2.li@intel.com>
12023 * config/riscv/riscv-vector-builtins-types.def
12024 (vfloat32mf2_t): New type for DEF_RVV_WEXTF_OPS.
12025 (vfloat32m1_t): Ditto.
12026 (vfloat32m2_t): Ditto.
12027 (vfloat32m4_t): Ditto.
12028 (vfloat32m8_t): Ditto.
12029 (vint16mf4_t): New type for DEF_RVV_CONVERT_I_OPS.
12030 (vint16mf2_t): Ditto.
12031 (vint16m1_t): Ditto.
12032 (vint16m2_t): Ditto.
12033 (vint16m4_t): Ditto.
12034 (vint16m8_t): Ditto.
12035 (vuint16mf4_t): New type for DEF_RVV_CONVERT_U_OPS.
12036 (vuint16mf2_t): Ditto.
12037 (vuint16m1_t): Ditto.
12038 (vuint16m2_t): Ditto.
12039 (vuint16m4_t): Ditto.
12040 (vuint16m8_t): Ditto.
12041 (vint32mf2_t): New type for DEF_RVV_WCONVERT_I_OPS.
12042 (vint32m1_t): Ditto.
12043 (vint32m2_t): Ditto.
12044 (vint32m4_t): Ditto.
12045 (vint32m8_t): Ditto.
12046 (vuint32mf2_t): New type for DEF_RVV_WCONVERT_U_OPS.
12047 (vuint32m1_t): Ditto.
12048 (vuint32m2_t): Ditto.
12049 (vuint32m4_t): Ditto.
12050 (vuint32m8_t): Ditto.
12051 * config/riscv/vector-iterators.md: Add FP=16 support for V,
12052 VWCONVERTI, VCONVERT, VNCONVERT, VMUL1 and vlmul1.
12054 2023-06-05 Andrew Pinski <apinski@marvell.com>
12056 PR bootstrap/110085
12057 * Makefile.in (clean): Remove the removing of
12058 MULTILIB_DIR/MULTILIB_OPTIONS directories.
12060 2023-06-05 YunQiang Su <yunqiang.su@cipunited.com>
12062 * config/mips/mips-protos.h (mips_emit_speculation_barrier): New
12064 * config/mips/mips.cc (speculation_barrier_libfunc): New static
12066 (mips_init_libfuncs): Initialize it.
12067 (mips_emit_speculation_barrier): New function.
12068 * config/mips/mips.md (speculation_barrier): Call
12069 mips_emit_speculation_barrier.
12071 2023-06-05 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12073 * config/riscv/riscv-v.cc (class rvv_builder): Reorganize functions.
12074 (rvv_builder::can_duplicate_repeating_sequence_p): Ditto.
12075 (rvv_builder::repeating_sequence_use_merge_profitable_p): Ditto.
12076 (rvv_builder::get_merged_repeating_sequence): Ditto.
12077 (rvv_builder::get_merge_scalar_mask): Ditto.
12078 (emit_scalar_move_insn): Ditto.
12079 (emit_vlmax_integer_move_insn): Ditto.
12080 (emit_nonvlmax_integer_move_insn): Ditto.
12081 (emit_vlmax_gather_insn): Ditto.
12082 (emit_vlmax_masked_gather_mu_insn): Ditto.
12083 (get_repeating_sequence_dup_machine_mode): Ditto.
12085 2023-06-05 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12087 * config/riscv/autovec.md: Split arguments.
12088 * config/riscv/riscv-protos.h (expand_vec_perm): Ditto.
12089 * config/riscv/riscv-v.cc (expand_vec_perm): Ditto.
12091 2023-06-04 Andrew Pinski <apinski@marvell.com>
12093 * expr.cc (do_store_flag): Improve for single bit testing
12094 not against zero but against that single bit.
12096 2023-06-04 Andrew Pinski <apinski@marvell.com>
12098 * expr.cc (do_store_flag): Extend the one bit checking case
12099 to handle the case where we don't have an and but rather still
12100 one bit is known to be non-zero.
12102 2023-06-04 Jeff Law <jlaw@ventanamicro.com>
12104 * config/h8300/constraints.md (Zz): Make this a normal
12106 * config/h8300/h8300.cc (TARGET_LRA_P): Remove.
12107 * config/h8300/logical.md (H8/SX bit patterns): Remove.
12109 2023-06-04 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12111 * config/xtensa/xtensa.md (*btrue_INT_MIN, *eqne_INT_MIN):
12112 New insn_and_split patterns.
12114 2023-06-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12117 * config/riscv/riscv-vector-builtins-bases.cc: Change expand approach.
12118 * config/riscv/vector.md (@vlmul_extx2<mode>): Remove it.
12119 (@vlmul_extx4<mode>): Ditto.
12120 (@vlmul_extx8<mode>): Ditto.
12121 (@vlmul_extx16<mode>): Ditto.
12122 (@vlmul_extx32<mode>): Ditto.
12123 (@vlmul_extx64<mode>): Ditto.
12124 (*vlmul_extx2<mode>): Ditto.
12125 (*vlmul_extx4<mode>): Ditto.
12126 (*vlmul_extx8<mode>): Ditto.
12127 (*vlmul_extx16<mode>): Ditto.
12128 (*vlmul_extx32<mode>): Ditto.
12129 (*vlmul_extx64<mode>): Ditto.
12131 2023-06-04 Pan Li <pan2.li@intel.com>
12133 * config/riscv/riscv-vector-builtins-types.def
12134 (vfloat32mf2_t): Add vfloat32mf2_t type to vfncvt.f.f.w operations.
12135 (vfloat32m1_t): Likewise.
12136 (vfloat32m2_t): Likewise.
12137 (vfloat32m4_t): Likewise.
12138 (vfloat32m8_t): Likewise.
12139 * config/riscv/riscv-vector-builtins.def: Fix typo in comments.
12140 * config/riscv/vector-iterators.md: Add single to half machine
12143 2023-06-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12145 * config/riscv/autovec-opt.md (*<optab>not<mode>): Move to autovec-opt.md.
12146 (*n<optab><mode>): Ditto.
12147 * config/riscv/autovec.md (*<optab>not<mode>): Ditto.
12148 (*n<optab><mode>): Ditto.
12149 * config/riscv/vector.md: Ditto.
12151 2023-06-04 Roger Sayle <roger@nextmovesoftware.com>
12154 * config/i386/i386-features.cc (scalar_chain::convert_compare):
12155 Update or delete REG_EQUAL notes, converting CONST_INT and
12156 CONST_WIDE_INT immediate operands to a suitable CONST_VECTOR.
12158 2023-06-04 Jason Merrill <jason@redhat.com>
12161 * tree-eh.cc (lower_resx): Pass the exception pointer to the
12163 * except.h: Tweak comment.
12165 2023-06-04 Hans-Peter Nilsson <hp@axis.com>
12167 * postreload.cc (move2add_use_add2_insn): Handle
12168 trivial single_sets. Rename variable PAT to SET.
12169 (move2add_use_add3_insn, reload_cse_move2add): Similar.
12171 2023-06-04 Pan Li <pan2.li@intel.com>
12173 * config/riscv/riscv-vector-builtins-types.def
12174 (vfloat16mf4_t): Add the float16 type to DEF_RVV_F_OPS.
12175 (vfloat16mf2_t): Likewise.
12176 (vfloat16m1_t): Likewise.
12177 (vfloat16m2_t): Likewise.
12178 (vfloat16m4_t): Likewise.
12179 (vfloat16m8_t): Likewise.
12180 * config/riscv/riscv.md: Add vfloat16*_t to attr mode.
12181 * config/riscv/vector-iterators.md: Add vfloat16*_t machine mode
12182 to V, V_WHOLE, V_FRACT, VINDEX, VM, VEL and sew.
12183 * config/riscv/vector.md: Add vfloat16*_t machine mode to sew,
12186 2023-06-03 Fei Gao <gaofei@eswincomputing.com>
12188 * config/riscv/riscv.cc (riscv_expand_epilogue): fix cfi issue with
12191 2023-06-03 Die Li <lidie@eswincomputing.com>
12193 * config/riscv/thead.md (*th_cond_gpr_mov<GPR:mode><GPR2:mode>): Delete.
12195 2023-06-03 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12197 * config/riscv/predicates.md: Change INTVAL into UINTVAL.
12199 2023-06-03 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12201 * config/riscv/vector.md: Add vector-opt.md.
12202 * config/riscv/autovec-opt.md: New file.
12204 2023-06-03 liuhongt <hongtao.liu@intel.com>
12206 PR tree-optimization/110067
12207 * gimple-ssa-store-merging.cc (find_bswap_or_nop): Don't try
12208 bswap + rotate when TYPE_PRECISION(n->type) > n->range.
12210 2023-06-03 liuhongt <hongtao.liu@intel.com>
12213 * config/i386/mmx.md (truncv2hiv2qi2): New define_insn.
12214 (truncv2si<mode>2): Ditto.
12216 2023-06-02 Andrew Pinski <apinski@marvell.com>
12218 PR rtl-optimization/102733
12219 * dse.cc (store_info): Add addrspace field.
12220 (record_store): Record the address space
12221 and check to make sure they are the same.
12223 2023-06-02 Andrew Pinski <apinski@marvell.com>
12225 PR rtl-optimization/110042
12226 * ifcvt.cc (bbs_ok_for_cmove_arith): Allow paradoxical subregs.
12227 (bb_valid_for_noce_process_p): Strip the subreg for the SET_DEST.
12229 2023-06-02 Iain Sandoe <iain@sandoe.co.uk>
12232 * config/rs6000/rs6000.cc (darwin_rs6000_special_round_type_align):
12233 Make sure that we do not have a cap on field alignment before altering
12234 the struct layout based on the type alignment of the first entry.
12236 2023-06-02 David Faust <david.faust@oracle.com>
12239 * btfout.cc (btf_absolute_func_id): New function.
12240 (btf_asm_func_type): Call it here. Change index parameter from
12241 size_t to ctf_id_t. Use PRIu64 formatter.
12243 2023-06-02 Alex Coplan <alex.coplan@arm.com>
12245 * btfout.cc (btf_asm_type): Use PRIu64 instead of %lu for uint64_t.
12246 (btf_asm_datasec_type): Likewise.
12248 2023-06-02 Carl Love <cel@us.ibm.com>
12250 * config/rs6000/rs6000-builtins.def (__builtin_altivec_tr_stxvrhx,
12251 __builtin_altivec_tr_stxvrwx): Fix type of third argument.
12253 2023-06-02 Jason Merrill <jason@redhat.com>
12257 * tree.h (DECL_MERGEABLE): New.
12258 * tree-core.h (struct tree_decl_common): Mention it.
12259 * gimplify.cc (gimplify_init_constructor): Check it.
12260 * cgraph.cc (symtab_node::address_can_be_compared_p): Likewise.
12261 * varasm.cc (categorize_decl_for_section): Likewise.
12263 2023-06-02 Uros Bizjak <ubizjak@gmail.com>
12265 * rtl.h (stack_regs_mentioned): Change return type from int to bool.
12266 * reg-stack.cc (struct_block_info_def): Change "done" to bool.
12267 (stack_regs_mentioned_p): Change return type from int to bool
12268 and adjust function body accordingly.
12269 (stack_regs_mentioned): Ditto.
12270 (check_asm_stack_operands): Ditto. Change "malformed_asm"
12272 (move_for_stack_reg): Recode handling of control_flow_insn_deleted.
12273 (swap_rtx_condition_1): Change return type from int to bool
12274 and adjust function body accordingly. Change "r" variable to bool.
12275 (swap_rtx_condition): Change return type from int to bool
12276 and adjust function body accordingly.
12277 (subst_stack_regs_pat): Recode handling of control_flow_insn_deleted.
12278 (subst_stack_regs): Ditto.
12279 (convert_regs_entry): Change return type from int to bool and adjust
12280 function body accordingly. Change "inserted" variable to bool.
12281 (convert_regs_1): Recode handling of control_flow_insn_deleted.
12282 (convert_regs_2): Recode handling of cfg_altered.
12283 (convert_regs): Ditto. Change "inserted" variable to bool.
12285 2023-06-02 Jason Merrill <jason@redhat.com>
12288 * varasm.cc (output_constant) [REAL_TYPE]: Check that sizes match.
12289 (initializer_constant_valid_p_1): Compare float precision.
12291 2023-06-02 Alexander Monakov <amonakov@ispras.ru>
12293 * doc/extend.texi (Vector Extensions): Clarify bitwise shift
12296 2023-06-02 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
12298 * tree-vect-loop-manip.cc (vect_set_loop_controls_directly): Change decrement IV flow.
12299 (vect_set_loop_condition_partial_vectors): Ditto.
12301 2023-06-02 Georg-Johann Lay <avr@gjlay.de>
12304 * config/avr/avr.md: Add an RTL peephole to optimize operations on
12305 non-LD_REGS after a move from LD_REGS.
12306 (piaop): New code iterator.
12308 2023-06-02 Thomas Schwinge <thomas@codesourcery.com>
12311 * doc/install.texi: Document (optional) Perl usage for parallel
12312 testing of libgomp.
12314 2023-06-02 Thomas Schwinge <thomas@codesourcery.com>
12317 * doc/install.texi (Perl): Back to requiring "Perl version 5.6.1 (or
12320 2023-06-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12321 KuanLin Chen <best124612@gmail.com>
12323 * config/riscv/riscv-vector-builtins-bases.cc: Add _mu overloaded intrinsics.
12324 * config/riscv/riscv-vector-builtins-shapes.cc (struct fault_load_def): Ditto.
12326 2023-06-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12328 * config/riscv/riscv-v.cc (expand_vec_series): Optimize reverse series index vector.
12330 2023-06-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12332 * config/riscv/predicates.md: Change INTVAL into UINTVAL.
12334 2023-06-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12336 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_VXRM_ENUM): Add
12338 (DEF_RVV_FRM_ENUM): Ditto.
12340 2023-06-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12342 * config/riscv/riscv-vector-builtins-bases.cc: Change vwadd.wv/vwsub.wv
12343 intrinsic API expander
12344 * config/riscv/vector.md
12345 (@pred_single_widen_<plus_minus:optab><any_extend:su><mode>): Remove it.
12346 (@pred_single_widen_sub<any_extend:su><mode>): New pattern.
12347 (@pred_single_widen_add<any_extend:su><mode>): New pattern.
12349 2023-06-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12351 * config/riscv/autovec.md (vec_perm<mode>): New pattern.
12352 * config/riscv/predicates.md (vector_perm_operand): New predicate.
12353 * config/riscv/riscv-protos.h (enum insn_type): New enum.
12354 (expand_vec_perm): New function.
12355 * config/riscv/riscv-v.cc (const_vec_all_in_range_p): Ditto.
12356 (gen_const_vector_dup): Ditto.
12357 (emit_vlmax_gather_insn): Ditto.
12358 (emit_vlmax_masked_gather_mu_insn): Ditto.
12359 (expand_vec_perm): Ditto.
12361 2023-06-01 Jason Merrill <jason@redhat.com>
12363 * doc/invoke.texi (-Wpedantic): Improve clarity.
12365 2023-06-01 Uros Bizjak <ubizjak@gmail.com>
12367 * rtl.h (exp_equiv_p): Change return type from int to bool.
12368 * cse.cc (mention_regs): Change return type from int to bool
12369 and adjust function body accordingly.
12370 (exp_equiv_p): Ditto.
12371 (insert_regs): Ditto. Change "modified" function argument to bool
12372 and update usage accordingly.
12373 (record_jump_cond): Remove always zero "reversed_nonequality"
12374 function argument and update usage accordingly.
12375 (fold_rtx): Change "changed" variable to bool.
12376 (record_jump_equiv): Remove unneeded "reversed_nonequality" variable.
12377 (is_dead_reg): Change return type from int to bool.
12379 2023-06-01 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12381 * config/xtensa/xtensa.md (adddi3, subdi3):
12382 New RTL generation patterns implemented according to the instruc-
12383 tion idioms described in the Xtensa ISA reference manual (p. 600).
12385 2023-06-01 Roger Sayle <roger@nextmovesoftware.com>
12386 Uros Bizjak <ubizjak@gmail.com>
12389 * config/i386/i386-builtin.def (__builtin_ia32_ptestz128): Use new
12390 CODE_for_sse4_1_ptestzv2di.
12391 (__builtin_ia32_ptestc128): Use new CODE_for_sse4_1_ptestcv2di.
12392 (__builtin_ia32_ptestz256): Use new CODE_for_avx_ptestzv4di.
12393 (__builtin_ia32_ptestc256): Use new CODE_for_avx_ptestcv4di.
12394 * config/i386/i386-expand.cc (ix86_expand_branch): Use CCZmode
12395 when expanding UNSPEC_PTEST to compare against zero.
12396 * config/i386/i386-features.cc (scalar_chain::convert_compare):
12397 Likewise generate CCZmode UNSPEC_PTESTs when converting comparisons.
12398 (general_scalar_chain::convert_insn): Use CCZmode for COMPARE result.
12399 (timode_scalar_chain::convert_insn): Use CCZmode for COMPARE result.
12400 * config/i386/i386-protos.h (ix86_match_ptest_ccmode): Prototype.
12401 * config/i386/i386.cc (ix86_match_ptest_ccmode): New predicate to
12402 check for suitable matching modes for the UNSPEC_PTEST pattern.
12403 * config/i386/sse.md (define_split): When splitting UNSPEC_MOVMSK
12404 to UNSPEC_PTEST, preserve the FLAG_REG mode as CCZ.
12405 (*<sse4_1>_ptest<mode>): Add asterisk to hide define_insn. Remove
12406 ":CC" mode of FLAGS_REG, instead use ix86_match_ptest_ccmode.
12407 (<sse4_1>_ptestz<mode>): New define_expand to specify CCZ.
12408 (<sse4_1>_ptestc<mode>): New define_expand to specify CCC.
12409 (<sse4_1>_ptest<mode>): A define_expand using CC to preserve the
12411 (*ptest<mode>_and): Specify CCZ to only perform this optimization
12412 when only the Z flag is required.
12414 2023-06-01 Jonathan Wakely <jwakely@redhat.com>
12417 * doc/invoke.texi (x86 Options): Fix description of -m32 option.
12419 2023-06-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12421 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<VDMOV:mode>):
12422 Add =r,m and =r,m alternatives.
12423 (load_pair<DREG:mode><DREG2:mode>): Likewise.
12424 (vec_store_pair<DREG:mode><DREG2:mode>): Likewise.
12426 2023-06-01 Pan Li <pan2.li@intel.com>
12428 * common/config/riscv/riscv-common.cc: Add FP_16 mask to zvfhmin
12430 * config/riscv/genrvv-type-indexer.cc (valid_type): Allow FP16.
12431 (main): Disable FP16 tuple.
12432 * config/riscv/riscv-opts.h (MASK_VECTOR_ELEN_FP_16): New macro.
12433 (TARGET_VECTOR_ELEN_FP_16): Ditto.
12434 * config/riscv/riscv-vector-builtins.cc (check_required_extensions):
12436 * config/riscv/riscv-vector-builtins.def (vfloat16mf4_t): New type.
12437 (vfloat16mf2_t): Ditto.
12438 (vfloat16m1_t): Ditto.
12439 (vfloat16m2_t): Ditto.
12440 (vfloat16m4_t): Ditto.
12441 (vfloat16m8_t): Ditto.
12442 * config/riscv/riscv-vector-builtins.h (RVV_REQUIRE_ELEN_FP_16):
12444 * config/riscv/riscv-vector-switch.def (ENTRY): Allow FP16
12445 machine mode based on TARGET_VECTOR_ELEN_FP_16.
12447 2023-06-01 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12449 * config/riscv/riscv-vector-builtins.cc (register_frm): New function.
12450 (DEF_RVV_FRM_ENUM): New macro.
12451 (handle_pragma_vector): Add FRM enum
12452 * config/riscv/riscv-vector-builtins.def (DEF_RVV_FRM_ENUM): New macro.
12459 2023-05-31 Roger Sayle <roger@nextmovesoftware.com>
12460 Richard Sandiford <richard.sandiford@arm.com>
12462 * fold-const-call.cc (fold_const_call_ss) <CFN_BUILT_IN_BSWAP*>:
12463 Update call to wi::bswap.
12464 * simplify-rtx.cc (simplify_const_unary_operation) <case BSWAP>:
12465 Update call to wi::bswap.
12466 * tree-ssa-ccp.cc (evaluate_stmt) <case BUILT_IN_BSWAP*>:
12467 Update calls to wi::bswap.
12468 * wide-int.cc (wide_int_storage::bswap): Remove/rename to...
12469 (wi::bswap_large): New function, with revised API.
12470 * wide-int.h (wi::bswap): New (template) function prototype.
12471 (wide_int_storage::bswap): Remove method.
12472 (sext_large, zext_large): Consistent indentation/line wrapping.
12473 (bswap_large): Prototype helper function containing implementation.
12474 (wi::bswap): New template wrapper around bswap_large.
12476 2023-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12479 * config/aarch64/aarch64-simd.md (<sur>dot_prod<vsi2qi>): Rename to...
12480 (<sur>dot_prod<vsi2qi><vczle><vczbe>): ... This.
12481 (usdot_prod<vsi2qi>): Rename to...
12482 (usdot_prod<vsi2qi><vczle><vczbe>): ... This.
12483 (aarch64_<sur>dot_lane<vsi2qi>): Rename to...
12484 (aarch64_<sur>dot_lane<vsi2qi><vczle><vczbe>): ... This.
12485 (aarch64_<sur>dot_laneq<vsi2qi>): Rename to...
12486 (aarch64_<sur>dot_laneq<vsi2qi><vczle><vczbe>): ... This.
12487 (aarch64_<DOTPROD_I8MM:sur>dot_lane<VB:isquadop><VS:vsi2qi>): Rename to...
12488 (aarch64_<DOTPROD_I8MM:sur>dot_lane<VB:isquadop><VS:vsi2qi><vczle><vczbe>):
12491 2023-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12494 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh<mode>): Rename to...
12495 (aarch64_sq<r>dmulh<mode><vczle><vczbe>): ... This.
12496 (aarch64_sq<r>dmulh_n<mode>): Rename to...
12497 (aarch64_sq<r>dmulh_n<mode><vczle><vczbe>): ... This.
12498 (aarch64_sq<r>dmulh_lane<mode>): Rename to...
12499 (aarch64_sq<r>dmulh_lane<mode><vczle><vczbe>): ... This.
12500 (aarch64_sq<r>dmulh_laneq<mode>): Rename to...
12501 (aarch64_sq<r>dmulh_laneq<mode><vczle><vczbe>): ... This.
12502 (aarch64_sqrdml<SQRDMLH_AS:rdma_as>h<mode>): Rename to...
12503 (aarch64_sqrdml<SQRDMLH_AS:rdma_as>h<mode><vczle><vczbe>): ... This.
12504 (aarch64_sqrdml<SQRDMLH_AS:rdma_as>h_lane<mode>): Rename to...
12505 (aarch64_sqrdml<SQRDMLH_AS:rdma_as>h_lane<mode><vczle><vczbe>): ... This.
12506 (aarch64_sqrdml<SQRDMLH_AS:rdma_as>h_laneq<mode>): Rename to...
12507 (aarch64_sqrdml<SQRDMLH_AS:rdma_as>h_laneq<mode><vczle><vczbe>): ... This.
12509 2023-05-31 David Faust <david.faust@oracle.com>
12511 * btfout.cc (btf_kind_names): New.
12512 (btf_kind_name): New.
12513 (btf_absolute_var_id): New utility function.
12514 (btf_relative_var_id): Likewise.
12515 (btf_relative_func_id): Likewise.
12516 (btf_absolute_datasec_id): Likewise.
12517 (btf_asm_type_ref): New.
12518 (btf_asm_type): Update asm comments and use btf_asm_type_ref ().
12519 (btf_asm_array): Likewise. Accept ctf_container_ref parameter.
12520 (btf_asm_varent): Likewise.
12521 (btf_asm_func_arg): Likewise.
12522 (btf_asm_datasec_entry): Likewise.
12523 (btf_asm_datasec_type): Likewise.
12524 (btf_asm_func_type): Likewise. Add index parameter.
12525 (btf_asm_enum_const): Likewise.
12526 (btf_asm_sou_member): Likewise.
12527 (output_btf_vars): Update btf_asm_* call accordingly.
12528 (output_asm_btf_sou_fields): Likewise.
12529 (output_asm_btf_enum_list): Likewise.
12530 (output_asm_btf_func_args_list): Likewise.
12531 (output_asm_btf_vlen_bytes): Likewise.
12532 (output_btf_func_types): Add ctf_container_ref parameter.
12533 Pass it to btf_asm_func_type.
12534 (output_btf_datasec_types): Update btf_asm_datsec_type call similarly.
12535 (btf_output): Update output_btf_func_types call similarly.
12537 2023-05-31 David Faust <david.faust@oracle.com>
12539 * btfout.cc (btf_asm_type): Add dedicated cases for BTF_KIND_ARRAY
12540 and BTF_KIND_FWD which do not use the size/type field at all.
12542 2023-05-31 Uros Bizjak <ubizjak@gmail.com>
12544 * rtl.h (subreg_lowpart_p): Change return type from int to bool.
12545 (active_insn_p): Ditto.
12546 (in_sequence_p): Ditto.
12547 (unshare_all_rtl): Change return type from int to void.
12548 * emit-rtl.h (mem_expr_equal_p): Change return type from int to bool.
12549 * emit-rtl.cc (subreg_lowpart_p): Change return type from int to bool
12550 and adjust function body accordingly.
12551 (mem_expr_equal_p): Ditto.
12552 (unshare_all_rtl): Change return type from int to void
12553 and adjust function body accordingly.
12554 (verify_rtx_sharing): Remove unneeded return.
12555 (active_insn_p): Change return type from int to bool
12556 and adjust function body accordingly.
12557 (in_sequence_p): Ditto.
12559 2023-05-31 Uros Bizjak <ubizjak@gmail.com>
12561 * rtl.h (true_dependence): Change return type from int to bool.
12562 (canon_true_dependence): Ditto.
12563 (read_dependence): Ditto.
12564 (anti_dependence): Ditto.
12565 (canon_anti_dependence): Ditto.
12566 (output_dependence): Ditto.
12567 (canon_output_dependence): Ditto.
12568 (may_alias_p): Ditto.
12569 * alias.h (alias_sets_conflict_p): Ditto.
12570 (alias_sets_must_conflict_p): Ditto.
12571 (objects_must_conflict_p): Ditto.
12572 (nonoverlapping_memrefs_p): Ditto.
12573 * alias.cc (rtx_equal_for_memref_p): Remove forward declaration.
12574 (record_set): Ditto.
12575 (base_alias_check): Ditto.
12576 (find_base_value): Ditto.
12577 (mems_in_disjoint_alias_sets_p): Ditto.
12578 (get_alias_set_entry): Ditto.
12579 (decl_for_component_ref): Ditto.
12580 (write_dependence_p): Ditto.
12581 (memory_modified_1): Ditto.
12582 (mems_in_disjoint_alias_set_p): Change return type from int to bool
12583 and adjust function body accordingly.
12584 (alias_sets_conflict_p): Ditto.
12585 (alias_sets_must_conflict_p): Ditto.
12586 (objects_must_conflict_p): Ditto.
12587 (rtx_equal_for_memref_p): Ditto.
12588 (base_alias_check): Ditto.
12589 (read_dependence): Ditto.
12590 (nonoverlapping_memrefs_p): Ditto.
12591 (true_dependence_1): Ditto.
12592 (true_dependence): Ditto.
12593 (canon_true_dependence): Ditto.
12594 (write_dependence_p): Ditto.
12595 (anti_dependence): Ditto.
12596 (canon_anti_dependence): Ditto.
12597 (output_dependence): Ditto.
12598 (canon_output_dependence): Ditto.
12599 (may_alias_p): Ditto.
12600 (init_alias_analysis): Change "changed" variable to bool.
12602 2023-05-31 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12604 * config/riscv/autovec.md (<optab><v_double_trunc><mode>2): Change
12605 expand into define_insn_and_split.
12607 2023-05-31 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12609 * config/riscv/vector.md: Remove FRM.
12611 2023-05-31 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12613 * config/riscv/vector.md: Remove FRM.
12615 2023-05-31 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12617 * config/riscv/vector.md: Remove FRM.
12619 2023-05-31 Christophe Lyon <christophe.lyon@linaro.org>
12622 * config/aarch64/aarch64.md (aarch64_rev16si2_alt3): New
12625 2023-05-31 Richard Biener <rguenther@suse.de>
12628 PR tree-optimization/109143
12629 * tree-ssa-structalias.cc (struct topo_info): Remove.
12630 (init_topo_info): Likewise.
12631 (free_topo_info): Likewise.
12632 (compute_topo_order): Simplify API, put the component
12633 with ESCAPED last so it's processed first.
12634 (topo_visit): Adjust.
12635 (solve_graph): Likewise.
12637 2023-05-31 Richard Biener <rguenther@suse.de>
12639 * tree-ssa-structalias.cc (constraint_stats::num_avoided_edges):
12641 (add_graph_edge): Count redundant edges we avoid to create.
12642 (dump_sa_stats): Dump them.
12643 (ipa_pta_execute): Do not dump generating constraints when
12644 we are not dumping them.
12646 2023-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12648 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<VDMOV:mode>): Rewrite
12649 output template to avoid explicit switch on which_alternative.
12650 (*aarch64_simd_mov<VQMOV:mode>): Likewise.
12651 (and<mode>3): Likewise.
12652 (ior<mode>3): Likewise.
12653 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Likewise.
12655 2023-05-31 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12657 * config/xtensa/predicates.md (xtensa_bit_join_operator):
12659 * config/xtensa/xtensa.md (ior_op): Remove.
12660 (*shlrd_reg): Rename from "*shlrd_reg_<code>", and add the
12661 insn_and_split pattern of the same name to express and capture
12662 the bit-combining operation with both sides swapped.
12663 In addition, replace use of code iterator with new operator
12665 (*shlrd_const, *shlrd_per_byte):
12666 Likewise regarding the code iterator.
12668 2023-05-31 Cui, Lili <lili.cui@intel.com>
12670 PR tree-optimization/110038
12671 * params.opt: Add a limit on tree-reassoc-width.
12672 * tree-ssa-reassoc.cc
12673 (rewrite_expr_tree_parallel): Add width limit.
12675 2023-05-31 Pan Li <pan2.li@intel.com>
12677 * common/config/riscv/riscv-common.cc:
12678 (riscv_implied_info): Add zvfh item.
12679 (riscv_ext_version_table): Ditto.
12680 (riscv_ext_flag_table): Ditto.
12681 * config/riscv/riscv-opts.h (MASK_ZVFH): New macro.
12682 (TARGET_ZVFH): Ditto.
12684 2023-05-30 liuhongt <hongtao.liu@intel.com>
12686 PR tree-optimization/108804
12687 * tree-vect-patterns.cc (vect_get_range_info): Remove static.
12688 * tree-vect-stmts.cc (vect_create_vectorized_demotion_stmts):
12689 Add new parameter narrow_src_p.
12690 (vectorizable_conversion): Enhance NARROW FLOAT_EXPR
12691 vectorization by truncating to lower precision.
12692 * tree-vectorizer.h (vect_get_range_info): New declare.
12694 2023-05-30 Vladimir N. Makarov <vmakarov@redhat.com>
12696 * lra-int.h (lra_update_sp_offset): Add the prototype.
12697 * lra.cc (setup_sp_offset): Change the return type. Use
12698 lra_update_sp_offset.
12699 * lra-eliminations.cc (lra_update_sp_offset): New function.
12700 (lra_process_new_insns): Push the current insn to reprocess if the
12701 input reload changes sp offset.
12703 2023-05-30 Uros Bizjak <ubizjak@gmail.com>
12706 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi2):
12707 Fix misleading identation.
12709 2023-05-30 Uros Bizjak <ubizjak@gmail.com>
12711 * rtl.h (comparison_dominates_p): Change return type from int to bool.
12712 (condjump_p): Ditto.
12713 (any_condjump_p): Ditto.
12714 (any_uncondjump_p): Ditto.
12715 (simplejump_p): Ditto.
12716 (returnjump_p): Ditto.
12717 (eh_returnjump_p): Ditto.
12718 (onlyjump_p): Ditto.
12719 (invert_jump_1): Ditto.
12720 (invert_jump): Ditto.
12721 (rtx_renumbered_equal_p): Ditto.
12722 (redirect_jump_1): Ditto.
12723 (redirect_jump): Ditto.
12724 (condjump_in_parallel_p): Ditto.
12725 * jump.cc (invert_exp_1): Adjust forward declaration.
12726 (comparison_dominates_p): Change return type from int to bool
12727 and adjust function body accordingly.
12728 (simplejump_p): Ditto.
12729 (condjump_p): Ditto.
12730 (condjump_in_parallel_p): Ditto.
12731 (any_uncondjump_p): Ditto.
12732 (any_condjump_p): Ditto.
12733 (returnjump_p): Ditto.
12734 (eh_returnjump_p): Ditto.
12735 (onlyjump_p): Ditto.
12736 (redirect_jump_1): Ditto.
12737 (redirect_jump): Ditto.
12738 (invert_exp_1): Ditto.
12739 (invert_jump_1): Ditto.
12740 (invert_jump): Ditto.
12741 (rtx_renumbered_equal_p): Ditto.
12743 2023-05-30 Andrew Pinski <apinski@marvell.com>
12745 * fold-const.cc (minmax_from_comparison): Add support for NE_EXPR.
12746 * match.pd ((cond (cmp (convert1? x) c1) (convert2? x) c2) pattern):
12747 Add ne as a possible cmp.
12748 ((a CMP b) ? minmax<a, c> : minmax<b, c> pattern): Likewise.
12750 2023-05-30 Andrew Pinski <apinski@marvell.com>
12752 * match.pd (`(a CMP CST1) ? max<a,CST2> : a`): New
12755 2023-05-30 Roger Sayle <roger@nextmovesoftware.com>
12757 * simplify-rtx.cc (simplify_binary_operation_1) <AND>: Use wide-int
12758 instead of HWI_COMPUTABLE_MODE_P and UINTVAL in transformation of
12759 (and (extend X) C) as (zero_extend (and X C)), to also optimize
12760 modes wider than HOST_WIDE_INT.
12762 2023-05-30 Roger Sayle <roger@nextmovesoftware.com>
12765 * simplify-rtx.cc (simplify_const_relational_operation): Return
12766 early if we have a MODE_CC comparison that isn't a COMPARE against
12769 2023-05-30 Robin Dapp <rdapp@ventanamicro.com>
12771 * config/riscv/riscv.cc (riscv_const_insns): Allow
12772 const_vec_duplicates.
12774 2023-05-30 liuhongt <hongtao.liu@intel.com>
12776 PR middle-end/108938
12777 * gimple-ssa-store-merging.cc (is_bswap_or_nop_p): New
12778 function, cut from original find_bswap_or_nop function.
12779 (find_bswap_or_nop): Add a new parameter, detect bswap +
12780 rotate and save rotate result in the new parameter.
12781 (bswap_replace): Add a new parameter to indicate rotate and
12782 generate rotate stmt if needed.
12783 (maybe_optimize_vector_constructor): Adjust for new rotate
12784 parameter in the upper 2 functions.
12785 (pass_optimize_bswap::execute): Ditto.
12786 (imm_store_chain_info::output_merged_store): Ditto.
12788 2023-05-30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12790 * config/aarch64/aarch64-simd.md (aarch64_<sur>adalp<mode>): Delete.
12791 (aarch64_<su>adalp<mode>): New define_expand.
12792 (*aarch64_<su>adalp<mode><vczle><vczbe>_insn): New define_insn.
12793 (aarch64_<su>addlp<mode>): Convert to define_expand.
12794 (*aarch64_<su>addlp<mode><vczle><vczbe>_insn): New define_insn.
12795 * config/aarch64/iterators.md (UNSPEC_SADDLP, UNSPEC_UADDLP): Delete.
12797 (USADDLP): Likewise.
12798 * config/aarch64/predicates.md (vect_par_cnst_even_or_odd_half): Define.
12800 2023-05-30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12802 * config/aarch64/aarch64-builtins.cc (VAR1): Move to after inclusion of
12803 aarch64-builtin-iterators.h. Add definition to remap shadd, uhadd,
12804 srhadd, urhadd builtin codes for standard optab ones.
12805 * config/aarch64/aarch64-simd.md (<u>avg<mode>3_floor): Rename to...
12806 (<su_optab>avg<mode>3_floor): ... This. Expand to RTL codes rather than
12808 (<u>avg<mode>3_ceil): Rename to...
12809 (<su_optab>avg<mode>3_ceil): ... This. Expand to RTL codes rather than
12811 (aarch64_<su>hsub<mode>): New define_expand.
12812 (aarch64_<sur>h<addsub><mode><vczle><vczbe>): Split into...
12813 (*aarch64_<su>h<ADDSUB:optab><mode><vczle><vczbe>_insn): ... This...
12814 (*aarch64_<su>rhadd<mode><vczle><vczbe>_insn): ... And this.
12816 2023-05-30 Andreas Schwab <schwab@suse.de>
12819 * config/riscv/riscv.cc (riscv_asan_shadow_offset): Update to
12820 match libsanitizer.
12822 2023-05-30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12824 * config/aarch64/aarch64-modes.def (V16HI, V8SI, V4DI, V2TI): New modes.
12825 * config/aarch64/aarch64-protos.h (aarch64_const_vec_rnd_cst_p):
12827 (aarch64_const_vec_rsra_rnd_imm_p): Likewise.
12828 * config/aarch64/aarch64-simd.md (*aarch64_simd_sra<mode>): Rename to...
12829 (aarch64_<sra_op>sra_n<mode>_insn): ... This.
12830 (aarch64_<sra_op>rsra_n<mode>_insn): New define_insn.
12831 (aarch64_<sra_op>sra_n<mode>): New define_expand.
12832 (aarch64_<sra_op>rsra_n<mode>): Likewise.
12833 (aarch64_<sur>sra_n<mode>): Rename to...
12834 (aarch64_<sur>sra_ndi): ... This.
12835 * config/aarch64/aarch64.cc (aarch64_classify_vector_mode): Add
12836 any_target_p argument.
12837 (aarch64_extract_vec_duplicate_wide_int): Define.
12838 (aarch64_const_vec_rsra_rnd_imm_p): Likewise.
12839 (aarch64_const_vec_rnd_cst_p): Likewise.
12840 (aarch64_vector_mode_supported_any_target_p): Likewise.
12841 (TARGET_VECTOR_MODE_SUPPORTED_ANY_TARGET_P): Likewise.
12842 * config/aarch64/iterators.md (UNSPEC_SRSRA, UNSPEC_URSRA): Delete.
12843 (VSRA): Adjust for the above.
12845 (V2XWIDE): New mode_attr.
12846 (vec_or_offset): Likewise.
12847 (SHIFTEXTEND): Likewise.
12848 * config/aarch64/predicates.md (aarch64_simd_rsra_rnd_imm_vec): New
12850 * doc/tm.texi (TARGET_VECTOR_MODE_SUPPORTED_P): Adjust description to
12851 clarify that it applies to current target options.
12852 (TARGET_VECTOR_MODE_SUPPORTED_ANY_TARGET_P): Document.
12853 * doc/tm.texi.in: Regenerate.
12854 * stor-layout.cc (mode_for_vector): Check
12855 vector_mode_supported_any_target_p when iterating through vector modes.
12856 * target.def (TARGET_VECTOR_MODE_SUPPORTED_P): Adjust description to
12857 clarify that it applies to current target options.
12858 (TARGET_VECTOR_MODE_SUPPORTED_ANY_TARGET_P): Define.
12860 2023-05-30 Lili Cui <lili.cui@intel.com>
12862 PR tree-optimization/98350
12863 * tree-ssa-reassoc.cc
12864 (rewrite_expr_tree_parallel): Rewrite this function.
12865 (rank_ops_for_fma): New.
12866 (reassociate_bb): Handle new function.
12868 2023-05-30 Uros Bizjak <ubizjak@gmail.com>
12870 * rtl.h (rtx_addr_can_trap_p): Change return type from int to bool.
12871 (rtx_unstable_p): Ditto.
12872 (reg_mentioned_p): Ditto.
12873 (reg_referenced_p): Ditto.
12874 (reg_used_between_p): Ditto.
12875 (reg_set_between_p): Ditto.
12876 (modified_between_p): Ditto.
12877 (no_labels_between_p): Ditto.
12878 (modified_in_p): Ditto.
12879 (reg_set_p): Ditto.
12880 (multiple_sets): Ditto.
12881 (set_noop_p): Ditto.
12882 (noop_move_p): Ditto.
12883 (reg_overlap_mentioned_p): Ditto.
12884 (dead_or_set_p): Ditto.
12885 (dead_or_set_regno_p): Ditto.
12886 (find_reg_fusage): Ditto.
12887 (find_regno_fusage): Ditto.
12888 (side_effects_p): Ditto.
12889 (volatile_refs_p): Ditto.
12890 (volatile_insn_p): Ditto.
12891 (may_trap_p_1): Ditto.
12892 (may_trap_p): Ditto.
12893 (may_trap_or_fault_p): Ditto.
12894 (computed_jump_p): Ditto.
12895 (auto_inc_p): Ditto.
12896 (loc_mentioned_in_p): Ditto.
12897 * rtlanal.cc (computed_jump_p_1): Adjust forward declaration.
12898 (rtx_unstable_p): Change return type from int to bool
12899 and adjust function body accordingly.
12900 (rtx_addr_can_trap_p): Ditto.
12901 (reg_mentioned_p): Ditto.
12902 (no_labels_between_p): Ditto.
12903 (reg_used_between_p): Ditto.
12904 (reg_referenced_p): Ditto.
12905 (reg_set_between_p): Ditto.
12906 (reg_set_p): Ditto.
12907 (modified_between_p): Ditto.
12908 (modified_in_p): Ditto.
12909 (multiple_sets): Ditto.
12910 (set_noop_p): Ditto.
12911 (noop_move_p): Ditto.
12912 (reg_overlap_mentioned_p): Ditto.
12913 (dead_or_set_p): Ditto.
12914 (dead_or_set_regno_p): Ditto.
12915 (find_reg_fusage): Ditto.
12916 (find_regno_fusage): Ditto.
12917 (remove_node_from_insn_list): Ditto.
12918 (volatile_insn_p): Ditto.
12919 (volatile_refs_p): Ditto.
12920 (side_effects_p): Ditto.
12921 (may_trap_p_1): Ditto.
12922 (may_trap_p): Ditto.
12923 (may_trap_or_fault_p): Ditto.
12924 (computed_jump_p): Ditto.
12925 (auto_inc_p): Ditto.
12926 (loc_mentioned_in_p): Ditto.
12927 * combine.cc (can_combine_p): Update indirect function.
12929 2023-05-30 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12931 * config/riscv/autovec.md (<optab><mode><vconvert>2): New pattern.
12932 * config/riscv/iterators.md: New attribute.
12933 * config/riscv/vector-iterators.md: New attribute.
12935 2023-05-30 From: Juzhe-Zhong <juzhe.zhong@rivai.ai>
12937 * config/riscv/riscv.md: Fix signed and unsigned comparison
12940 2023-05-30 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12942 * config/riscv/autovec.md (fnma<mode>4): New pattern.
12943 (*fnma<mode>): Ditto.
12945 2023-05-29 Die Li <lidie@eswincomputing.com>
12947 * config/riscv/riscv.cc (riscv_expand_conditional_move_onesided):
12949 (riscv_expand_conditional_move): Reuse the TARGET_SFB_ALU expand
12950 process for TARGET_XTHEADCONDMOV
12952 2023-05-29 Uros Bizjak <ubizjak@gmail.com>
12955 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi2): Also require
12956 TARGET_AVX512BW to generate truncv16hiv16qi2.
12958 2023-05-29 Jivan Hakobyan <jivanhakobyan9@gmail.com>
12960 * config/riscv/riscv.md (and<mode>3): New expander.
12961 (*and<mode>3) New pattern.
12962 * config/riscv/predicates.md (arith_operand_or_mode_mask): New
12965 2023-05-29 Pan Li <pan2.li@intel.com>
12967 * config/riscv/riscv-v.cc (emit_vlmax_insn): Remove unnecessary
12968 comments and rename local variables.
12969 (emit_nonvlmax_insn): Diito.
12970 (emit_vlmax_merge_insn): Ditto.
12971 (emit_vlmax_cmp_insn): Ditto.
12972 (emit_vlmax_cmp_mu_insn): Ditto.
12973 (emit_scalar_move_insn): Ditto.
12975 2023-05-29 Pan Li <pan2.li@intel.com>
12977 * config/riscv/riscv-v.cc (emit_vlmax_insn): Eliminate the
12979 (emit_nonvlmax_insn): Ditto.
12980 (emit_vlmax_merge_insn): Ditto.
12981 (emit_vlmax_cmp_insn): Ditto.
12982 (emit_vlmax_cmp_mu_insn): Ditto.
12983 (expand_vec_series): Ditto.
12985 2023-05-29 Pan Li <pan2.li@intel.com>
12987 * config/riscv/riscv-protos.h (enum insn_type): New type.
12988 * config/riscv/riscv-v.cc (RVV_INSN_OPERANDS_MAX): New macro.
12989 (rvv_builder::can_duplicate_repeating_sequence_p): Align the referenced
12991 (rvv_builder::get_merged_repeating_sequence): Ditto.
12992 (rvv_builder::repeating_sequence_use_merge_profitable_p): New function
12993 to evaluate the optimization cost.
12994 (rvv_builder::get_merge_scalar_mask): New function to get the merge
12996 (emit_scalar_move_insn): New function to emit vmv.s.x.
12997 (emit_vlmax_integer_move_insn): New function to emit vlmax vmv.v.x.
12998 (emit_nonvlmax_integer_move_insn): New function to emit nonvlmax
13000 (get_repeating_sequence_dup_machine_mode): New function to get the dup
13002 (expand_vector_init_merge_repeating_sequence): New function to perform
13004 (expand_vec_init): Add this vector init optimization.
13005 * config/riscv/riscv.h (BITS_PER_WORD): New macro.
13007 2023-05-29 Eric Botcazou <ebotcazou@adacore.com>
13009 * tree-ssa-loop-manip.cc (create_iv): Try harder to find a SLOC to
13010 put onto the increment when it is inserted after the position.
13012 2023-05-29 Eric Botcazou <ebotcazou@adacore.com>
13014 * match.pd ((T)P - (T)(P + A) -> -(T) A): Avoid artificial overflow
13017 2023-05-29 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13019 * config/riscv/riscv-vsetvl.cc (source_equal_p): Fix ICE.
13021 2023-05-29 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13023 * config/riscv/autovec.md (fma<mode>4): New pattern.
13024 (*fma<mode>): Ditto.
13025 * config/riscv/riscv-protos.h (enum insn_type): New enum.
13026 (emit_vlmax_ternary_insn): New function.
13027 * config/riscv/riscv-v.cc (emit_vlmax_ternary_insn): Ditto.
13029 2023-05-29 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13031 * config/riscv/vector.md: Fix vimuladd instruction bug.
13033 2023-05-29 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13035 * config/riscv/riscv.cc (global_state_unknown_p): New function.
13036 (riscv_mode_after): Fix incorrect VXM.
13038 2023-05-29 Pan Li <pan2.li@intel.com>
13040 * common/config/riscv/riscv-common.cc:
13041 (riscv_implied_info): Add zvfhmin item.
13042 (riscv_ext_version_table): Ditto.
13043 (riscv_ext_flag_table): Ditto.
13044 * config/riscv/riscv-opts.h (MASK_ZVFHMIN): New macro.
13045 (TARGET_ZFHMIN): Align indent.
13046 (TARGET_ZFH): Ditto.
13047 (TARGET_ZVFHMIN): New macro.
13049 2023-05-27 liuhongt <hongtao.liu@intel.com>
13052 * config/i386/sse.md (*andnot<mode>3): Extend below splitter
13053 to VI_AVX2 to cover more modes.
13055 2023-05-27 liuhongt <hongtao.liu@intel.com>
13057 * config/i386/x86-tune.def (X86_TUNE_AVOID_FALSE_DEP_FOR_BMI):
13058 Remove ATOM and ICELAKE(and later) core processors.
13060 2023-05-26 Robin Dapp <rdapp@ventanamicro.com>
13062 * config/riscv/autovec.md (<optab><mode>2): Add vneg/vnot.
13064 * config/riscv/riscv-protos.h (emit_vlmax_masked_mu_insn):
13066 * config/riscv/riscv-v.cc (emit_vlmax_masked_mu_insn): New
13069 2023-05-26 Robin Dapp <rdapp@ventanamicro.com>
13070 Juzhe Zhong <juzhe.zhong@rivai.ai>
13072 * config/riscv/autovec.md (<optab><v_double_trunc><mode>2): New
13074 (<optab><v_quad_trunc><mode>2): Dito.
13075 (<optab><v_oct_trunc><mode>2): Dito.
13076 (trunc<mode><v_double_trunc>2): Dito.
13077 (trunc<mode><v_quad_trunc>2): Dito.
13078 (trunc<mode><v_oct_trunc>2): Dito.
13079 * config/riscv/riscv-protos.h (vectorize_related_mode): Define.
13080 (autovectorize_vector_modes): Define.
13081 * config/riscv/riscv-v.cc (vectorize_related_mode): Implement
13083 (autovectorize_vector_modes): Implement hook.
13084 * config/riscv/riscv.cc (riscv_autovectorize_vector_modes):
13085 Implement target hook.
13086 (riscv_vectorize_related_mode): Implement target hook.
13087 (TARGET_VECTORIZE_AUTOVECTORIZE_VECTOR_MODES): Define.
13088 (TARGET_VECTORIZE_RELATED_MODE): Define.
13089 * config/riscv/vector-iterators.md: Add lowercase versions of
13090 mode_attr iterators.
13092 2023-05-26 Andrew Stubbs <ams@codesourcery.com>
13093 Tobias Burnus <tobias@codesourcery.com>
13095 * config/gcn/gcn-hsa.h (XNACKOPT): New macro.
13096 (ASM_SPEC): Use XNACKOPT.
13097 * config/gcn/gcn-opts.h (enum sram_ecc_type): Rename to ...
13098 (enum hsaco_attr_type): ... this, and generalize the names.
13099 (TARGET_XNACK): New macro.
13100 * config/gcn/gcn.cc (gcn_option_override): Update to sorry for all
13102 (output_file_start): Update xnack handling.
13103 (gcn_hsa_declare_function_name): Use TARGET_XNACK.
13104 * config/gcn/gcn.opt (-mxnack): Add the "on/off/any" syntax.
13105 (sram_ecc_type): Rename to ...
13106 (hsaco_attr_type: ... this.)
13107 * config/gcn/mkoffload.cc (SET_XNACK_ANY): New macro.
13108 (TEST_XNACK): Delete.
13109 (TEST_XNACK_ANY): New macro.
13110 (TEST_XNACK_ON): New macro.
13111 (main): Support the new -mxnack=on/off/any syntax.
13112 * doc/invoke.texi (-mxnack): Update for new syntax.
13114 2023-05-26 Andrew Pinski <apinski@marvell.com>
13116 * genmatch.cc (emit_debug_printf): New function.
13117 (dt_simplify::gen_1): Emit printf into the code
13118 before the `return true` or returning the folded result
13119 instead of emitting it always.
13121 2023-05-26 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
13123 * config/xtensa/xtensa-protos.h
13124 (xtensa_expand_block_set_unrolled_loop,
13125 xtensa_expand_block_set_small_loop): Remove.
13126 (xtensa_expand_block_set): New prototype.
13127 * config/xtensa/xtensa.cc
13128 (xtensa_expand_block_set_libcall): New subfunction.
13129 (xtensa_expand_block_set_unrolled_loop,
13130 xtensa_expand_block_set_small_loop): Rewrite as subfunctions.
13131 (xtensa_expand_block_set): New function that calls the above
13133 * config/xtensa/xtensa.md (memsetsi): Change to invoke only
13134 xtensa_expand_block_set().
13136 2023-05-26 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
13138 * config/xtensa/xtensa-protos.h (xtensa_m1_or_1_thru_15):
13140 * config/xtensa/xtensa.cc (xtensa_m1_or_1_thru_15):
13142 * config/xtensa/constraints.md (O):
13143 Change to use the above function.
13144 * config/xtensa/xtensa.md (*subsi3_from_const):
13145 New insn_and_split pattern.
13147 2023-05-26 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
13149 * config/xtensa/xtensa.md (*extzvsi-1bit_ashlsi3):
13150 Retract excessive line folding, and correct the value of
13151 the "length" insn attribute related to TARGET_DENSITY.
13152 (*extzvsi-1bit_addsubx): Ditto.
13154 2023-05-26 Uros Bizjak <ubizjak@gmail.com>
13156 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi):
13157 Do not disable call to ix86_expand_vecop_qihi2.
13159 2023-05-26 liuhongt <hongtao.liu@intel.com>
13163 * ira-costs.cc (scan_one_insn): Only use NO_REGS in cost
13164 calculation when !hard_regno_mode_ok for GENERAL_REGS and
13165 mode, otherwise still use GENERAL_REGS.
13167 2023-05-26 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13169 * config/riscv/riscv.cc (vector_zero_call_used_regs): Add
13170 explict VL and drop VL in ops.
13172 2023-05-25 Jin Ma <jinma@linux.alibaba.com>
13174 * sched-deps.cc (sched_macro_fuse_insns): Insns should not be fusion
13175 in different BB blocks.
13177 2023-05-25 Uros Bizjak <ubizjak@gmail.com>
13179 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi2):
13180 Rewrite to expand to 2x-wider (e.g. V16QI -> V16HImode)
13181 instructions when available. Emulate truncation via
13182 ix86_expand_vec_perm_const_1 when native truncate insn
13184 (ix86_expand_vecop_qihi_partial) <case MULT>: Use pmovzx
13185 when available. Trivially rename some variables.
13186 (ix86_expand_vecop_qihi): Unconditionally call ix86_expand_vecop_qihi2.
13187 * config/i386/i386.cc (ix86_multiplication_cost): Rewrite cost
13188 calculation of V*QImode emulations to account for generation of
13189 2x-wider mode instructions.
13190 (ix86_shift_rotate_cost): Update cost calculation of V*QImode
13191 emulations to account for generation of 2x-wider mode instructions.
13193 2023-05-25 Georg-Johann Lay <avr@gjlay.de>
13196 * config/avr/avr.cc (avr_can_inline_p): New static function.
13197 (TARGET_CAN_INLINE_P): Define to that function.
13199 2023-05-25 Georg-Johann Lay <avr@gjlay.de>
13202 * config/avr/avr.md (*movbitqi.0): Rename to *movbit<mode>.0-6.
13203 Handle any bit position and use mode QISI.
13204 * config/avr/avr.cc (avr_rtx_costs_1) [IOR]: Return a cost
13205 of 2 insns for bit-transfer of respective style.
13207 2023-05-25 Christophe Lyon <christophe.lyon@linaro.org>
13209 * config/arm/iterators.md (MVE_6): Remove.
13210 * config/arm/mve.md: Replace MVE_6 with MVE_5.
13212 2023-05-25 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
13213 Richard Sandiford <richard.sandiford@arm.com>
13215 * tree-vect-loop-manip.cc (vect_adjust_loop_lens_control): New
13217 (vect_set_loop_controls_directly): Add decrement IV support.
13218 (vect_set_loop_condition_partial_vectors): Ditto.
13219 * tree-vect-loop.cc (_loop_vec_info::_loop_vec_info): New
13221 * tree-vectorizer.h (LOOP_VINFO_USING_DECREMENTING_IV_P): New
13224 2023-05-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13227 * config/aarch64/aarch64-simd.md (aarch64_fcadd<rot><mode>): Rename to...
13228 (aarch64_fcadd<rot><mode><vczle><vczbe>): ... This.
13229 Fix canonicalization of PLUS operands.
13230 (aarch64_fcmla<rot><mode>): Rename to...
13231 (aarch64_fcmla<rot><mode><vczle><vczbe>): ... This.
13232 Fix canonicalization of PLUS operands.
13233 (aarch64_fcmla_lane<rot><mode>): Rename to...
13234 (aarch64_fcmla_lane<rot><mode><vczle><vczbe>): ... This.
13235 Fix canonicalization of PLUS operands.
13236 (aarch64_fcmla_laneq<rot>v4hf): Rename to...
13237 (aarch64_fcmla_laneq<rot>v4hf<vczle><vczbe>): ... This.
13238 Fix canonicalization of PLUS operands.
13239 (aarch64_fcmlaq_lane<rot><mode>): Fix canonicalization of PLUS operands.
13241 2023-05-25 Chris Sidebottom <chris.sidebottom@arm.com>
13243 * config/arm/arm.md (rbitsi2): Rename to...
13244 (arm_rbit): ... This.
13245 (ctzsi2): Adjust for the above.
13246 (arm_rev16si2): Convert to define_expand.
13247 (arm_rev16si2_alt1): New pattern.
13248 (arm_rev16si2_alt): Rename to...
13249 (*arm_rev16si2_alt2): ... This.
13250 * config/arm/arm_acle.h (__ror, __rorl, __rorll, __clz, __clzl, __clzll,
13251 __cls, __clsl, __clsll, __revsh, __rev, __revl, __revll, __rev16,
13252 __rev16l, __rev16ll, __rbit, __rbitl, __rbitll): Define intrinsics.
13253 * config/arm/arm_acle_builtins.def (rbit, rev16si2): Define builtins.
13255 2023-05-25 Alex Coplan <alex.coplan@arm.com>
13258 * config/arm/arm.md (movdf): Generate temporary pseudo in DImode
13260 * config/arm/vfp.md (no_literal_pool_df_immediate): Rather than punning an
13261 lvalue DFmode pseudo into DImode, use a DImode pseudo and pun it into
13262 DFmode as an rvalue.
13264 2023-05-25 Richard Biener <rguenther@suse.de>
13267 * tree-vect-stmts.cc (vectorizable_condition): For
13268 embedded comparisons also handle the case when the target
13269 only provides vec_cmp and vcond_mask.
13271 2023-05-25 Claudiu Zissulescu <claziss@gmail.com>
13273 * config/arc/arc.cc (arc_call_tls_get_addr): Simplify access using
13276 2023-05-25 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
13278 * config/aarch64/aarch64.cc (scalar_move_insn_p): New function.
13279 (seq_cost_ignoring_scalar_moves): Likewise.
13280 (aarch64_expand_vector_init): Call seq_cost_ignoring_scalar_moves.
13282 2023-05-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13284 * config/aarch64/arm_neon.h (vcage_f64): Reimplement with builtins.
13285 (vcage_f32): Likewise.
13286 (vcages_f32): Likewise.
13287 (vcageq_f32): Likewise.
13288 (vcaged_f64): Likewise.
13289 (vcageq_f64): Likewise.
13290 (vcagts_f32): Likewise.
13291 (vcagt_f32): Likewise.
13292 (vcagt_f64): Likewise.
13293 (vcagtq_f32): Likewise.
13294 (vcagtd_f64): Likewise.
13295 (vcagtq_f64): Likewise.
13296 (vcale_f32): Likewise.
13297 (vcale_f64): Likewise.
13298 (vcaled_f64): Likewise.
13299 (vcales_f32): Likewise.
13300 (vcaleq_f32): Likewise.
13301 (vcaleq_f64): Likewise.
13302 (vcalt_f32): Likewise.
13303 (vcalt_f64): Likewise.
13304 (vcaltd_f64): Likewise.
13305 (vcaltq_f32): Likewise.
13306 (vcaltq_f64): Likewise.
13307 (vcalts_f32): Likewise.
13309 2023-05-25 Hu, Lin1 <lin1.hu@intel.com>
13313 * config/i386/avx512bwintrin.h (_mm512_srli_epi16): Change type from
13314 int to const int or const int to const unsigned int.
13315 (_mm512_mask_srli_epi16): Ditto.
13316 (_mm512_slli_epi16): Ditto.
13317 (_mm512_mask_slli_epi16): Ditto.
13318 (_mm512_maskz_slli_epi16): Ditto.
13319 (_mm512_srai_epi16): Ditto.
13320 (_mm512_mask_srai_epi16): Ditto.
13321 (_mm512_maskz_srai_epi16): Ditto.
13322 * config/i386/avx512fintrin.h (_mm512_slli_epi64): Ditto.
13323 (_mm512_mask_slli_epi64): Ditto.
13324 (_mm512_maskz_slli_epi64): Ditto.
13325 (_mm512_srli_epi64): Ditto.
13326 (_mm512_mask_srli_epi64): Ditto.
13327 (_mm512_maskz_srli_epi64): Ditto.
13328 (_mm512_srai_epi64): Ditto.
13329 (_mm512_mask_srai_epi64): Ditto.
13330 (_mm512_maskz_srai_epi64): Ditto.
13331 (_mm512_slli_epi32): Ditto.
13332 (_mm512_mask_slli_epi32): Ditto.
13333 (_mm512_maskz_slli_epi32): Ditto.
13334 (_mm512_srli_epi32): Ditto.
13335 (_mm512_mask_srli_epi32): Ditto.
13336 (_mm512_maskz_srli_epi32): Ditto.
13337 (_mm512_srai_epi32): Ditto.
13338 (_mm512_mask_srai_epi32): Ditto.
13339 (_mm512_maskz_srai_epi32): Ditto.
13340 * config/i386/avx512vlbwintrin.h (_mm256_mask_srai_epi16): Ditto.
13341 (_mm256_maskz_srai_epi16): Ditto.
13342 (_mm_mask_srai_epi16): Ditto.
13343 (_mm_maskz_srai_epi16): Ditto.
13344 (_mm256_mask_slli_epi16): Ditto.
13345 (_mm256_maskz_slli_epi16): Ditto.
13346 (_mm_mask_slli_epi16): Ditto.
13347 (_mm_maskz_slli_epi16): Ditto.
13348 (_mm_maskz_srli_epi16): Ditto.
13349 * config/i386/avx512vlintrin.h (_mm256_mask_srli_epi32): Ditto.
13350 (_mm256_maskz_srli_epi32): Ditto.
13351 (_mm_mask_srli_epi32): Ditto.
13352 (_mm_maskz_srli_epi32): Ditto.
13353 (_mm256_mask_srli_epi64): Ditto.
13354 (_mm256_maskz_srli_epi64): Ditto.
13355 (_mm_mask_srli_epi64): Ditto.
13356 (_mm_maskz_srli_epi64): Ditto.
13357 (_mm256_mask_srai_epi32): Ditto.
13358 (_mm256_maskz_srai_epi32): Ditto.
13359 (_mm_mask_srai_epi32): Ditto.
13360 (_mm_maskz_srai_epi32): Ditto.
13361 (_mm256_srai_epi64): Ditto.
13362 (_mm256_mask_srai_epi64): Ditto.
13363 (_mm256_maskz_srai_epi64): Ditto.
13364 (_mm_srai_epi64): Ditto.
13365 (_mm_mask_srai_epi64): Ditto.
13366 (_mm_maskz_srai_epi64): Ditto.
13367 (_mm_mask_slli_epi32): Ditto.
13368 (_mm_maskz_slli_epi32): Ditto.
13369 (_mm_mask_slli_epi64): Ditto.
13370 (_mm_maskz_slli_epi64): Ditto.
13371 (_mm256_mask_slli_epi32): Ditto.
13372 (_mm256_maskz_slli_epi32): Ditto.
13373 (_mm256_mask_slli_epi64): Ditto.
13374 (_mm256_maskz_slli_epi64): Ditto.
13376 2023-05-25 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13378 * config/riscv/vector.md: Remove FRM_REGNUM dependency in rtz
13381 2023-05-25 Aldy Hernandez <aldyh@redhat.com>
13383 * data-streamer-in.cc (streamer_read_value_range): Handle NANs.
13384 * data-streamer-out.cc (streamer_write_vrange): Same.
13385 * value-range.h (class vrange): Make streamer_write_vrange a friend.
13387 2023-05-25 Aldy Hernandez <aldyh@redhat.com>
13389 * value-query.cc (range_query::get_tree_range): Set NAN directly
13391 * value-range.cc (frange::set): Assert that bounds are not NAN.
13393 2023-05-25 Aldy Hernandez <aldyh@redhat.com>
13395 * value-range.cc (add_vrange): Handle known NANs.
13397 2023-05-25 Aldy Hernandez <aldyh@redhat.com>
13399 * value-range.h (frange::set_nan): New.
13401 2023-05-25 Alexandre Oliva <oliva@adacore.com>
13404 * emit-rtl.cc (validate_subreg): Reject a SUBREG of a MEM that
13405 requires stricter alignment than MEM's.
13407 2023-05-24 Andrew MacLeod <amacleod@redhat.com>
13409 PR tree-optimization/107822
13410 PR tree-optimization/107986
13411 * Makefile.in (OBJS): Add gimple-range-phi.o.
13412 * gimple-range-cache.h (ranger_cache::m_estimate): New
13413 phi_analyzer pointer member.
13414 * gimple-range-fold.cc (fold_using_range::range_of_phi): Use
13415 phi_analyzer if no loop info is available.
13416 * gimple-range-phi.cc: New file.
13417 * gimple-range-phi.h: New file.
13418 * tree-vrp.cc (execute_ranger_vrp): Utililze a phi_analyzer.
13420 2023-05-24 Andrew MacLeod <amacleod@redhat.com>
13422 * gimple-range-fold.cc (fur_list::fur_list): Add range_query param
13424 (fold_range): Add range_query parameter.
13425 (fur_relation::fur_relation): New.
13426 (fur_relation::trio): New.
13427 (fur_relation::register_relation): New.
13428 (fold_relations): New.
13429 * gimple-range-fold.h (fold_range): Adjust prototypes.
13430 (fold_relations): New.
13432 2023-05-24 Andrew MacLeod <amacleod@redhat.com>
13434 * gimple-range-cache.cc (ssa_cache::range_of_expr): New.
13435 * gimple-range-cache.h (class ssa_cache): Inherit from range_query.
13436 (ranger_cache::const_query): New.
13437 * gimple-range.cc (gimple_ranger::const_query): New.
13438 * gimple-range.h (gimple_ranger::const_query): New prototype.
13440 2023-05-24 Andrew MacLeod <amacleod@redhat.com>
13442 * gimple-range-cache.cc (ssa_cache::dump): Use get_range.
13443 (ssa_cache::dump_range_query): Delete.
13444 (ssa_lazy_cache::dump_range_query): Delete.
13445 (ssa_lazy_cache::get_range): Move from header file.
13446 (ssa_lazy_cache::clear_range): ditto.
13447 (ssa_lazy_cache::clear): Ditto.
13448 * gimple-range-cache.h (class ssa_cache): Virtualize.
13449 (class ssa_lazy_cache): Inherit and virtualize.
13451 2023-05-24 Aldy Hernandez <aldyh@redhat.com>
13453 * value-range.h (vrange::kind): Remove.
13455 2023-05-24 Roger Sayle <roger@nextmovesoftware.com>
13457 PR middle-end/109840
13458 * match.pd <popcount optimizations>: Preserve zero-extension when
13459 optimizing popcount((T)bswap(x)) and popcount((T)rotate(x,y)) as
13460 popcount((T)x), so the popcount's argument keeps the same type.
13461 <parity optimizations>: Likewise preserve extensions when
13462 simplifying parity((T)bswap(x)) and parity((T)rotate(x,y)) as
13463 parity((T)x), so that the parity's argument type is the same.
13465 2023-05-24 Aldy Hernandez <aldyh@redhat.com>
13467 * ipa-cp.cc (ipa_value_range_from_jfunc): Use new ipa_vr API.
13468 (ipcp_store_vr_results): Same.
13469 * ipa-prop.cc (ipa_vr::ipa_vr): New.
13470 (ipa_vr::get_vrange): New.
13471 (ipa_vr::set_unknown): New.
13472 (ipa_vr::streamer_read): New.
13473 (ipa_vr::streamer_write): New.
13474 (write_ipcp_transformation_info): Use new ipa_vr API.
13475 (read_ipcp_transformation_info): Same.
13476 (ipa_vr::nonzero_p): Delete.
13477 (ipcp_update_vr): Use new ipa_vr API.
13478 * ipa-prop.h (class ipa_vr): Provide an API and hide internals.
13479 * ipa-sra.cc (zap_useless_ipcp_results): Use new ipa_vr API.
13481 2023-05-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13483 * config/mcore/mcore.cc (output_inline_const) Make buffer smaller to
13484 silence overflow warnings later on.
13486 2023-05-24 Uros Bizjak <ubizjak@gmail.com>
13488 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi2):
13489 Remove handling of V8QImode.
13490 * config/i386/mmx.md (v<insn>v8qi3): Move from sse.md.
13491 Call ix86_expand_vecop_qihi_partial. Enable for TARGET_MMX_WITH_SSE.
13492 (v<insn>v4qi3): Ditto.
13493 * config/i386/sse.md (v<insn>v8qi3): Remove.
13495 2023-05-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13498 * config/aarch64/aarch64-simd.md (aarch64_simd_lshr<mode>): Rename to...
13499 (aarch64_simd_lshr<mode><vczle><vczbe>): ... This.
13500 (aarch64_simd_ashr<mode>): Rename to...
13501 (aarch64_simd_ashr<mode><vczle><vczbe>): ... This.
13502 (aarch64_simd_imm_shl<mode>): Rename to...
13503 (aarch64_simd_imm_shl<mode><vczle><vczbe>): ... This.
13504 (aarch64_simd_reg_sshl<mode>): Rename to...
13505 (aarch64_simd_reg_sshl<mode><vczle><vczbe>): ... This.
13506 (aarch64_simd_reg_shl<mode>_unsigned): Rename to...
13507 (aarch64_simd_reg_shl<mode>_unsigned<vczle><vczbe>): ... This.
13508 (aarch64_simd_reg_shl<mode>_signed): Rename to...
13509 (aarch64_simd_reg_shl<mode>_signed<vczle><vczbe>): ... This.
13510 (vec_shr_<mode>): Rename to...
13511 (vec_shr_<mode><vczle><vczbe>): ... This.
13512 (aarch64_<sur>shl<mode>): Rename to...
13513 (aarch64_<sur>shl<mode><vczle><vczbe>): ... This.
13514 (aarch64_<sur>q<r>shl<mode>): Rename to...
13515 (aarch64_<sur>q<r>shl<mode><vczle><vczbe>): ... This.
13517 2023-05-24 Richard Biener <rguenther@suse.de>
13520 * config/i386/i386-expand.cc (ix86_expand_vector_init_general):
13521 Perform final vector composition using
13522 ix86_expand_vector_init_general instead of setting
13523 the highpart and lowpart which causes spilling.
13525 2023-05-24 Andrew MacLeod <amacleod@redhat.com>
13527 PR tree-optimization/109695
13528 * gimple-range-cache.cc (ranger_cache::get_global_range): Add
13530 * gimple-range-cache.h (ranger_cache::get_global_range): Ditto.
13531 * gimple-range.cc (gimple_ranger::range_of_stmt): Pass changed
13532 flag to set_global_range.
13533 (gimple_ranger::prefill_stmt_dependencies): Ditto.
13535 2023-05-24 Andrew MacLeod <amacleod@redhat.com>
13537 PR tree-optimization/109695
13538 * gimple-range-cache.cc (temporal_cache::temporal_value): Return
13540 (temporal_cache::current_p): Check always_current method.
13541 (temporal_cache::set_always_current): Add param and set value
13543 (temporal_cache::always_current_p): New.
13544 (ranger_cache::get_global_range): Adjust.
13545 (ranger_cache::set_global_range): set always current first.
13547 2023-05-24 Andrew MacLeod <amacleod@redhat.com>
13549 PR tree-optimization/109695
13550 * gimple-range-cache.cc (ranger_cache::get_global_range): Call
13551 fold_range with global query to choose an initial value.
13553 2023-05-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13555 * config/riscv/riscv-protos.h (enum frm_field_enum): Add FRM_
13558 2023-05-24 Richard Biener <rguenther@suse.de>
13560 PR tree-optimization/109849
13561 * tree-ssa-pre.cc (do_hoist_insertion): Do not intersect
13562 expressions but take the first sets.
13564 2023-05-24 Gaius Mulley <gaiusmod2@gmail.com>
13567 * doc/gm2.texi (High procedure function): New node.
13568 (Using): New menu entry for High procedure function.
13570 2023-05-24 Richard Sandiford <richard.sandiford@arm.com>
13572 PR rtl-optimization/109940
13573 * early-remat.cc (postorder_index): Rename to...
13574 (rpo_index): ...this.
13575 (compare_candidates): Sort by decreasing rpo_index rather than
13576 increasing postorder_index.
13577 (early_remat::sort_candidates): Calculate the forward RPO from
13579 (early_remat::local_phase): Follow forward RPO using DF_FORWARD,
13580 rather than DF_BACKWARD in reverse.
13582 2023-05-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13585 * config/arm/arm-builtins.cc (SAT_BINOP_UNSIGNED_IMM_QUALIFIERS): Use
13586 qualifier_none for the return operand.
13588 2023-05-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13590 * config/riscv/autovec.md (<optab><mode>3): New pattern.
13591 (one_cmpl<mode>2): Ditto.
13592 (*<optab>not<mode>): Ditto.
13593 (*n<optab><mode>): Ditto.
13594 * config/riscv/riscv-v.cc (expand_vec_cmp_float): Change to
13597 2023-05-24 Kewen Lin <linkw@linux.ibm.com>
13599 * tree-vect-slp.cc (vect_transform_slp_perm_load_1): Adjust the
13600 calculation on n_perms by considering nvectors_per_build.
13602 2023-05-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13603 Richard Sandiford <richard.sandiford@arm.com>
13605 * config/riscv/autovec.md (@vcond_mask_<mode><vm>): New pattern.
13606 (vec_cmp<mode><vm>): New pattern.
13607 (vec_cmpu<mode><vm>): New pattern.
13608 (vcond<V:mode><VI:mode>): New pattern.
13609 (vcondu<V:mode><VI:mode>): New pattern.
13610 * config/riscv/riscv-protos.h (enum insn_type): Add new enum.
13611 (emit_vlmax_merge_insn): New function.
13612 (emit_vlmax_cmp_insn): Ditto.
13613 (emit_vlmax_cmp_mu_insn): Ditto.
13614 (expand_vec_cmp): Ditto.
13615 (expand_vec_cmp_float): Ditto.
13616 (expand_vcond): Ditto.
13617 * config/riscv/riscv-v.cc (emit_vlmax_merge_insn): Ditto.
13618 (emit_vlmax_cmp_insn): Ditto.
13619 (emit_vlmax_cmp_mu_insn): Ditto.
13620 (get_cmp_insn_code): Ditto.
13621 (expand_vec_cmp): Ditto.
13622 (expand_vec_cmp_float): Ditto.
13623 (expand_vcond): Ditto.
13625 2023-05-24 Pan Li <pan2.li@intel.com>
13627 * config/riscv/genrvv-type-indexer.cc (main): Add
13628 unsigned_eew*_lmul1_interpret for indexer.
13629 * config/riscv/riscv-vector-builtins-functions.def (vreinterpret):
13630 Register vuint*m1_t interpret function.
13631 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_UNSIGNED_EEW8_LMUL1_INTERPRET_OPS):
13632 New macro for vuint8m1_t.
13633 (DEF_RVV_UNSIGNED_EEW16_LMUL1_INTERPRET_OPS): Likewise.
13634 (DEF_RVV_UNSIGNED_EEW32_LMUL1_INTERPRET_OPS): Likewise.
13635 (DEF_RVV_UNSIGNED_EEW64_LMUL1_INTERPRET_OPS): Likewise.
13636 (vbool1_t): Add to unsigned_eew*_interpret_ops.
13637 (vbool2_t): Likewise.
13638 (vbool4_t): Likewise.
13639 (vbool8_t): Likewise.
13640 (vbool16_t): Likewise.
13641 (vbool32_t): Likewise.
13642 (vbool64_t): Likewise.
13643 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_UNSIGNED_EEW8_LMUL1_INTERPRET_OPS):
13644 New macro for vuint*m1_t.
13645 (DEF_RVV_UNSIGNED_EEW16_LMUL1_INTERPRET_OPS): Likewise.
13646 (DEF_RVV_UNSIGNED_EEW32_LMUL1_INTERPRET_OPS): Likewise.
13647 (DEF_RVV_UNSIGNED_EEW64_LMUL1_INTERPRET_OPS): Likewise.
13648 (required_extensions_p): Add vuint*m1_t interpret case.
13649 * config/riscv/riscv-vector-builtins.def (unsigned_eew8_lmul1_interpret):
13650 Add vuint*m1_t interpret to base type.
13651 (unsigned_eew16_lmul1_interpret): Likewise.
13652 (unsigned_eew32_lmul1_interpret): Likewise.
13653 (unsigned_eew64_lmul1_interpret): Likewise.
13655 2023-05-24 Pan Li <pan2.li@intel.com>
13657 * config/riscv/genrvv-type-indexer.cc (EEW_SIZE_LIST): New macro
13658 for the eew size list.
13659 (LMUL1_LOG2): New macro for the log2 value of lmul=1.
13660 (main): Add signed_eew*_lmul1_interpret for indexer.
13661 * config/riscv/riscv-vector-builtins-functions.def (vreinterpret):
13662 Register vint*m1_t interpret function.
13663 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_SIGNED_EEW8_LMUL1_INTERPRET_OPS):
13664 New macro for vint8m1_t.
13665 (DEF_RVV_SIGNED_EEW16_LMUL1_INTERPRET_OPS): Likewise.
13666 (DEF_RVV_SIGNED_EEW32_LMUL1_INTERPRET_OPS): Likewise.
13667 (DEF_RVV_SIGNED_EEW64_LMUL1_INTERPRET_OPS): Likewise.
13668 (vbool1_t): Add to signed_eew*_interpret_ops.
13669 (vbool2_t): Likewise.
13670 (vbool4_t): Likewise.
13671 (vbool8_t): Likewise.
13672 (vbool16_t): Likewise.
13673 (vbool32_t): Likewise.
13674 (vbool64_t): Likewise.
13675 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_SIGNED_EEW8_LMUL1_INTERPRET_OPS):
13676 New macro for vint*m1_t.
13677 (DEF_RVV_SIGNED_EEW16_LMUL1_INTERPRET_OPS): Likewise.
13678 (DEF_RVV_SIGNED_EEW32_LMUL1_INTERPRET_OPS): Likewise.
13679 (DEF_RVV_SIGNED_EEW64_LMUL1_INTERPRET_OPS): Likewise.
13680 (required_extensions_p): Add vint8m1_t interpret case.
13681 * config/riscv/riscv-vector-builtins.def (signed_eew8_lmul1_interpret):
13682 Add vint*m1_t interpret to base type.
13683 (signed_eew16_lmul1_interpret): Likewise.
13684 (signed_eew32_lmul1_interpret): Likewise.
13685 (signed_eew64_lmul1_interpret): Likewise.
13687 2023-05-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13689 * config/riscv/autovec.md: Adjust for new interface.
13690 * config/riscv/riscv-protos.h (emit_vlmax_insn): Add VL operand.
13691 (emit_nonvlmax_insn): Add AVL operand.
13692 * config/riscv/riscv-v.cc (emit_vlmax_insn): Add VL operand.
13693 (emit_nonvlmax_insn): Add AVL operand.
13694 (sew64_scalar_helper): Adjust for new interface.
13695 (expand_tuple_move): Ditto.
13696 * config/riscv/vector.md: Ditto.
13698 2023-05-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13700 * config/riscv/riscv-v.cc (expand_vec_series): Remove magic number.
13701 (expand_const_vector): Ditto.
13702 (legitimize_move): Ditto.
13703 (sew64_scalar_helper): Ditto.
13704 (expand_tuple_move): Ditto.
13705 (expand_vector_init_insert_elems): Ditto.
13706 * config/riscv/riscv.cc (vector_zero_call_used_regs): Ditto.
13708 2023-05-24 liuhongt <hongtao.liu@intel.com>
13711 * config/i386/i386.cc (ix86_gimple_fold_builtin): Fold
13712 _mm{,256,512}_abs_{epi8,epi16,epi32,epi64} and
13713 _mm_abs_{pi8,pi16,pi32} into gimple ABS_EXPR.
13714 (ix86_masked_all_ones): Handle 64-bit mask.
13715 * config/i386/i386-builtin.def: Replace icode of related
13716 non-mask simd abs builtins with CODE_FOR_nothing.
13718 2023-05-23 Martin Uecker <uecker@tugraz.at>
13721 * function.cc (gimplify_parm_type): Remove function.
13722 (gimplify_parameters): Call gimplify_type_sizes.
13724 2023-05-23 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
13726 * config/xtensa/xtensa.md (*addsubx): Rename from '*addx',
13727 and change to also accept '*subx' pattern.
13730 2023-05-23 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
13732 * config/xtensa/predicates.md (addsub_operator): New.
13733 * config/xtensa/xtensa.md (*extzvsi-1bit_ashlsi3,
13734 *extzvsi-1bit_addsubx): New insn_and_split patterns.
13735 * config/xtensa/xtensa.cc (xtensa_rtx_costs):
13736 Add a special case about ifcvt 'noce_try_cmove()' to handle
13737 constant loads that do not fit into signed 12 bits in the
13738 patterns added above.
13740 2023-05-23 Richard Biener <rguenther@suse.de>
13742 PR tree-optimization/109747
13743 * tree-vect-slp.cc (vect_prologue_cost_for_slp): Pass down
13744 the SLP node only once to the cost hook.
13746 2023-05-23 Georg-Johann Lay <avr@gjlay.de>
13748 * config/avr/avr.cc (avr_insn_cost): New static function.
13749 (TARGET_INSN_COST): Define to that function.
13751 2023-05-23 Richard Biener <rguenther@suse.de>
13754 * config/i386/i386.cc (ix86_vector_costs::add_stmt_cost):
13755 For vector construction or splats apply GPR->XMM move
13756 costing. QImode memory can be handled directly only
13757 with SSE4.1 pinsrb.
13759 2023-05-23 Richard Biener <rguenther@suse.de>
13761 PR tree-optimization/108752
13762 * tree-vect-stmts.cc (vectorizable_operation): For bit
13763 operations with generic word_mode vectors do not cost
13764 an extra stmt. For plus, minus and negate also cost the
13765 constant materialization.
13767 2023-05-23 Uros Bizjak <ubizjak@gmail.com>
13769 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi_partial):
13770 Call ix86_expand_vec_shift_qihi_constant for shifts
13771 with constant count operand.
13772 * config/i386/i386.cc (ix86_shift_rotate_cost):
13773 Handle V4QImode and V8QImode.
13774 * config/i386/mmx.md (<insn>v8qi3): New insn pattern.
13775 (<insn>v4qi3): Ditto.
13777 2023-05-23 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13779 * config/riscv/vector.md: Add mode.
13781 2023-05-23 Aldy Hernandez <aldyh@redhat.com>
13783 PR tree-optimization/109934
13784 * value-range.cc (irange::invert): Remove buggy special case.
13786 2023-05-23 Richard Biener <rguenther@suse.de>
13788 * tree-ssa-pre.cc (compute_antic_aux): Dump the correct
13791 2023-05-23 Richard Sandiford <richard.sandiford@arm.com>
13794 * config/aarch64/aarch64.cc (aarch64_modes_tieable_p): Allow
13795 subregs between any scalars that are 64 bits or smaller.
13796 * config/aarch64/iterators.md (SUBDI_BITS): New int iterator.
13797 (bits_etype): New int attribute.
13798 * config/aarch64/aarch64.md (*insv_reg<mode>_<SUBDI_BITS>)
13799 (*aarch64_bfi<GPI:mode><ALLX:mode>_<SUBDI_BITS>): New patterns.
13800 (*aarch64_bfidi<ALLX:mode>_subreg_<SUBDI_BITS>): Likewise.
13802 2023-05-23 Richard Sandiford <richard.sandiford@arm.com>
13804 * doc/md.texi: Document that <FOO> can be used to refer to the
13805 numerical value of an int iterator FOO. Tweak other parts of
13806 the int iterator documentation.
13807 * read-rtl.cc (iterator_group::has_self_attr): New field.
13808 (map_attr_string): When has_self_attr is true, make <FOO>
13809 expand to the current value of iterator FOO.
13810 (initialize_iterators): Set has_self_attr for int iterators.
13812 2023-05-23 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13814 * config/riscv/autovec.md: Refactor the framework of RVV auto-vectorization.
13815 * config/riscv/riscv-protos.h (RVV_MISC_OP_NUM): Ditto.
13816 (RVV_UNOP_NUM): New macro.
13817 (RVV_BINOP_NUM): Ditto.
13818 (legitimize_move): Refactor the framework of RVV auto-vectorization.
13819 (emit_vlmax_op): Ditto.
13820 (emit_vlmax_reg_op): Ditto.
13821 (emit_len_op): Ditto.
13822 (emit_len_binop): Ditto.
13823 (emit_vlmax_tany_many): Ditto.
13824 (emit_nonvlmax_tany_many): Ditto.
13825 (sew64_scalar_helper): Ditto.
13826 (expand_tuple_move): Ditto.
13827 * config/riscv/riscv-v.cc (emit_pred_op): Ditto.
13828 (emit_pred_binop): Ditto.
13829 (emit_vlmax_op): Ditto.
13830 (emit_vlmax_tany_many): New function.
13831 (emit_len_op): Remove.
13832 (emit_nonvlmax_tany_many): New function.
13833 (emit_vlmax_reg_op): Remove.
13834 (emit_len_binop): Ditto.
13835 (emit_index_op): Ditto.
13836 (expand_vec_series): Refactor the framework of RVV auto-vectorization.
13837 (expand_const_vector): Ditto.
13838 (legitimize_move): Ditto.
13839 (sew64_scalar_helper): Ditto.
13840 (expand_tuple_move): Ditto.
13841 (expand_vector_init_insert_elems): Ditto.
13842 * config/riscv/riscv.cc (vector_zero_call_used_regs): Ditto.
13843 * config/riscv/vector.md: Ditto.
13845 2023-05-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13848 * config/aarch64/aarch64-simd.md (add_vec_concat_subst_le): Add predicate
13849 and constraint for operand 0.
13850 (add_vec_concat_subst_be): Likewise.
13852 2023-05-23 Richard Biener <rguenther@suse.de>
13854 PR tree-optimization/109849
13855 * tree-ssa-pre.cc (do_hoist_insertion): Compute ANTIC_OUT
13856 and use that to determine what to hoist.
13858 2023-05-23 Eric Botcazou <ebotcazou@adacore.com>
13860 * fold-const.cc (native_encode_initializer) <CONSTRUCTOR>: Apply the
13861 specific treatment for bit-fields only if they have an integral type
13862 and filter out non-integral bit-fields that do not start and end on
13865 2023-05-23 Aldy Hernandez <aldyh@redhat.com>
13867 PR tree-optimization/109920
13868 * value-range.h (RESIZABLE>::~int_range): Use delete[].
13870 2023-05-22 Uros Bizjak <ubizjak@gmail.com>
13872 * config/i386/i386.cc (ix86_shift_rotate_cost): Correct
13873 calcuation of integer vector mode costs to reflect generated
13874 instruction sequences of different integer vector modes and
13875 different target ABIs. Remove "speed" function argument.
13876 (ix86_rtx_costs): Update call for removed function argument.
13877 (ix86_vector_costs::add_stmt_cost): Ditto.
13879 2023-05-22 Aldy Hernandez <aldyh@redhat.com>
13881 * value-range.h (class Value_Range): Implement set_zero,
13882 set_nonzero, and nonzero_p.
13884 2023-05-22 Uros Bizjak <ubizjak@gmail.com>
13886 * config/i386/i386.cc (ix86_multiplication_cost): Add
13887 the cost of a memory read to the cost of V?QImode sequences.
13889 2023-05-22 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13891 * config/riscv/riscv-v.cc: Add "m_" prefix.
13893 2023-05-22 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
13895 * tree-vect-loop.cc (vect_get_loop_len): Fix issue for
13896 multiple-rgroup of length.
13897 * tree-vect-stmts.cc (vectorizable_store): Ditto.
13898 (vectorizable_load): Ditto.
13899 * tree-vectorizer.h (vect_get_loop_len): Ditto.
13901 2023-05-22 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13903 * config/riscv/riscv.cc (riscv_const_insns): Reorganize the
13906 2023-05-22 Kewen Lin <linkw@linux.ibm.com>
13908 * tree-vect-slp.cc (vect_transform_slp_perm_load_1): Refactor the
13909 handling for the case index == count.
13911 2023-05-21 Georg-Johann Lay <avr@gjlay.de>
13914 * config/avr/avr.cc (avr_fold_builtin) [AVR_BUILTIN_INSERT_BITS]:
13915 Don't fold to XOR / AND / XOR if just one bit is copied to the
13918 2023-05-21 Roger Sayle <roger@nextmovesoftware.com>
13920 * config/nvptx/nvptx.cc (nvptx_expand_brev): Expand target
13921 builtin for bit reversal using brev instruction.
13922 (enum nvptx_builtins): Add NVPTX_BUILTIN_BREV and
13923 NVPTX_BUILTIN_BREVLL.
13924 (nvptx_init_builtins): Define "brev" and "brevll".
13925 (nvptx_expand_builtin): Expand NVPTX_BUILTIN_BREV and
13926 NVPTX_BUILTIN_BREVLL via nvptx_expand_brev function.
13927 * doc/extend.texi (Nvidia PTX Builtin-in Functions): New
13928 section, document __builtin_nvptx_brev{,ll}.
13930 2023-05-21 Jakub Jelinek <jakub@redhat.com>
13932 PR tree-optimization/109505
13933 * match.pd ((x | CST1) & CST2 -> (x & CST2) | (CST1 & CST2),
13934 Combine successive equal operations with constants,
13935 (A +- CST1) +- CST2 -> A + CST3, (CST1 - A) +- CST2 -> CST3 - A,
13936 CST1 - (CST2 - A) -> CST3 + A): Use ! on ops with 2 CONSTANT_CLASS_P
13939 2023-05-21 Andrew Pinski <apinski@marvell.com>
13941 * expr.cc (expand_single_bit_test): Correct bitpos for big-endian.
13943 2023-05-21 Pan Li <pan2.li@intel.com>
13945 * config/riscv/genrvv-type-indexer.cc (BOOL_SIZE_LIST): Add the
13946 rest bool size, aka 2, 4, 8, 16, 32, 64.
13947 * config/riscv/riscv-vector-builtins-functions.def (vreinterpret):
13948 Register vbool[2|4|8|16|32|64] interpret function.
13949 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_BOOL2_INTERPRET_OPS):
13950 New macro for vbool2_t.
13951 (DEF_RVV_BOOL4_INTERPRET_OPS): Likewise.
13952 (DEF_RVV_BOOL8_INTERPRET_OPS): Likewise.
13953 (DEF_RVV_BOOL16_INTERPRET_OPS): Likewise.
13954 (DEF_RVV_BOOL32_INTERPRET_OPS): Likewise.
13955 (DEF_RVV_BOOL64_INTERPRET_OPS): Likewise.
13956 (vint8m1_t): Add the type to bool[2|4|8|16|32|64]_interpret_ops.
13957 (vint16m1_t): Likewise.
13958 (vint32m1_t): Likewise.
13959 (vint64m1_t): Likewise.
13960 (vuint8m1_t): Likewise.
13961 (vuint16m1_t): Likewise.
13962 (vuint32m1_t): Likewise.
13963 (vuint64m1_t): Likewise.
13964 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_BOOL2_INTERPRET_OPS):
13965 New macro for vbool2_t.
13966 (DEF_RVV_BOOL4_INTERPRET_OPS): Likewise.
13967 (DEF_RVV_BOOL8_INTERPRET_OPS): Likewise.
13968 (DEF_RVV_BOOL16_INTERPRET_OPS): Likewise.
13969 (DEF_RVV_BOOL32_INTERPRET_OPS): Likewise.
13970 (DEF_RVV_BOOL64_INTERPRET_OPS): Likewise.
13971 (required_extensions_p): Add vbool[2|4|8|16|32|64] interpret case.
13972 * config/riscv/riscv-vector-builtins.def (bool2_interpret): Add
13973 vbool2_t interprect to base type.
13974 (bool4_interpret): Likewise.
13975 (bool8_interpret): Likewise.
13976 (bool16_interpret): Likewise.
13977 (bool32_interpret): Likewise.
13978 (bool64_interpret): Likewise.
13980 2023-05-21 Andrew Pinski <apinski@marvell.com>
13982 PR middle-end/109919
13983 * expr.cc (expand_single_bit_test): Don't use the
13984 target for expand_expr.
13986 2023-05-20 Gerald Pfeifer <gerald@pfeifer.com>
13988 * doc/install.texi (Specific): Remove de facto empty alpha*-*-*
13991 2023-05-20 Pan Li <pan2.li@intel.com>
13993 * mode-switching.cc (entity_map): Initialize the array to zero.
13996 2023-05-20 Triffid Hunter <triffid.hunter@gmail.com>
13999 * config/avr/avr.md (divmodpsi, udivmodpsi, divmodsi, udivmodsi):
14000 Remove superfluous "parallel" in insn pattern.
14001 ([u]divmod<mode>4): Tidy code. Use gcc_unreachable() instead of
14002 printing error text to assembly.
14004 2023-05-20 Andrew Pinski <apinski@marvell.com>
14006 * expr.cc (fold_single_bit_test): Rename to ...
14007 (expand_single_bit_test): This and expand directly.
14008 (do_store_flag): Update for the rename function.
14010 2023-05-20 Andrew Pinski <apinski@marvell.com>
14012 * expr.cc (fold_single_bit_test): Use BIT_FIELD_REF
14013 instead of shift/and.
14015 2023-05-20 Andrew Pinski <apinski@marvell.com>
14017 * expr.cc (fold_single_bit_test): Add an assert
14018 and simplify based on code being NE_EXPR or EQ_EXPR.
14020 2023-05-20 Andrew Pinski <apinski@marvell.com>
14022 * expr.cc (fold_single_bit_test): Take inner and bitnum
14023 instead of arg0 and arg1. Update the code.
14024 (do_store_flag): Don't create a tree when calling
14025 fold_single_bit_test instead just call it with the bitnum
14026 and the inner tree.
14028 2023-05-20 Andrew Pinski <apinski@marvell.com>
14030 * expr.cc (fold_single_bit_test): Use get_def_for_expr
14031 instead of checking the inner's code.
14033 2023-05-20 Andrew Pinski <apinski@marvell.com>
14035 * expr.cc (fold_single_bit_test_into_sign_test): Inline into ...
14036 (fold_single_bit_test): This and simplify.
14038 2023-05-20 Andrew Pinski <apinski@marvell.com>
14040 * fold-const.cc (fold_single_bit_test_into_sign_test): Move to
14042 (fold_single_bit_test): Likewise.
14043 * expr.cc (fold_single_bit_test_into_sign_test): Move from fold-const.cc
14044 (fold_single_bit_test): Likewise and make static.
14045 * fold-const.h (fold_single_bit_test): Remove declaration.
14047 2023-05-20 Die Li <lidie@eswincomputing.com>
14049 * config/riscv/riscv.cc (riscv_expand_conditional_move): Fix mode
14052 2023-05-20 Raphael Moreira Zinsly <rzinsly@ventanamicro.com>
14054 * config/riscv/bitmanip.md (branch<X:mode>_bext): New split pattern.
14056 2023-05-20 Raphael Moreira Zinsly <rzinsly@ventanamicro.com>
14059 * config/riscv/bitmanip.md
14060 (<bitmanip_optab>disi2): Match with any_extend.
14061 (<bitmanip_optab>disi2_sext): New pattern to match
14062 with sign extend using an ANDI instruction.
14064 2023-05-19 Nathan Sidwell <nathan@acm.org>
14067 * opts.h (handle_deferred_dump_options): Declare.
14068 * opts-global.cc (handle_common_deferred_options): Do not handle
14070 (handle_deferred_dump_options): New.
14071 * toplev.cc (toplev::main): Call it after plugin init.
14073 2023-05-19 Joern Rennecke <joern.rennecke@embecosm.com>
14075 * config/riscv/constraints.md (DsS, DsD): Restore agreement
14076 with shiftm1 mode attribute.
14078 2023-05-19 Andrew Pinski <apinski@marvell.com>
14081 * gcc.cc (default_compilers["@c-header"]): Add %w
14082 after the --output-pch.
14084 2023-05-19 Vineet Gupta <vineetg@rivosinc.com>
14086 * config/riscv/riscv.cc (riscv_split_integer): if loval is equal
14087 to hival, ASHIFT the corresponding regs.
14089 2023-05-19 Robin Dapp <rdapp@ventanamicro.com>
14091 * config/riscv/riscv.cc (riscv_const_insns): Remove else.
14093 2023-05-19 Jakub Jelinek <jakub@redhat.com>
14095 PR tree-optimization/105776
14096 * tree-ssa-math-opts.cc (arith_overflow_check_p): If cast_stmt is
14097 non-NULL, allow division statement to have a cast as single imm use
14098 rather than comparison/condition.
14099 (match_arith_overflow): In that case remove the cast stmt in addition
14100 to the division statement.
14102 2023-05-19 Jakub Jelinek <jakub@redhat.com>
14104 PR tree-optimization/101856
14105 * tree-ssa-math-opts.cc (match_arith_overflow): Pattern detect
14106 unsigned __builtin_mul_overflow_p even when umulv4_optab doesn't
14107 support it but umul_highpart_optab does.
14109 2023-05-19 Eric Botcazou <ebotcazou@adacore.com>
14111 * varasm.cc (output_constructor_bitfield): Call tree_to_uhwi instead
14112 of tree_to_shwi on array indices. Minor tweaks.
14114 2023-05-18 Bernhard Reutner-Fischer <aldot@gcc.gnu.org>
14116 * alias.cc (ref_all_alias_ptr_type_p): Use _P() defines from tree.h.
14117 * attribs.cc (diag_attr_exclusions): Ditto.
14118 (decl_attributes): Ditto.
14119 (build_type_attribute_qual_variant): Ditto.
14120 * builtins.cc (fold_builtin_carg): Ditto.
14121 (fold_builtin_next_arg): Ditto.
14122 (do_mpc_arg2): Ditto.
14123 * cfgexpand.cc (expand_return): Ditto.
14124 * cgraph.h (decl_in_symtab_p): Ditto.
14125 (symtab_node::get_create): Ditto.
14126 * dwarf2out.cc (base_type_die): Ditto.
14127 (implicit_ptr_descriptor): Ditto.
14128 (gen_array_type_die): Ditto.
14129 (gen_type_die_with_usage): Ditto.
14130 (optimize_location_into_implicit_ptr): Ditto.
14131 * expr.cc (do_store_flag): Ditto.
14132 * fold-const.cc (negate_expr_p): Ditto.
14133 (fold_negate_expr_1): Ditto.
14134 (fold_convert_const): Ditto.
14135 (fold_convert_loc): Ditto.
14136 (constant_boolean_node): Ditto.
14137 (fold_binary_op_with_conditional_arg): Ditto.
14138 (build_fold_addr_expr_with_type_loc): Ditto.
14139 (fold_comparison): Ditto.
14140 (fold_checksum_tree): Ditto.
14141 (tree_unary_nonnegative_warnv_p): Ditto.
14142 (integer_valued_real_unary_p): Ditto.
14143 (fold_read_from_constant_string): Ditto.
14144 * gcc-rich-location.cc (maybe_range_label_for_tree_type_mismatch::get_text): Ditto.
14145 * gimple-expr.cc (useless_type_conversion_p): Ditto.
14146 (is_gimple_reg): Ditto.
14147 (is_gimple_asm_val): Ditto.
14148 (mark_addressable): Ditto.
14149 * gimple-expr.h (is_gimple_variable): Ditto.
14150 (virtual_operand_p): Ditto.
14151 * gimple-ssa-warn-access.cc (pass_waccess::check_dangling_stores): Ditto.
14152 * gimplify.cc (gimplify_bind_expr): Ditto.
14153 (gimplify_return_expr): Ditto.
14154 (gimple_add_padding_init_for_auto_var): Ditto.
14155 (gimplify_addr_expr): Ditto.
14156 (omp_add_variable): Ditto.
14157 (omp_notice_variable): Ditto.
14158 (omp_get_base_pointer): Ditto.
14159 (omp_strip_components_and_deref): Ditto.
14160 (omp_strip_indirections): Ditto.
14161 (omp_accumulate_sibling_list): Ditto.
14162 (omp_build_struct_sibling_lists): Ditto.
14163 (gimplify_adjust_omp_clauses_1): Ditto.
14164 (gimplify_adjust_omp_clauses): Ditto.
14165 (gimplify_omp_for): Ditto.
14166 (goa_lhs_expr_p): Ditto.
14167 (gimplify_one_sizepos): Ditto.
14168 * graphite-scop-detection.cc (scop_detection::graphite_can_represent_scev): Ditto.
14169 * ipa-devirt.cc (odr_types_equivalent_p): Ditto.
14170 * ipa-prop.cc (ipa_set_jf_constant): Ditto.
14171 (propagate_controlled_uses): Ditto.
14172 * ipa-sra.cc (type_prevails_p): Ditto.
14173 (scan_expr_access): Ditto.
14174 * optabs-tree.cc (optab_for_tree_code): Ditto.
14175 * toplev.cc (wrapup_global_declaration_1): Ditto.
14176 * trans-mem.cc (transaction_invariant_address_p): Ditto.
14177 * tree-cfg.cc (verify_types_in_gimple_reference): Ditto.
14178 (verify_gimple_comparison): Ditto.
14179 (verify_gimple_assign_binary): Ditto.
14180 (verify_gimple_assign_single): Ditto.
14181 * tree-complex.cc (get_component_ssa_name): Ditto.
14182 * tree-emutls.cc (lower_emutls_2): Ditto.
14183 * tree-inline.cc (copy_tree_body_r): Ditto.
14184 (estimate_move_cost): Ditto.
14185 (copy_decl_for_dup_finish): Ditto.
14186 * tree-nested.cc (convert_nonlocal_omp_clauses): Ditto.
14187 (note_nonlocal_vla_type): Ditto.
14188 (convert_local_omp_clauses): Ditto.
14189 (remap_vla_decls): Ditto.
14190 (fixup_vla_decls): Ditto.
14191 * tree-parloops.cc (loop_has_vector_phi_nodes): Ditto.
14192 * tree-pretty-print.cc (print_declaration): Ditto.
14193 (print_call_name): Ditto.
14194 * tree-sra.cc (compare_access_positions): Ditto.
14195 * tree-ssa-alias.cc (compare_type_sizes): Ditto.
14196 * tree-ssa-ccp.cc (get_default_value): Ditto.
14197 * tree-ssa-coalesce.cc (populate_coalesce_list_for_outofssa): Ditto.
14198 * tree-ssa-dom.cc (reduce_vector_comparison_to_scalar_comparison): Ditto.
14199 * tree-ssa-forwprop.cc (can_propagate_from): Ditto.
14200 * tree-ssa-propagate.cc (may_propagate_copy): Ditto.
14201 * tree-ssa-sccvn.cc (fully_constant_vn_reference_p): Ditto.
14202 * tree-ssa-sink.cc (statement_sink_location): Ditto.
14203 * tree-ssa-structalias.cc (type_must_have_pointers): Ditto.
14204 * tree-ssa-ter.cc (find_replaceable_in_bb): Ditto.
14205 * tree-ssa-uninit.cc (warn_uninit): Ditto.
14206 * tree-ssa.cc (maybe_rewrite_mem_ref_base): Ditto.
14207 (non_rewritable_mem_ref_base): Ditto.
14208 * tree-streamer-in.cc (lto_input_ts_type_non_common_tree_pointers): Ditto.
14209 * tree-streamer-out.cc (write_ts_type_non_common_tree_pointers): Ditto.
14210 * tree-vect-generic.cc (do_binop): Ditto.
14212 * tree-vect-stmts.cc (vect_init_vector): Ditto.
14213 * tree-vector-builder.h (tree_vector_builder::note_representative): Ditto.
14214 * tree.cc (sign_mask_for): Ditto.
14215 (verify_type_variant): Ditto.
14216 (gimple_canonical_types_compatible_p): Ditto.
14217 (verify_type): Ditto.
14218 * ubsan.cc (get_ubsan_type_info_for_type): Ditto.
14219 * var-tracking.cc (prepare_call_arguments): Ditto.
14220 (vt_add_function_parameters): Ditto.
14221 * varasm.cc (decode_addr_const): Ditto.
14223 2023-05-18 Bernhard Reutner-Fischer <aldot@gcc.gnu.org>
14225 * omp-low.cc (scan_sharing_clauses): Use _P() defines from tree.h.
14226 (lower_reduction_clauses): Ditto.
14227 (lower_send_clauses): Ditto.
14228 (lower_omp_task_reductions): Ditto.
14229 * omp-oacc-neuter-broadcast.cc (install_var_field): Ditto.
14230 (worker_single_copy): Ditto.
14231 * omp-offload.cc (oacc_rewrite_var_decl): Ditto.
14232 * omp-simd-clone.cc (plausible_type_for_simd_clone): Ditto.
14234 2023-05-18 Bernhard Reutner-Fischer <aldot@gcc.gnu.org>
14236 * lto-streamer-in.cc (lto_input_var_decl_ref): Use _P defines from
14238 (lto_read_body_or_constructor): Ditto.
14239 * lto-streamer-out.cc (tree_is_indexable): Ditto.
14240 (lto_output_var_decl_ref): Ditto.
14241 (DFS::DFS_write_tree_body): Ditto.
14242 (wrap_refs): Ditto.
14243 (write_symbol_extension_info): Ditto.
14245 2023-05-18 Bernhard Reutner-Fischer <aldot@gcc.gnu.org>
14247 * config/aarch64/aarch64.cc (aarch64_short_vector_p): Use _P
14248 defines from tree.h.
14249 (aarch64_mangle_type): Ditto.
14250 * config/alpha/alpha.cc (alpha_in_small_data_p): Ditto.
14251 (alpha_gimplify_va_arg_1): Ditto.
14252 * config/arc/arc.cc (arc_encode_section_info): Ditto.
14253 (arc_is_aux_reg_p): Ditto.
14254 (arc_is_uncached_mem_p): Ditto.
14255 (arc_handle_aux_attribute): Ditto.
14256 * config/arm/arm.cc (arm_handle_isr_attribute): Ditto.
14257 (arm_handle_cmse_nonsecure_call): Ditto.
14258 (arm_set_default_type_attributes): Ditto.
14259 (arm_is_segment_info_known): Ditto.
14260 (arm_mangle_type): Ditto.
14261 * config/arm/unknown-elf.h (IN_NAMED_SECTION_P): Ditto.
14262 * config/avr/avr.cc (avr_lookup_function_attribute1): Ditto.
14263 (avr_decl_absdata_p): Ditto.
14264 (avr_insert_attributes): Ditto.
14265 (avr_section_type_flags): Ditto.
14266 (avr_encode_section_info): Ditto.
14267 * config/bfin/bfin.cc (bfin_handle_l2_attribute): Ditto.
14268 * config/bpf/bpf.cc (bpf_core_compute): Ditto.
14269 * config/c6x/c6x.cc (c6x_in_small_data_p): Ditto.
14270 * config/csky/csky.cc (csky_handle_isr_attribute): Ditto.
14271 (csky_mangle_type): Ditto.
14272 * config/darwin-c.cc (darwin_pragma_unused): Ditto.
14273 * config/darwin.cc (is_objc_metadata): Ditto.
14274 * config/epiphany/epiphany.cc (epiphany_function_ok_for_sibcall): Ditto.
14275 * config/epiphany/epiphany.h (ROUND_TYPE_ALIGN): Ditto.
14276 * config/frv/frv.cc (frv_emit_movsi): Ditto.
14277 * config/gcn/gcn-tree.cc (gcn_lockless_update): Ditto.
14278 * config/gcn/gcn.cc (gcn_asm_output_symbol_ref): Ditto.
14279 * config/h8300/h8300.cc (h8300_encode_section_info): Ditto.
14280 * config/i386/i386-expand.cc: Ditto.
14281 * config/i386/i386.cc (type_natural_mode): Ditto.
14282 (ix86_function_arg): Ditto.
14283 (ix86_data_alignment): Ditto.
14284 (ix86_local_alignment): Ditto.
14285 (ix86_simd_clone_compute_vecsize_and_simdlen): Ditto.
14286 * config/i386/winnt-cxx.cc (i386_pe_type_dllimport_p): Ditto.
14287 (i386_pe_type_dllexport_p): Ditto.
14288 (i386_pe_adjust_class_at_definition): Ditto.
14289 * config/i386/winnt.cc (i386_pe_determine_dllimport_p): Ditto.
14290 (i386_pe_binds_local_p): Ditto.
14291 (i386_pe_section_type_flags): Ditto.
14292 * config/ia64/ia64.cc (ia64_encode_section_info): Ditto.
14293 (ia64_gimplify_va_arg): Ditto.
14294 (ia64_in_small_data_p): Ditto.
14295 * config/iq2000/iq2000.cc (iq2000_function_arg): Ditto.
14296 * config/lm32/lm32.cc (lm32_in_small_data_p): Ditto.
14297 * config/loongarch/loongarch.cc (loongarch_handle_model_attribute): Ditto.
14298 * config/m32c/m32c.cc (m32c_insert_attributes): Ditto.
14299 * config/mcore/mcore.cc (mcore_mark_dllimport): Ditto.
14300 (mcore_encode_section_info): Ditto.
14301 * config/microblaze/microblaze.cc (microblaze_elf_in_small_data_p): Ditto.
14302 * config/mips/mips.cc (mips_output_aligned_decl_common): Ditto.
14303 * config/mmix/mmix.cc (mmix_encode_section_info): Ditto.
14304 * config/nvptx/nvptx.cc (nvptx_encode_section_info): Ditto.
14305 (pass_in_memory): Ditto.
14306 (nvptx_generate_vector_shuffle): Ditto.
14307 (nvptx_lockless_update): Ditto.
14308 * config/pa/pa.cc (pa_function_arg_padding): Ditto.
14309 (pa_function_value): Ditto.
14310 (pa_function_arg): Ditto.
14311 * config/pa/pa.h (IN_NAMED_SECTION_P): Ditto.
14312 (TEXT_SPACE_P): Ditto.
14313 * config/pa/som.h (MAKE_DECL_ONE_ONLY): Ditto.
14314 * config/pdp11/pdp11.cc (pdp11_return_in_memory): Ditto.
14315 * config/riscv/riscv.cc (riscv_in_small_data_p): Ditto.
14316 (riscv_mangle_type): Ditto.
14317 * config/rl78/rl78.cc (rl78_insert_attributes): Ditto.
14318 (rl78_addsi3_internal): Ditto.
14319 * config/rs6000/aix.h (ROUND_TYPE_ALIGN): Ditto.
14320 * config/rs6000/darwin.h (ROUND_TYPE_ALIGN): Ditto.
14321 * config/rs6000/freebsd64.h (ROUND_TYPE_ALIGN): Ditto.
14322 * config/rs6000/linux64.h (ROUND_TYPE_ALIGN): Ditto.
14323 * config/rs6000/rs6000-call.cc (rs6000_function_arg_boundary): Ditto.
14324 (rs6000_function_arg_advance_1): Ditto.
14325 (rs6000_function_arg): Ditto.
14326 (rs6000_pass_by_reference): Ditto.
14327 * config/rs6000/rs6000-logue.cc (rs6000_function_ok_for_sibcall): Ditto.
14328 * config/rs6000/rs6000.cc (rs6000_data_alignment): Ditto.
14329 (rs6000_set_default_type_attributes): Ditto.
14330 (rs6000_elf_in_small_data_p): Ditto.
14331 (IN_NAMED_SECTION): Ditto.
14332 (rs6000_xcoff_encode_section_info): Ditto.
14333 (rs6000_function_value): Ditto.
14334 (invalid_arg_for_unprototyped_fn): Ditto.
14335 * config/s390/s390-c.cc (s390_fn_types_compatible): Ditto.
14336 (s390_vec_n_elem): Ditto.
14337 * config/s390/s390.cc (s390_check_type_for_vector_abi): Ditto.
14338 (s390_function_arg_integer): Ditto.
14339 (s390_return_in_memory): Ditto.
14340 (s390_encode_section_info): Ditto.
14341 * config/sh/sh.cc (sh_gimplify_va_arg_expr): Ditto.
14342 (sh_function_value): Ditto.
14343 * config/sol2.cc (solaris_insert_attributes): Ditto.
14344 * config/sparc/sparc.cc (function_arg_slotno): Ditto.
14345 * config/sparc/sparc.h (ROUND_TYPE_ALIGN): Ditto.
14346 * config/stormy16/stormy16.cc (xstormy16_encode_section_info): Ditto.
14347 (xstormy16_handle_below100_attribute): Ditto.
14348 * config/v850/v850.cc (v850_encode_section_info): Ditto.
14349 (v850_insert_attributes): Ditto.
14350 * config/visium/visium.cc (visium_pass_by_reference): Ditto.
14351 (visium_return_in_memory): Ditto.
14352 * config/xtensa/xtensa.cc (xtensa_multibss_section_type_flags): Ditto.
14354 2023-05-18 Uros Bizjak <ubizjak@gmail.com>
14356 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi_partial): New.
14357 (ix86_expand_vecop_qihi): Add op2vec bool variable.
14358 Do not set REG_EQUAL note.
14359 * config/i386/i386-protos.h (ix86_expand_vecop_qihi_partial):
14361 * config/i386/i386.cc (ix86_multiplication_cost): Handle
14362 V4QImode and V8QImode.
14363 * config/i386/mmx.md (mulv8qi3): New expander.
14365 * config/i386/sse.md (mulv8qi3): Remove.
14367 2023-05-18 Georg-Johann Lay <avr@gjlay.de>
14369 * config/avr/gen-avr-mmcu-specs.cc: Remove stale */ after // comment.
14371 2023-05-18 Jonathan Wakely <jwakely@redhat.com>
14373 PR bootstrap/105831
14374 * config.gcc: Use = operator instead of ==.
14376 2023-05-18 Michael Bäuerle <micha@NetBSD.org>
14378 PR bootstrap/105831
14379 * config/nvptx/gen-opt.sh: Use = operator instead of ==.
14380 * configure.ac: Likewise.
14381 * configure: Regenerate.
14383 2023-05-18 Stam Markianos-Wright <stam.markianos-wright@arm.com>
14385 * config/arm/arm_mve.h: (__ARM_mve_typeid): Add more pointer types.
14386 (__ARM_mve_coerce1): Remove.
14387 (__ARM_mve_coerce2): Remove.
14388 (__ARM_mve_coerce3): Remove.
14389 (__ARM_mve_coerce_i_scalar): New.
14390 (__ARM_mve_coerce_s8_ptr): New.
14391 (__ARM_mve_coerce_u8_ptr): New.
14392 (__ARM_mve_coerce_s16_ptr): New.
14393 (__ARM_mve_coerce_u16_ptr): New.
14394 (__ARM_mve_coerce_s32_ptr): New.
14395 (__ARM_mve_coerce_u32_ptr): New.
14396 (__ARM_mve_coerce_s64_ptr): New.
14397 (__ARM_mve_coerce_u64_ptr): New.
14398 (__ARM_mve_coerce_f_scalar): New.
14399 (__ARM_mve_coerce_f16_ptr): New.
14400 (__ARM_mve_coerce_f32_ptr): New.
14401 (__arm_vst4q): Change _coerce_ overloads.
14402 (__arm_vbicq): Change _coerce_ overloads.
14403 (__arm_vld1q): Change _coerce_ overloads.
14404 (__arm_vld1q_z): Change _coerce_ overloads.
14405 (__arm_vld2q): Change _coerce_ overloads.
14406 (__arm_vld4q): Change _coerce_ overloads.
14407 (__arm_vldrhq_gather_offset): Change _coerce_ overloads.
14408 (__arm_vldrhq_gather_offset_z): Change _coerce_ overloads.
14409 (__arm_vldrhq_gather_shifted_offset): Change _coerce_ overloads.
14410 (__arm_vldrhq_gather_shifted_offset_z): Change _coerce_ overloads.
14411 (__arm_vldrwq_gather_offset): Change _coerce_ overloads.
14412 (__arm_vldrwq_gather_offset_z): Change _coerce_ overloads.
14413 (__arm_vldrwq_gather_shifted_offset): Change _coerce_ overloads.
14414 (__arm_vldrwq_gather_shifted_offset_z): Change _coerce_ overloads.
14415 (__arm_vst1q_p): Change _coerce_ overloads.
14416 (__arm_vst2q): Change _coerce_ overloads.
14417 (__arm_vst1q): Change _coerce_ overloads.
14418 (__arm_vstrhq): Change _coerce_ overloads.
14419 (__arm_vstrhq_p): Change _coerce_ overloads.
14420 (__arm_vstrhq_scatter_offset_p): Change _coerce_ overloads.
14421 (__arm_vstrhq_scatter_offset): Change _coerce_ overloads.
14422 (__arm_vstrhq_scatter_shifted_offset_p): Change _coerce_ overloads.
14423 (__arm_vstrhq_scatter_shifted_offset): Change _coerce_ overloads.
14424 (__arm_vstrwq_p): Change _coerce_ overloads.
14425 (__arm_vstrwq): Change _coerce_ overloads.
14426 (__arm_vstrwq_scatter_offset): Change _coerce_ overloads.
14427 (__arm_vstrwq_scatter_offset_p): Change _coerce_ overloads.
14428 (__arm_vstrwq_scatter_shifted_offset): Change _coerce_ overloads.
14429 (__arm_vstrwq_scatter_shifted_offset_p): Change _coerce_ overloads.
14430 (__arm_vsetq_lane): Change _coerce_ overloads.
14431 (__arm_vldrbq_gather_offset): Change _coerce_ overloads.
14432 (__arm_vdwdupq_x_u8): Change _coerce_ overloads.
14433 (__arm_vdwdupq_x_u16): Change _coerce_ overloads.
14434 (__arm_vdwdupq_x_u32): Change _coerce_ overloads.
14435 (__arm_viwdupq_x_u8): Change _coerce_ overloads.
14436 (__arm_viwdupq_x_u16): Change _coerce_ overloads.
14437 (__arm_viwdupq_x_u32): Change _coerce_ overloads.
14438 (__arm_vidupq_x_u8): Change _coerce_ overloads.
14439 (__arm_vddupq_x_u8): Change _coerce_ overloads.
14440 (__arm_vidupq_x_u16): Change _coerce_ overloads.
14441 (__arm_vddupq_x_u16): Change _coerce_ overloads.
14442 (__arm_vidupq_x_u32): Change _coerce_ overloads.
14443 (__arm_vddupq_x_u32): Change _coerce_ overloads.
14444 (__arm_vldrdq_gather_offset): Change _coerce_ overloads.
14445 (__arm_vldrdq_gather_offset_z): Change _coerce_ overloads.
14446 (__arm_vldrdq_gather_shifted_offset): Change _coerce_ overloads.
14447 (__arm_vldrdq_gather_shifted_offset_z): Change _coerce_ overloads.
14448 (__arm_vldrbq_gather_offset_z): Change _coerce_ overloads.
14449 (__arm_vidupq_u16): Change _coerce_ overloads.
14450 (__arm_vidupq_u32): Change _coerce_ overloads.
14451 (__arm_vidupq_u8): Change _coerce_ overloads.
14452 (__arm_vddupq_u16): Change _coerce_ overloads.
14453 (__arm_vddupq_u32): Change _coerce_ overloads.
14454 (__arm_vddupq_u8): Change _coerce_ overloads.
14455 (__arm_viwdupq_m): Change _coerce_ overloads.
14456 (__arm_viwdupq_u16): Change _coerce_ overloads.
14457 (__arm_viwdupq_u32): Change _coerce_ overloads.
14458 (__arm_viwdupq_u8): Change _coerce_ overloads.
14459 (__arm_vdwdupq_m): Change _coerce_ overloads.
14460 (__arm_vdwdupq_u16): Change _coerce_ overloads.
14461 (__arm_vdwdupq_u32): Change _coerce_ overloads.
14462 (__arm_vdwdupq_u8): Change _coerce_ overloads.
14463 (__arm_vstrbq): Change _coerce_ overloads.
14464 (__arm_vstrbq_p): Change _coerce_ overloads.
14465 (__arm_vstrbq_scatter_offset_p): Change _coerce_ overloads.
14466 (__arm_vstrdq_scatter_offset_p): Change _coerce_ overloads.
14467 (__arm_vstrdq_scatter_offset): Change _coerce_ overloads.
14468 (__arm_vstrdq_scatter_shifted_offset_p): Change _coerce_ overloads.
14469 (__arm_vstrdq_scatter_shifted_offset): Change _coerce_ overloads.
14471 2023-05-18 Stam Markianos-Wright <stam.markianos-wright@arm.com>
14473 * config/arm/arm_mve.h (__arm_vbicq): Change coerce on
14476 2023-05-18 Stam Markianos-Wright <stam.markianos-wright@arm.com>
14478 * config/arm/arm_mve.h (__arm_vadcq_s32): Fix arithmetic.
14479 (__arm_vadcq_u32): Likewise.
14480 (__arm_vadcq_m_s32): Likewise.
14481 (__arm_vadcq_m_u32): Likewise.
14482 (__arm_vsbcq_s32): Likewise.
14483 (__arm_vsbcq_u32): Likewise.
14484 (__arm_vsbcq_m_s32): Likewise.
14485 (__arm_vsbcq_m_u32): Likewise.
14486 * config/arm/mve.md (get_fpscr_nzcvqc): Make unspec_volatile.
14488 2023-05-18 Andrea Corallo <andrea.corallo@arm.com>
14490 * config/arm/mve.md (mve_vrndq_m_f<mode>, mve_vrev64q_f<mode>)
14491 (mve_vrev32q_fv8hf, mve_vcvttq_f32_f16v4sf)
14492 (mve_vcvtbq_f32_f16v4sf, mve_vcvtq_to_f_<supf><mode>)
14493 (mve_vrev64q_<supf><mode>, mve_vcvtq_from_f_<supf><mode>)
14494 (mve_vmovltq_<supf><mode>, mve_vmovlbq_<supf><mode>)
14495 (mve_vcvtpq_<supf><mode>, mve_vcvtnq_<supf><mode>)
14496 (mve_vcvtmq_<supf><mode>, mve_vcvtaq_<supf><mode>)
14497 (mve_vmvnq_n_<supf><mode>, mve_vrev16q_<supf>v16qi)
14498 (mve_vctp<MVE_vctp>q<MVE_vpred>, mve_vbrsrq_n_f<mode>)
14499 (mve_vbrsrq_n_<supf><mode>, mve_vandq_f<mode>, mve_vbicq_f<mode>)
14500 (mve_vctp<MVE_vctp>q_m<MVE_vpred>, mve_vcvtbq_f16_f32v8hf)
14501 (mve_vcvttq_f16_f32v8hf, mve_veorq_f<mode>)
14502 (mve_vmlaldavxq_s<mode>, mve_vmlsldavq_s<mode>)
14503 (mve_vmlsldavxq_s<mode>, mve_vornq_f<mode>, mve_vorrq_f<mode>)
14504 (mve_vrmlaldavhxq_sv4si, mve_vcvtq_m_to_f_<supf><mode>)
14505 (mve_vshlcq_<supf><mode>, mve_vmvnq_m_<supf><mode>)
14506 (mve_vpselq_<supf><mode>, mve_vcvtbq_m_f16_f32v8hf)
14507 (mve_vcvtbq_m_f32_f16v4sf, mve_vcvttq_m_f16_f32v8hf)
14508 (mve_vcvttq_m_f32_f16v4sf, mve_vmlaldavq_p_<supf><mode>)
14509 (mve_vmlsldavaq_s<mode>, mve_vmlsldavaxq_s<mode>)
14510 (mve_vmlsldavq_p_s<mode>, mve_vmlsldavxq_p_s<mode>)
14511 (mve_vmvnq_m_n_<supf><mode>, mve_vorrq_m_n_<supf><mode>)
14512 (mve_vpselq_f<mode>, mve_vrev32q_m_fv8hf)
14513 (mve_vrev32q_m_<supf><mode>, mve_vrev64q_m_f<mode>)
14514 (mve_vrmlaldavhaxq_sv4si, mve_vrmlaldavhxq_p_sv4si)
14515 (mve_vrmlsldavhaxq_sv4si, mve_vrmlsldavhq_p_sv4si)
14516 (mve_vrmlsldavhxq_p_sv4si, mve_vrev16q_m_<supf>v16qi)
14517 (mve_vrmlaldavhq_p_<supf>v4si, mve_vrmlsldavhaq_sv4si)
14518 (mve_vandq_m_<supf><mode>, mve_vbicq_m_<supf><mode>)
14519 (mve_veorq_m_<supf><mode>, mve_vornq_m_<supf><mode>)
14520 (mve_vorrq_m_<supf><mode>, mve_vandq_m_f<mode>)
14521 (mve_vbicq_m_f<mode>, mve_veorq_m_f<mode>, mve_vornq_m_f<mode>)
14522 (mve_vorrq_m_f<mode>)
14523 (mve_vstrdq_scatter_shifted_offset_p_<supf>v2di_insn)
14524 (mve_vstrdq_scatter_shifted_offset_<supf>v2di_insn)
14525 (mve_vstrdq_scatter_base_wb_p_<supf>v2di) : Fix spacing and
14526 capitalization in the emitted asm.
14528 2023-05-18 Andrea Corallo <andrea.corallo@arm.com>
14530 * config/arm/constraints.md (mve_vldrd_immediate): Move it to
14532 (Ri): Move constraint definition from predicates.md.
14533 (Rl): Define new constraint.
14534 * config/arm/mve.md (mve_vstrwq_scatter_base_wb_p_<supf>v4si): Add
14535 missing constraint.
14536 (mve_vstrwq_scatter_base_wb_p_fv4sf): Add missing Up constraint
14537 for op 1, use mve_vstrw_immediate predicate and Rl constraint for
14538 op 2. Fix asm output spacing.
14539 (mve_vstrdq_scatter_base_wb_p_<supf>v2di): Add missing constraint.
14540 * config/arm/predicates.md (Ri) Move constraint to constraints.md
14541 (mve_vldrd_immediate): Move it from
14543 (mve_vstrw_immediate): New predicate.
14545 2023-05-18 Pan Li <pan2.li@intel.com>
14546 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
14547 Kito Cheng <kito.cheng@sifive.com>
14548 Richard Biener <rguenther@suse.de>
14549 Richard Sandiford <richard.sandiford@arm.com>
14551 * combine.cc (struct reg_stat_type): Extend machine_mode to 16 bits.
14552 * cse.cc (struct qty_table_elem): Extend machine_mode to 16 bits
14553 (struct table_elt): Extend machine_mode to 16 bits.
14554 (struct set): Ditto.
14555 * genmodes.cc (emit_mode_wider): Extend type from char to short.
14556 (emit_mode_complex): Ditto.
14557 (emit_mode_inner): Ditto.
14558 (emit_class_narrowest_mode): Ditto.
14559 * genopinit.cc (main): Extend the machine_mode limit.
14560 * ira-int.h (struct ira_allocno): Extend machine_mode to 16 bits and
14561 re-ordered the struct fields for padding.
14562 * machmode.h (MACHINE_MODE_BITSIZE): New macro.
14563 (GET_MODE_2XWIDER_MODE): Extend type from char to short.
14564 (get_mode_alignment): Extend type from char to short.
14565 * ree.cc (struct ext_modified): Extend machine_mode to 16 bits and
14566 removed the ATTRIBUTE_PACKED.
14567 * rtl-ssa/accesses.h: Extend machine_mode to 16 bits, narrow
14568 * rtl-ssa/internals.inl (rtl_ssa::access_info): Adjust the assignment.
14569 m_kind to 2 bits and remove m_spare.
14570 * rtl.h (RTX_CODE_BITSIZE): New macro.
14571 (struct rtx_def): Swap both the bit size and location between the
14572 rtx_code and the machine_mode.
14573 (subreg_shape::unique_id): Extend the machine_mode limit.
14574 * rtlanal.h: Extend machine_mode to 16 bits.
14575 * tree-core.h (struct tree_type_common): Extend machine_mode to 16
14576 bits and re-ordered the struct fields for padding.
14577 (struct tree_decl_common): Extend machine_mode to 16 bits.
14579 2023-05-17 Jin Ma <jinma@linux.alibaba.com>
14581 * genrecog.cc (print_nonbool_test): Fix type error of
14582 switch (SUBREG_BYTE (op))'.
14584 2023-05-17 Jin Ma <jinma@linux.alibaba.com>
14586 * common/config/riscv/riscv-common.cc: Remove
14587 trailing spaces on lines.
14588 * config/riscv/riscv.cc (riscv_legitimize_move): Likewise.
14589 * config/riscv/riscv.h (enum reg_class): Likewise.
14590 * config/riscv/riscv.md: Likewise.
14592 2023-05-17 John David Anglin <danglin@gcc.gnu.org>
14594 * config/pa/pa.md (clear_cache): New.
14596 2023-05-17 Arsen Arsenović <arsen@aarsen.me>
14598 * doc/extend.texi (C++ Concepts) <forall>: Remove extraneous
14599 parenthesis. Fix misnamed index entry.
14600 <concept>: Fix misnamed index entry.
14602 2023-05-17 Jivan Hakobyan <jivanhakobyan9@gmail.com>
14604 * config/riscv/riscv.md (*<optab><GPR:mode>3_mask): New pattern,
14606 (*<optab>si3_mask, *<optab>di3_mask): Here.
14607 (*<optab>si3_mask_1, *<optab>di3_mask_1): And here.
14608 * config/riscv/bitmanip.md (*<bitmanip_optab><GPR:mode>3_mask): New
14610 (*<bitmanip_optab>si3_sext_mask): Likewise.
14611 * config/riscv/iterators.md (shiftm1): Use const_si_mask_operand
14612 and const_di_mask_operand.
14613 (bitmanip_rotate): New iterator.
14614 (bitmanip_optab): Add rotates.
14615 * config/riscv/predicates.md (const_si_mask_operand): Renamed
14616 from const31_operand. Generalize to handle more mask constants.
14617 (const_di_mask_operand): Similarly.
14619 2023-05-17 Jakub Jelinek <jakub@redhat.com>
14622 * config/i386/i386-builtin-types.def (FLOAT128): Use
14623 float128t_type_node rather than float128_type_node.
14625 2023-05-17 Alexander Monakov <amonakov@ispras.ru>
14627 * tree-ssa-math-opts.cc (convert_mult_to_fma): Enable only for
14628 FP_CONTRACT_FAST (no functional change).
14630 2023-05-17 Uros Bizjak <ubizjak@gmail.com>
14632 * config/i386/i386.cc (ix86_multiplication_cost): Correct
14633 calcuation of integer vector mode costs to reflect generated
14634 instruction sequences of different integer vector modes and
14635 different target ABIs.
14637 2023-05-17 Juzhe-Zhong <juzhe.zhong@rivai.ai>
14639 * config/riscv/riscv-opts.h (enum riscv_entity): New enum.
14640 * config/riscv/riscv.cc (riscv_emit_mode_set): New function.
14641 (riscv_mode_needed): Ditto.
14642 (riscv_mode_after): Ditto.
14643 (riscv_mode_entry): Ditto.
14644 (riscv_mode_exit): Ditto.
14645 (riscv_mode_priority): Ditto.
14646 (TARGET_MODE_EMIT): New target hook.
14647 (TARGET_MODE_NEEDED): Ditto.
14648 (TARGET_MODE_AFTER): Ditto.
14649 (TARGET_MODE_ENTRY): Ditto.
14650 (TARGET_MODE_EXIT): Ditto.
14651 (TARGET_MODE_PRIORITY): Ditto.
14652 * config/riscv/riscv.h (OPTIMIZE_MODE_SWITCHING): Ditto.
14653 (NUM_MODES_FOR_MODE_SWITCHING): Ditto.
14654 * config/riscv/riscv.md: Add csrwvxrm.
14655 * config/riscv/vector.md (rnu,rne,rdn,rod,none): New attribute.
14656 (vxrmsi): New pattern.
14658 2023-05-17 Juzhe-Zhong <juzhe.zhong@rivai.ai>
14660 * config/riscv/riscv-vector-builtins-bases.cc: Introduce rounding mode.
14661 * config/riscv/riscv-vector-builtins-shapes.cc (struct alu_def): Ditto.
14662 (struct narrow_alu_def): Ditto.
14663 * config/riscv/riscv-vector-builtins.cc (function_builder::apply_predication): Ditto.
14664 (function_expander::use_exact_insn): Ditto.
14665 * config/riscv/riscv-vector-builtins.h (function_checker::arg_num): New function.
14666 (function_base::has_rounding_mode_operand_p): New function.
14668 2023-05-17 Andrew Pinski <apinski@marvell.com>
14670 * tree-ssa-forwprop.cc (simplify_builtin_call): Check
14671 against 0 instead of calling integer_zerop.
14673 2023-05-17 Juzhe-Zhong <juzhe.zhong@rivai.ai>
14675 * config/riscv/riscv-vector-builtins.cc (register_vxrm): New function.
14676 (DEF_RVV_VXRM_ENUM): New macro.
14677 (handle_pragma_vector): Add vxrm enum register.
14678 * config/riscv/riscv-vector-builtins.def (DEF_RVV_VXRM_ENUM): New macro.
14684 2023-05-17 Aldy Hernandez <aldyh@redhat.com>
14686 * value-range.h (Value_Range::operator=): New.
14688 2023-05-17 Aldy Hernandez <aldyh@redhat.com>
14690 * value-range.cc (vrange::operator=): Add a stub to copy
14691 unsupported ranges.
14692 * value-range.h (is_a <unsupported_range>): New.
14693 (Value_Range::operator=): Support copying unsupported ranges.
14695 2023-05-17 Aldy Hernandez <aldyh@redhat.com>
14697 * data-streamer-in.cc (streamer_read_real_value): New.
14698 (streamer_read_value_range): New.
14699 * data-streamer-out.cc (streamer_write_real_value): New.
14700 (streamer_write_vrange): New.
14701 * data-streamer.h (streamer_write_vrange): New.
14702 (streamer_read_value_range): New.
14704 2023-05-17 Jonathan Wakely <jwakely@redhat.com>
14707 * doc/invoke.texi (Code Gen Options): Note that -fshort-enums
14708 is ignored for a fixed underlying type.
14709 (C++ Dialect Options): Likewise for -fstrict-enums.
14711 2023-05-17 Tobias Burnus <tobias@codesourcery.com>
14713 * gimplify.cc (gimplify_scan_omp_clauses): Remove Fortran
14716 2023-05-17 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
14718 * config/s390/s390.cc (TARGET_ATOMIC_ALIGN_FOR_MODE):
14720 (s390_atomic_align_for_mode): New.
14722 2023-05-17 Jakub Jelinek <jakub@redhat.com>
14724 * wide-int.cc (wi::from_array): Add missing closing paren in function
14727 2023-05-17 Kewen Lin <linkw@linux.ibm.com>
14729 * tree-vect-loop.cc (vect_analyze_loop_1): Don't retry analysis with
14730 suggested unroll factor once the previous analysis fails.
14732 2023-05-17 Pan Li <pan2.li@intel.com>
14734 * config/riscv/genrvv-type-indexer.cc (BOOL_SIZE_LIST): New
14736 (main): Add bool1 to the type indexer.
14737 * config/riscv/riscv-vector-builtins-functions.def
14738 (vreinterpret): Register vbool1 interpret function.
14739 * config/riscv/riscv-vector-builtins-types.def
14740 (DEF_RVV_BOOL1_INTERPRET_OPS): New macro.
14741 (vint8m1_t): Add the type to bool1_interpret_ops.
14742 (vint16m1_t): Ditto.
14743 (vint32m1_t): Ditto.
14744 (vint64m1_t): Ditto.
14745 (vuint8m1_t): Ditto.
14746 (vuint16m1_t): Ditto.
14747 (vuint32m1_t): Ditto.
14748 (vuint64m1_t): Ditto.
14749 * config/riscv/riscv-vector-builtins.cc
14750 (DEF_RVV_BOOL1_INTERPRET_OPS): New macro.
14751 (required_extensions_p): Add bool1 interpret case.
14752 * config/riscv/riscv-vector-builtins.def
14753 (bool1_interpret): Add bool1 interpret to base type.
14754 * config/riscv/vector.md (@vreinterpret<mode>): Add new expand
14755 with VB dest for vreinterpret.
14757 2023-05-17 Jiufu Guo <guojiufu@linux.ibm.com>
14760 * config/rs6000/rs6000.cc (rs6000_emit_set_long_const): Support building
14761 constants through "lis; xoris".
14763 2023-05-16 Ajit Kumar Agarwal <aagarwa1@linux.ibm.com>
14765 * common/config/rs6000/rs6000-common.cc: Add REE pass as a
14766 default rs6000 target pass for O2 and above.
14767 * doc/invoke.texi: Document -free
14769 2023-05-16 Kito Cheng <kito.cheng@sifive.com>
14771 * common/config/riscv/riscv-common.cc (riscv_compute_multilib):
14772 Fix wrong select_kind...
14774 2023-05-16 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
14776 * config/s390/s390-protos.h (s390_expand_setmem): Change
14777 function signature.
14778 * config/s390/s390.cc (s390_expand_setmem): For memset's less
14779 than or equal to 256 byte do not perform a libc call.
14780 * config/s390/s390.md: Change expander into a version which
14783 2023-05-16 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
14785 * config/s390/s390-protos.h (s390_expand_movmem): New.
14786 * config/s390/s390.cc (s390_expand_movmem): New.
14787 * config/s390/s390.md (movmem<mode>): New.
14791 2023-05-16 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
14793 * config/s390/s390-protos.h (s390_expand_cpymem): Change
14794 function signature.
14795 * config/s390/s390.cc (s390_expand_cpymem): For memcpy's less
14796 than or equal to 256 byte do not perform a libc call.
14797 (s390_expand_insv): Adapt new function signature of
14798 s390_expand_cpymem.
14799 * config/s390/s390.md: Change expander into a version which
14802 2023-05-16 Andrew Pinski <apinski@marvell.com>
14804 PR tree-optimization/109424
14805 * match.pd: Add patterns for min/max of zero_one_valued
14808 2023-05-16 Juzhe-Zhong <juzhe.zhong@rivai.ai>
14810 * config/riscv/riscv-protos.h (enum frm_field_enum): New enum.
14811 * config/riscv/riscv-vector-builtins.cc
14812 (function_expander::use_ternop_insn): Add default rounding mode.
14813 (function_expander::use_widen_ternop_insn): Ditto.
14814 * config/riscv/riscv.cc (riscv_hard_regno_nregs): Add FRM REGNUM.
14815 (riscv_hard_regno_mode_ok): Ditto.
14816 (riscv_conditional_register_usage): Ditto.
14817 * config/riscv/riscv.h (DWARF_FRAME_REGNUM): Ditto.
14818 (FRM_REG_P): Ditto.
14819 (RISCV_DWARF_FRM): Ditto.
14820 * config/riscv/riscv.md: Ditto.
14821 * config/riscv/vector-iterators.md: split no frm and has frm operations.
14822 * config/riscv/vector.md (@pred_<optab><mode>_scalar): New pattern.
14823 (@pred_<optab><mode>): Ditto.
14825 2023-05-15 Aldy Hernandez <aldyh@redhat.com>
14827 PR tree-optimization/109695
14828 * value-range.cc (irange::operator=): Resize range.
14829 (irange::union_): Same.
14830 (irange::intersect): Same.
14831 (irange::invert): Same.
14832 (int_range_max): Default to 3 sub-ranges and resize as needed.
14833 * value-range.h (irange::maybe_resize): New.
14835 (int_range::int_range): Adjust for resizing.
14836 (int_range::operator=): Same.
14838 2023-05-15 Aldy Hernandez <aldyh@redhat.com>
14840 * ipa-cp.cc (ipcp_vr_lattice::meet_with_1): Avoid unnecessary
14842 * value-range.cc (irange::union_nonzero_bits): Return TRUE only
14843 when range changed.
14845 2023-05-15 Juzhe-Zhong <juzhe.zhong@rivai.ai>
14847 * config/riscv/riscv-protos.h (enum vxrm_field_enum): New enum.
14848 * config/riscv/riscv-vector-builtins.cc
14849 (function_expander::use_exact_insn): Add default rounding mode operand.
14850 * config/riscv/riscv.cc (riscv_hard_regno_nregs): Add VXRM_REGNUM.
14851 (riscv_hard_regno_mode_ok): Ditto.
14852 (riscv_conditional_register_usage): Ditto.
14853 * config/riscv/riscv.h (DWARF_FRAME_REGNUM): Ditto.
14854 (VXRM_REG_P): Ditto.
14855 (RISCV_DWARF_VXRM): Ditto.
14856 * config/riscv/riscv.md: Ditto.
14857 * config/riscv/vector.md: Ditto
14859 2023-05-15 Pan Li <pan2.li@intel.com>
14861 * optabs.cc (maybe_gen_insn): Add case to generate instruction
14862 that has 11 operands.
14864 2023-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14866 * config/aarch64/aarch64.cc (aarch64_rtx_costs, NEG case): Add costing
14867 logic for vector modes.
14869 2023-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14872 * config/aarch64/aarch64-simd.md (aarch64_cm<optab><mode>): Rename to...
14873 (aarch64_cm<optab><mode><vczle><vczbe>): ... This.
14874 (aarch64_cmtst<mode>): Rename to...
14875 (aarch64_cmtst<mode><vczle><vczbe>): ... This.
14876 (*aarch64_cmtst_same_<mode>): Rename to...
14877 (*aarch64_cmtst_same_<mode><vczle><vczbe>): ... This.
14878 (*aarch64_cmtstdi): Rename to...
14879 (*aarch64_cmtstdi<vczle><vczbe>): ... This.
14880 (aarch64_fac<optab><mode>): Rename to...
14881 (aarch64_fac<optab><mode><vczle><vczbe>): ... This.
14883 2023-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14886 * config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>): Rename to...
14887 (aarch64_s<optab><mode><vczle><vczbe>): ... This.
14889 2023-05-15 Pan Li <pan2.li@intel.com>
14890 Juzhe-Zhong <juzhe.zhong@rivai.ai>
14891 kito-cheng <kito.cheng@sifive.com>
14893 * config/riscv/riscv-v.cc (const_vlmax_p): New function for
14894 deciding the mode is constant or not.
14895 (set_len_and_policy): Optimize VLS-VLMAX code gen to vsetivli.
14897 2023-05-15 Richard Biener <rguenther@suse.de>
14899 PR tree-optimization/109848
14900 * tree-ssa-forwprop.cc (pass_forwprop::execute): Put the
14901 TARGET_MEM_REF address preparation before the store, not
14904 2023-05-15 Juzhe-Zhong <juzhe.zhong@rivai.ai>
14906 * config/riscv/riscv.cc
14907 (riscv_vectorize_preferred_vector_alignment): New function.
14908 (TARGET_VECTORIZE_PREFERRED_VECTOR_ALIGNMENT): New target hook.
14910 2023-05-14 Andrew Pinski <apinski@marvell.com>
14912 PR tree-optimization/109829
14913 * match.pd: Add pattern for `signbit(x) !=/== 0 ? x : -x`.
14915 2023-05-14 Uros Bizjak <ubizjak@gmail.com>
14918 * config/i386/i386.cc: Revert the 2023-05-11 change.
14919 (ix86_widen_mult_cost): Return high value instead of
14920 ICEing for unsupported modes.
14922 2023-05-14 Ard Biesheuvel <ardb@kernel.org>
14924 * config/i386/i386.cc (x86_function_profiler): Take
14925 ix86_direct_extern_access into account when generating calls
14928 2023-05-14 Pan Li <pan2.li@intel.com>
14930 * config/riscv/riscv-vector-builtins.cc (required_extensions_p):
14931 Refactor the or pattern to switch cases.
14933 2023-05-13 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
14935 * config/aarch64/aarch64.cc (aarch64_expand_vector_init_fallback): Rename
14936 aarch64_expand_vector_init to this, and remove interleaving case.
14937 Recursively call aarch64_expand_vector_init_fallback, instead of
14938 aarch64_expand_vector_init.
14939 (aarch64_unzip_vector_init): New function.
14940 (aarch64_expand_vector_init): Likewise.
14942 2023-05-13 Kito Cheng <kito.cheng@sifive.com>
14944 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::cleanup_insns):
14945 Pull out function call from the gcc_assert.
14947 2023-05-13 Kito Cheng <kito.cheng@sifive.com>
14949 * config/riscv/riscv-vsetvl.cc (vlmul_to_str): New.
14950 (policy_to_str): New.
14951 (vector_insn_info::dump): Use vlmul_to_str and policy_to_str.
14953 2023-05-13 Andrew Pinski <apinski@marvell.com>
14955 PR tree-optimization/109834
14956 * match.pd (popcount(bswap(x))->popcount(x)): Fix up unsigned type checking.
14957 (popcount(rotate(x,y))->popcount(x)): Likewise.
14959 2023-05-12 Uros Bizjak <ubizjak@gmail.com>
14961 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi2): Also
14962 reject ymm instructions for TARGET_PREFER_AVX128. Use generic
14963 gen_extend_insn to generate zero/sign extension instructions.
14965 (ix86_expand_vecop_qihi): Initialize interleave functions
14966 for MULT code only. Fix comments.
14968 2023-05-12 Uros Bizjak <ubizjak@gmail.com>
14971 * config/i386/mmx.md (mulv2si3): Remove expander.
14972 (mulv2si3): Rename insn pattern from *mulv2si.
14974 2023-05-12 Tobias Burnus <tobias@codesourcery.com>
14976 PR libstdc++/109816
14977 * lto-cgraph.cc (output_symtab): Guard lto_output_toplevel_asms by
14978 '!lto_stream_offload_p'.
14980 2023-05-12 Kito Cheng <kito.cheng@sifive.com>
14981 Juzhe-Zhong <juzhe.zhong@rivai.ai>
14984 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::get_vsetvl_at_end): New.
14985 (local_avl_compatible_p): New.
14986 (pass_vsetvl::local_eliminate_vsetvl_insn): Enhance local optimizations
14987 for LCM, rewrite as a backward algorithm.
14988 (pass_vsetvl::cleanup_insns): Use new local_eliminate_vsetvl_insn
14989 interface, handle a BB at once.
14991 2023-05-12 Richard Biener <rguenther@suse.de>
14993 PR tree-optimization/64731
14994 * tree-ssa-forwprop.cc (pass_forwprop::execute): Also
14995 handle TARGET_MEM_REF destinations of stores from vector
14998 2023-05-12 Richard Biener <rguenther@suse.de>
15000 PR tree-optimization/109791
15001 * match.pd (minus (convert ADDR_EXPR@0) (convert (pointer_plus @1 @2))):
15003 (minus (convert (pointer_plus @1 @2)) (convert ADDR_EXPR@0)):
15006 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15008 * config/arm/arm-mve-builtins-base.cc (vsriq): New.
15009 * config/arm/arm-mve-builtins-base.def (vsriq): New.
15010 * config/arm/arm-mve-builtins-base.h (vsriq): New.
15011 * config/arm/arm-mve-builtins.cc
15012 (function_instance::has_inactive_argument): Handle vsriq.
15013 * config/arm/arm_mve.h (vsriq): Remove.
15015 (vsriq_n_u8): Remove.
15016 (vsriq_n_s8): Remove.
15017 (vsriq_n_u16): Remove.
15018 (vsriq_n_s16): Remove.
15019 (vsriq_n_u32): Remove.
15020 (vsriq_n_s32): Remove.
15021 (vsriq_m_n_s8): Remove.
15022 (vsriq_m_n_u8): Remove.
15023 (vsriq_m_n_s16): Remove.
15024 (vsriq_m_n_u16): Remove.
15025 (vsriq_m_n_s32): Remove.
15026 (vsriq_m_n_u32): Remove.
15027 (__arm_vsriq_n_u8): Remove.
15028 (__arm_vsriq_n_s8): Remove.
15029 (__arm_vsriq_n_u16): Remove.
15030 (__arm_vsriq_n_s16): Remove.
15031 (__arm_vsriq_n_u32): Remove.
15032 (__arm_vsriq_n_s32): Remove.
15033 (__arm_vsriq_m_n_s8): Remove.
15034 (__arm_vsriq_m_n_u8): Remove.
15035 (__arm_vsriq_m_n_s16): Remove.
15036 (__arm_vsriq_m_n_u16): Remove.
15037 (__arm_vsriq_m_n_s32): Remove.
15038 (__arm_vsriq_m_n_u32): Remove.
15039 (__arm_vsriq): Remove.
15040 (__arm_vsriq_m): Remove.
15042 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15044 * config/arm/iterators.md (mve_insn): Add vsri.
15045 * config/arm/mve.md (mve_vsriq_n_<supf><mode>): Rename into ...
15046 (@mve_<mve_insn>q_n_<supf><mode>): .,. this.
15047 (mve_vsriq_m_n_<supf><mode>): Rename into ...
15048 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
15050 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15052 * config/arm/arm-mve-builtins-shapes.cc (ternary_rshift): New.
15053 * config/arm/arm-mve-builtins-shapes.h (ternary_rshift): New.
15055 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15057 * config/arm/arm-mve-builtins-base.cc (vsliq): New.
15058 * config/arm/arm-mve-builtins-base.def (vsliq): New.
15059 * config/arm/arm-mve-builtins-base.h (vsliq): New.
15060 * config/arm/arm-mve-builtins.cc
15061 (function_instance::has_inactive_argument): Handle vsliq.
15062 * config/arm/arm_mve.h (vsliq): Remove.
15064 (vsliq_n_u8): Remove.
15065 (vsliq_n_s8): Remove.
15066 (vsliq_n_u16): Remove.
15067 (vsliq_n_s16): Remove.
15068 (vsliq_n_u32): Remove.
15069 (vsliq_n_s32): Remove.
15070 (vsliq_m_n_s8): Remove.
15071 (vsliq_m_n_s32): Remove.
15072 (vsliq_m_n_s16): Remove.
15073 (vsliq_m_n_u8): Remove.
15074 (vsliq_m_n_u32): Remove.
15075 (vsliq_m_n_u16): Remove.
15076 (__arm_vsliq_n_u8): Remove.
15077 (__arm_vsliq_n_s8): Remove.
15078 (__arm_vsliq_n_u16): Remove.
15079 (__arm_vsliq_n_s16): Remove.
15080 (__arm_vsliq_n_u32): Remove.
15081 (__arm_vsliq_n_s32): Remove.
15082 (__arm_vsliq_m_n_s8): Remove.
15083 (__arm_vsliq_m_n_s32): Remove.
15084 (__arm_vsliq_m_n_s16): Remove.
15085 (__arm_vsliq_m_n_u8): Remove.
15086 (__arm_vsliq_m_n_u32): Remove.
15087 (__arm_vsliq_m_n_u16): Remove.
15088 (__arm_vsliq): Remove.
15089 (__arm_vsliq_m): Remove.
15091 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15093 * config/arm/iterators.md (mve_insn>): Add vsli.
15094 * config/arm/mve.md (mve_vsliq_n_<supf><mode>): Rename into ...
15095 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
15096 (mve_vsliq_m_n_<supf><mode>): Rename into ...
15097 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
15099 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15101 * config/arm/arm-mve-builtins-shapes.cc (ternary_lshift): New.
15102 * config/arm/arm-mve-builtins-shapes.h (ternary_lshift): New.
15104 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15106 * config/arm/arm-mve-builtins-base.cc (vpselq): New.
15107 * config/arm/arm-mve-builtins-base.def (vpselq): New.
15108 * config/arm/arm-mve-builtins-base.h (vpselq): New.
15109 * config/arm/arm_mve.h (vpselq): Remove.
15110 (vpselq_u8): Remove.
15111 (vpselq_s8): Remove.
15112 (vpselq_u16): Remove.
15113 (vpselq_s16): Remove.
15114 (vpselq_u32): Remove.
15115 (vpselq_s32): Remove.
15116 (vpselq_u64): Remove.
15117 (vpselq_s64): Remove.
15118 (vpselq_f16): Remove.
15119 (vpselq_f32): Remove.
15120 (__arm_vpselq_u8): Remove.
15121 (__arm_vpselq_s8): Remove.
15122 (__arm_vpselq_u16): Remove.
15123 (__arm_vpselq_s16): Remove.
15124 (__arm_vpselq_u32): Remove.
15125 (__arm_vpselq_s32): Remove.
15126 (__arm_vpselq_u64): Remove.
15127 (__arm_vpselq_s64): Remove.
15128 (__arm_vpselq_f16): Remove.
15129 (__arm_vpselq_f32): Remove.
15130 (__arm_vpselq): Remove.
15132 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15134 * config/arm/arm-mve-builtins-shapes.cc (vpsel): New.
15135 * config/arm/arm-mve-builtins-shapes.h (vpsel): New.
15137 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15139 * config/arm/arm.cc (arm_expand_vcond): Use gen_mve_q instead of
15141 * config/arm/iterators.md (MVE_VPSELQ_F): New.
15142 (mve_insn): Add vpsel.
15143 * config/arm/mve.md (@mve_vpselq_<supf><mode>): Rename into ...
15144 (@mve_<mve_insn>q_<supf><mode>): ... this.
15145 (@mve_vpselq_f<mode>): Rename into ...
15146 (@mve_<mve_insn>q_f<mode>): ... this.
15148 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15150 * config/arm/arm-mve-builtins-base.cc (vfmaq, vfmasq, vfmsq): New.
15151 * config/arm/arm-mve-builtins-base.def (vfmaq, vfmasq, vfmsq): New.
15152 * config/arm/arm-mve-builtins-base.h (vfmaq, vfmasq, vfmsq): New.
15153 * config/arm/arm-mve-builtins.cc
15154 (function_instance::has_inactive_argument): Handle vfmaq, vfmasq,
15156 * config/arm/arm_mve.h (vfmaq): Remove.
15160 (vfmasq_m): Remove.
15162 (vfmaq_f16): Remove.
15163 (vfmaq_n_f16): Remove.
15164 (vfmasq_n_f16): Remove.
15165 (vfmsq_f16): Remove.
15166 (vfmaq_f32): Remove.
15167 (vfmaq_n_f32): Remove.
15168 (vfmasq_n_f32): Remove.
15169 (vfmsq_f32): Remove.
15170 (vfmaq_m_f32): Remove.
15171 (vfmaq_m_f16): Remove.
15172 (vfmaq_m_n_f32): Remove.
15173 (vfmaq_m_n_f16): Remove.
15174 (vfmasq_m_n_f32): Remove.
15175 (vfmasq_m_n_f16): Remove.
15176 (vfmsq_m_f32): Remove.
15177 (vfmsq_m_f16): Remove.
15178 (__arm_vfmaq_f16): Remove.
15179 (__arm_vfmaq_n_f16): Remove.
15180 (__arm_vfmasq_n_f16): Remove.
15181 (__arm_vfmsq_f16): Remove.
15182 (__arm_vfmaq_f32): Remove.
15183 (__arm_vfmaq_n_f32): Remove.
15184 (__arm_vfmasq_n_f32): Remove.
15185 (__arm_vfmsq_f32): Remove.
15186 (__arm_vfmaq_m_f32): Remove.
15187 (__arm_vfmaq_m_f16): Remove.
15188 (__arm_vfmaq_m_n_f32): Remove.
15189 (__arm_vfmaq_m_n_f16): Remove.
15190 (__arm_vfmasq_m_n_f32): Remove.
15191 (__arm_vfmasq_m_n_f16): Remove.
15192 (__arm_vfmsq_m_f32): Remove.
15193 (__arm_vfmsq_m_f16): Remove.
15194 (__arm_vfmaq): Remove.
15195 (__arm_vfmasq): Remove.
15196 (__arm_vfmsq): Remove.
15197 (__arm_vfmaq_m): Remove.
15198 (__arm_vfmasq_m): Remove.
15199 (__arm_vfmsq_m): Remove.
15201 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15203 * config/arm/iterators.md (MVE_FP_M_BINARY): Add VFMAQ_M_F,
15205 (MVE_FP_M_N_BINARY): Add VFMAQ_M_N_F, VFMASQ_M_N_F.
15206 (MVE_VFMxQ_F, MVE_VFMAxQ_N_F): New.
15207 (mve_insn): Add vfma, vfmas, vfms.
15208 * config/arm/mve.md (mve_vfmaq_f<mode>, mve_vfmsq_f<mode>): Merge
15210 (@mve_<mve_insn>q_f<mode>): ... this.
15211 (mve_vfmaq_n_f<mode>, mve_vfmasq_n_f<mode>): Merge into ...
15212 (@mve_<mve_insn>q_n_f<mode>): ... this.
15213 (mve_vfmaq_m_f<mode>, mve_vfmsq_m_f<mode>): Merge into
15214 @mve_<mve_insn>q_m_f<mode>.
15215 (mve_vfmaq_m_n_f<mode>, mve_vfmasq_m_n_f<mode>): Merge into
15216 @mve_<mve_insn>q_m_n_f<mode>.
15218 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15220 * config/arm/arm-mve-builtins-shapes.cc (ternary_opt_n): New.
15221 * config/arm/arm-mve-builtins-shapes.h (ternary_opt_n): New.
15223 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15225 * config/arm/arm-mve-builtins-base.cc
15226 (FUNCTION_WITH_RTX_M_N_NO_F): New.
15228 * config/arm/arm-mve-builtins-base.def (vmvnq): New.
15229 * config/arm/arm-mve-builtins-base.h (vmvnq): New.
15230 * config/arm/arm_mve.h (vmvnq): Remove.
15233 (vmvnq_s8): Remove.
15234 (vmvnq_s16): Remove.
15235 (vmvnq_s32): Remove.
15236 (vmvnq_n_s16): Remove.
15237 (vmvnq_n_s32): Remove.
15238 (vmvnq_u8): Remove.
15239 (vmvnq_u16): Remove.
15240 (vmvnq_u32): Remove.
15241 (vmvnq_n_u16): Remove.
15242 (vmvnq_n_u32): Remove.
15243 (vmvnq_m_u8): Remove.
15244 (vmvnq_m_s8): Remove.
15245 (vmvnq_m_u16): Remove.
15246 (vmvnq_m_s16): Remove.
15247 (vmvnq_m_u32): Remove.
15248 (vmvnq_m_s32): Remove.
15249 (vmvnq_m_n_s16): Remove.
15250 (vmvnq_m_n_u16): Remove.
15251 (vmvnq_m_n_s32): Remove.
15252 (vmvnq_m_n_u32): Remove.
15253 (vmvnq_x_s8): Remove.
15254 (vmvnq_x_s16): Remove.
15255 (vmvnq_x_s32): Remove.
15256 (vmvnq_x_u8): Remove.
15257 (vmvnq_x_u16): Remove.
15258 (vmvnq_x_u32): Remove.
15259 (vmvnq_x_n_s16): Remove.
15260 (vmvnq_x_n_s32): Remove.
15261 (vmvnq_x_n_u16): Remove.
15262 (vmvnq_x_n_u32): Remove.
15263 (__arm_vmvnq_s8): Remove.
15264 (__arm_vmvnq_s16): Remove.
15265 (__arm_vmvnq_s32): Remove.
15266 (__arm_vmvnq_n_s16): Remove.
15267 (__arm_vmvnq_n_s32): Remove.
15268 (__arm_vmvnq_u8): Remove.
15269 (__arm_vmvnq_u16): Remove.
15270 (__arm_vmvnq_u32): Remove.
15271 (__arm_vmvnq_n_u16): Remove.
15272 (__arm_vmvnq_n_u32): Remove.
15273 (__arm_vmvnq_m_u8): Remove.
15274 (__arm_vmvnq_m_s8): Remove.
15275 (__arm_vmvnq_m_u16): Remove.
15276 (__arm_vmvnq_m_s16): Remove.
15277 (__arm_vmvnq_m_u32): Remove.
15278 (__arm_vmvnq_m_s32): Remove.
15279 (__arm_vmvnq_m_n_s16): Remove.
15280 (__arm_vmvnq_m_n_u16): Remove.
15281 (__arm_vmvnq_m_n_s32): Remove.
15282 (__arm_vmvnq_m_n_u32): Remove.
15283 (__arm_vmvnq_x_s8): Remove.
15284 (__arm_vmvnq_x_s16): Remove.
15285 (__arm_vmvnq_x_s32): Remove.
15286 (__arm_vmvnq_x_u8): Remove.
15287 (__arm_vmvnq_x_u16): Remove.
15288 (__arm_vmvnq_x_u32): Remove.
15289 (__arm_vmvnq_x_n_s16): Remove.
15290 (__arm_vmvnq_x_n_s32): Remove.
15291 (__arm_vmvnq_x_n_u16): Remove.
15292 (__arm_vmvnq_x_n_u32): Remove.
15293 (__arm_vmvnq): Remove.
15294 (__arm_vmvnq_m): Remove.
15295 (__arm_vmvnq_x): Remove.
15297 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15299 * config/arm/iterators.md (mve_insn): Add vmvn.
15300 * config/arm/mve.md (mve_vmvnq_n_<supf><mode>): Rename into ...
15301 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
15302 (mve_vmvnq_m_<supf><mode>): Rename into ...
15303 (@mve_<mve_insn>q_m_<supf><mode>): ... this.
15304 (mve_vmvnq_m_n_<supf><mode>): Rename into ...
15305 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
15307 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15309 * config/arm/arm-mve-builtins-shapes.cc (mvn): New.
15310 * config/arm/arm-mve-builtins-shapes.h (mvn): New.
15312 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15314 * config/arm/arm-mve-builtins-base.cc (vbrsrq): New.
15315 * config/arm/arm-mve-builtins-base.def (vbrsrq): New.
15316 * config/arm/arm-mve-builtins-base.h (vbrsrq): New.
15317 * config/arm/arm_mve.h (vbrsrq): Remove.
15318 (vbrsrq_m): Remove.
15319 (vbrsrq_x): Remove.
15320 (vbrsrq_n_f16): Remove.
15321 (vbrsrq_n_f32): Remove.
15322 (vbrsrq_n_u8): Remove.
15323 (vbrsrq_n_s8): Remove.
15324 (vbrsrq_n_u16): Remove.
15325 (vbrsrq_n_s16): Remove.
15326 (vbrsrq_n_u32): Remove.
15327 (vbrsrq_n_s32): Remove.
15328 (vbrsrq_m_n_s8): Remove.
15329 (vbrsrq_m_n_s32): Remove.
15330 (vbrsrq_m_n_s16): Remove.
15331 (vbrsrq_m_n_u8): Remove.
15332 (vbrsrq_m_n_u32): Remove.
15333 (vbrsrq_m_n_u16): Remove.
15334 (vbrsrq_m_n_f32): Remove.
15335 (vbrsrq_m_n_f16): Remove.
15336 (vbrsrq_x_n_s8): Remove.
15337 (vbrsrq_x_n_s16): Remove.
15338 (vbrsrq_x_n_s32): Remove.
15339 (vbrsrq_x_n_u8): Remove.
15340 (vbrsrq_x_n_u16): Remove.
15341 (vbrsrq_x_n_u32): Remove.
15342 (vbrsrq_x_n_f16): Remove.
15343 (vbrsrq_x_n_f32): Remove.
15344 (__arm_vbrsrq_n_u8): Remove.
15345 (__arm_vbrsrq_n_s8): Remove.
15346 (__arm_vbrsrq_n_u16): Remove.
15347 (__arm_vbrsrq_n_s16): Remove.
15348 (__arm_vbrsrq_n_u32): Remove.
15349 (__arm_vbrsrq_n_s32): Remove.
15350 (__arm_vbrsrq_m_n_s8): Remove.
15351 (__arm_vbrsrq_m_n_s32): Remove.
15352 (__arm_vbrsrq_m_n_s16): Remove.
15353 (__arm_vbrsrq_m_n_u8): Remove.
15354 (__arm_vbrsrq_m_n_u32): Remove.
15355 (__arm_vbrsrq_m_n_u16): Remove.
15356 (__arm_vbrsrq_x_n_s8): Remove.
15357 (__arm_vbrsrq_x_n_s16): Remove.
15358 (__arm_vbrsrq_x_n_s32): Remove.
15359 (__arm_vbrsrq_x_n_u8): Remove.
15360 (__arm_vbrsrq_x_n_u16): Remove.
15361 (__arm_vbrsrq_x_n_u32): Remove.
15362 (__arm_vbrsrq_n_f16): Remove.
15363 (__arm_vbrsrq_n_f32): Remove.
15364 (__arm_vbrsrq_m_n_f32): Remove.
15365 (__arm_vbrsrq_m_n_f16): Remove.
15366 (__arm_vbrsrq_x_n_f16): Remove.
15367 (__arm_vbrsrq_x_n_f32): Remove.
15368 (__arm_vbrsrq): Remove.
15369 (__arm_vbrsrq_m): Remove.
15370 (__arm_vbrsrq_x): Remove.
15372 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15374 * config/arm/iterators.md (MVE_VBRSR_M_N_FP, MVE_VBRSR_N_FP): New.
15375 (mve_insn): Add vbrsr.
15376 * config/arm/mve.md (mve_vbrsrq_n_f<mode>): Rename into ...
15377 (@mve_<mve_insn>q_n_f<mode>): ... this.
15378 (mve_vbrsrq_n_<supf><mode>): Rename into ...
15379 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
15380 (mve_vbrsrq_m_n_<supf><mode>): Rename into ...
15381 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
15382 (mve_vbrsrq_m_n_f<mode>): Rename into ...
15383 (@mve_<mve_insn>q_m_n_f<mode>): ... this.
15385 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15387 * config/arm/arm-mve-builtins-shapes.cc (binary_imm32): New.
15388 * config/arm/arm-mve-builtins-shapes.h (binary_imm32): New.
15390 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15392 * config/arm/arm-mve-builtins-base.cc (vqshluq): New.
15393 * config/arm/arm-mve-builtins-base.def (vqshluq): New.
15394 * config/arm/arm-mve-builtins-base.h (vqshluq): New.
15395 * config/arm/arm_mve.h (vqshluq): Remove.
15396 (vqshluq_m): Remove.
15397 (vqshluq_n_s8): Remove.
15398 (vqshluq_n_s16): Remove.
15399 (vqshluq_n_s32): Remove.
15400 (vqshluq_m_n_s8): Remove.
15401 (vqshluq_m_n_s16): Remove.
15402 (vqshluq_m_n_s32): Remove.
15403 (__arm_vqshluq_n_s8): Remove.
15404 (__arm_vqshluq_n_s16): Remove.
15405 (__arm_vqshluq_n_s32): Remove.
15406 (__arm_vqshluq_m_n_s8): Remove.
15407 (__arm_vqshluq_m_n_s16): Remove.
15408 (__arm_vqshluq_m_n_s32): Remove.
15409 (__arm_vqshluq): Remove.
15410 (__arm_vqshluq_m): Remove.
15412 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15414 * config/arm/iterators.md (mve_insn): Add vqshlu.
15415 (supf): Add VQSHLUQ_M_N_S, VQSHLUQ_N_S.
15416 (VQSHLUQ_M_N, VQSHLUQ_N): New.
15417 * config/arm/mve.md (mve_vqshluq_n_s<mode>): Change name into ...
15418 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
15419 (mve_vqshluq_m_n_s<mode>): Change name into ...
15420 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
15422 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15424 * config/arm/arm-mve-builtins-shapes.cc
15425 (binary_lshift_unsigned): New.
15426 * config/arm/arm-mve-builtins-shapes.h
15427 (binary_lshift_unsigned): New.
15429 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15431 * config/arm/arm-mve-builtins-base.cc (vrmlaldavhaq)
15432 (vrmlaldavhaxq, vrmlsldavhaq, vrmlsldavhaxq): New.
15433 * config/arm/arm-mve-builtins-base.def (vrmlaldavhaq)
15434 (vrmlaldavhaxq, vrmlsldavhaq, vrmlsldavhaxq): New.
15435 * config/arm/arm-mve-builtins-base.h (vrmlaldavhaq)
15436 (vrmlaldavhaxq, vrmlsldavhaq, vrmlsldavhaxq): New.
15437 * config/arm/arm-mve-builtins-functions.h: Handle vrmlaldavhaq,
15438 vrmlaldavhaxq, vrmlsldavhaq, vrmlsldavhaxq.
15439 * config/arm/arm_mve.h (vrmlaldavhaq): Remove.
15440 (vrmlaldavhaxq): Remove.
15441 (vrmlsldavhaq): Remove.
15442 (vrmlsldavhaxq): Remove.
15443 (vrmlaldavhaq_p): Remove.
15444 (vrmlaldavhaxq_p): Remove.
15445 (vrmlsldavhaq_p): Remove.
15446 (vrmlsldavhaxq_p): Remove.
15447 (vrmlaldavhaq_s32): Remove.
15448 (vrmlaldavhaq_u32): Remove.
15449 (vrmlaldavhaxq_s32): Remove.
15450 (vrmlsldavhaq_s32): Remove.
15451 (vrmlsldavhaxq_s32): Remove.
15452 (vrmlaldavhaq_p_s32): Remove.
15453 (vrmlaldavhaq_p_u32): Remove.
15454 (vrmlaldavhaxq_p_s32): Remove.
15455 (vrmlsldavhaq_p_s32): Remove.
15456 (vrmlsldavhaxq_p_s32): Remove.
15457 (__arm_vrmlaldavhaq_s32): Remove.
15458 (__arm_vrmlaldavhaq_u32): Remove.
15459 (__arm_vrmlaldavhaxq_s32): Remove.
15460 (__arm_vrmlsldavhaq_s32): Remove.
15461 (__arm_vrmlsldavhaxq_s32): Remove.
15462 (__arm_vrmlaldavhaq_p_s32): Remove.
15463 (__arm_vrmlaldavhaq_p_u32): Remove.
15464 (__arm_vrmlaldavhaxq_p_s32): Remove.
15465 (__arm_vrmlsldavhaq_p_s32): Remove.
15466 (__arm_vrmlsldavhaxq_p_s32): Remove.
15467 (__arm_vrmlaldavhaq): Remove.
15468 (__arm_vrmlaldavhaxq): Remove.
15469 (__arm_vrmlsldavhaq): Remove.
15470 (__arm_vrmlsldavhaxq): Remove.
15471 (__arm_vrmlaldavhaq_p): Remove.
15472 (__arm_vrmlaldavhaxq_p): Remove.
15473 (__arm_vrmlsldavhaq_p): Remove.
15474 (__arm_vrmlsldavhaxq_p): Remove.
15476 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15478 * config/arm/iterators.md (MVE_VRMLxLDAVHAxQ)
15479 (MVE_VRMLxLDAVHAxQ_P): New.
15480 (mve_insn): Add vrmlaldavha, vrmlaldavhax, vrmlsldavha,
15482 (supf): Add VRMLALDAVHAXQ_P_S, VRMLALDAVHAXQ_S, VRMLSLDAVHAQ_P_S,
15483 VRMLSLDAVHAQ_S, VRMLSLDAVHAXQ_P_S, VRMLSLDAVHAXQ_S,
15485 * config/arm/mve.md (mve_vrmlaldavhaq_<supf>v4si)
15486 (mve_vrmlaldavhaxq_sv4si, mve_vrmlsldavhaxq_sv4si)
15487 (mve_vrmlsldavhaq_sv4si): Merge into ...
15488 (@mve_<mve_insn>q_<supf>v4si): ... this.
15489 (mve_vrmlaldavhaq_p_sv4si, mve_vrmlaldavhaq_p_uv4si)
15490 (mve_vrmlaldavhaxq_p_sv4si, mve_vrmlsldavhaq_p_sv4si)
15491 (mve_vrmlsldavhaxq_p_sv4si): Merge into ...
15492 (@mve_<mve_insn>q_p_<supf>v4si): ... this.
15494 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15496 * config/arm/arm-mve-builtins-base.cc (vqdmullbq, vqdmulltq): New.
15497 * config/arm/arm-mve-builtins-base.def (vqdmullbq, vqdmulltq):
15499 * config/arm/arm-mve-builtins-base.h (vqdmullbq, vqdmulltq): New.
15500 * config/arm/arm_mve.h (vqdmulltq): Remove.
15501 (vqdmullbq): Remove.
15502 (vqdmullbq_m): Remove.
15503 (vqdmulltq_m): Remove.
15504 (vqdmulltq_s16): Remove.
15505 (vqdmulltq_n_s16): Remove.
15506 (vqdmullbq_s16): Remove.
15507 (vqdmullbq_n_s16): Remove.
15508 (vqdmulltq_s32): Remove.
15509 (vqdmulltq_n_s32): Remove.
15510 (vqdmullbq_s32): Remove.
15511 (vqdmullbq_n_s32): Remove.
15512 (vqdmullbq_m_n_s32): Remove.
15513 (vqdmullbq_m_n_s16): Remove.
15514 (vqdmullbq_m_s32): Remove.
15515 (vqdmullbq_m_s16): Remove.
15516 (vqdmulltq_m_n_s32): Remove.
15517 (vqdmulltq_m_n_s16): Remove.
15518 (vqdmulltq_m_s32): Remove.
15519 (vqdmulltq_m_s16): Remove.
15520 (__arm_vqdmulltq_s16): Remove.
15521 (__arm_vqdmulltq_n_s16): Remove.
15522 (__arm_vqdmullbq_s16): Remove.
15523 (__arm_vqdmullbq_n_s16): Remove.
15524 (__arm_vqdmulltq_s32): Remove.
15525 (__arm_vqdmulltq_n_s32): Remove.
15526 (__arm_vqdmullbq_s32): Remove.
15527 (__arm_vqdmullbq_n_s32): Remove.
15528 (__arm_vqdmullbq_m_n_s32): Remove.
15529 (__arm_vqdmullbq_m_n_s16): Remove.
15530 (__arm_vqdmullbq_m_s32): Remove.
15531 (__arm_vqdmullbq_m_s16): Remove.
15532 (__arm_vqdmulltq_m_n_s32): Remove.
15533 (__arm_vqdmulltq_m_n_s16): Remove.
15534 (__arm_vqdmulltq_m_s32): Remove.
15535 (__arm_vqdmulltq_m_s16): Remove.
15536 (__arm_vqdmulltq): Remove.
15537 (__arm_vqdmullbq): Remove.
15538 (__arm_vqdmullbq_m): Remove.
15539 (__arm_vqdmulltq_m): Remove.
15541 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15543 * config/arm/iterators.md (MVE_VQDMULLxQ, MVE_VQDMULLxQ_M)
15544 (MVE_VQDMULLxQ_M_N, MVE_VQDMULLxQ_N): New.
15545 (mve_insn): Add vqdmullb, vqdmullt.
15546 (supf): Add VQDMULLBQ_S, VQDMULLBQ_M_S, VQDMULLBQ_M_N_S,
15547 VQDMULLBQ_N_S, VQDMULLTQ_S, VQDMULLTQ_M_S, VQDMULLTQ_M_N_S,
15549 * config/arm/mve.md (mve_vqdmullbq_n_s<mode>)
15550 (mve_vqdmulltq_n_s<mode>): Merge into ...
15551 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
15552 (mve_vqdmullbq_s<mode>, mve_vqdmulltq_s<mode>): Merge into ...
15553 (@mve_<mve_insn>q_<supf><mode>): ... this.
15554 (mve_vqdmullbq_m_n_s<mode>, mve_vqdmulltq_m_n_s<mode>): Merge into
15556 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
15557 (mve_vqdmullbq_m_s<mode>, mve_vqdmulltq_m_s<mode>): Merge into ...
15558 (@mve_<mve_insn>q_m_<supf><mode>): ... this.
15560 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
15562 * config/arm/arm-mve-builtins-shapes.cc (binary_widen_opt_n): New.
15563 * config/arm/arm-mve-builtins-shapes.h (binary_widen_opt_n): New.
15565 2023-05-12 Kito Cheng <kito.cheng@sifive.com>
15567 * common/config/riscv/riscv-common.cc (riscv_select_multilib_by_abi):
15568 Drop unused parameter.
15569 (riscv_select_multilib): Ditto.
15570 (riscv_compute_multilib): Update call site of
15571 riscv_select_multilib_by_abi and riscv_select_multilib_by_abi.
15573 2023-05-12 Juzhe Zhong <juzhe.zhong@rivai.ai>
15575 * config/riscv/autovec.md (vec_init<mode><vel>): New pattern.
15576 * config/riscv/riscv-protos.h (expand_vec_init): New function.
15577 * config/riscv/riscv-v.cc (class rvv_builder): New class.
15578 (rvv_builder::can_duplicate_repeating_sequence_p): New function.
15579 (rvv_builder::get_merged_repeating_sequence): Ditto.
15580 (expand_vector_init_insert_elems): Ditto.
15581 (expand_vec_init): Ditto.
15582 * config/riscv/vector-iterators.md: New attribute.
15584 2023-05-12 Haochen Gui <guihaoc@gcc.gnu.org>
15586 * config/rs6000/rs6000-builtins.def
15587 (__builtin_vsx_scalar_insert_exp): Replace bif-pattern from xsiexpdp
15589 (__builtin_vsx_scalar_insert_exp_dp): Replace bif-pattern from
15590 xsiexpdpf to xsiexpdpf_di.
15591 * config/rs6000/vsx.md (xsiexpdp): Rename to...
15592 (xsiexpdp_<mode>): ..., set the mode of second operand to GPR and
15593 replace TARGET_64BIT with TARGET_POWERPC64.
15594 (xsiexpdpf): Rename to...
15595 (xsiexpdpf_<mode>): ..., set the mode of second operand to GPR and
15596 replace TARGET_64BIT with TARGET_POWERPC64.
15598 2023-05-12 Haochen Gui <guihaoc@gcc.gnu.org>
15600 * config/rs6000/rs6000-builtins.def
15601 (__builtin_vsx_scalar_extract_sig): Set return type to const signed
15603 * config/rs6000/vsx.md (xsxsigdp): Replace TARGET_64BIT with
15606 2023-05-12 Haochen Gui <guihaoc@gcc.gnu.org>
15608 * config/rs6000/rs6000-builtins.def
15609 (__builtin_vsx_scalar_extract_exp): Set return type to const signed
15610 int and set its bif-pattern to xsxexpdp_si, move it from power9-64
15612 * config/rs6000/vsx.md (xsxexpdp): Rename to ...
15613 (xsxexpdp_<mode>): ..., set mode of operand 0 to GPR and remove
15614 TARGET_64BIT check.
15615 * doc/extend.texi (scalar_extract_exp): Remove 64-bit environment
15616 requirement when it has a 64-bit argument.
15618 2023-05-12 Pan Li <pan2.li@intel.com>
15619 Richard Sandiford <richard.sandiford@arm.com>
15620 Richard Biener <rguenther@suse.de>
15621 Jakub Jelinek <jakub@redhat.com>
15623 * mux-utils.h: Add overload operator == and != for pointer_mux.
15624 * var-tracking.cc: Included mux-utils.h for pointer_tmux.
15625 (decl_or_value): Changed from void * to pointer_mux<tree_node, rtx_def>.
15626 (dv_is_decl_p): Reconciled to the new type, aka pointer_mux.
15627 (dv_as_decl): Ditto.
15628 (dv_as_opaque): Removed due to unnecessary.
15629 (struct variable_hasher): Take decl_or_value as compare_type.
15630 (variable_hasher::equal): Diito.
15631 (dv_from_decl): Reconciled to the new type, aka pointer_mux.
15632 (dv_from_value): Ditto.
15633 (attrs_list_member): Ditto.
15634 (vars_copy): Ditto.
15635 (var_reg_decl_set): Ditto.
15636 (var_reg_delete_and_set): Ditto.
15637 (find_loc_in_1pdv): Ditto.
15638 (canonicalize_values_star): Ditto.
15639 (variable_post_merge_new_vals): Ditto.
15640 (dump_onepart_variable_differences): Ditto.
15641 (variable_different_p): Ditto.
15642 (set_slot_part): Ditto.
15643 (clobber_slot_part): Ditto.
15644 (clobber_variable_part): Ditto.
15646 2023-05-11 mtsamis <manolis.tsamis@vrull.eu>
15648 * match.pd: simplify vector shift + bit_and + multiply.
15650 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15652 * config/arm/arm-mve-builtins-base.cc (vmlaq, vmlasq, vqdmlahq)
15653 (vqdmlashq, vqrdmlahq, vqrdmlashq): New.
15654 * config/arm/arm-mve-builtins-base.def (vmlaq, vmlasq, vqdmlahq)
15655 (vqdmlashq, vqrdmlahq, vqrdmlashq): New.
15656 * config/arm/arm-mve-builtins-base.h (vmlaq, vmlasq, vqdmlahq)
15657 (vqdmlashq, vqrdmlahq, vqrdmlashq): New.
15658 * config/arm/arm-mve-builtins.cc
15659 (function_instance::has_inactive_argument): Handle vmlaq, vmlasq,
15660 vqdmlahq, vqdmlashq, vqrdmlahq, vqrdmlashq.
15661 * config/arm/arm_mve.h (vqrdmlashq): Remove.
15662 (vqrdmlahq): Remove.
15663 (vqdmlashq): Remove.
15664 (vqdmlahq): Remove.
15668 (vmlasq_m): Remove.
15669 (vqdmlashq_m): Remove.
15670 (vqdmlahq_m): Remove.
15671 (vqrdmlahq_m): Remove.
15672 (vqrdmlashq_m): Remove.
15673 (vmlasq_n_u8): Remove.
15674 (vmlaq_n_u8): Remove.
15675 (vqrdmlashq_n_s8): Remove.
15676 (vqrdmlahq_n_s8): Remove.
15677 (vqdmlahq_n_s8): Remove.
15678 (vqdmlashq_n_s8): Remove.
15679 (vmlasq_n_s8): Remove.
15680 (vmlaq_n_s8): Remove.
15681 (vmlasq_n_u16): Remove.
15682 (vmlaq_n_u16): Remove.
15683 (vqrdmlashq_n_s16): Remove.
15684 (vqrdmlahq_n_s16): Remove.
15685 (vqdmlashq_n_s16): Remove.
15686 (vqdmlahq_n_s16): Remove.
15687 (vmlasq_n_s16): Remove.
15688 (vmlaq_n_s16): Remove.
15689 (vmlasq_n_u32): Remove.
15690 (vmlaq_n_u32): Remove.
15691 (vqrdmlashq_n_s32): Remove.
15692 (vqrdmlahq_n_s32): Remove.
15693 (vqdmlashq_n_s32): Remove.
15694 (vqdmlahq_n_s32): Remove.
15695 (vmlasq_n_s32): Remove.
15696 (vmlaq_n_s32): Remove.
15697 (vmlaq_m_n_s8): Remove.
15698 (vmlaq_m_n_s32): Remove.
15699 (vmlaq_m_n_s16): Remove.
15700 (vmlaq_m_n_u8): Remove.
15701 (vmlaq_m_n_u32): Remove.
15702 (vmlaq_m_n_u16): Remove.
15703 (vmlasq_m_n_s8): Remove.
15704 (vmlasq_m_n_s32): Remove.
15705 (vmlasq_m_n_s16): Remove.
15706 (vmlasq_m_n_u8): Remove.
15707 (vmlasq_m_n_u32): Remove.
15708 (vmlasq_m_n_u16): Remove.
15709 (vqdmlashq_m_n_s8): Remove.
15710 (vqdmlashq_m_n_s32): Remove.
15711 (vqdmlashq_m_n_s16): Remove.
15712 (vqdmlahq_m_n_s8): Remove.
15713 (vqdmlahq_m_n_s32): Remove.
15714 (vqdmlahq_m_n_s16): Remove.
15715 (vqrdmlahq_m_n_s8): Remove.
15716 (vqrdmlahq_m_n_s32): Remove.
15717 (vqrdmlahq_m_n_s16): Remove.
15718 (vqrdmlashq_m_n_s8): Remove.
15719 (vqrdmlashq_m_n_s32): Remove.
15720 (vqrdmlashq_m_n_s16): Remove.
15721 (__arm_vmlasq_n_u8): Remove.
15722 (__arm_vmlaq_n_u8): Remove.
15723 (__arm_vqrdmlashq_n_s8): Remove.
15724 (__arm_vqdmlashq_n_s8): Remove.
15725 (__arm_vqrdmlahq_n_s8): Remove.
15726 (__arm_vqdmlahq_n_s8): Remove.
15727 (__arm_vmlasq_n_s8): Remove.
15728 (__arm_vmlaq_n_s8): Remove.
15729 (__arm_vmlasq_n_u16): Remove.
15730 (__arm_vmlaq_n_u16): Remove.
15731 (__arm_vqrdmlashq_n_s16): Remove.
15732 (__arm_vqdmlashq_n_s16): Remove.
15733 (__arm_vqrdmlahq_n_s16): Remove.
15734 (__arm_vqdmlahq_n_s16): Remove.
15735 (__arm_vmlasq_n_s16): Remove.
15736 (__arm_vmlaq_n_s16): Remove.
15737 (__arm_vmlasq_n_u32): Remove.
15738 (__arm_vmlaq_n_u32): Remove.
15739 (__arm_vqrdmlashq_n_s32): Remove.
15740 (__arm_vqdmlashq_n_s32): Remove.
15741 (__arm_vqrdmlahq_n_s32): Remove.
15742 (__arm_vqdmlahq_n_s32): Remove.
15743 (__arm_vmlasq_n_s32): Remove.
15744 (__arm_vmlaq_n_s32): Remove.
15745 (__arm_vmlaq_m_n_s8): Remove.
15746 (__arm_vmlaq_m_n_s32): Remove.
15747 (__arm_vmlaq_m_n_s16): Remove.
15748 (__arm_vmlaq_m_n_u8): Remove.
15749 (__arm_vmlaq_m_n_u32): Remove.
15750 (__arm_vmlaq_m_n_u16): Remove.
15751 (__arm_vmlasq_m_n_s8): Remove.
15752 (__arm_vmlasq_m_n_s32): Remove.
15753 (__arm_vmlasq_m_n_s16): Remove.
15754 (__arm_vmlasq_m_n_u8): Remove.
15755 (__arm_vmlasq_m_n_u32): Remove.
15756 (__arm_vmlasq_m_n_u16): Remove.
15757 (__arm_vqdmlahq_m_n_s8): Remove.
15758 (__arm_vqdmlahq_m_n_s32): Remove.
15759 (__arm_vqdmlahq_m_n_s16): Remove.
15760 (__arm_vqrdmlahq_m_n_s8): Remove.
15761 (__arm_vqrdmlahq_m_n_s32): Remove.
15762 (__arm_vqrdmlahq_m_n_s16): Remove.
15763 (__arm_vqrdmlashq_m_n_s8): Remove.
15764 (__arm_vqrdmlashq_m_n_s32): Remove.
15765 (__arm_vqrdmlashq_m_n_s16): Remove.
15766 (__arm_vqdmlashq_m_n_s8): Remove.
15767 (__arm_vqdmlashq_m_n_s16): Remove.
15768 (__arm_vqdmlashq_m_n_s32): Remove.
15769 (__arm_vmlasq): Remove.
15770 (__arm_vmlaq): Remove.
15771 (__arm_vqrdmlashq): Remove.
15772 (__arm_vqdmlashq): Remove.
15773 (__arm_vqrdmlahq): Remove.
15774 (__arm_vqdmlahq): Remove.
15775 (__arm_vmlaq_m): Remove.
15776 (__arm_vmlasq_m): Remove.
15777 (__arm_vqdmlahq_m): Remove.
15778 (__arm_vqrdmlahq_m): Remove.
15779 (__arm_vqrdmlashq_m): Remove.
15780 (__arm_vqdmlashq_m): Remove.
15782 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15784 * config/arm/iterators.md (MVE_VMLxQ_N): New.
15785 (mve_insn): Add vmla, vmlas, vqdmlah, vqdmlash, vqrdmlah,
15787 (supf): Add VQDMLAHQ_N_S, VQDMLASHQ_N_S, VQRDMLAHQ_N_S,
15789 * config/arm/mve.md (mve_vmlaq_n_<supf><mode>)
15790 (mve_vmlasq_n_<supf><mode>, mve_vqdmlahq_n_<supf><mode>)
15791 (mve_vqdmlashq_n_<supf><mode>, mve_vqrdmlahq_n_<supf><mode>)
15792 (mve_vqrdmlashq_n_<supf><mode>): Merge into ...
15793 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
15795 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15797 * config/arm/arm-mve-builtins-shapes.cc (ternary_n): New.
15798 * config/arm/arm-mve-builtins-shapes.h (ternary_n): New.
15800 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15802 * config/arm/arm-mve-builtins-base.cc (vqdmladhq, vqdmladhxq)
15803 (vqdmlsdhq, vqdmlsdhxq, vqrdmladhq, vqrdmladhxq, vqrdmlsdhq)
15804 (vqrdmlsdhxq): New.
15805 * config/arm/arm-mve-builtins-base.def (vqdmladhq, vqdmladhxq)
15806 (vqdmlsdhq, vqdmlsdhxq, vqrdmladhq, vqrdmladhxq, vqrdmlsdhq)
15807 (vqrdmlsdhxq): New.
15808 * config/arm/arm-mve-builtins-base.h (vqdmladhq, vqdmladhxq)
15809 (vqdmlsdhq, vqdmlsdhxq, vqrdmladhq, vqrdmladhxq, vqrdmlsdhq)
15810 (vqrdmlsdhxq): New.
15811 * config/arm/arm-mve-builtins.cc
15812 (function_instance::has_inactive_argument): Handle vqrdmladhq,
15813 vqrdmladhxq, vqrdmlsdhq, vqrdmlsdhxq vqdmladhq, vqdmladhxq,
15814 vqdmlsdhq, vqdmlsdhxq.
15815 * config/arm/arm_mve.h (vqrdmlsdhxq): Remove.
15816 (vqrdmlsdhq): Remove.
15817 (vqrdmladhxq): Remove.
15818 (vqrdmladhq): Remove.
15819 (vqdmlsdhxq): Remove.
15820 (vqdmlsdhq): Remove.
15821 (vqdmladhxq): Remove.
15822 (vqdmladhq): Remove.
15823 (vqdmladhq_m): Remove.
15824 (vqdmladhxq_m): Remove.
15825 (vqdmlsdhq_m): Remove.
15826 (vqdmlsdhxq_m): Remove.
15827 (vqrdmladhq_m): Remove.
15828 (vqrdmladhxq_m): Remove.
15829 (vqrdmlsdhq_m): Remove.
15830 (vqrdmlsdhxq_m): Remove.
15831 (vqrdmlsdhxq_s8): Remove.
15832 (vqrdmlsdhq_s8): Remove.
15833 (vqrdmladhxq_s8): Remove.
15834 (vqrdmladhq_s8): Remove.
15835 (vqdmlsdhxq_s8): Remove.
15836 (vqdmlsdhq_s8): Remove.
15837 (vqdmladhxq_s8): Remove.
15838 (vqdmladhq_s8): Remove.
15839 (vqrdmlsdhxq_s16): Remove.
15840 (vqrdmlsdhq_s16): Remove.
15841 (vqrdmladhxq_s16): Remove.
15842 (vqrdmladhq_s16): Remove.
15843 (vqdmlsdhxq_s16): Remove.
15844 (vqdmlsdhq_s16): Remove.
15845 (vqdmladhxq_s16): Remove.
15846 (vqdmladhq_s16): Remove.
15847 (vqrdmlsdhxq_s32): Remove.
15848 (vqrdmlsdhq_s32): Remove.
15849 (vqrdmladhxq_s32): Remove.
15850 (vqrdmladhq_s32): Remove.
15851 (vqdmlsdhxq_s32): Remove.
15852 (vqdmlsdhq_s32): Remove.
15853 (vqdmladhxq_s32): Remove.
15854 (vqdmladhq_s32): Remove.
15855 (vqdmladhq_m_s8): Remove.
15856 (vqdmladhq_m_s32): Remove.
15857 (vqdmladhq_m_s16): Remove.
15858 (vqdmladhxq_m_s8): Remove.
15859 (vqdmladhxq_m_s32): Remove.
15860 (vqdmladhxq_m_s16): Remove.
15861 (vqdmlsdhq_m_s8): Remove.
15862 (vqdmlsdhq_m_s32): Remove.
15863 (vqdmlsdhq_m_s16): Remove.
15864 (vqdmlsdhxq_m_s8): Remove.
15865 (vqdmlsdhxq_m_s32): Remove.
15866 (vqdmlsdhxq_m_s16): Remove.
15867 (vqrdmladhq_m_s8): Remove.
15868 (vqrdmladhq_m_s32): Remove.
15869 (vqrdmladhq_m_s16): Remove.
15870 (vqrdmladhxq_m_s8): Remove.
15871 (vqrdmladhxq_m_s32): Remove.
15872 (vqrdmladhxq_m_s16): Remove.
15873 (vqrdmlsdhq_m_s8): Remove.
15874 (vqrdmlsdhq_m_s32): Remove.
15875 (vqrdmlsdhq_m_s16): Remove.
15876 (vqrdmlsdhxq_m_s8): Remove.
15877 (vqrdmlsdhxq_m_s32): Remove.
15878 (vqrdmlsdhxq_m_s16): Remove.
15879 (__arm_vqrdmlsdhxq_s8): Remove.
15880 (__arm_vqrdmlsdhq_s8): Remove.
15881 (__arm_vqrdmladhxq_s8): Remove.
15882 (__arm_vqrdmladhq_s8): Remove.
15883 (__arm_vqdmlsdhxq_s8): Remove.
15884 (__arm_vqdmlsdhq_s8): Remove.
15885 (__arm_vqdmladhxq_s8): Remove.
15886 (__arm_vqdmladhq_s8): Remove.
15887 (__arm_vqrdmlsdhxq_s16): Remove.
15888 (__arm_vqrdmlsdhq_s16): Remove.
15889 (__arm_vqrdmladhxq_s16): Remove.
15890 (__arm_vqrdmladhq_s16): Remove.
15891 (__arm_vqdmlsdhxq_s16): Remove.
15892 (__arm_vqdmlsdhq_s16): Remove.
15893 (__arm_vqdmladhxq_s16): Remove.
15894 (__arm_vqdmladhq_s16): Remove.
15895 (__arm_vqrdmlsdhxq_s32): Remove.
15896 (__arm_vqrdmlsdhq_s32): Remove.
15897 (__arm_vqrdmladhxq_s32): Remove.
15898 (__arm_vqrdmladhq_s32): Remove.
15899 (__arm_vqdmlsdhxq_s32): Remove.
15900 (__arm_vqdmlsdhq_s32): Remove.
15901 (__arm_vqdmladhxq_s32): Remove.
15902 (__arm_vqdmladhq_s32): Remove.
15903 (__arm_vqdmladhq_m_s8): Remove.
15904 (__arm_vqdmladhq_m_s32): Remove.
15905 (__arm_vqdmladhq_m_s16): Remove.
15906 (__arm_vqdmladhxq_m_s8): Remove.
15907 (__arm_vqdmladhxq_m_s32): Remove.
15908 (__arm_vqdmladhxq_m_s16): Remove.
15909 (__arm_vqdmlsdhq_m_s8): Remove.
15910 (__arm_vqdmlsdhq_m_s32): Remove.
15911 (__arm_vqdmlsdhq_m_s16): Remove.
15912 (__arm_vqdmlsdhxq_m_s8): Remove.
15913 (__arm_vqdmlsdhxq_m_s32): Remove.
15914 (__arm_vqdmlsdhxq_m_s16): Remove.
15915 (__arm_vqrdmladhq_m_s8): Remove.
15916 (__arm_vqrdmladhq_m_s32): Remove.
15917 (__arm_vqrdmladhq_m_s16): Remove.
15918 (__arm_vqrdmladhxq_m_s8): Remove.
15919 (__arm_vqrdmladhxq_m_s32): Remove.
15920 (__arm_vqrdmladhxq_m_s16): Remove.
15921 (__arm_vqrdmlsdhq_m_s8): Remove.
15922 (__arm_vqrdmlsdhq_m_s32): Remove.
15923 (__arm_vqrdmlsdhq_m_s16): Remove.
15924 (__arm_vqrdmlsdhxq_m_s8): Remove.
15925 (__arm_vqrdmlsdhxq_m_s32): Remove.
15926 (__arm_vqrdmlsdhxq_m_s16): Remove.
15927 (__arm_vqrdmlsdhxq): Remove.
15928 (__arm_vqrdmlsdhq): Remove.
15929 (__arm_vqrdmladhxq): Remove.
15930 (__arm_vqrdmladhq): Remove.
15931 (__arm_vqdmlsdhxq): Remove.
15932 (__arm_vqdmlsdhq): Remove.
15933 (__arm_vqdmladhxq): Remove.
15934 (__arm_vqdmladhq): Remove.
15935 (__arm_vqdmladhq_m): Remove.
15936 (__arm_vqdmladhxq_m): Remove.
15937 (__arm_vqdmlsdhq_m): Remove.
15938 (__arm_vqdmlsdhxq_m): Remove.
15939 (__arm_vqrdmladhq_m): Remove.
15940 (__arm_vqrdmladhxq_m): Remove.
15941 (__arm_vqrdmlsdhq_m): Remove.
15942 (__arm_vqrdmlsdhxq_m): Remove.
15944 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15946 * config/arm/iterators.md (MVE_VQxDMLxDHxQ_S): New.
15947 (mve_insn): Add vqdmladh, vqdmladhx, vqdmlsdh, vqdmlsdhx,
15948 vqrdmladh, vqrdmladhx, vqrdmlsdh, vqrdmlsdhx.
15949 (supf): Add VQDMLADHQ_S, VQDMLADHXQ_S, VQDMLSDHQ_S, VQDMLSDHXQ_S,
15950 VQRDMLADHQ_S,VQRDMLADHXQ_S, VQRDMLSDHQ_S, VQRDMLSDHXQ_S.
15951 * config/arm/mve.md (mve_vqrdmladhq_s<mode>)
15952 (mve_vqrdmladhxq_s<mode>, mve_vqrdmlsdhq_s<mode>)
15953 (mve_vqrdmlsdhxq_s<mode>, mve_vqdmlsdhxq_s<mode>)
15954 (mve_vqdmlsdhq_s<mode>, mve_vqdmladhxq_s<mode>)
15955 (mve_vqdmladhq_s<mode>): Merge into ...
15956 (@mve_<mve_insn>q_<supf><mode>): ... this.
15958 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15960 * config/arm/arm-mve-builtins-shapes.cc (ternary): New.
15961 * config/arm/arm-mve-builtins-shapes.h (ternary): New.
15963 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15965 * config/arm/arm-mve-builtins-base.cc (vmlaldavaq, vmlaldavaxq)
15966 (vmlsldavaq, vmlsldavaxq): New.
15967 * config/arm/arm-mve-builtins-base.def (vmlaldavaq, vmlaldavaxq)
15968 (vmlsldavaq, vmlsldavaxq): New.
15969 * config/arm/arm-mve-builtins-base.h (vmlaldavaq, vmlaldavaxq)
15970 (vmlsldavaq, vmlsldavaxq): New.
15971 * config/arm/arm_mve.h (vmlaldavaq): Remove.
15972 (vmlaldavaxq): Remove.
15973 (vmlsldavaq): Remove.
15974 (vmlsldavaxq): Remove.
15975 (vmlaldavaq_p): Remove.
15976 (vmlaldavaxq_p): Remove.
15977 (vmlsldavaq_p): Remove.
15978 (vmlsldavaxq_p): Remove.
15979 (vmlaldavaq_s16): Remove.
15980 (vmlaldavaxq_s16): Remove.
15981 (vmlsldavaq_s16): Remove.
15982 (vmlsldavaxq_s16): Remove.
15983 (vmlaldavaq_u16): Remove.
15984 (vmlaldavaq_s32): Remove.
15985 (vmlaldavaxq_s32): Remove.
15986 (vmlsldavaq_s32): Remove.
15987 (vmlsldavaxq_s32): Remove.
15988 (vmlaldavaq_u32): Remove.
15989 (vmlaldavaq_p_s32): Remove.
15990 (vmlaldavaq_p_s16): Remove.
15991 (vmlaldavaq_p_u32): Remove.
15992 (vmlaldavaq_p_u16): Remove.
15993 (vmlaldavaxq_p_s32): Remove.
15994 (vmlaldavaxq_p_s16): Remove.
15995 (vmlsldavaq_p_s32): Remove.
15996 (vmlsldavaq_p_s16): Remove.
15997 (vmlsldavaxq_p_s32): Remove.
15998 (vmlsldavaxq_p_s16): Remove.
15999 (__arm_vmlaldavaq_s16): Remove.
16000 (__arm_vmlaldavaxq_s16): Remove.
16001 (__arm_vmlsldavaq_s16): Remove.
16002 (__arm_vmlsldavaxq_s16): Remove.
16003 (__arm_vmlaldavaq_u16): Remove.
16004 (__arm_vmlaldavaq_s32): Remove.
16005 (__arm_vmlaldavaxq_s32): Remove.
16006 (__arm_vmlsldavaq_s32): Remove.
16007 (__arm_vmlsldavaxq_s32): Remove.
16008 (__arm_vmlaldavaq_u32): Remove.
16009 (__arm_vmlaldavaq_p_s32): Remove.
16010 (__arm_vmlaldavaq_p_s16): Remove.
16011 (__arm_vmlaldavaq_p_u32): Remove.
16012 (__arm_vmlaldavaq_p_u16): Remove.
16013 (__arm_vmlaldavaxq_p_s32): Remove.
16014 (__arm_vmlaldavaxq_p_s16): Remove.
16015 (__arm_vmlsldavaq_p_s32): Remove.
16016 (__arm_vmlsldavaq_p_s16): Remove.
16017 (__arm_vmlsldavaxq_p_s32): Remove.
16018 (__arm_vmlsldavaxq_p_s16): Remove.
16019 (__arm_vmlaldavaq): Remove.
16020 (__arm_vmlaldavaxq): Remove.
16021 (__arm_vmlsldavaq): Remove.
16022 (__arm_vmlsldavaxq): Remove.
16023 (__arm_vmlaldavaq_p): Remove.
16024 (__arm_vmlaldavaxq_p): Remove.
16025 (__arm_vmlsldavaq_p): Remove.
16026 (__arm_vmlsldavaxq_p): Remove.
16028 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16030 * config/arm/iterators.md (MVE_VMLxLDAVAxQ, MVE_VMLxLDAVAxQ_P):
16032 (mve_insn): Add vmlaldava, vmlaldavax, vmlsldava, vmlsldavax.
16033 (supf): Add VMLALDAVAXQ_P_S, VMLALDAVAXQ_S, VMLSLDAVAQ_P_S,
16034 VMLSLDAVAQ_S, VMLSLDAVAXQ_P_S, VMLSLDAVAXQ_S.
16035 * config/arm/mve.md (mve_vmlaldavaq_<supf><mode>)
16036 (mve_vmlsldavaq_s<mode>, mve_vmlsldavaxq_s<mode>)
16037 (mve_vmlaldavaxq_s<mode>): Merge into ...
16038 (@mve_<mve_insn>q_<supf><mode>): ... this.
16039 (mve_vmlaldavaq_p_<supf><mode>, mve_vmlaldavaxq_p_<supf><mode>)
16040 (mve_vmlsldavaq_p_s<mode>, mve_vmlsldavaxq_p_s<mode>): Merge into
16042 (@mve_<mve_insn>q_p_<supf><mode>): ... this.
16044 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16046 * config/arm/arm-mve-builtins-shapes.cc (binary_acca_int64): New.
16047 * config/arm/arm-mve-builtins-shapes.h (binary_acca_int64): New.
16049 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16051 * config/arm/arm-mve-builtins-base.cc (vrmlaldavhq, vrmlaldavhxq)
16052 (vrmlsldavhq, vrmlsldavhxq): New.
16053 * config/arm/arm-mve-builtins-base.def (vrmlaldavhq, vrmlaldavhxq)
16054 (vrmlsldavhq, vrmlsldavhxq): New.
16055 * config/arm/arm-mve-builtins-base.h (vrmlaldavhq, vrmlaldavhxq)
16056 (vrmlsldavhq, vrmlsldavhxq): New.
16057 * config/arm/arm-mve-builtins-functions.h
16058 (unspec_mve_function_exact_insn_pred_p): Handle vrmlaldavhq,
16059 vrmlaldavhxq, vrmlsldavhq, vrmlsldavhxq.
16060 * config/arm/arm_mve.h (vrmlaldavhq): Remove.
16061 (vrmlsldavhxq): Remove.
16062 (vrmlsldavhq): Remove.
16063 (vrmlaldavhxq): Remove.
16064 (vrmlaldavhq_p): Remove.
16065 (vrmlaldavhxq_p): Remove.
16066 (vrmlsldavhq_p): Remove.
16067 (vrmlsldavhxq_p): Remove.
16068 (vrmlaldavhq_u32): Remove.
16069 (vrmlsldavhxq_s32): Remove.
16070 (vrmlsldavhq_s32): Remove.
16071 (vrmlaldavhxq_s32): Remove.
16072 (vrmlaldavhq_s32): Remove.
16073 (vrmlaldavhq_p_s32): Remove.
16074 (vrmlaldavhxq_p_s32): Remove.
16075 (vrmlsldavhq_p_s32): Remove.
16076 (vrmlsldavhxq_p_s32): Remove.
16077 (vrmlaldavhq_p_u32): Remove.
16078 (__arm_vrmlaldavhq_u32): Remove.
16079 (__arm_vrmlsldavhxq_s32): Remove.
16080 (__arm_vrmlsldavhq_s32): Remove.
16081 (__arm_vrmlaldavhxq_s32): Remove.
16082 (__arm_vrmlaldavhq_s32): Remove.
16083 (__arm_vrmlaldavhq_p_s32): Remove.
16084 (__arm_vrmlaldavhxq_p_s32): Remove.
16085 (__arm_vrmlsldavhq_p_s32): Remove.
16086 (__arm_vrmlsldavhxq_p_s32): Remove.
16087 (__arm_vrmlaldavhq_p_u32): Remove.
16088 (__arm_vrmlaldavhq): Remove.
16089 (__arm_vrmlsldavhxq): Remove.
16090 (__arm_vrmlsldavhq): Remove.
16091 (__arm_vrmlaldavhxq): Remove.
16092 (__arm_vrmlaldavhq_p): Remove.
16093 (__arm_vrmlaldavhxq_p): Remove.
16094 (__arm_vrmlsldavhq_p): Remove.
16095 (__arm_vrmlsldavhxq_p): Remove.
16097 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16099 * config/arm/iterators.md (MVE_VRMLxLDAVxQ, MVE_VRMLxLDAVHxQ_P):
16101 (mve_insn): Add vrmlaldavh, vrmlaldavhx, vrmlsldavh, vrmlsldavhx.
16102 (supf): Add VRMLALDAVHXQ_P_S, VRMLALDAVHXQ_S, VRMLSLDAVHQ_P_S,
16103 VRMLSLDAVHQ_S, VRMLSLDAVHXQ_P_S, VRMLSLDAVHXQ_S.
16104 * config/arm/mve.md (mve_vrmlaldavhxq_sv4si)
16105 (mve_vrmlsldavhq_sv4si, mve_vrmlsldavhxq_sv4si)
16106 (mve_vrmlaldavhq_<supf>v4si): Merge into ...
16107 (@mve_<mve_insn>q_<supf>v4si): ... this.
16108 (mve_vrmlaldavhxq_p_sv4si, mve_vrmlsldavhq_p_sv4si)
16109 (mve_vrmlsldavhxq_p_sv4si, mve_vrmlaldavhq_p_<supf>v4si): Merge
16111 (@mve_<mve_insn>q_p_<supf>v4si): ... this.
16113 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16115 * config/arm/arm-mve-builtins-base.cc (vmlaldavq, vmlaldavxq)
16116 (vmlsldavq, vmlsldavxq): New.
16117 * config/arm/arm-mve-builtins-base.def (vmlaldavq, vmlaldavxq)
16118 (vmlsldavq, vmlsldavxq): New.
16119 * config/arm/arm-mve-builtins-base.h (vmlaldavq, vmlaldavxq)
16120 (vmlsldavq, vmlsldavxq): New.
16121 * config/arm/arm_mve.h (vmlaldavq): Remove.
16122 (vmlsldavxq): Remove.
16123 (vmlsldavq): Remove.
16124 (vmlaldavxq): Remove.
16125 (vmlaldavq_p): Remove.
16126 (vmlaldavxq_p): Remove.
16127 (vmlsldavq_p): Remove.
16128 (vmlsldavxq_p): Remove.
16129 (vmlaldavq_u16): Remove.
16130 (vmlsldavxq_s16): Remove.
16131 (vmlsldavq_s16): Remove.
16132 (vmlaldavxq_s16): Remove.
16133 (vmlaldavq_s16): Remove.
16134 (vmlaldavq_u32): Remove.
16135 (vmlsldavxq_s32): Remove.
16136 (vmlsldavq_s32): Remove.
16137 (vmlaldavxq_s32): Remove.
16138 (vmlaldavq_s32): Remove.
16139 (vmlaldavq_p_s16): Remove.
16140 (vmlaldavxq_p_s16): Remove.
16141 (vmlsldavq_p_s16): Remove.
16142 (vmlsldavxq_p_s16): Remove.
16143 (vmlaldavq_p_u16): Remove.
16144 (vmlaldavq_p_s32): Remove.
16145 (vmlaldavxq_p_s32): Remove.
16146 (vmlsldavq_p_s32): Remove.
16147 (vmlsldavxq_p_s32): Remove.
16148 (vmlaldavq_p_u32): Remove.
16149 (__arm_vmlaldavq_u16): Remove.
16150 (__arm_vmlsldavxq_s16): Remove.
16151 (__arm_vmlsldavq_s16): Remove.
16152 (__arm_vmlaldavxq_s16): Remove.
16153 (__arm_vmlaldavq_s16): Remove.
16154 (__arm_vmlaldavq_u32): Remove.
16155 (__arm_vmlsldavxq_s32): Remove.
16156 (__arm_vmlsldavq_s32): Remove.
16157 (__arm_vmlaldavxq_s32): Remove.
16158 (__arm_vmlaldavq_s32): Remove.
16159 (__arm_vmlaldavq_p_s16): Remove.
16160 (__arm_vmlaldavxq_p_s16): Remove.
16161 (__arm_vmlsldavq_p_s16): Remove.
16162 (__arm_vmlsldavxq_p_s16): Remove.
16163 (__arm_vmlaldavq_p_u16): Remove.
16164 (__arm_vmlaldavq_p_s32): Remove.
16165 (__arm_vmlaldavxq_p_s32): Remove.
16166 (__arm_vmlsldavq_p_s32): Remove.
16167 (__arm_vmlsldavxq_p_s32): Remove.
16168 (__arm_vmlaldavq_p_u32): Remove.
16169 (__arm_vmlaldavq): Remove.
16170 (__arm_vmlsldavxq): Remove.
16171 (__arm_vmlsldavq): Remove.
16172 (__arm_vmlaldavxq): Remove.
16173 (__arm_vmlaldavq_p): Remove.
16174 (__arm_vmlaldavxq_p): Remove.
16175 (__arm_vmlsldavq_p): Remove.
16176 (__arm_vmlsldavxq_p): Remove.
16178 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16180 * config/arm/iterators.md (MVE_VMLxLDAVxQ, MVE_VMLxLDAVxQ_P): New.
16181 (mve_insn): Add vmlaldav, vmlaldavx, vmlsldav, vmlsldavx.
16182 (supf): Add VMLALDAVXQ_S, VMLSLDAVQ_S, VMLSLDAVXQ_S,
16183 VMLALDAVXQ_P_S, VMLSLDAVQ_P_S, VMLSLDAVXQ_P_S.
16184 * config/arm/mve.md (mve_vmlaldavq_<supf><mode>)
16185 (mve_vmlaldavxq_s<mode>, mve_vmlsldavq_s<mode>)
16186 (mve_vmlsldavxq_s<mode>): Merge into ...
16187 (@mve_<mve_insn>q_<supf><mode>): ... this.
16188 (mve_vmlaldavq_p_<supf><mode>, mve_vmlaldavxq_p_s<mode>)
16189 (mve_vmlsldavq_p_s<mode>, mve_vmlsldavxq_p_s<mode>): Merge into
16191 (@mve_<mve_insn>q_p_<supf><mode>): ... this.
16193 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16195 * config/arm/arm-mve-builtins-shapes.cc (binary_acc_int64): New.
16196 * config/arm/arm-mve-builtins-shapes.h (binary_acc_int64): New.
16198 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16200 * config/arm/arm-mve-builtins-base.cc (vabavq): New.
16201 * config/arm/arm-mve-builtins-base.def (vabavq): New.
16202 * config/arm/arm-mve-builtins-base.h (vabavq): New.
16203 * config/arm/arm_mve.h (vabavq): Remove.
16204 (vabavq_p): Remove.
16205 (vabavq_s8): Remove.
16206 (vabavq_s16): Remove.
16207 (vabavq_s32): Remove.
16208 (vabavq_u8): Remove.
16209 (vabavq_u16): Remove.
16210 (vabavq_u32): Remove.
16211 (vabavq_p_s8): Remove.
16212 (vabavq_p_u8): Remove.
16213 (vabavq_p_s16): Remove.
16214 (vabavq_p_u16): Remove.
16215 (vabavq_p_s32): Remove.
16216 (vabavq_p_u32): Remove.
16217 (__arm_vabavq_s8): Remove.
16218 (__arm_vabavq_s16): Remove.
16219 (__arm_vabavq_s32): Remove.
16220 (__arm_vabavq_u8): Remove.
16221 (__arm_vabavq_u16): Remove.
16222 (__arm_vabavq_u32): Remove.
16223 (__arm_vabavq_p_s8): Remove.
16224 (__arm_vabavq_p_u8): Remove.
16225 (__arm_vabavq_p_s16): Remove.
16226 (__arm_vabavq_p_u16): Remove.
16227 (__arm_vabavq_p_s32): Remove.
16228 (__arm_vabavq_p_u32): Remove.
16229 (__arm_vabavq): Remove.
16230 (__arm_vabavq_p): Remove.
16232 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16234 * config/arm/iterators.md (mve_insn): Add vabav.
16235 * config/arm/mve.md (mve_vabavq_<supf><mode>): Rename into ...
16236 (@mve_<mve_insn>q_<supf><mode>): ... this,.
16237 (mve_vabavq_p_<supf><mode>): Rename into ...
16238 (@mve_<mve_insn>q_p_<supf><mode>): ... this,.
16240 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16242 * config/arm/arm-mve-builtins-base.cc (vmladavaxq, vmladavaq)
16243 (vmlsdavaq, vmlsdavaxq): New.
16244 * config/arm/arm-mve-builtins-base.def (vmladavaxq, vmladavaq)
16245 (vmlsdavaq, vmlsdavaxq): New.
16246 * config/arm/arm-mve-builtins-base.h (vmladavaxq, vmladavaq)
16247 (vmlsdavaq, vmlsdavaxq): New.
16248 * config/arm/arm_mve.h (vmladavaq): Remove.
16249 (vmlsdavaxq): Remove.
16250 (vmlsdavaq): Remove.
16251 (vmladavaxq): Remove.
16252 (vmladavaq_p): Remove.
16253 (vmladavaxq_p): Remove.
16254 (vmlsdavaq_p): Remove.
16255 (vmlsdavaxq_p): Remove.
16256 (vmladavaq_u8): Remove.
16257 (vmlsdavaxq_s8): Remove.
16258 (vmlsdavaq_s8): Remove.
16259 (vmladavaxq_s8): Remove.
16260 (vmladavaq_s8): Remove.
16261 (vmladavaq_u16): Remove.
16262 (vmlsdavaxq_s16): Remove.
16263 (vmlsdavaq_s16): Remove.
16264 (vmladavaxq_s16): Remove.
16265 (vmladavaq_s16): Remove.
16266 (vmladavaq_u32): Remove.
16267 (vmlsdavaxq_s32): Remove.
16268 (vmlsdavaq_s32): Remove.
16269 (vmladavaxq_s32): Remove.
16270 (vmladavaq_s32): Remove.
16271 (vmladavaq_p_s8): Remove.
16272 (vmladavaq_p_s32): Remove.
16273 (vmladavaq_p_s16): Remove.
16274 (vmladavaq_p_u8): Remove.
16275 (vmladavaq_p_u32): Remove.
16276 (vmladavaq_p_u16): Remove.
16277 (vmladavaxq_p_s8): Remove.
16278 (vmladavaxq_p_s32): Remove.
16279 (vmladavaxq_p_s16): Remove.
16280 (vmlsdavaq_p_s8): Remove.
16281 (vmlsdavaq_p_s32): Remove.
16282 (vmlsdavaq_p_s16): Remove.
16283 (vmlsdavaxq_p_s8): Remove.
16284 (vmlsdavaxq_p_s32): Remove.
16285 (vmlsdavaxq_p_s16): Remove.
16286 (__arm_vmladavaq_u8): Remove.
16287 (__arm_vmlsdavaxq_s8): Remove.
16288 (__arm_vmlsdavaq_s8): Remove.
16289 (__arm_vmladavaxq_s8): Remove.
16290 (__arm_vmladavaq_s8): Remove.
16291 (__arm_vmladavaq_u16): Remove.
16292 (__arm_vmlsdavaxq_s16): Remove.
16293 (__arm_vmlsdavaq_s16): Remove.
16294 (__arm_vmladavaxq_s16): Remove.
16295 (__arm_vmladavaq_s16): Remove.
16296 (__arm_vmladavaq_u32): Remove.
16297 (__arm_vmlsdavaxq_s32): Remove.
16298 (__arm_vmlsdavaq_s32): Remove.
16299 (__arm_vmladavaxq_s32): Remove.
16300 (__arm_vmladavaq_s32): Remove.
16301 (__arm_vmladavaq_p_s8): Remove.
16302 (__arm_vmladavaq_p_s32): Remove.
16303 (__arm_vmladavaq_p_s16): Remove.
16304 (__arm_vmladavaq_p_u8): Remove.
16305 (__arm_vmladavaq_p_u32): Remove.
16306 (__arm_vmladavaq_p_u16): Remove.
16307 (__arm_vmladavaxq_p_s8): Remove.
16308 (__arm_vmladavaxq_p_s32): Remove.
16309 (__arm_vmladavaxq_p_s16): Remove.
16310 (__arm_vmlsdavaq_p_s8): Remove.
16311 (__arm_vmlsdavaq_p_s32): Remove.
16312 (__arm_vmlsdavaq_p_s16): Remove.
16313 (__arm_vmlsdavaxq_p_s8): Remove.
16314 (__arm_vmlsdavaxq_p_s32): Remove.
16315 (__arm_vmlsdavaxq_p_s16): Remove.
16316 (__arm_vmladavaq): Remove.
16317 (__arm_vmlsdavaxq): Remove.
16318 (__arm_vmlsdavaq): Remove.
16319 (__arm_vmladavaxq): Remove.
16320 (__arm_vmladavaq_p): Remove.
16321 (__arm_vmladavaxq_p): Remove.
16322 (__arm_vmlsdavaq_p): Remove.
16323 (__arm_vmlsdavaxq_p): Remove.
16325 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16327 * config/arm/arm-mve-builtins-shapes.cc (binary_acca_int32): New.
16328 * config/arm/arm-mve-builtins-shapes.h (binary_acca_int32): New.
16330 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16332 * config/arm/arm-mve-builtins-base.cc (vmladavq, vmladavxq)
16333 (vmlsdavq, vmlsdavxq): New.
16334 * config/arm/arm-mve-builtins-base.def (vmladavq, vmladavxq)
16335 (vmlsdavq, vmlsdavxq): New.
16336 * config/arm/arm-mve-builtins-base.h (vmladavq, vmladavxq)
16337 (vmlsdavq, vmlsdavxq): New.
16338 * config/arm/arm_mve.h (vmladavq): Remove.
16339 (vmlsdavxq): Remove.
16340 (vmlsdavq): Remove.
16341 (vmladavxq): Remove.
16342 (vmladavq_p): Remove.
16343 (vmlsdavxq_p): Remove.
16344 (vmlsdavq_p): Remove.
16345 (vmladavxq_p): Remove.
16346 (vmladavq_u8): Remove.
16347 (vmlsdavxq_s8): Remove.
16348 (vmlsdavq_s8): Remove.
16349 (vmladavxq_s8): Remove.
16350 (vmladavq_s8): Remove.
16351 (vmladavq_u16): Remove.
16352 (vmlsdavxq_s16): Remove.
16353 (vmlsdavq_s16): Remove.
16354 (vmladavxq_s16): Remove.
16355 (vmladavq_s16): Remove.
16356 (vmladavq_u32): Remove.
16357 (vmlsdavxq_s32): Remove.
16358 (vmlsdavq_s32): Remove.
16359 (vmladavxq_s32): Remove.
16360 (vmladavq_s32): Remove.
16361 (vmladavq_p_u8): Remove.
16362 (vmlsdavxq_p_s8): Remove.
16363 (vmlsdavq_p_s8): Remove.
16364 (vmladavxq_p_s8): Remove.
16365 (vmladavq_p_s8): Remove.
16366 (vmladavq_p_u16): Remove.
16367 (vmlsdavxq_p_s16): Remove.
16368 (vmlsdavq_p_s16): Remove.
16369 (vmladavxq_p_s16): Remove.
16370 (vmladavq_p_s16): Remove.
16371 (vmladavq_p_u32): Remove.
16372 (vmlsdavxq_p_s32): Remove.
16373 (vmlsdavq_p_s32): Remove.
16374 (vmladavxq_p_s32): Remove.
16375 (vmladavq_p_s32): Remove.
16376 (__arm_vmladavq_u8): Remove.
16377 (__arm_vmlsdavxq_s8): Remove.
16378 (__arm_vmlsdavq_s8): Remove.
16379 (__arm_vmladavxq_s8): Remove.
16380 (__arm_vmladavq_s8): Remove.
16381 (__arm_vmladavq_u16): Remove.
16382 (__arm_vmlsdavxq_s16): Remove.
16383 (__arm_vmlsdavq_s16): Remove.
16384 (__arm_vmladavxq_s16): Remove.
16385 (__arm_vmladavq_s16): Remove.
16386 (__arm_vmladavq_u32): Remove.
16387 (__arm_vmlsdavxq_s32): Remove.
16388 (__arm_vmlsdavq_s32): Remove.
16389 (__arm_vmladavxq_s32): Remove.
16390 (__arm_vmladavq_s32): Remove.
16391 (__arm_vmladavq_p_u8): Remove.
16392 (__arm_vmlsdavxq_p_s8): Remove.
16393 (__arm_vmlsdavq_p_s8): Remove.
16394 (__arm_vmladavxq_p_s8): Remove.
16395 (__arm_vmladavq_p_s8): Remove.
16396 (__arm_vmladavq_p_u16): Remove.
16397 (__arm_vmlsdavxq_p_s16): Remove.
16398 (__arm_vmlsdavq_p_s16): Remove.
16399 (__arm_vmladavxq_p_s16): Remove.
16400 (__arm_vmladavq_p_s16): Remove.
16401 (__arm_vmladavq_p_u32): Remove.
16402 (__arm_vmlsdavxq_p_s32): Remove.
16403 (__arm_vmlsdavq_p_s32): Remove.
16404 (__arm_vmladavxq_p_s32): Remove.
16405 (__arm_vmladavq_p_s32): Remove.
16406 (__arm_vmladavq): Remove.
16407 (__arm_vmlsdavxq): Remove.
16408 (__arm_vmlsdavq): Remove.
16409 (__arm_vmladavxq): Remove.
16410 (__arm_vmladavq_p): Remove.
16411 (__arm_vmlsdavxq_p): Remove.
16412 (__arm_vmlsdavq_p): Remove.
16413 (__arm_vmladavxq_p): Remove.
16415 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16417 * config/arm/iterators.md (MVE_VMLxDAVQ, MVE_VMLxDAVQ_P)
16418 (MVE_VMLxDAVAQ, MVE_VMLxDAVAQ_P): New.
16419 (mve_insn): Add vmladava, vmladavax, vmladav, vmladavx, vmlsdava,
16420 vmlsdavax, vmlsdav, vmlsdavx.
16421 (supf): Add VMLADAVAXQ_P_S, VMLADAVAXQ_S, VMLADAVXQ_P_S,
16422 VMLADAVXQ_S, VMLSDAVAQ_P_S, VMLSDAVAQ_S, VMLSDAVAXQ_P_S,
16423 VMLSDAVAXQ_S, VMLSDAVQ_P_S, VMLSDAVQ_S, VMLSDAVXQ_P_S,
16425 * config/arm/mve.md (mve_vmladavq_<supf><mode>)
16426 (mve_vmladavxq_s<mode>, mve_vmlsdavq_s<mode>)
16427 (mve_vmlsdavxq_s<mode>): Merge into ...
16428 (@mve_<mve_insn>q_<supf><mode>): ... this.
16429 (mve_vmlsdavaq_s<mode>, mve_vmladavaxq_s<mode>)
16430 (mve_vmlsdavaxq_s<mode>, mve_vmladavaq_<supf><mode>): Merge into
16432 (@mve_<mve_insn>q_<supf><mode>): ... this.
16433 (mve_vmladavq_p_<supf><mode>, mve_vmladavxq_p_s<mode>)
16434 (mve_vmlsdavq_p_s<mode>, mve_vmlsdavxq_p_s<mode>): Merge into ...
16435 (@mve_<mve_insn>q_p_<supf><mode>): ... this.
16436 (mve_vmladavaq_p_<supf><mode>, mve_vmladavaxq_p_s<mode>)
16437 (mve_vmlsdavaq_p_s<mode>, mve_vmlsdavaxq_p_s<mode>): Merge into
16439 (@mve_<mve_insn>q_p_<supf><mode>): ... this.
16441 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16443 * config/arm/arm-mve-builtins-shapes.cc (binary_acc_int32): New.
16444 * config/arm/arm-mve-builtins-shapes.h (binary_acc_int32): New.
16446 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16448 * config/arm/arm-mve-builtins-base.cc (vaddlvaq): New.
16449 * config/arm/arm-mve-builtins-base.def (vaddlvaq): New.
16450 * config/arm/arm-mve-builtins-base.h (vaddlvaq): New.
16451 * config/arm/arm_mve.h (vaddlvaq): Remove.
16452 (vaddlvaq_p): Remove.
16453 (vaddlvaq_u32): Remove.
16454 (vaddlvaq_s32): Remove.
16455 (vaddlvaq_p_s32): Remove.
16456 (vaddlvaq_p_u32): Remove.
16457 (__arm_vaddlvaq_u32): Remove.
16458 (__arm_vaddlvaq_s32): Remove.
16459 (__arm_vaddlvaq_p_s32): Remove.
16460 (__arm_vaddlvaq_p_u32): Remove.
16461 (__arm_vaddlvaq): Remove.
16462 (__arm_vaddlvaq_p): Remove.
16464 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16466 * config/arm/arm-mve-builtins-shapes.cc (unary_widen_acc): New.
16467 * config/arm/arm-mve-builtins-shapes.h (unary_widen_acc): New.
16469 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16471 * config/arm/iterators.md (mve_insn): Add vaddlva.
16472 * config/arm/mve.md (mve_vaddlvaq_<supf>v4si): Rename into ...
16473 (@mve_<mve_insn>q_<supf>v4si): ... this.
16474 (mve_vaddlvaq_p_<supf>v4si): Rename into ...
16475 (@mve_<mve_insn>q_p_<supf>v4si): ... this.
16477 2023-05-11 Uros Bizjak <ubizjak@gmail.com>
16480 * config/i386/i386.cc (ix86_widen_mult_cost):
16481 Handle V4HImode and V2SImode.
16483 2023-05-11 Andrew Pinski <apinski@marvell.com>
16485 * tree-ssa-dce.cc (simple_dce_from_worklist): For ssa names
16486 defined by a phi node with more than one uses, allow for the
16487 only uses are in that same defining statement.
16489 2023-05-11 Robin Dapp <rdapp@ventanamicro.com>
16491 * config/riscv/riscv.cc (riscv_const_insns): Add permissible
16494 2023-05-11 Pan Li <pan2.li@intel.com>
16496 * config/riscv/vector.md: Add comments for simplifying to vmset.
16498 2023-05-11 Robin Dapp <rdapp@ventanamicro.com>
16500 * config/riscv/autovec.md (<optab><mode>3): Add scalar shift
16502 (v<optab><mode>3): Add vector shift pattern.
16503 * config/riscv/vector-iterators.md: New iterator.
16505 2023-05-11 Robin Dapp <rdapp@ventanamicro.com>
16507 * config/riscv/autovec.md: Use renamed functions.
16508 * config/riscv/riscv-protos.h (emit_vlmax_op): Rename.
16509 (emit_vlmax_reg_op): To this.
16510 (emit_nonvlmax_op): Rename.
16511 (emit_len_op): To this.
16512 (emit_nonvlmax_binop): Rename.
16513 (emit_len_binop): To this.
16514 * config/riscv/riscv-v.cc (emit_pred_op): Add default parameter.
16515 (emit_pred_binop): Remove vlmax_p.
16516 (emit_vlmax_op): Rename.
16517 (emit_vlmax_reg_op): To this.
16518 (emit_nonvlmax_op): Rename.
16519 (emit_len_op): To this.
16520 (emit_nonvlmax_binop): Rename.
16521 (emit_len_binop): To this.
16522 (sew64_scalar_helper): Use renamed functions.
16523 (expand_tuple_move): Use renamed functions.
16524 * config/riscv/riscv.cc (vector_zero_call_used_regs): Use
16526 * config/riscv/vector.md: Use renamed functions.
16528 2023-05-11 Robin Dapp <rdapp@ventanamicro.com>
16529 Michael Collison <collison@rivosinc.com>
16531 * config/riscv/autovec.md (<optab><mode>3): Add integer binops.
16532 * config/riscv/riscv-protos.h (emit_nonvlmax_binop): Declare.
16533 * config/riscv/riscv-v.cc (emit_pred_op): New function.
16534 (set_expander_dest_and_mask): New function.
16535 (emit_pred_binop): New function.
16536 (emit_nonvlmax_binop): New function.
16538 2023-05-11 Pan Li <pan2.li@intel.com>
16540 * cfgloopmanip.cc (create_empty_loop_on_edge): Add PLUS_EXPR.
16541 * gimple-loop-interchange.cc
16542 (tree_loop_interchange::map_inductions_to_loop): Ditto.
16543 * tree-ssa-loop-ivcanon.cc (create_canonical_iv): Ditto.
16544 * tree-ssa-loop-ivopts.cc (create_new_iv): Ditto.
16545 * tree-ssa-loop-manip.cc (create_iv): Ditto.
16546 (tree_transform_and_unroll_loop): Ditto.
16547 (canonicalize_loop_ivs): Ditto.
16548 * tree-ssa-loop-manip.h (create_iv): Ditto.
16549 * tree-vect-data-refs.cc (vect_create_data_ref_ptr): Ditto.
16550 * tree-vect-loop-manip.cc (vect_set_loop_controls_directly):
16552 (vect_set_loop_condition_normal): Ditto.
16553 * tree-vect-loop.cc (vect_create_epilog_for_reduction): Ditto.
16554 * tree-vect-stmts.cc (vectorizable_store): Ditto.
16555 (vectorizable_load): Ditto.
16557 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16559 * config/arm/arm-mve-builtins-base.cc (vmovlbq, vmovltq): New.
16560 * config/arm/arm-mve-builtins-base.def (vmovlbq, vmovltq): New.
16561 * config/arm/arm-mve-builtins-base.h (vmovlbq, vmovltq): New.
16562 * config/arm/arm_mve.h (vmovlbq): Remove.
16564 (vmovlbq_m): Remove.
16565 (vmovltq_m): Remove.
16566 (vmovlbq_x): Remove.
16567 (vmovltq_x): Remove.
16568 (vmovlbq_s8): Remove.
16569 (vmovlbq_s16): Remove.
16570 (vmovltq_s8): Remove.
16571 (vmovltq_s16): Remove.
16572 (vmovltq_u8): Remove.
16573 (vmovltq_u16): Remove.
16574 (vmovlbq_u8): Remove.
16575 (vmovlbq_u16): Remove.
16576 (vmovlbq_m_s8): Remove.
16577 (vmovltq_m_s8): Remove.
16578 (vmovlbq_m_u8): Remove.
16579 (vmovltq_m_u8): Remove.
16580 (vmovlbq_m_s16): Remove.
16581 (vmovltq_m_s16): Remove.
16582 (vmovlbq_m_u16): Remove.
16583 (vmovltq_m_u16): Remove.
16584 (vmovlbq_x_s8): Remove.
16585 (vmovlbq_x_s16): Remove.
16586 (vmovlbq_x_u8): Remove.
16587 (vmovlbq_x_u16): Remove.
16588 (vmovltq_x_s8): Remove.
16589 (vmovltq_x_s16): Remove.
16590 (vmovltq_x_u8): Remove.
16591 (vmovltq_x_u16): Remove.
16592 (__arm_vmovlbq_s8): Remove.
16593 (__arm_vmovlbq_s16): Remove.
16594 (__arm_vmovltq_s8): Remove.
16595 (__arm_vmovltq_s16): Remove.
16596 (__arm_vmovltq_u8): Remove.
16597 (__arm_vmovltq_u16): Remove.
16598 (__arm_vmovlbq_u8): Remove.
16599 (__arm_vmovlbq_u16): Remove.
16600 (__arm_vmovlbq_m_s8): Remove.
16601 (__arm_vmovltq_m_s8): Remove.
16602 (__arm_vmovlbq_m_u8): Remove.
16603 (__arm_vmovltq_m_u8): Remove.
16604 (__arm_vmovlbq_m_s16): Remove.
16605 (__arm_vmovltq_m_s16): Remove.
16606 (__arm_vmovlbq_m_u16): Remove.
16607 (__arm_vmovltq_m_u16): Remove.
16608 (__arm_vmovlbq_x_s8): Remove.
16609 (__arm_vmovlbq_x_s16): Remove.
16610 (__arm_vmovlbq_x_u8): Remove.
16611 (__arm_vmovlbq_x_u16): Remove.
16612 (__arm_vmovltq_x_s8): Remove.
16613 (__arm_vmovltq_x_s16): Remove.
16614 (__arm_vmovltq_x_u8): Remove.
16615 (__arm_vmovltq_x_u16): Remove.
16616 (__arm_vmovlbq): Remove.
16617 (__arm_vmovltq): Remove.
16618 (__arm_vmovlbq_m): Remove.
16619 (__arm_vmovltq_m): Remove.
16620 (__arm_vmovlbq_x): Remove.
16621 (__arm_vmovltq_x): Remove.
16623 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16625 * config/arm/arm-mve-builtins-shapes.cc (unary_widen): New.
16626 * config/arm/arm-mve-builtins-shapes.h (unary_widen): New.
16628 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16630 * config/arm/iterators.md (mve_insn): Add vmovlb, vmovlt.
16631 (VMOVLBQ, VMOVLTQ): Merge into ...
16632 (VMOVLxQ): ... this.
16633 (VMOVLTQ_M, VMOVLBQ_M): Merge into ...
16634 (VMOVLxQ_M): ... this.
16635 * config/arm/mve.md (mve_vmovltq_<supf><mode>)
16636 (mve_vmovlbq_<supf><mode>): Merge into ...
16637 (@mve_<mve_insn>q_<supf><mode>): ... this.
16638 (mve_vmovlbq_m_<supf><mode>, mve_vmovltq_m_<supf><mode>): Merge
16640 (@mve_<mve_insn>q_m_<supf><mode>): ... this.
16642 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16644 * config/arm/arm-mve-builtins-base.cc (vaddlvq): New.
16645 * config/arm/arm-mve-builtins-base.def (vaddlvq): New.
16646 * config/arm/arm-mve-builtins-base.h (vaddlvq): New.
16647 * config/arm/arm-mve-builtins-functions.h
16648 (unspec_mve_function_exact_insn_pred_p): Handle vaddlvq.
16649 * config/arm/arm_mve.h (vaddlvq): Remove.
16650 (vaddlvq_p): Remove.
16651 (vaddlvq_s32): Remove.
16652 (vaddlvq_u32): Remove.
16653 (vaddlvq_p_s32): Remove.
16654 (vaddlvq_p_u32): Remove.
16655 (__arm_vaddlvq_s32): Remove.
16656 (__arm_vaddlvq_u32): Remove.
16657 (__arm_vaddlvq_p_s32): Remove.
16658 (__arm_vaddlvq_p_u32): Remove.
16659 (__arm_vaddlvq): Remove.
16660 (__arm_vaddlvq_p): Remove.
16662 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16664 * config/arm/iterators.md (mve_insn): Add vaddlv.
16665 * config/arm/mve.md (mve_vaddlvq_<supf>v4si): Rename into ...
16666 (@mve_<mve_insn>q_<supf>v4si): ... this.
16667 (mve_vaddlvq_p_<supf>v4si): Rename into ...
16668 (@mve_<mve_insn>q_p_<supf>v4si): ... this.
16670 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16672 * config/arm/arm-mve-builtins-shapes.cc (unary_acc): New.
16673 * config/arm/arm-mve-builtins-shapes.h (unary_acc): New.
16675 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16677 * config/arm/arm-mve-builtins-base.cc (vaddvaq): New.
16678 * config/arm/arm-mve-builtins-base.def (vaddvaq): New.
16679 * config/arm/arm-mve-builtins-base.h (vaddvaq): New.
16680 * config/arm/arm_mve.h (vaddvaq): Remove.
16681 (vaddvaq_p): Remove.
16682 (vaddvaq_u8): Remove.
16683 (vaddvaq_s8): Remove.
16684 (vaddvaq_u16): Remove.
16685 (vaddvaq_s16): Remove.
16686 (vaddvaq_u32): Remove.
16687 (vaddvaq_s32): Remove.
16688 (vaddvaq_p_u8): Remove.
16689 (vaddvaq_p_s8): Remove.
16690 (vaddvaq_p_u16): Remove.
16691 (vaddvaq_p_s16): Remove.
16692 (vaddvaq_p_u32): Remove.
16693 (vaddvaq_p_s32): Remove.
16694 (__arm_vaddvaq_u8): Remove.
16695 (__arm_vaddvaq_s8): Remove.
16696 (__arm_vaddvaq_u16): Remove.
16697 (__arm_vaddvaq_s16): Remove.
16698 (__arm_vaddvaq_u32): Remove.
16699 (__arm_vaddvaq_s32): Remove.
16700 (__arm_vaddvaq_p_u8): Remove.
16701 (__arm_vaddvaq_p_s8): Remove.
16702 (__arm_vaddvaq_p_u16): Remove.
16703 (__arm_vaddvaq_p_s16): Remove.
16704 (__arm_vaddvaq_p_u32): Remove.
16705 (__arm_vaddvaq_p_s32): Remove.
16706 (__arm_vaddvaq): Remove.
16707 (__arm_vaddvaq_p): Remove.
16709 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16711 * config/arm/arm-mve-builtins-shapes.cc (unary_int32_acc): New.
16712 * config/arm/arm-mve-builtins-shapes.h (unary_int32_acc): New.
16714 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16716 * config/arm/iterators.md (mve_insn): Add vaddva.
16717 * config/arm/mve.md (mve_vaddvaq_<supf><mode>): Rename into ...
16718 (@mve_<mve_insn>q_<supf><mode>): ... this.
16719 (mve_vaddvaq_p_<supf><mode>): Rename into ...
16720 (@mve_<mve_insn>q_p_<supf><mode>): ... this.
16722 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16724 * config/arm/arm-mve-builtins-base.cc (vaddvq): New.
16725 * config/arm/arm-mve-builtins-base.def (vaddvq): New.
16726 * config/arm/arm-mve-builtins-base.h (vaddvq): New.
16727 * config/arm/arm_mve.h (vaddvq): Remove.
16728 (vaddvq_p): Remove.
16729 (vaddvq_s8): Remove.
16730 (vaddvq_s16): Remove.
16731 (vaddvq_s32): Remove.
16732 (vaddvq_u8): Remove.
16733 (vaddvq_u16): Remove.
16734 (vaddvq_u32): Remove.
16735 (vaddvq_p_u8): Remove.
16736 (vaddvq_p_s8): Remove.
16737 (vaddvq_p_u16): Remove.
16738 (vaddvq_p_s16): Remove.
16739 (vaddvq_p_u32): Remove.
16740 (vaddvq_p_s32): Remove.
16741 (__arm_vaddvq_s8): Remove.
16742 (__arm_vaddvq_s16): Remove.
16743 (__arm_vaddvq_s32): Remove.
16744 (__arm_vaddvq_u8): Remove.
16745 (__arm_vaddvq_u16): Remove.
16746 (__arm_vaddvq_u32): Remove.
16747 (__arm_vaddvq_p_u8): Remove.
16748 (__arm_vaddvq_p_s8): Remove.
16749 (__arm_vaddvq_p_u16): Remove.
16750 (__arm_vaddvq_p_s16): Remove.
16751 (__arm_vaddvq_p_u32): Remove.
16752 (__arm_vaddvq_p_s32): Remove.
16753 (__arm_vaddvq): Remove.
16754 (__arm_vaddvq_p): Remove.
16756 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16758 * config/arm/arm-mve-builtins-shapes.cc (unary_int32): New.
16759 * config/arm/arm-mve-builtins-shapes.h (unary_int32): New.
16761 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16763 * config/arm/iterators.md (mve_insn): Add vaddv.
16764 * config/arm/mve.md (@mve_vaddvq_<supf><mode>): Rename into ...
16765 (@mve_<mve_insn>q_<supf><mode>): ... this.
16766 (mve_vaddvq_p_<supf><mode>): Rename into ...
16767 (@mve_<mve_insn>q_p_<supf><mode>): ... this.
16768 * config/arm/vec-common.md: Use gen_mve_q instead of
16771 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16773 * config/arm/arm-mve-builtins-base.cc (FUNCTION_ONLY_N): New.
16775 * config/arm/arm-mve-builtins-base.def (vdupq): New.
16776 * config/arm/arm-mve-builtins-base.h: (vdupq): New.
16777 * config/arm/arm_mve.h (vdupq_n): Remove.
16779 (vdupq_n_f16): Remove.
16780 (vdupq_n_f32): Remove.
16781 (vdupq_n_s8): Remove.
16782 (vdupq_n_s16): Remove.
16783 (vdupq_n_s32): Remove.
16784 (vdupq_n_u8): Remove.
16785 (vdupq_n_u16): Remove.
16786 (vdupq_n_u32): Remove.
16787 (vdupq_m_n_u8): Remove.
16788 (vdupq_m_n_s8): Remove.
16789 (vdupq_m_n_u16): Remove.
16790 (vdupq_m_n_s16): Remove.
16791 (vdupq_m_n_u32): Remove.
16792 (vdupq_m_n_s32): Remove.
16793 (vdupq_m_n_f16): Remove.
16794 (vdupq_m_n_f32): Remove.
16795 (vdupq_x_n_s8): Remove.
16796 (vdupq_x_n_s16): Remove.
16797 (vdupq_x_n_s32): Remove.
16798 (vdupq_x_n_u8): Remove.
16799 (vdupq_x_n_u16): Remove.
16800 (vdupq_x_n_u32): Remove.
16801 (vdupq_x_n_f16): Remove.
16802 (vdupq_x_n_f32): Remove.
16803 (__arm_vdupq_n_s8): Remove.
16804 (__arm_vdupq_n_s16): Remove.
16805 (__arm_vdupq_n_s32): Remove.
16806 (__arm_vdupq_n_u8): Remove.
16807 (__arm_vdupq_n_u16): Remove.
16808 (__arm_vdupq_n_u32): Remove.
16809 (__arm_vdupq_m_n_u8): Remove.
16810 (__arm_vdupq_m_n_s8): Remove.
16811 (__arm_vdupq_m_n_u16): Remove.
16812 (__arm_vdupq_m_n_s16): Remove.
16813 (__arm_vdupq_m_n_u32): Remove.
16814 (__arm_vdupq_m_n_s32): Remove.
16815 (__arm_vdupq_x_n_s8): Remove.
16816 (__arm_vdupq_x_n_s16): Remove.
16817 (__arm_vdupq_x_n_s32): Remove.
16818 (__arm_vdupq_x_n_u8): Remove.
16819 (__arm_vdupq_x_n_u16): Remove.
16820 (__arm_vdupq_x_n_u32): Remove.
16821 (__arm_vdupq_n_f16): Remove.
16822 (__arm_vdupq_n_f32): Remove.
16823 (__arm_vdupq_m_n_f16): Remove.
16824 (__arm_vdupq_m_n_f32): Remove.
16825 (__arm_vdupq_x_n_f16): Remove.
16826 (__arm_vdupq_x_n_f32): Remove.
16827 (__arm_vdupq_n): Remove.
16828 (__arm_vdupq_m): Remove.
16830 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16832 * config/arm/arm-mve-builtins-shapes.cc (unary_n): New.
16833 * config/arm/arm-mve-builtins-shapes.h (unary_n): New.
16835 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16837 * config/arm/iterators.md (MVE_FP_M_N_VDUPQ_ONLY)
16838 (MVE_FP_N_VDUPQ_ONLY): New.
16839 (mve_insn): Add vdupq.
16840 * config/arm/mve.md (mve_vdupq_n_f<mode>): Rename into ...
16841 (@mve_<mve_insn>q_n_f<mode>): ... this.
16842 (mve_vdupq_n_<supf><mode>): Rename into ...
16843 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
16844 (mve_vdupq_m_n_<supf><mode>): Rename into ...
16845 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
16846 (mve_vdupq_m_n_f<mode>): Rename into ...
16847 (@mve_<mve_insn>q_m_n_f<mode>): ... this.
16849 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16851 * config/arm/arm-mve-builtins-base.cc (vrev16q, vrev32q, vrev64q):
16853 * config/arm/arm-mve-builtins-base.def (vrev16q, vrev32q)
16855 * config/arm/arm-mve-builtins-base.h (vrev16q, vrev32q)
16857 * config/arm/arm_mve.h (vrev16q): Remove.
16860 (vrev64q_m): Remove.
16861 (vrev16q_m): Remove.
16862 (vrev32q_m): Remove.
16863 (vrev16q_x): Remove.
16864 (vrev32q_x): Remove.
16865 (vrev64q_x): Remove.
16866 (vrev64q_f16): Remove.
16867 (vrev64q_f32): Remove.
16868 (vrev32q_f16): Remove.
16869 (vrev16q_s8): Remove.
16870 (vrev32q_s8): Remove.
16871 (vrev32q_s16): Remove.
16872 (vrev64q_s8): Remove.
16873 (vrev64q_s16): Remove.
16874 (vrev64q_s32): Remove.
16875 (vrev64q_u8): Remove.
16876 (vrev64q_u16): Remove.
16877 (vrev64q_u32): Remove.
16878 (vrev32q_u8): Remove.
16879 (vrev32q_u16): Remove.
16880 (vrev16q_u8): Remove.
16881 (vrev64q_m_u8): Remove.
16882 (vrev64q_m_s8): Remove.
16883 (vrev64q_m_u16): Remove.
16884 (vrev64q_m_s16): Remove.
16885 (vrev64q_m_u32): Remove.
16886 (vrev64q_m_s32): Remove.
16887 (vrev16q_m_s8): Remove.
16888 (vrev32q_m_f16): Remove.
16889 (vrev16q_m_u8): Remove.
16890 (vrev32q_m_s8): Remove.
16891 (vrev64q_m_f16): Remove.
16892 (vrev32q_m_u8): Remove.
16893 (vrev32q_m_s16): Remove.
16894 (vrev64q_m_f32): Remove.
16895 (vrev32q_m_u16): Remove.
16896 (vrev16q_x_s8): Remove.
16897 (vrev16q_x_u8): Remove.
16898 (vrev32q_x_s8): Remove.
16899 (vrev32q_x_s16): Remove.
16900 (vrev32q_x_u8): Remove.
16901 (vrev32q_x_u16): Remove.
16902 (vrev64q_x_s8): Remove.
16903 (vrev64q_x_s16): Remove.
16904 (vrev64q_x_s32): Remove.
16905 (vrev64q_x_u8): Remove.
16906 (vrev64q_x_u16): Remove.
16907 (vrev64q_x_u32): Remove.
16908 (vrev32q_x_f16): Remove.
16909 (vrev64q_x_f16): Remove.
16910 (vrev64q_x_f32): Remove.
16911 (__arm_vrev16q_s8): Remove.
16912 (__arm_vrev32q_s8): Remove.
16913 (__arm_vrev32q_s16): Remove.
16914 (__arm_vrev64q_s8): Remove.
16915 (__arm_vrev64q_s16): Remove.
16916 (__arm_vrev64q_s32): Remove.
16917 (__arm_vrev64q_u8): Remove.
16918 (__arm_vrev64q_u16): Remove.
16919 (__arm_vrev64q_u32): Remove.
16920 (__arm_vrev32q_u8): Remove.
16921 (__arm_vrev32q_u16): Remove.
16922 (__arm_vrev16q_u8): Remove.
16923 (__arm_vrev64q_m_u8): Remove.
16924 (__arm_vrev64q_m_s8): Remove.
16925 (__arm_vrev64q_m_u16): Remove.
16926 (__arm_vrev64q_m_s16): Remove.
16927 (__arm_vrev64q_m_u32): Remove.
16928 (__arm_vrev64q_m_s32): Remove.
16929 (__arm_vrev16q_m_s8): Remove.
16930 (__arm_vrev16q_m_u8): Remove.
16931 (__arm_vrev32q_m_s8): Remove.
16932 (__arm_vrev32q_m_u8): Remove.
16933 (__arm_vrev32q_m_s16): Remove.
16934 (__arm_vrev32q_m_u16): Remove.
16935 (__arm_vrev16q_x_s8): Remove.
16936 (__arm_vrev16q_x_u8): Remove.
16937 (__arm_vrev32q_x_s8): Remove.
16938 (__arm_vrev32q_x_s16): Remove.
16939 (__arm_vrev32q_x_u8): Remove.
16940 (__arm_vrev32q_x_u16): Remove.
16941 (__arm_vrev64q_x_s8): Remove.
16942 (__arm_vrev64q_x_s16): Remove.
16943 (__arm_vrev64q_x_s32): Remove.
16944 (__arm_vrev64q_x_u8): Remove.
16945 (__arm_vrev64q_x_u16): Remove.
16946 (__arm_vrev64q_x_u32): Remove.
16947 (__arm_vrev64q_f16): Remove.
16948 (__arm_vrev64q_f32): Remove.
16949 (__arm_vrev32q_f16): Remove.
16950 (__arm_vrev32q_m_f16): Remove.
16951 (__arm_vrev64q_m_f16): Remove.
16952 (__arm_vrev64q_m_f32): Remove.
16953 (__arm_vrev32q_x_f16): Remove.
16954 (__arm_vrev64q_x_f16): Remove.
16955 (__arm_vrev64q_x_f32): Remove.
16956 (__arm_vrev16q): Remove.
16957 (__arm_vrev32q): Remove.
16958 (__arm_vrev64q): Remove.
16959 (__arm_vrev64q_m): Remove.
16960 (__arm_vrev16q_m): Remove.
16961 (__arm_vrev32q_m): Remove.
16962 (__arm_vrev16q_x): Remove.
16963 (__arm_vrev32q_x): Remove.
16964 (__arm_vrev64q_x): Remove.
16966 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16968 * config/arm/iterators.md (MVE_V8HF, MVE_V16QI)
16969 (MVE_FP_VREV64Q_ONLY, MVE_FP_M_VREV64Q_ONLY, MVE_FP_VREV32Q_ONLY)
16970 (MVE_FP_M_VREV32Q_ONLY): New iterators.
16971 (mve_insn): Add vrev16q, vrev32q, vrev64q.
16972 * config/arm/mve.md (mve_vrev64q_f<mode>): Rename into ...
16973 (@mve_<mve_insn>q_f<mode>): ... this
16974 (mve_vrev32q_fv8hf): Rename into @mve_<mve_insn>q_f<mode>.
16975 (mve_vrev64q_<supf><mode>): Rename into ...
16976 (@mve_<mve_insn>q_<supf><mode>): ... this.
16977 (mve_vrev32q_<supf><mode>): Rename into
16978 @mve_<mve_insn>q_<supf><mode>.
16979 (mve_vrev16q_<supf>v16qi): Rename into
16980 @mve_<mve_insn>q_<supf><mode>.
16981 (mve_vrev64q_m_<supf><mode>): Rename into
16982 @mve_<mve_insn>q_m_<supf><mode>.
16983 (mve_vrev32q_m_fv8hf): Rename into @mve_<mve_insn>q_m_f<mode>.
16984 (mve_vrev32q_m_<supf><mode>): Rename into
16985 @mve_<mve_insn>q_m_<supf><mode>.
16986 (mve_vrev64q_m_f<mode>): Rename into @mve_<mve_insn>q_m_f<mode>.
16987 (mve_vrev16q_m_<supf>v16qi): Rename into
16988 @mve_<mve_insn>q_m_<supf><mode>.
16990 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16992 * config/arm/arm-mve-builtins-base.cc (vcmpeqq, vcmpneq, vcmpgeq)
16993 (vcmpgtq, vcmpleq, vcmpltq, vcmpcsq, vcmphiq): New.
16994 * config/arm/arm-mve-builtins-base.def (vcmpeqq, vcmpneq, vcmpgeq)
16995 (vcmpgtq, vcmpleq, vcmpltq, vcmpcsq, vcmphiq): New.
16996 * config/arm/arm-mve-builtins-base.h (vcmpeqq, vcmpneq, vcmpgeq)
16997 (vcmpgtq, vcmpleq, vcmpltq, vcmpcsq, vcmphiq): New.
16998 * config/arm/arm-mve-builtins-functions.h (class
16999 unspec_based_mve_function_exact_insn_vcmp): New.
17000 * config/arm/arm-mve-builtins.cc
17001 (function_instance::has_inactive_argument): Handle vcmp.
17002 * config/arm/arm_mve.h (vcmpneq): Remove.
17010 (vcmpneq_m): Remove.
17011 (vcmphiq_m): Remove.
17012 (vcmpeqq_m): Remove.
17013 (vcmpcsq_m): Remove.
17014 (vcmpcsq_m_n): Remove.
17015 (vcmpltq_m): Remove.
17016 (vcmpleq_m): Remove.
17017 (vcmpgtq_m): Remove.
17018 (vcmpgeq_m): Remove.
17019 (vcmpneq_s8): Remove.
17020 (vcmpneq_s16): Remove.
17021 (vcmpneq_s32): Remove.
17022 (vcmpneq_u8): Remove.
17023 (vcmpneq_u16): Remove.
17024 (vcmpneq_u32): Remove.
17025 (vcmpneq_n_u8): Remove.
17026 (vcmphiq_u8): Remove.
17027 (vcmphiq_n_u8): Remove.
17028 (vcmpeqq_u8): Remove.
17029 (vcmpeqq_n_u8): Remove.
17030 (vcmpcsq_u8): Remove.
17031 (vcmpcsq_n_u8): Remove.
17032 (vcmpneq_n_s8): Remove.
17033 (vcmpltq_s8): Remove.
17034 (vcmpltq_n_s8): Remove.
17035 (vcmpleq_s8): Remove.
17036 (vcmpleq_n_s8): Remove.
17037 (vcmpgtq_s8): Remove.
17038 (vcmpgtq_n_s8): Remove.
17039 (vcmpgeq_s8): Remove.
17040 (vcmpgeq_n_s8): Remove.
17041 (vcmpeqq_s8): Remove.
17042 (vcmpeqq_n_s8): Remove.
17043 (vcmpneq_n_u16): Remove.
17044 (vcmphiq_u16): Remove.
17045 (vcmphiq_n_u16): Remove.
17046 (vcmpeqq_u16): Remove.
17047 (vcmpeqq_n_u16): Remove.
17048 (vcmpcsq_u16): Remove.
17049 (vcmpcsq_n_u16): Remove.
17050 (vcmpneq_n_s16): Remove.
17051 (vcmpltq_s16): Remove.
17052 (vcmpltq_n_s16): Remove.
17053 (vcmpleq_s16): Remove.
17054 (vcmpleq_n_s16): Remove.
17055 (vcmpgtq_s16): Remove.
17056 (vcmpgtq_n_s16): Remove.
17057 (vcmpgeq_s16): Remove.
17058 (vcmpgeq_n_s16): Remove.
17059 (vcmpeqq_s16): Remove.
17060 (vcmpeqq_n_s16): Remove.
17061 (vcmpneq_n_u32): Remove.
17062 (vcmphiq_u32): Remove.
17063 (vcmphiq_n_u32): Remove.
17064 (vcmpeqq_u32): Remove.
17065 (vcmpeqq_n_u32): Remove.
17066 (vcmpcsq_u32): Remove.
17067 (vcmpcsq_n_u32): Remove.
17068 (vcmpneq_n_s32): Remove.
17069 (vcmpltq_s32): Remove.
17070 (vcmpltq_n_s32): Remove.
17071 (vcmpleq_s32): Remove.
17072 (vcmpleq_n_s32): Remove.
17073 (vcmpgtq_s32): Remove.
17074 (vcmpgtq_n_s32): Remove.
17075 (vcmpgeq_s32): Remove.
17076 (vcmpgeq_n_s32): Remove.
17077 (vcmpeqq_s32): Remove.
17078 (vcmpeqq_n_s32): Remove.
17079 (vcmpneq_n_f16): Remove.
17080 (vcmpneq_f16): Remove.
17081 (vcmpltq_n_f16): Remove.
17082 (vcmpltq_f16): Remove.
17083 (vcmpleq_n_f16): Remove.
17084 (vcmpleq_f16): Remove.
17085 (vcmpgtq_n_f16): Remove.
17086 (vcmpgtq_f16): Remove.
17087 (vcmpgeq_n_f16): Remove.
17088 (vcmpgeq_f16): Remove.
17089 (vcmpeqq_n_f16): Remove.
17090 (vcmpeqq_f16): Remove.
17091 (vcmpneq_n_f32): Remove.
17092 (vcmpneq_f32): Remove.
17093 (vcmpltq_n_f32): Remove.
17094 (vcmpltq_f32): Remove.
17095 (vcmpleq_n_f32): Remove.
17096 (vcmpleq_f32): Remove.
17097 (vcmpgtq_n_f32): Remove.
17098 (vcmpgtq_f32): Remove.
17099 (vcmpgeq_n_f32): Remove.
17100 (vcmpgeq_f32): Remove.
17101 (vcmpeqq_n_f32): Remove.
17102 (vcmpeqq_f32): Remove.
17103 (vcmpeqq_m_f16): Remove.
17104 (vcmpeqq_m_f32): Remove.
17105 (vcmpneq_m_u8): Remove.
17106 (vcmpneq_m_n_u8): Remove.
17107 (vcmphiq_m_u8): Remove.
17108 (vcmphiq_m_n_u8): Remove.
17109 (vcmpeqq_m_u8): Remove.
17110 (vcmpeqq_m_n_u8): Remove.
17111 (vcmpcsq_m_u8): Remove.
17112 (vcmpcsq_m_n_u8): Remove.
17113 (vcmpneq_m_s8): Remove.
17114 (vcmpneq_m_n_s8): Remove.
17115 (vcmpltq_m_s8): Remove.
17116 (vcmpltq_m_n_s8): Remove.
17117 (vcmpleq_m_s8): Remove.
17118 (vcmpleq_m_n_s8): Remove.
17119 (vcmpgtq_m_s8): Remove.
17120 (vcmpgtq_m_n_s8): Remove.
17121 (vcmpgeq_m_s8): Remove.
17122 (vcmpgeq_m_n_s8): Remove.
17123 (vcmpeqq_m_s8): Remove.
17124 (vcmpeqq_m_n_s8): Remove.
17125 (vcmpneq_m_u16): Remove.
17126 (vcmpneq_m_n_u16): Remove.
17127 (vcmphiq_m_u16): Remove.
17128 (vcmphiq_m_n_u16): Remove.
17129 (vcmpeqq_m_u16): Remove.
17130 (vcmpeqq_m_n_u16): Remove.
17131 (vcmpcsq_m_u16): Remove.
17132 (vcmpcsq_m_n_u16): Remove.
17133 (vcmpneq_m_s16): Remove.
17134 (vcmpneq_m_n_s16): Remove.
17135 (vcmpltq_m_s16): Remove.
17136 (vcmpltq_m_n_s16): Remove.
17137 (vcmpleq_m_s16): Remove.
17138 (vcmpleq_m_n_s16): Remove.
17139 (vcmpgtq_m_s16): Remove.
17140 (vcmpgtq_m_n_s16): Remove.
17141 (vcmpgeq_m_s16): Remove.
17142 (vcmpgeq_m_n_s16): Remove.
17143 (vcmpeqq_m_s16): Remove.
17144 (vcmpeqq_m_n_s16): Remove.
17145 (vcmpneq_m_u32): Remove.
17146 (vcmpneq_m_n_u32): Remove.
17147 (vcmphiq_m_u32): Remove.
17148 (vcmphiq_m_n_u32): Remove.
17149 (vcmpeqq_m_u32): Remove.
17150 (vcmpeqq_m_n_u32): Remove.
17151 (vcmpcsq_m_u32): Remove.
17152 (vcmpcsq_m_n_u32): Remove.
17153 (vcmpneq_m_s32): Remove.
17154 (vcmpneq_m_n_s32): Remove.
17155 (vcmpltq_m_s32): Remove.
17156 (vcmpltq_m_n_s32): Remove.
17157 (vcmpleq_m_s32): Remove.
17158 (vcmpleq_m_n_s32): Remove.
17159 (vcmpgtq_m_s32): Remove.
17160 (vcmpgtq_m_n_s32): Remove.
17161 (vcmpgeq_m_s32): Remove.
17162 (vcmpgeq_m_n_s32): Remove.
17163 (vcmpeqq_m_s32): Remove.
17164 (vcmpeqq_m_n_s32): Remove.
17165 (vcmpeqq_m_n_f16): Remove.
17166 (vcmpgeq_m_f16): Remove.
17167 (vcmpgeq_m_n_f16): Remove.
17168 (vcmpgtq_m_f16): Remove.
17169 (vcmpgtq_m_n_f16): Remove.
17170 (vcmpleq_m_f16): Remove.
17171 (vcmpleq_m_n_f16): Remove.
17172 (vcmpltq_m_f16): Remove.
17173 (vcmpltq_m_n_f16): Remove.
17174 (vcmpneq_m_f16): Remove.
17175 (vcmpneq_m_n_f16): Remove.
17176 (vcmpeqq_m_n_f32): Remove.
17177 (vcmpgeq_m_f32): Remove.
17178 (vcmpgeq_m_n_f32): Remove.
17179 (vcmpgtq_m_f32): Remove.
17180 (vcmpgtq_m_n_f32): Remove.
17181 (vcmpleq_m_f32): Remove.
17182 (vcmpleq_m_n_f32): Remove.
17183 (vcmpltq_m_f32): Remove.
17184 (vcmpltq_m_n_f32): Remove.
17185 (vcmpneq_m_f32): Remove.
17186 (vcmpneq_m_n_f32): Remove.
17187 (__arm_vcmpneq_s8): Remove.
17188 (__arm_vcmpneq_s16): Remove.
17189 (__arm_vcmpneq_s32): Remove.
17190 (__arm_vcmpneq_u8): Remove.
17191 (__arm_vcmpneq_u16): Remove.
17192 (__arm_vcmpneq_u32): Remove.
17193 (__arm_vcmpneq_n_u8): Remove.
17194 (__arm_vcmphiq_u8): Remove.
17195 (__arm_vcmphiq_n_u8): Remove.
17196 (__arm_vcmpeqq_u8): Remove.
17197 (__arm_vcmpeqq_n_u8): Remove.
17198 (__arm_vcmpcsq_u8): Remove.
17199 (__arm_vcmpcsq_n_u8): Remove.
17200 (__arm_vcmpneq_n_s8): Remove.
17201 (__arm_vcmpltq_s8): Remove.
17202 (__arm_vcmpltq_n_s8): Remove.
17203 (__arm_vcmpleq_s8): Remove.
17204 (__arm_vcmpleq_n_s8): Remove.
17205 (__arm_vcmpgtq_s8): Remove.
17206 (__arm_vcmpgtq_n_s8): Remove.
17207 (__arm_vcmpgeq_s8): Remove.
17208 (__arm_vcmpgeq_n_s8): Remove.
17209 (__arm_vcmpeqq_s8): Remove.
17210 (__arm_vcmpeqq_n_s8): Remove.
17211 (__arm_vcmpneq_n_u16): Remove.
17212 (__arm_vcmphiq_u16): Remove.
17213 (__arm_vcmphiq_n_u16): Remove.
17214 (__arm_vcmpeqq_u16): Remove.
17215 (__arm_vcmpeqq_n_u16): Remove.
17216 (__arm_vcmpcsq_u16): Remove.
17217 (__arm_vcmpcsq_n_u16): Remove.
17218 (__arm_vcmpneq_n_s16): Remove.
17219 (__arm_vcmpltq_s16): Remove.
17220 (__arm_vcmpltq_n_s16): Remove.
17221 (__arm_vcmpleq_s16): Remove.
17222 (__arm_vcmpleq_n_s16): Remove.
17223 (__arm_vcmpgtq_s16): Remove.
17224 (__arm_vcmpgtq_n_s16): Remove.
17225 (__arm_vcmpgeq_s16): Remove.
17226 (__arm_vcmpgeq_n_s16): Remove.
17227 (__arm_vcmpeqq_s16): Remove.
17228 (__arm_vcmpeqq_n_s16): Remove.
17229 (__arm_vcmpneq_n_u32): Remove.
17230 (__arm_vcmphiq_u32): Remove.
17231 (__arm_vcmphiq_n_u32): Remove.
17232 (__arm_vcmpeqq_u32): Remove.
17233 (__arm_vcmpeqq_n_u32): Remove.
17234 (__arm_vcmpcsq_u32): Remove.
17235 (__arm_vcmpcsq_n_u32): Remove.
17236 (__arm_vcmpneq_n_s32): Remove.
17237 (__arm_vcmpltq_s32): Remove.
17238 (__arm_vcmpltq_n_s32): Remove.
17239 (__arm_vcmpleq_s32): Remove.
17240 (__arm_vcmpleq_n_s32): Remove.
17241 (__arm_vcmpgtq_s32): Remove.
17242 (__arm_vcmpgtq_n_s32): Remove.
17243 (__arm_vcmpgeq_s32): Remove.
17244 (__arm_vcmpgeq_n_s32): Remove.
17245 (__arm_vcmpeqq_s32): Remove.
17246 (__arm_vcmpeqq_n_s32): Remove.
17247 (__arm_vcmpneq_m_u8): Remove.
17248 (__arm_vcmpneq_m_n_u8): Remove.
17249 (__arm_vcmphiq_m_u8): Remove.
17250 (__arm_vcmphiq_m_n_u8): Remove.
17251 (__arm_vcmpeqq_m_u8): Remove.
17252 (__arm_vcmpeqq_m_n_u8): Remove.
17253 (__arm_vcmpcsq_m_u8): Remove.
17254 (__arm_vcmpcsq_m_n_u8): Remove.
17255 (__arm_vcmpneq_m_s8): Remove.
17256 (__arm_vcmpneq_m_n_s8): Remove.
17257 (__arm_vcmpltq_m_s8): Remove.
17258 (__arm_vcmpltq_m_n_s8): Remove.
17259 (__arm_vcmpleq_m_s8): Remove.
17260 (__arm_vcmpleq_m_n_s8): Remove.
17261 (__arm_vcmpgtq_m_s8): Remove.
17262 (__arm_vcmpgtq_m_n_s8): Remove.
17263 (__arm_vcmpgeq_m_s8): Remove.
17264 (__arm_vcmpgeq_m_n_s8): Remove.
17265 (__arm_vcmpeqq_m_s8): Remove.
17266 (__arm_vcmpeqq_m_n_s8): Remove.
17267 (__arm_vcmpneq_m_u16): Remove.
17268 (__arm_vcmpneq_m_n_u16): Remove.
17269 (__arm_vcmphiq_m_u16): Remove.
17270 (__arm_vcmphiq_m_n_u16): Remove.
17271 (__arm_vcmpeqq_m_u16): Remove.
17272 (__arm_vcmpeqq_m_n_u16): Remove.
17273 (__arm_vcmpcsq_m_u16): Remove.
17274 (__arm_vcmpcsq_m_n_u16): Remove.
17275 (__arm_vcmpneq_m_s16): Remove.
17276 (__arm_vcmpneq_m_n_s16): Remove.
17277 (__arm_vcmpltq_m_s16): Remove.
17278 (__arm_vcmpltq_m_n_s16): Remove.
17279 (__arm_vcmpleq_m_s16): Remove.
17280 (__arm_vcmpleq_m_n_s16): Remove.
17281 (__arm_vcmpgtq_m_s16): Remove.
17282 (__arm_vcmpgtq_m_n_s16): Remove.
17283 (__arm_vcmpgeq_m_s16): Remove.
17284 (__arm_vcmpgeq_m_n_s16): Remove.
17285 (__arm_vcmpeqq_m_s16): Remove.
17286 (__arm_vcmpeqq_m_n_s16): Remove.
17287 (__arm_vcmpneq_m_u32): Remove.
17288 (__arm_vcmpneq_m_n_u32): Remove.
17289 (__arm_vcmphiq_m_u32): Remove.
17290 (__arm_vcmphiq_m_n_u32): Remove.
17291 (__arm_vcmpeqq_m_u32): Remove.
17292 (__arm_vcmpeqq_m_n_u32): Remove.
17293 (__arm_vcmpcsq_m_u32): Remove.
17294 (__arm_vcmpcsq_m_n_u32): Remove.
17295 (__arm_vcmpneq_m_s32): Remove.
17296 (__arm_vcmpneq_m_n_s32): Remove.
17297 (__arm_vcmpltq_m_s32): Remove.
17298 (__arm_vcmpltq_m_n_s32): Remove.
17299 (__arm_vcmpleq_m_s32): Remove.
17300 (__arm_vcmpleq_m_n_s32): Remove.
17301 (__arm_vcmpgtq_m_s32): Remove.
17302 (__arm_vcmpgtq_m_n_s32): Remove.
17303 (__arm_vcmpgeq_m_s32): Remove.
17304 (__arm_vcmpgeq_m_n_s32): Remove.
17305 (__arm_vcmpeqq_m_s32): Remove.
17306 (__arm_vcmpeqq_m_n_s32): Remove.
17307 (__arm_vcmpneq_n_f16): Remove.
17308 (__arm_vcmpneq_f16): Remove.
17309 (__arm_vcmpltq_n_f16): Remove.
17310 (__arm_vcmpltq_f16): Remove.
17311 (__arm_vcmpleq_n_f16): Remove.
17312 (__arm_vcmpleq_f16): Remove.
17313 (__arm_vcmpgtq_n_f16): Remove.
17314 (__arm_vcmpgtq_f16): Remove.
17315 (__arm_vcmpgeq_n_f16): Remove.
17316 (__arm_vcmpgeq_f16): Remove.
17317 (__arm_vcmpeqq_n_f16): Remove.
17318 (__arm_vcmpeqq_f16): Remove.
17319 (__arm_vcmpneq_n_f32): Remove.
17320 (__arm_vcmpneq_f32): Remove.
17321 (__arm_vcmpltq_n_f32): Remove.
17322 (__arm_vcmpltq_f32): Remove.
17323 (__arm_vcmpleq_n_f32): Remove.
17324 (__arm_vcmpleq_f32): Remove.
17325 (__arm_vcmpgtq_n_f32): Remove.
17326 (__arm_vcmpgtq_f32): Remove.
17327 (__arm_vcmpgeq_n_f32): Remove.
17328 (__arm_vcmpgeq_f32): Remove.
17329 (__arm_vcmpeqq_n_f32): Remove.
17330 (__arm_vcmpeqq_f32): Remove.
17331 (__arm_vcmpeqq_m_f16): Remove.
17332 (__arm_vcmpeqq_m_f32): Remove.
17333 (__arm_vcmpeqq_m_n_f16): Remove.
17334 (__arm_vcmpgeq_m_f16): Remove.
17335 (__arm_vcmpgeq_m_n_f16): Remove.
17336 (__arm_vcmpgtq_m_f16): Remove.
17337 (__arm_vcmpgtq_m_n_f16): Remove.
17338 (__arm_vcmpleq_m_f16): Remove.
17339 (__arm_vcmpleq_m_n_f16): Remove.
17340 (__arm_vcmpltq_m_f16): Remove.
17341 (__arm_vcmpltq_m_n_f16): Remove.
17342 (__arm_vcmpneq_m_f16): Remove.
17343 (__arm_vcmpneq_m_n_f16): Remove.
17344 (__arm_vcmpeqq_m_n_f32): Remove.
17345 (__arm_vcmpgeq_m_f32): Remove.
17346 (__arm_vcmpgeq_m_n_f32): Remove.
17347 (__arm_vcmpgtq_m_f32): Remove.
17348 (__arm_vcmpgtq_m_n_f32): Remove.
17349 (__arm_vcmpleq_m_f32): Remove.
17350 (__arm_vcmpleq_m_n_f32): Remove.
17351 (__arm_vcmpltq_m_f32): Remove.
17352 (__arm_vcmpltq_m_n_f32): Remove.
17353 (__arm_vcmpneq_m_f32): Remove.
17354 (__arm_vcmpneq_m_n_f32): Remove.
17355 (__arm_vcmpneq): Remove.
17356 (__arm_vcmphiq): Remove.
17357 (__arm_vcmpeqq): Remove.
17358 (__arm_vcmpcsq): Remove.
17359 (__arm_vcmpltq): Remove.
17360 (__arm_vcmpleq): Remove.
17361 (__arm_vcmpgtq): Remove.
17362 (__arm_vcmpgeq): Remove.
17363 (__arm_vcmpneq_m): Remove.
17364 (__arm_vcmphiq_m): Remove.
17365 (__arm_vcmpeqq_m): Remove.
17366 (__arm_vcmpcsq_m): Remove.
17367 (__arm_vcmpltq_m): Remove.
17368 (__arm_vcmpleq_m): Remove.
17369 (__arm_vcmpgtq_m): Remove.
17370 (__arm_vcmpgeq_m): Remove.
17372 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
17374 * config/arm/arm-mve-builtins-shapes.cc (cmp): New.
17375 * config/arm/arm-mve-builtins-shapes.h (cmp): New.
17377 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
17379 * config/arm/iterators.md (MVE_CMP_M, MVE_CMP_M_F, MVE_CMP_M_N)
17380 (MVE_CMP_M_N_F, mve_cmp_op1): New.
17383 * config/arm/mve.md (mve_vcmp<mve_cmp_op>q_n_<mode>): Rename into ...
17384 (@mve_vcmp<mve_cmp_op>q_n_<mode>): ... this.
17385 (mve_vcmpeqq_m_f<mode>, mve_vcmpgeq_m_f<mode>)
17386 (mve_vcmpgtq_m_f<mode>, mve_vcmpleq_m_f<mode>)
17387 (mve_vcmpltq_m_f<mode>, mve_vcmpneq_m_f<mode>): Merge into ...
17388 (@mve_vcmp<mve_cmp_op1>q_m_f<mode>): ... this.
17389 (mve_vcmpcsq_m_u<mode>, mve_vcmpeqq_m_<supf><mode>)
17390 (mve_vcmpgeq_m_s<mode>, mve_vcmpgtq_m_s<mode>)
17391 (mve_vcmphiq_m_u<mode>, mve_vcmpleq_m_s<mode>)
17392 (mve_vcmpltq_m_s<mode>, mve_vcmpneq_m_<supf><mode>): Merge into
17394 (@mve_vcmp<mve_cmp_op1>q_m_<supf><mode>): ... this.
17395 (mve_vcmpcsq_m_n_u<mode>, mve_vcmpeqq_m_n_<supf><mode>)
17396 (mve_vcmpgeq_m_n_s<mode>, mve_vcmpgtq_m_n_s<mode>)
17397 (mve_vcmphiq_m_n_u<mode>, mve_vcmpleq_m_n_s<mode>)
17398 (mve_vcmpltq_m_n_s<mode>, mve_vcmpneq_m_n_<supf><mode>): Merge
17400 (@mve_vcmp<mve_cmp_op1>q_m_n_<supf><mode>): ... this.
17401 (mve_vcmpeqq_m_n_f<mode>, mve_vcmpgeq_m_n_f<mode>)
17402 (mve_vcmpgtq_m_n_f<mode>, mve_vcmpleq_m_n_f<mode>)
17403 (mve_vcmpltq_m_n_f<mode>, mve_vcmpneq_m_n_f<mode>): Merge into ...
17404 (@mve_vcmp<mve_cmp_op1>q_m_n_f<mode>): ... this.
17406 2023-05-11 Roger Sayle <roger@nextmovesoftware.com>
17408 * match.pd <popcount optimizations>: Simplify popcount(X|Y) +
17409 popcount(X&Y) as popcount(X)+popcount(Y). Likewise, simplify
17410 popcount(X)+popcount(Y)-popcount(X&Y) as popcount(X|Y), and
17413 2023-05-11 Roger Sayle <roger@nextmovesoftware.com>
17415 * match.pd <popcount optimizations>: Simplify popcount(bswap(x))
17416 as popcount(x). Simplify popcount(rotate(x,y)) as popcount(x).
17417 <parity optimizations>: Simplify parity(bswap(x)) as parity(x).
17418 Simplify parity(rotate(x,y)) as parity(x).
17420 2023-05-11 Juzhe-Zhong <juzhe.zhong@rivai.ai>
17422 * config/riscv/autovec.md (@vec_series<mode>): New pattern
17423 * config/riscv/riscv-protos.h (expand_vec_series): New function.
17424 * config/riscv/riscv-v.cc (emit_binop): Ditto.
17425 (emit_index_op): Ditto.
17426 (expand_vec_series): Ditto.
17427 (expand_const_vector): Add series vector handling.
17428 * config/riscv/riscv.cc (riscv_const_insns): Enable series vector for testing.
17430 2023-05-10 Roger Sayle <roger@nextmovesoftware.com>
17432 * config/i386/i386.md (*concat<mode><dwi>3_1): Use preferred
17433 [(const_int 0)] idiom, instead of [(clobber (const_int 0))].
17434 (*concat<mode><dwi>3_2): Likewise.
17435 (*concat<mode><dwi>3_3): Likewise.
17436 (*concat<mode><dwi>3_4): Likewise.
17437 (*concat<mode><dwi>3_5): Likewise.
17438 (*concat<mode><dwi>3_6): Likewise.
17439 (*concat<mode><dwi>3_7): Likewise.
17441 2023-05-10 Uros Bizjak <ubizjak@gmail.com>
17444 * config/i386/mmx.md (sse4_1_<code>v2qiv2si2): New insn pattern.
17445 (<insn>v4qiv4hi2): New expander.
17446 (<insn>v2hiv2si2): Ditto.
17447 (<insn>v2qiv2si2): Ditto.
17448 (<insn>v2qiv2hi2): Ditto.
17450 2023-05-10 Jeff Law <jlaw@ventanamicro>
17452 * config/h8300/constraints.md (Q): Make this a special memory
17456 2023-05-10 Jakub Jelinek <jakub@redhat.com>
17459 * ipa-prop.cc (ipa_get_callee_param_type): Don't return TREE_VALUE (t)
17460 if t is void_list_node.
17462 2023-05-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17464 * config/aarch64/aarch64-simd.md (aarch64_sqmovun<mode>_insn_le): Delete.
17465 (aarch64_sqmovun<mode>_insn_be): Delete.
17466 (aarch64_sqmovun<mode><vczle><vczbe>): New define_insn.
17467 (aarch64_sqmovun<mode>): Delete expander.
17469 2023-05-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17472 * config/aarch64/aarch64-simd.md (aarch64_<PERMUTE:perm_insn><mode>):
17474 (aarch64_<PERMUTE:perm_insn><mode><vczle><vczbe>): ... This.
17475 (aarch64_rev<REVERSE:rev_op><mode>): Rename to...
17476 (aarch64_rev<REVERSE:rev_op><mode><vczle><vczbe>): ... This.
17478 2023-05-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17481 * config/aarch64/aarch64-simd.md (aarch64_<su_optab>q<addsub><mode>):
17483 (aarch64_<su_optab>q<addsub><mode><vczle><vczbe>): ... This.
17484 (aarch64_<sur>qadd<mode>): Rename to...
17485 (aarch64_<sur>qadd<mode><vczle><vczbe>): ... This.
17487 2023-05-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17489 * config/aarch64/aarch64-simd.md
17490 (aarch64_<sur>q<r>shr<u>n_n<mode>_insn_le): Delete.
17491 (aarch64_<sur>q<r>shr<u>n_n<mode>_insn_be): Delete.
17492 (aarch64_<sur>q<r>shr<u>n_n<mode>_insn<vczle><vczbe>): New define_insn.
17493 (aarch64_<sur>q<r>shr<u>n_n<mode>): Simplify expander.
17495 2023-05-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17498 * config/aarch64/aarch64-simd.md (aarch64_xtn<mode>_insn_le): Delete.
17499 (aarch64_xtn<mode>_insn_be): Likewise.
17500 (trunc<mode><Vnarrowq>2): Rename to...
17501 (trunc<mode><Vnarrowq>2<vczle><vczbe>): ... This.
17502 (aarch64_xtn<mode>): Move under the above. Just emit the truncate RTL.
17503 (aarch64_<su>qmovn<mode>): Likewise.
17504 (aarch64_<su>qmovn<mode><vczle><vczbe>): New define_insn.
17505 (aarch64_<su>qmovn<mode>_insn_le): Delete.
17506 (aarch64_<su>qmovn<mode>_insn_be): Likewise.
17508 2023-05-10 Li Xu <xuli1@eswincomputing.com>
17510 * config/riscv/riscv-vsetvl.cc (gen_vsetvl_pat): For vfmv.f.s/vmv.x.s
17511 intruction replace null avl with (const_int 0).
17513 2023-05-10 Juzhe-Zhong <juzhe.zhong@rivai.ai>
17515 * config/riscv/riscv.cc (riscv_support_vector_misalignment): Fix
17518 2023-05-10 Juzhe-Zhong <juzhe.zhong@rivai.ai>
17521 * config/riscv/riscv-vsetvl.cc (avl_source_has_vsetvl_p): New function.
17522 (source_equal_p): Fix dead loop in vsetvl avl checking.
17524 2023-05-10 Hans-Peter Nilsson <hp@axis.com>
17526 * config/cris/cris.cc (cris_postdbr_cmpelim): Correct mode
17527 of modeadjusted_dccr.
17529 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17531 * config/arm/arm-mve-builtins-base.cc (vmaxaq, vminaq): New.
17532 * config/arm/arm-mve-builtins-base.def (vmaxaq, vminaq): New.
17533 * config/arm/arm-mve-builtins-base.h (vmaxaq, vminaq): New.
17534 * config/arm/arm-mve-builtins.cc
17535 (function_instance::has_inactive_argument): Handle vmaxaq and
17537 * config/arm/arm_mve.h (vminaq): Remove.
17539 (vminaq_m): Remove.
17540 (vmaxaq_m): Remove.
17541 (vminaq_s8): Remove.
17542 (vmaxaq_s8): Remove.
17543 (vminaq_s16): Remove.
17544 (vmaxaq_s16): Remove.
17545 (vminaq_s32): Remove.
17546 (vmaxaq_s32): Remove.
17547 (vminaq_m_s8): Remove.
17548 (vmaxaq_m_s8): Remove.
17549 (vminaq_m_s16): Remove.
17550 (vmaxaq_m_s16): Remove.
17551 (vminaq_m_s32): Remove.
17552 (vmaxaq_m_s32): Remove.
17553 (__arm_vminaq_s8): Remove.
17554 (__arm_vmaxaq_s8): Remove.
17555 (__arm_vminaq_s16): Remove.
17556 (__arm_vmaxaq_s16): Remove.
17557 (__arm_vminaq_s32): Remove.
17558 (__arm_vmaxaq_s32): Remove.
17559 (__arm_vminaq_m_s8): Remove.
17560 (__arm_vmaxaq_m_s8): Remove.
17561 (__arm_vminaq_m_s16): Remove.
17562 (__arm_vmaxaq_m_s16): Remove.
17563 (__arm_vminaq_m_s32): Remove.
17564 (__arm_vmaxaq_m_s32): Remove.
17565 (__arm_vminaq): Remove.
17566 (__arm_vmaxaq): Remove.
17567 (__arm_vminaq_m): Remove.
17568 (__arm_vmaxaq_m): Remove.
17570 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17572 * config/arm/iterators.md (MVE_VMAXAVMINAQ, MVE_VMAXAVMINAQ_M):
17574 (mve_insn): Add vmaxa, vmina.
17575 (supf): Add VMAXAQ_S, VMAXAQ_M_S, VMINAQ_S, VMINAQ_M_S.
17576 * config/arm/mve.md (mve_vmaxaq_s<mode>, mve_vminaq_s<mode>):
17578 (@mve_<mve_insn>q_<supf><mode>): ... this.
17579 (mve_vmaxaq_m_s<mode>, mve_vminaq_m_s<mode>): Merge into ...
17580 (@mve_<mve_insn>q_m_<supf><mode>): ... this.
17582 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17584 * config/arm/arm-mve-builtins-shapes.cc (binary_maxamina): New.
17585 * config/arm/arm-mve-builtins-shapes.h (binary_maxamina): New.
17587 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17589 * config/arm/arm-mve-builtins-base.cc (vmaxnmaq, vminnmaq): New.
17590 * config/arm/arm-mve-builtins-base.def (vmaxnmaq, vminnmaq): New.
17591 * config/arm/arm-mve-builtins-base.h (vmaxnmaq, vminnmaq): New.
17592 * config/arm/arm-mve-builtins.cc
17593 (function_instance::has_inactive_argument): Handle vmaxnmaq and
17595 * config/arm/arm_mve.h (vminnmaq): Remove.
17596 (vmaxnmaq): Remove.
17597 (vmaxnmaq_m): Remove.
17598 (vminnmaq_m): Remove.
17599 (vminnmaq_f16): Remove.
17600 (vmaxnmaq_f16): Remove.
17601 (vminnmaq_f32): Remove.
17602 (vmaxnmaq_f32): Remove.
17603 (vmaxnmaq_m_f16): Remove.
17604 (vminnmaq_m_f16): Remove.
17605 (vmaxnmaq_m_f32): Remove.
17606 (vminnmaq_m_f32): Remove.
17607 (__arm_vminnmaq_f16): Remove.
17608 (__arm_vmaxnmaq_f16): Remove.
17609 (__arm_vminnmaq_f32): Remove.
17610 (__arm_vmaxnmaq_f32): Remove.
17611 (__arm_vmaxnmaq_m_f16): Remove.
17612 (__arm_vminnmaq_m_f16): Remove.
17613 (__arm_vmaxnmaq_m_f32): Remove.
17614 (__arm_vminnmaq_m_f32): Remove.
17615 (__arm_vminnmaq): Remove.
17616 (__arm_vmaxnmaq): Remove.
17617 (__arm_vmaxnmaq_m): Remove.
17618 (__arm_vminnmaq_m): Remove.
17620 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17622 * config/arm/iterators.md (MVE_VMAXNMA_VMINNMAQ)
17623 (MVE_VMAXNMA_VMINNMAQ_M): New.
17624 (mve_insn): Add vmaxnma, vminnma.
17625 * config/arm/mve.md (mve_vmaxnmaq_f<mode>, mve_vminnmaq_f<mode>):
17627 (@mve_<mve_insn>q_f<mode>): ... this.
17628 (mve_vmaxnmaq_m_f<mode>, mve_vminnmaq_m_f<mode>): Merge into ...
17629 (@mve_<mve_insn>q_m_f<mode>): ... this.
17631 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17633 * config/arm/arm-mve-builtins-base.cc (FUNCTION_PRED_P_F): New.
17634 (vmaxnmavq, vmaxnmvq, vminnmavq, vminnmvq): New.
17635 * config/arm/arm-mve-builtins-base.def (vmaxnmavq, vmaxnmvq)
17636 (vminnmavq, vminnmvq): New.
17637 * config/arm/arm-mve-builtins-base.h (vmaxnmavq, vmaxnmvq)
17638 (vminnmavq, vminnmvq): New.
17639 * config/arm/arm_mve.h (vminnmvq): Remove.
17640 (vminnmavq): Remove.
17641 (vmaxnmvq): Remove.
17642 (vmaxnmavq): Remove.
17643 (vmaxnmavq_p): Remove.
17644 (vmaxnmvq_p): Remove.
17645 (vminnmavq_p): Remove.
17646 (vminnmvq_p): Remove.
17647 (vminnmvq_f16): Remove.
17648 (vminnmavq_f16): Remove.
17649 (vmaxnmvq_f16): Remove.
17650 (vmaxnmavq_f16): Remove.
17651 (vminnmvq_f32): Remove.
17652 (vminnmavq_f32): Remove.
17653 (vmaxnmvq_f32): Remove.
17654 (vmaxnmavq_f32): Remove.
17655 (vmaxnmavq_p_f16): Remove.
17656 (vmaxnmvq_p_f16): Remove.
17657 (vminnmavq_p_f16): Remove.
17658 (vminnmvq_p_f16): Remove.
17659 (vmaxnmavq_p_f32): Remove.
17660 (vmaxnmvq_p_f32): Remove.
17661 (vminnmavq_p_f32): Remove.
17662 (vminnmvq_p_f32): Remove.
17663 (__arm_vminnmvq_f16): Remove.
17664 (__arm_vminnmavq_f16): Remove.
17665 (__arm_vmaxnmvq_f16): Remove.
17666 (__arm_vmaxnmavq_f16): Remove.
17667 (__arm_vminnmvq_f32): Remove.
17668 (__arm_vminnmavq_f32): Remove.
17669 (__arm_vmaxnmvq_f32): Remove.
17670 (__arm_vmaxnmavq_f32): Remove.
17671 (__arm_vmaxnmavq_p_f16): Remove.
17672 (__arm_vmaxnmvq_p_f16): Remove.
17673 (__arm_vminnmavq_p_f16): Remove.
17674 (__arm_vminnmvq_p_f16): Remove.
17675 (__arm_vmaxnmavq_p_f32): Remove.
17676 (__arm_vmaxnmvq_p_f32): Remove.
17677 (__arm_vminnmavq_p_f32): Remove.
17678 (__arm_vminnmvq_p_f32): Remove.
17679 (__arm_vminnmvq): Remove.
17680 (__arm_vminnmavq): Remove.
17681 (__arm_vmaxnmvq): Remove.
17682 (__arm_vmaxnmavq): Remove.
17683 (__arm_vmaxnmavq_p): Remove.
17684 (__arm_vmaxnmvq_p): Remove.
17685 (__arm_vminnmavq_p): Remove.
17686 (__arm_vminnmvq_p): Remove.
17687 (__arm_vmaxnmavq_m): Remove.
17688 (__arm_vmaxnmvq_m): Remove.
17690 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17692 * config/arm/arm-mve-builtins-functions.h
17693 (unspec_mve_function_exact_insn_pred_p): Use code_for_mve_q_p_f.
17695 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17697 * config/arm/iterators.md (MVE_VMAXNMxV_MINNMxVQ)
17698 (MVE_VMAXNMxV_MINNMxVQ_P): New.
17699 (mve_insn): Add vmaxnmav, vmaxnmv, vminnmav, vminnmv.
17700 * config/arm/mve.md (mve_vmaxnmavq_f<mode>, mve_vmaxnmvq_f<mode>)
17701 (mve_vminnmavq_f<mode>, mve_vminnmvq_f<mode>): Merge into ...
17702 (@mve_<mve_insn>q_f<mode>): ... this.
17703 (mve_vmaxnmavq_p_f<mode>, mve_vmaxnmvq_p_f<mode>)
17704 (mve_vminnmavq_p_f<mode>, mve_vminnmvq_p_f<mode>): Merge into ...
17705 (@mve_<mve_insn>q_p_f<mode>): ... this.
17707 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17709 * config/arm/arm-mve-builtins-base.cc (vmaxnmq, vminnmq): New.
17710 * config/arm/arm-mve-builtins-base.def (vmaxnmq, vminnmq): New.
17711 * config/arm/arm-mve-builtins-base.h (vmaxnmq, vminnmq): New.
17712 * config/arm/arm_mve.h (vminnmq): Remove.
17714 (vmaxnmq_m): Remove.
17715 (vminnmq_m): Remove.
17716 (vminnmq_x): Remove.
17717 (vmaxnmq_x): Remove.
17718 (vminnmq_f16): Remove.
17719 (vmaxnmq_f16): Remove.
17720 (vminnmq_f32): Remove.
17721 (vmaxnmq_f32): Remove.
17722 (vmaxnmq_m_f32): Remove.
17723 (vmaxnmq_m_f16): Remove.
17724 (vminnmq_m_f32): Remove.
17725 (vminnmq_m_f16): Remove.
17726 (vminnmq_x_f16): Remove.
17727 (vminnmq_x_f32): Remove.
17728 (vmaxnmq_x_f16): Remove.
17729 (vmaxnmq_x_f32): Remove.
17730 (__arm_vminnmq_f16): Remove.
17731 (__arm_vmaxnmq_f16): Remove.
17732 (__arm_vminnmq_f32): Remove.
17733 (__arm_vmaxnmq_f32): Remove.
17734 (__arm_vmaxnmq_m_f32): Remove.
17735 (__arm_vmaxnmq_m_f16): Remove.
17736 (__arm_vminnmq_m_f32): Remove.
17737 (__arm_vminnmq_m_f16): Remove.
17738 (__arm_vminnmq_x_f16): Remove.
17739 (__arm_vminnmq_x_f32): Remove.
17740 (__arm_vmaxnmq_x_f16): Remove.
17741 (__arm_vmaxnmq_x_f32): Remove.
17742 (__arm_vminnmq): Remove.
17743 (__arm_vmaxnmq): Remove.
17744 (__arm_vmaxnmq_m): Remove.
17745 (__arm_vminnmq_m): Remove.
17746 (__arm_vminnmq_x): Remove.
17747 (__arm_vmaxnmq_x): Remove.
17749 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17751 * config/arm/iterators.md (MAX_MIN_F): New.
17752 (MVE_FP_M_BINARY): Add VMAXNMQ_M_F, VMINNMQ_M_F.
17753 (mve_insn): Add vmaxnm, vminnm.
17754 (max_min_f_str): New.
17755 * config/arm/mve.md (mve_vmaxnmq_f<mode>, mve_vminnmq_f<mode>):
17757 (@mve_<max_min_f_str>q_f<mode>): ... this.
17758 (mve_vmaxnmq_m_f<mode>, mve_vminnmq_m_f<mode>): Merge into ...
17759 (@mve_<mve_insn>q_m_f<mode>): ... this.
17761 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17763 * config/arm/vec-common.md (smin<mode>3): Use VDQWH iterator.
17764 (smax<mode>3): Likewise.
17766 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17768 * config/arm/arm-mve-builtins-base.cc (FUNCTION_PRED_P_S_U)
17769 (FUNCTION_PRED_P_S): New.
17770 (vmaxavq, vminavq, vmaxvq, vminvq): New.
17771 * config/arm/arm-mve-builtins-base.def (vmaxavq, vminavq, vmaxvq)
17773 * config/arm/arm-mve-builtins-base.h (vmaxavq, vminavq, vmaxvq)
17775 * config/arm/arm_mve.h (vminvq): Remove.
17777 (vminvq_p): Remove.
17778 (vmaxvq_p): Remove.
17779 (vminvq_u8): Remove.
17780 (vmaxvq_u8): Remove.
17781 (vminvq_s8): Remove.
17782 (vmaxvq_s8): Remove.
17783 (vminvq_u16): Remove.
17784 (vmaxvq_u16): Remove.
17785 (vminvq_s16): Remove.
17786 (vmaxvq_s16): Remove.
17787 (vminvq_u32): Remove.
17788 (vmaxvq_u32): Remove.
17789 (vminvq_s32): Remove.
17790 (vmaxvq_s32): Remove.
17791 (vminvq_p_u8): Remove.
17792 (vmaxvq_p_u8): Remove.
17793 (vminvq_p_s8): Remove.
17794 (vmaxvq_p_s8): Remove.
17795 (vminvq_p_u16): Remove.
17796 (vmaxvq_p_u16): Remove.
17797 (vminvq_p_s16): Remove.
17798 (vmaxvq_p_s16): Remove.
17799 (vminvq_p_u32): Remove.
17800 (vmaxvq_p_u32): Remove.
17801 (vminvq_p_s32): Remove.
17802 (vmaxvq_p_s32): Remove.
17803 (__arm_vminvq_u8): Remove.
17804 (__arm_vmaxvq_u8): Remove.
17805 (__arm_vminvq_s8): Remove.
17806 (__arm_vmaxvq_s8): Remove.
17807 (__arm_vminvq_u16): Remove.
17808 (__arm_vmaxvq_u16): Remove.
17809 (__arm_vminvq_s16): Remove.
17810 (__arm_vmaxvq_s16): Remove.
17811 (__arm_vminvq_u32): Remove.
17812 (__arm_vmaxvq_u32): Remove.
17813 (__arm_vminvq_s32): Remove.
17814 (__arm_vmaxvq_s32): Remove.
17815 (__arm_vminvq_p_u8): Remove.
17816 (__arm_vmaxvq_p_u8): Remove.
17817 (__arm_vminvq_p_s8): Remove.
17818 (__arm_vmaxvq_p_s8): Remove.
17819 (__arm_vminvq_p_u16): Remove.
17820 (__arm_vmaxvq_p_u16): Remove.
17821 (__arm_vminvq_p_s16): Remove.
17822 (__arm_vmaxvq_p_s16): Remove.
17823 (__arm_vminvq_p_u32): Remove.
17824 (__arm_vmaxvq_p_u32): Remove.
17825 (__arm_vminvq_p_s32): Remove.
17826 (__arm_vmaxvq_p_s32): Remove.
17827 (__arm_vminvq): Remove.
17828 (__arm_vmaxvq): Remove.
17829 (__arm_vminvq_p): Remove.
17830 (__arm_vmaxvq_p): Remove.
17833 (vminavq_p): Remove.
17834 (vmaxavq_p): Remove.
17835 (vminavq_s8): Remove.
17836 (vmaxavq_s8): Remove.
17837 (vminavq_s16): Remove.
17838 (vmaxavq_s16): Remove.
17839 (vminavq_s32): Remove.
17840 (vmaxavq_s32): Remove.
17841 (vminavq_p_s8): Remove.
17842 (vmaxavq_p_s8): Remove.
17843 (vminavq_p_s16): Remove.
17844 (vmaxavq_p_s16): Remove.
17845 (vminavq_p_s32): Remove.
17846 (vmaxavq_p_s32): Remove.
17847 (__arm_vminavq_s8): Remove.
17848 (__arm_vmaxavq_s8): Remove.
17849 (__arm_vminavq_s16): Remove.
17850 (__arm_vmaxavq_s16): Remove.
17851 (__arm_vminavq_s32): Remove.
17852 (__arm_vmaxavq_s32): Remove.
17853 (__arm_vminavq_p_s8): Remove.
17854 (__arm_vmaxavq_p_s8): Remove.
17855 (__arm_vminavq_p_s16): Remove.
17856 (__arm_vmaxavq_p_s16): Remove.
17857 (__arm_vminavq_p_s32): Remove.
17858 (__arm_vmaxavq_p_s32): Remove.
17859 (__arm_vminavq): Remove.
17860 (__arm_vmaxavq): Remove.
17861 (__arm_vminavq_p): Remove.
17862 (__arm_vmaxavq_p): Remove.
17864 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17866 * config/arm/iterators.md (MVE_VMAXVQ_VMINVQ, MVE_VMAXVQ_VMINVQ_P): New.
17867 (mve_insn): Add vmaxav, vmaxv, vminav, vminv.
17868 (supf): Add VMAXAVQ_S, VMAXAVQ_P_S, VMINAVQ_S, VMINAVQ_P_S.
17869 * config/arm/mve.md (mve_vmaxavq_s<mode>, mve_vmaxvq_<supf><mode>)
17870 (mve_vminavq_s<mode>, mve_vminvq_<supf><mode>): Merge into ...
17871 (@mve_<mve_insn>q_<supf><mode>): ... this.
17872 (mve_vmaxavq_p_s<mode>, mve_vmaxvq_p_<supf><mode>)
17873 (mve_vminavq_p_s<mode>, mve_vminvq_p_<supf><mode>): Merge into ...
17874 (@mve_<mve_insn>q_p_<supf><mode>): ... this.
17876 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17878 * config/arm/arm-mve-builtins-functions.h (class
17879 unspec_mve_function_exact_insn_pred_p): New.
17881 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17883 * config/arm/arm-mve-builtins-shapes.cc (binary_maxavminav): New.
17884 * config/arm/arm-mve-builtins-shapes.h (binary_maxavminav): New.
17886 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17888 * config/arm/arm-mve-builtins-shapes.cc (binary_maxvminv): New.
17889 * config/arm/arm-mve-builtins-shapes.h (binary_maxvminv): New.
17891 2023-05-09 Richard Sandiford <richard.sandiford@arm.com>
17893 * config/aarch64/aarch64-protos.h (aarch64_adjust_reg_alloc_order):
17895 * config/aarch64/aarch64.h (REG_ALLOC_ORDER): Define.
17896 (ADJUST_REG_ALLOC_ORDER): Likewise.
17897 * config/aarch64/aarch64.cc (aarch64_adjust_reg_alloc_order): New
17899 * config/aarch64/aarch64-sve.md (*vcond_mask_<mode><vpred>): Use
17900 Upa rather than Upl for unpredicated movprfx alternatives.
17902 2023-05-09 Jeff Law <jlaw@ventanamicro>
17904 * config/h8300/testcompare.md: Add peephole2 which uses a memory
17905 load to set flags, thus eliminating a compare against zero.
17907 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17909 * config/arm/arm-mve-builtins-base.cc (vshllbq, vshlltq): New.
17910 * config/arm/arm-mve-builtins-base.def (vshllbq, vshlltq): New.
17911 * config/arm/arm-mve-builtins-base.h (vshllbq, vshlltq): New.
17912 * config/arm/arm_mve.h (vshlltq): Remove.
17914 (vshllbq_m): Remove.
17915 (vshlltq_m): Remove.
17916 (vshllbq_x): Remove.
17917 (vshlltq_x): Remove.
17918 (vshlltq_n_u8): Remove.
17919 (vshllbq_n_u8): Remove.
17920 (vshlltq_n_s8): Remove.
17921 (vshllbq_n_s8): Remove.
17922 (vshlltq_n_u16): Remove.
17923 (vshllbq_n_u16): Remove.
17924 (vshlltq_n_s16): Remove.
17925 (vshllbq_n_s16): Remove.
17926 (vshllbq_m_n_s8): Remove.
17927 (vshllbq_m_n_s16): Remove.
17928 (vshllbq_m_n_u8): Remove.
17929 (vshllbq_m_n_u16): Remove.
17930 (vshlltq_m_n_s8): Remove.
17931 (vshlltq_m_n_s16): Remove.
17932 (vshlltq_m_n_u8): Remove.
17933 (vshlltq_m_n_u16): Remove.
17934 (vshllbq_x_n_s8): Remove.
17935 (vshllbq_x_n_s16): Remove.
17936 (vshllbq_x_n_u8): Remove.
17937 (vshllbq_x_n_u16): Remove.
17938 (vshlltq_x_n_s8): Remove.
17939 (vshlltq_x_n_s16): Remove.
17940 (vshlltq_x_n_u8): Remove.
17941 (vshlltq_x_n_u16): Remove.
17942 (__arm_vshlltq_n_u8): Remove.
17943 (__arm_vshllbq_n_u8): Remove.
17944 (__arm_vshlltq_n_s8): Remove.
17945 (__arm_vshllbq_n_s8): Remove.
17946 (__arm_vshlltq_n_u16): Remove.
17947 (__arm_vshllbq_n_u16): Remove.
17948 (__arm_vshlltq_n_s16): Remove.
17949 (__arm_vshllbq_n_s16): Remove.
17950 (__arm_vshllbq_m_n_s8): Remove.
17951 (__arm_vshllbq_m_n_s16): Remove.
17952 (__arm_vshllbq_m_n_u8): Remove.
17953 (__arm_vshllbq_m_n_u16): Remove.
17954 (__arm_vshlltq_m_n_s8): Remove.
17955 (__arm_vshlltq_m_n_s16): Remove.
17956 (__arm_vshlltq_m_n_u8): Remove.
17957 (__arm_vshlltq_m_n_u16): Remove.
17958 (__arm_vshllbq_x_n_s8): Remove.
17959 (__arm_vshllbq_x_n_s16): Remove.
17960 (__arm_vshllbq_x_n_u8): Remove.
17961 (__arm_vshllbq_x_n_u16): Remove.
17962 (__arm_vshlltq_x_n_s8): Remove.
17963 (__arm_vshlltq_x_n_s16): Remove.
17964 (__arm_vshlltq_x_n_u8): Remove.
17965 (__arm_vshlltq_x_n_u16): Remove.
17966 (__arm_vshlltq): Remove.
17967 (__arm_vshllbq): Remove.
17968 (__arm_vshllbq_m): Remove.
17969 (__arm_vshlltq_m): Remove.
17970 (__arm_vshllbq_x): Remove.
17971 (__arm_vshlltq_x): Remove.
17973 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17975 * config/arm/iterators.md (mve_insn): Add vshllb, vshllt.
17976 (VSHLLBQ_N, VSHLLTQ_N): Remove.
17978 (VSHLLBQ_M_N, VSHLLTQ_M_N): Remove.
17979 (VSHLLxQ_M_N): New.
17980 * config/arm/mve.md (mve_vshllbq_n_<supf><mode>)
17981 (mve_vshlltq_n_<supf><mode>): Merge into ...
17982 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
17983 (mve_vshllbq_m_n_<supf><mode>, mve_vshlltq_m_n_<supf><mode>):
17985 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
17987 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17989 * config/arm/arm-mve-builtins-shapes.cc (binary_widen_n): New.
17990 * config/arm/arm-mve-builtins-shapes.h (binary_widen_n): New.
17992 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17994 * config/arm/arm-mve-builtins-base.cc (vmovnbq, vmovntq, vqmovnbq)
17995 (vqmovntq, vqmovunbq, vqmovuntq): New.
17996 * config/arm/arm-mve-builtins-base.def (vmovnbq, vmovntq)
17997 (vqmovnbq, vqmovntq, vqmovunbq, vqmovuntq): New.
17998 * config/arm/arm-mve-builtins-base.h (vmovnbq, vmovntq, vqmovnbq)
17999 (vqmovntq, vqmovunbq, vqmovuntq): New.
18000 * config/arm/arm-mve-builtins.cc
18001 (function_instance::has_inactive_argument): Handle vmovnbq,
18002 vmovntq, vqmovnbq, vqmovntq, vqmovunbq, vqmovuntq.
18003 * config/arm/arm_mve.h (vqmovntq): Remove.
18004 (vqmovnbq): Remove.
18005 (vqmovnbq_m): Remove.
18006 (vqmovntq_m): Remove.
18007 (vqmovntq_u16): Remove.
18008 (vqmovnbq_u16): Remove.
18009 (vqmovntq_s16): Remove.
18010 (vqmovnbq_s16): Remove.
18011 (vqmovntq_u32): Remove.
18012 (vqmovnbq_u32): Remove.
18013 (vqmovntq_s32): Remove.
18014 (vqmovnbq_s32): Remove.
18015 (vqmovnbq_m_s16): Remove.
18016 (vqmovntq_m_s16): Remove.
18017 (vqmovnbq_m_u16): Remove.
18018 (vqmovntq_m_u16): Remove.
18019 (vqmovnbq_m_s32): Remove.
18020 (vqmovntq_m_s32): Remove.
18021 (vqmovnbq_m_u32): Remove.
18022 (vqmovntq_m_u32): Remove.
18023 (__arm_vqmovntq_u16): Remove.
18024 (__arm_vqmovnbq_u16): Remove.
18025 (__arm_vqmovntq_s16): Remove.
18026 (__arm_vqmovnbq_s16): Remove.
18027 (__arm_vqmovntq_u32): Remove.
18028 (__arm_vqmovnbq_u32): Remove.
18029 (__arm_vqmovntq_s32): Remove.
18030 (__arm_vqmovnbq_s32): Remove.
18031 (__arm_vqmovnbq_m_s16): Remove.
18032 (__arm_vqmovntq_m_s16): Remove.
18033 (__arm_vqmovnbq_m_u16): Remove.
18034 (__arm_vqmovntq_m_u16): Remove.
18035 (__arm_vqmovnbq_m_s32): Remove.
18036 (__arm_vqmovntq_m_s32): Remove.
18037 (__arm_vqmovnbq_m_u32): Remove.
18038 (__arm_vqmovntq_m_u32): Remove.
18039 (__arm_vqmovntq): Remove.
18040 (__arm_vqmovnbq): Remove.
18041 (__arm_vqmovnbq_m): Remove.
18042 (__arm_vqmovntq_m): Remove.
18045 (vmovnbq_m): Remove.
18046 (vmovntq_m): Remove.
18047 (vmovntq_u16): Remove.
18048 (vmovnbq_u16): Remove.
18049 (vmovntq_s16): Remove.
18050 (vmovnbq_s16): Remove.
18051 (vmovntq_u32): Remove.
18052 (vmovnbq_u32): Remove.
18053 (vmovntq_s32): Remove.
18054 (vmovnbq_s32): Remove.
18055 (vmovnbq_m_s16): Remove.
18056 (vmovntq_m_s16): Remove.
18057 (vmovnbq_m_u16): Remove.
18058 (vmovntq_m_u16): Remove.
18059 (vmovnbq_m_s32): Remove.
18060 (vmovntq_m_s32): Remove.
18061 (vmovnbq_m_u32): Remove.
18062 (vmovntq_m_u32): Remove.
18063 (__arm_vmovntq_u16): Remove.
18064 (__arm_vmovnbq_u16): Remove.
18065 (__arm_vmovntq_s16): Remove.
18066 (__arm_vmovnbq_s16): Remove.
18067 (__arm_vmovntq_u32): Remove.
18068 (__arm_vmovnbq_u32): Remove.
18069 (__arm_vmovntq_s32): Remove.
18070 (__arm_vmovnbq_s32): Remove.
18071 (__arm_vmovnbq_m_s16): Remove.
18072 (__arm_vmovntq_m_s16): Remove.
18073 (__arm_vmovnbq_m_u16): Remove.
18074 (__arm_vmovntq_m_u16): Remove.
18075 (__arm_vmovnbq_m_s32): Remove.
18076 (__arm_vmovntq_m_s32): Remove.
18077 (__arm_vmovnbq_m_u32): Remove.
18078 (__arm_vmovntq_m_u32): Remove.
18079 (__arm_vmovntq): Remove.
18080 (__arm_vmovnbq): Remove.
18081 (__arm_vmovnbq_m): Remove.
18082 (__arm_vmovntq_m): Remove.
18083 (vqmovuntq): Remove.
18084 (vqmovunbq): Remove.
18085 (vqmovunbq_m): Remove.
18086 (vqmovuntq_m): Remove.
18087 (vqmovuntq_s16): Remove.
18088 (vqmovunbq_s16): Remove.
18089 (vqmovuntq_s32): Remove.
18090 (vqmovunbq_s32): Remove.
18091 (vqmovunbq_m_s16): Remove.
18092 (vqmovuntq_m_s16): Remove.
18093 (vqmovunbq_m_s32): Remove.
18094 (vqmovuntq_m_s32): Remove.
18095 (__arm_vqmovuntq_s16): Remove.
18096 (__arm_vqmovunbq_s16): Remove.
18097 (__arm_vqmovuntq_s32): Remove.
18098 (__arm_vqmovunbq_s32): Remove.
18099 (__arm_vqmovunbq_m_s16): Remove.
18100 (__arm_vqmovuntq_m_s16): Remove.
18101 (__arm_vqmovunbq_m_s32): Remove.
18102 (__arm_vqmovuntq_m_s32): Remove.
18103 (__arm_vqmovuntq): Remove.
18104 (__arm_vqmovunbq): Remove.
18105 (__arm_vqmovunbq_m): Remove.
18106 (__arm_vqmovuntq_m): Remove.
18108 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
18110 * config/arm/iterators.md (MVE_MOVN, MVE_MOVN_M): New.
18111 (mve_insn): Add vmovnb, vmovnt, vqmovnb, vqmovnt, vqmovunb,
18114 (supf): Add VQMOVUNBQ_M_S, VQMOVUNBQ_S, VQMOVUNTQ_M_S,
18116 * config/arm/mve.md (mve_vmovnbq_<supf><mode>)
18117 (mve_vmovntq_<supf><mode>, mve_vqmovnbq_<supf><mode>)
18118 (mve_vqmovntq_<supf><mode>, mve_vqmovunbq_s<mode>)
18119 (mve_vqmovuntq_s<mode>): Merge into ...
18120 (@mve_<mve_insn>q_<supf><mode>): ... this.
18121 (mve_vmovnbq_m_<supf><mode>, mve_vmovntq_m_<supf><mode>)
18122 (mve_vqmovnbq_m_<supf><mode>, mve_vqmovntq_m_<supf><mode>)
18123 (mve_vqmovunbq_m_s<mode>, mve_vqmovuntq_m_s<mode>): Merge into ...
18124 (@mve_<mve_insn>q_m_<supf><mode>): ... this.
18126 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
18128 * config/arm/arm-mve-builtins-shapes.cc (binary_move_narrow): New.
18129 (binary_move_narrow_unsigned): New.
18130 * config/arm/arm-mve-builtins-shapes.h (binary_move_narrow): New.
18131 (binary_move_narrow_unsigned): New.
18133 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
18135 * config/arm/arm-mve-builtins-base.cc (FUNCTION_ONLY_F): New.
18136 (vrndaq, vrndmq, vrndnq, vrndpq, vrndq, vrndxq): New.
18137 * config/arm/arm-mve-builtins-base.def (vrndaq, vrndmq, vrndnq)
18138 (vrndpq, vrndq, vrndxq): New.
18139 * config/arm/arm-mve-builtins-base.h (vrndaq, vrndmq, vrndnq)
18140 (vrndpq, vrndq, vrndxq): New.
18141 * config/arm/arm_mve.h (vrndxq): Remove.
18147 (vrndaq_m): Remove.
18148 (vrndmq_m): Remove.
18149 (vrndnq_m): Remove.
18150 (vrndpq_m): Remove.
18152 (vrndxq_m): Remove.
18154 (vrndnq_x): Remove.
18155 (vrndmq_x): Remove.
18156 (vrndpq_x): Remove.
18157 (vrndaq_x): Remove.
18158 (vrndxq_x): Remove.
18159 (vrndxq_f16): Remove.
18160 (vrndxq_f32): Remove.
18161 (vrndq_f16): Remove.
18162 (vrndq_f32): Remove.
18163 (vrndpq_f16): Remove.
18164 (vrndpq_f32): Remove.
18165 (vrndnq_f16): Remove.
18166 (vrndnq_f32): Remove.
18167 (vrndmq_f16): Remove.
18168 (vrndmq_f32): Remove.
18169 (vrndaq_f16): Remove.
18170 (vrndaq_f32): Remove.
18171 (vrndaq_m_f16): Remove.
18172 (vrndmq_m_f16): Remove.
18173 (vrndnq_m_f16): Remove.
18174 (vrndpq_m_f16): Remove.
18175 (vrndq_m_f16): Remove.
18176 (vrndxq_m_f16): Remove.
18177 (vrndaq_m_f32): Remove.
18178 (vrndmq_m_f32): Remove.
18179 (vrndnq_m_f32): Remove.
18180 (vrndpq_m_f32): Remove.
18181 (vrndq_m_f32): Remove.
18182 (vrndxq_m_f32): Remove.
18183 (vrndq_x_f16): Remove.
18184 (vrndq_x_f32): Remove.
18185 (vrndnq_x_f16): Remove.
18186 (vrndnq_x_f32): Remove.
18187 (vrndmq_x_f16): Remove.
18188 (vrndmq_x_f32): Remove.
18189 (vrndpq_x_f16): Remove.
18190 (vrndpq_x_f32): Remove.
18191 (vrndaq_x_f16): Remove.
18192 (vrndaq_x_f32): Remove.
18193 (vrndxq_x_f16): Remove.
18194 (vrndxq_x_f32): Remove.
18195 (__arm_vrndxq_f16): Remove.
18196 (__arm_vrndxq_f32): Remove.
18197 (__arm_vrndq_f16): Remove.
18198 (__arm_vrndq_f32): Remove.
18199 (__arm_vrndpq_f16): Remove.
18200 (__arm_vrndpq_f32): Remove.
18201 (__arm_vrndnq_f16): Remove.
18202 (__arm_vrndnq_f32): Remove.
18203 (__arm_vrndmq_f16): Remove.
18204 (__arm_vrndmq_f32): Remove.
18205 (__arm_vrndaq_f16): Remove.
18206 (__arm_vrndaq_f32): Remove.
18207 (__arm_vrndaq_m_f16): Remove.
18208 (__arm_vrndmq_m_f16): Remove.
18209 (__arm_vrndnq_m_f16): Remove.
18210 (__arm_vrndpq_m_f16): Remove.
18211 (__arm_vrndq_m_f16): Remove.
18212 (__arm_vrndxq_m_f16): Remove.
18213 (__arm_vrndaq_m_f32): Remove.
18214 (__arm_vrndmq_m_f32): Remove.
18215 (__arm_vrndnq_m_f32): Remove.
18216 (__arm_vrndpq_m_f32): Remove.
18217 (__arm_vrndq_m_f32): Remove.
18218 (__arm_vrndxq_m_f32): Remove.
18219 (__arm_vrndq_x_f16): Remove.
18220 (__arm_vrndq_x_f32): Remove.
18221 (__arm_vrndnq_x_f16): Remove.
18222 (__arm_vrndnq_x_f32): Remove.
18223 (__arm_vrndmq_x_f16): Remove.
18224 (__arm_vrndmq_x_f32): Remove.
18225 (__arm_vrndpq_x_f16): Remove.
18226 (__arm_vrndpq_x_f32): Remove.
18227 (__arm_vrndaq_x_f16): Remove.
18228 (__arm_vrndaq_x_f32): Remove.
18229 (__arm_vrndxq_x_f16): Remove.
18230 (__arm_vrndxq_x_f32): Remove.
18231 (__arm_vrndxq): Remove.
18232 (__arm_vrndq): Remove.
18233 (__arm_vrndpq): Remove.
18234 (__arm_vrndnq): Remove.
18235 (__arm_vrndmq): Remove.
18236 (__arm_vrndaq): Remove.
18237 (__arm_vrndaq_m): Remove.
18238 (__arm_vrndmq_m): Remove.
18239 (__arm_vrndnq_m): Remove.
18240 (__arm_vrndpq_m): Remove.
18241 (__arm_vrndq_m): Remove.
18242 (__arm_vrndxq_m): Remove.
18243 (__arm_vrndq_x): Remove.
18244 (__arm_vrndnq_x): Remove.
18245 (__arm_vrndmq_x): Remove.
18246 (__arm_vrndpq_x): Remove.
18247 (__arm_vrndaq_x): Remove.
18248 (__arm_vrndxq_x): Remove.
18250 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
18252 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITHOUT_N_NO_U_F): New.
18253 (vabsq, vnegq, vclsq, vclzq, vqabsq, vqnegq): New.
18254 * config/arm/arm-mve-builtins-base.def (vabsq, vnegq, vclsq)
18255 (vclzq, vqabsq, vqnegq): New.
18256 * config/arm/arm-mve-builtins-base.h (vabsq, vnegq, vclsq, vclzq)
18257 (vqabsq, vqnegq): New.
18258 * config/arm/arm_mve.h (vabsq): Remove.
18261 (vabsq_f16): Remove.
18262 (vabsq_f32): Remove.
18263 (vabsq_s8): Remove.
18264 (vabsq_s16): Remove.
18265 (vabsq_s32): Remove.
18266 (vabsq_m_s8): Remove.
18267 (vabsq_m_s16): Remove.
18268 (vabsq_m_s32): Remove.
18269 (vabsq_m_f16): Remove.
18270 (vabsq_m_f32): Remove.
18271 (vabsq_x_s8): Remove.
18272 (vabsq_x_s16): Remove.
18273 (vabsq_x_s32): Remove.
18274 (vabsq_x_f16): Remove.
18275 (vabsq_x_f32): Remove.
18276 (__arm_vabsq_s8): Remove.
18277 (__arm_vabsq_s16): Remove.
18278 (__arm_vabsq_s32): Remove.
18279 (__arm_vabsq_m_s8): Remove.
18280 (__arm_vabsq_m_s16): Remove.
18281 (__arm_vabsq_m_s32): Remove.
18282 (__arm_vabsq_x_s8): Remove.
18283 (__arm_vabsq_x_s16): Remove.
18284 (__arm_vabsq_x_s32): Remove.
18285 (__arm_vabsq_f16): Remove.
18286 (__arm_vabsq_f32): Remove.
18287 (__arm_vabsq_m_f16): Remove.
18288 (__arm_vabsq_m_f32): Remove.
18289 (__arm_vabsq_x_f16): Remove.
18290 (__arm_vabsq_x_f32): Remove.
18291 (__arm_vabsq): Remove.
18292 (__arm_vabsq_m): Remove.
18293 (__arm_vabsq_x): Remove.
18297 (vnegq_f16): Remove.
18298 (vnegq_f32): Remove.
18299 (vnegq_s8): Remove.
18300 (vnegq_s16): Remove.
18301 (vnegq_s32): Remove.
18302 (vnegq_m_s8): Remove.
18303 (vnegq_m_s16): Remove.
18304 (vnegq_m_s32): Remove.
18305 (vnegq_m_f16): Remove.
18306 (vnegq_m_f32): Remove.
18307 (vnegq_x_s8): Remove.
18308 (vnegq_x_s16): Remove.
18309 (vnegq_x_s32): Remove.
18310 (vnegq_x_f16): Remove.
18311 (vnegq_x_f32): Remove.
18312 (__arm_vnegq_s8): Remove.
18313 (__arm_vnegq_s16): Remove.
18314 (__arm_vnegq_s32): Remove.
18315 (__arm_vnegq_m_s8): Remove.
18316 (__arm_vnegq_m_s16): Remove.
18317 (__arm_vnegq_m_s32): Remove.
18318 (__arm_vnegq_x_s8): Remove.
18319 (__arm_vnegq_x_s16): Remove.
18320 (__arm_vnegq_x_s32): Remove.
18321 (__arm_vnegq_f16): Remove.
18322 (__arm_vnegq_f32): Remove.
18323 (__arm_vnegq_m_f16): Remove.
18324 (__arm_vnegq_m_f32): Remove.
18325 (__arm_vnegq_x_f16): Remove.
18326 (__arm_vnegq_x_f32): Remove.
18327 (__arm_vnegq): Remove.
18328 (__arm_vnegq_m): Remove.
18329 (__arm_vnegq_x): Remove.
18333 (vclsq_s8): Remove.
18334 (vclsq_s16): Remove.
18335 (vclsq_s32): Remove.
18336 (vclsq_m_s8): Remove.
18337 (vclsq_m_s16): Remove.
18338 (vclsq_m_s32): Remove.
18339 (vclsq_x_s8): Remove.
18340 (vclsq_x_s16): Remove.
18341 (vclsq_x_s32): Remove.
18342 (__arm_vclsq_s8): Remove.
18343 (__arm_vclsq_s16): Remove.
18344 (__arm_vclsq_s32): Remove.
18345 (__arm_vclsq_m_s8): Remove.
18346 (__arm_vclsq_m_s16): Remove.
18347 (__arm_vclsq_m_s32): Remove.
18348 (__arm_vclsq_x_s8): Remove.
18349 (__arm_vclsq_x_s16): Remove.
18350 (__arm_vclsq_x_s32): Remove.
18351 (__arm_vclsq): Remove.
18352 (__arm_vclsq_m): Remove.
18353 (__arm_vclsq_x): Remove.
18357 (vclzq_s8): Remove.
18358 (vclzq_s16): Remove.
18359 (vclzq_s32): Remove.
18360 (vclzq_u8): Remove.
18361 (vclzq_u16): Remove.
18362 (vclzq_u32): Remove.
18363 (vclzq_m_u8): Remove.
18364 (vclzq_m_s8): Remove.
18365 (vclzq_m_u16): Remove.
18366 (vclzq_m_s16): Remove.
18367 (vclzq_m_u32): Remove.
18368 (vclzq_m_s32): Remove.
18369 (vclzq_x_s8): Remove.
18370 (vclzq_x_s16): Remove.
18371 (vclzq_x_s32): Remove.
18372 (vclzq_x_u8): Remove.
18373 (vclzq_x_u16): Remove.
18374 (vclzq_x_u32): Remove.
18375 (__arm_vclzq_s8): Remove.
18376 (__arm_vclzq_s16): Remove.
18377 (__arm_vclzq_s32): Remove.
18378 (__arm_vclzq_u8): Remove.
18379 (__arm_vclzq_u16): Remove.
18380 (__arm_vclzq_u32): Remove.
18381 (__arm_vclzq_m_u8): Remove.
18382 (__arm_vclzq_m_s8): Remove.
18383 (__arm_vclzq_m_u16): Remove.
18384 (__arm_vclzq_m_s16): Remove.
18385 (__arm_vclzq_m_u32): Remove.
18386 (__arm_vclzq_m_s32): Remove.
18387 (__arm_vclzq_x_s8): Remove.
18388 (__arm_vclzq_x_s16): Remove.
18389 (__arm_vclzq_x_s32): Remove.
18390 (__arm_vclzq_x_u8): Remove.
18391 (__arm_vclzq_x_u16): Remove.
18392 (__arm_vclzq_x_u32): Remove.
18393 (__arm_vclzq): Remove.
18394 (__arm_vclzq_m): Remove.
18395 (__arm_vclzq_x): Remove.
18398 (vqnegq_m): Remove.
18399 (vqabsq_m): Remove.
18400 (vqabsq_s8): Remove.
18401 (vqabsq_s16): Remove.
18402 (vqabsq_s32): Remove.
18403 (vqnegq_s8): Remove.
18404 (vqnegq_s16): Remove.
18405 (vqnegq_s32): Remove.
18406 (vqnegq_m_s8): Remove.
18407 (vqabsq_m_s8): Remove.
18408 (vqnegq_m_s16): Remove.
18409 (vqabsq_m_s16): Remove.
18410 (vqnegq_m_s32): Remove.
18411 (vqabsq_m_s32): Remove.
18412 (__arm_vqabsq_s8): Remove.
18413 (__arm_vqabsq_s16): Remove.
18414 (__arm_vqabsq_s32): Remove.
18415 (__arm_vqnegq_s8): Remove.
18416 (__arm_vqnegq_s16): Remove.
18417 (__arm_vqnegq_s32): Remove.
18418 (__arm_vqnegq_m_s8): Remove.
18419 (__arm_vqabsq_m_s8): Remove.
18420 (__arm_vqnegq_m_s16): Remove.
18421 (__arm_vqabsq_m_s16): Remove.
18422 (__arm_vqnegq_m_s32): Remove.
18423 (__arm_vqabsq_m_s32): Remove.
18424 (__arm_vqabsq): Remove.
18425 (__arm_vqnegq): Remove.
18426 (__arm_vqnegq_m): Remove.
18427 (__arm_vqabsq_m): Remove.
18429 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
18431 * config/arm/iterators.md (MVE_INT_M_UNARY, MVE_INT_UNARY)
18432 (MVE_FP_UNARY, MVE_FP_M_UNARY): New.
18433 (mve_insn): Add vabs, vcls, vclz, vneg, vqabs, vqneg, vrnda,
18434 vrndm, vrndn, vrndp, vrnd, vrndx.
18435 (isu): Add VABSQ_M_S, VCLSQ_M_S, VCLZQ_M_S, VCLZQ_M_U, VNEGQ_M_S,
18436 VQABSQ_M_S, VQNEGQ_M_S.
18438 * config/arm/mve.md (mve_vrndq_m_f<mode>, mve_vrndxq_f<mode>)
18439 (mve_vrndq_f<mode>, mve_vrndpq_f<mode>, mve_vrndnq_f<mode>)
18440 (mve_vrndmq_f<mode>, mve_vrndaq_f<mode>): Merge into ...
18441 (@mve_<mve_insn>q_f<mode>): ... this.
18442 (mve_vnegq_f<mode>, mve_vabsq_f<mode>): Merge into ...
18443 (mve_v<absneg_str>q_f<mode>): ... this.
18444 (mve_vnegq_s<mode>, mve_vabsq_s<mode>): Merge into ...
18445 (mve_v<absneg_str>q_s<mode>): ... this.
18446 (mve_vclsq_s<mode>, mve_vqnegq_s<mode>, mve_vqabsq_s<mode>): Merge into ...
18447 (@mve_<mve_insn>q_<supf><mode>): ... this.
18448 (mve_vabsq_m_s<mode>, mve_vclsq_m_s<mode>)
18449 (mve_vclzq_m_<supf><mode>, mve_vnegq_m_s<mode>)
18450 (mve_vqabsq_m_s<mode>, mve_vqnegq_m_s<mode>): Merge into ...
18451 (@mve_<mve_insn>q_m_<supf><mode>): ... this.
18452 (mve_vabsq_m_f<mode>, mve_vnegq_m_f<mode>, mve_vrndaq_m_f<mode>)
18453 (mve_vrndmq_m_f<mode>, mve_vrndnq_m_f<mode>, mve_vrndpq_m_f<mode>)
18454 (mve_vrndxq_m_f<mode>): Merge into ...
18455 (@mve_<mve_insn>q_m_f<mode>): ... this.
18457 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
18459 * config/arm/arm-mve-builtins-shapes.cc (unary): New.
18460 * config/arm/arm-mve-builtins-shapes.h (unary): New.
18462 2023-05-09 Jakub Jelinek <jakub@redhat.com>
18464 * mux-utils.h: Fix comment typo, avoides -> avoids.
18466 2023-05-09 Jakub Jelinek <jakub@redhat.com>
18468 PR tree-optimization/109778
18469 * wide-int.h (wi::lrotate, wi::rrotate): Call wi::lrshift on
18470 wi::zext (x, width) rather than x if width != precision, rather
18471 than using wi::zext (right, width) after the shift.
18472 * tree-ssa-ccp.cc (bit_value_binop): Call wi::ext on the results
18473 of wi::lrotate or wi::rrotate.
18475 2023-05-09 Alexander Monakov <amonakov@ispras.ru>
18477 * genmatch.cc (get_out_file): Make static and rename to ...
18478 (choose_output): ... this. Reimplement. Update all uses ...
18479 (decision_tree::gen): ... here and ...
18482 2023-05-09 Alexander Monakov <amonakov@ispras.ru>
18484 * genmatch.cc (showUsage): Reimplement as ...
18485 (usage): ...this. Adjust all uses.
18486 (main): Print usage when no arguments. Add missing 'return 1'.
18488 2023-05-09 Alexander Monakov <amonakov@ispras.ru>
18490 * genmatch.cc (header_file): Make static.
18491 (emit_func): Rename to...
18492 (fp_decl): ... this. Adjust all uses.
18493 (fp_decl_done): New function. Use it...
18494 (decision_tree::gen): ... here and...
18495 (write_predicate): ... here.
18498 2023-05-09 Richard Sandiford <richard.sandiford@arm.com>
18500 * ira-conflicts.cc (can_use_same_reg_p): Skip over non-matching
18503 2023-05-08 Roger Sayle <roger@nextmovesoftware.com>
18504 Uros Bizjak <ubizjak@gmail.com>
18506 * config/i386/i386.md (any_or_plus): Move definition earlier.
18507 (*insvti_highpart_1): New define_insn_and_split to overwrite
18508 (insv) the highpart of a TImode register/memory.
18510 2023-05-08 Eugene Rozenfeld <erozen@microsoft.com>
18512 * auto-profile.cc (auto_profile): Check todo from early_inline
18513 to see if cleanup_tree_vfg needs to be called.
18514 (early_inline): Return todo from early_inliner.
18516 2023-05-08 Kito Cheng <kito.cheng@sifive.com>
18518 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::get_vector_info):
18520 (pass_vsetvl::get_block_info): New.
18521 (pass_vsetvl::update_vector_info): New.
18522 (pass_vsetvl::simple_vsetvl): Use get_vector_info.
18523 (pass_vsetvl::compute_local_backward_infos): Ditto.
18524 (pass_vsetvl::transfer_before): Ditto.
18525 (pass_vsetvl::transfer_after): Ditto.
18526 (pass_vsetvl::emit_local_forward_vsetvls): Ditto.
18527 (pass_vsetvl::local_eliminate_vsetvl_insn): Ditto.
18528 (pass_vsetvl::cleanup_insns): Ditto.
18529 (pass_vsetvl::compute_local_backward_infos): Use
18530 update_vector_info.
18532 2023-05-08 Jeff Law <jlaw@ventanamicro>
18534 * config/stormy16/stormy16.md (zero_extendhisi2): Fix length.
18536 2023-05-08 Richard Biener <rguenther@suse.de>
18537 Michael Meissner <meissner@linux.ibm.com>
18539 PR middle-end/108623
18540 * tree-core.h (tree_type_common): Bump up precision field to 16 bits.
18541 Align bit fields > 1 bit to at least an 8-bit boundary.
18543 2023-05-08 Andrew Pinski <apinski@marvell.com>
18545 PR tree-optimization/109424
18546 PR tree-optimization/59424
18547 * tree-ssa-phiopt.cc (factor_out_conditional_conversion): Rename to ...
18548 (factor_out_conditional_operation): This and add support for all unary
18550 (pass_phiopt::execute): Update call to factor_out_conditional_conversion
18551 to call factor_out_conditional_operation instead.
18553 2023-05-08 Andrew Pinski <apinski@marvell.com>
18555 * tree-ssa-phiopt.cc (pass_phiopt::execute): Loop
18556 over factor_out_conditional_conversion.
18558 2023-05-08 Andrew Pinski <apinski@marvell.com>
18560 PR tree-optimization/49959
18561 PR tree-optimization/103771
18562 * tree-ssa-phiopt.cc (pass_phiopt::execute): Support
18563 Diamond shapped bb form for factor_out_conditional_conversion.
18565 2023-05-08 Juzhe-Zhong <juzhe.zhong@rivai.ai>
18567 * config/riscv/autovec.md (movmisalign<mode>): New pattern.
18568 * config/riscv/riscv-protos.h (riscv_vector_mask_mode_p): Delete.
18569 (riscv_vector_get_mask_mode): Ditto.
18570 (get_mask_policy_no_pred): Ditto.
18571 (get_tail_policy_no_pred): Ditto.
18572 (get_mask_mode): New function.
18573 * config/riscv/riscv-v.cc (get_mask_policy_no_pred): Delete.
18574 (get_tail_policy_no_pred): Ditto.
18575 (riscv_vector_mask_mode_p): Ditto.
18576 (riscv_vector_get_mask_mode): Ditto.
18577 (get_mask_mode): New function.
18578 * config/riscv/riscv-vector-builtins.cc (use_real_merge_p): Remove
18580 (get_tail_policy_for_pred): Ditto.
18581 * config/riscv/riscv-vector-builtins.h (get_tail_policy_for_pred): Ditto.
18582 (get_mask_policy_for_pred): Ditto
18583 * config/riscv/riscv.cc (riscv_get_mask_mode): Refine codes.
18585 2023-05-08 Kito Cheng <kito.cheng@sifive.com>
18587 * common/config/riscv/riscv-common.cc (riscv_select_multilib_by_abi): New.
18588 (riscv_select_multilib): New.
18589 (riscv_compute_multilib): Extract logic to riscv_select_multilib and
18590 also handle select_by_abi.
18591 * config/riscv/elf.h (RISCV_USE_CUSTOMISED_MULTI_LIB): Change it
18592 to select_by_abi_arch_cmodel from 1.
18593 * config/riscv/linux.h (RISCV_USE_CUSTOMISED_MULTI_LIB): Define.
18594 * config/riscv/riscv-opts.h (enum riscv_multilib_select_kind): New.
18596 2023-05-08 Alexander Monakov <amonakov@ispras.ru>
18598 * Makefile.in: (gimple-match-head.o-warn): Remove.
18599 (GIMPLE_MATCH_PD_SEQ_SRC): Do not depend on
18600 gimple-match-exports.cc.
18601 (gimple-match-auto.h): Only depend on s-gimple-match.
18602 (generic-match-auto.h): Likewise.
18604 2023-05-08 Andrew Pinski <apinski@marvell.com>
18606 PR tree-optimization/109691
18607 * tree-ssa-dce.cc (simple_dce_from_worklist): Add need_eh_cleanup
18609 If the removed statement can throw, have need_eh_cleanup
18610 include the bb of that statement.
18611 * tree-ssa-dce.h (simple_dce_from_worklist): Update declaration.
18612 * tree-ssa-propagate.cc (struct prop_stats_d): Remove
18614 (substitute_and_fold_dom_walker::substitute_and_fold_dom_walker):
18615 Initialize dceworklist instead of stmts_to_remove.
18616 (substitute_and_fold_dom_walker::~substitute_and_fold_dom_walker):
18617 Destore dceworklist instead of stmts_to_remove.
18618 (substitute_and_fold_dom_walker::before_dom_children):
18619 Set dceworklist instead of adding to stmts_to_remove.
18620 (substitute_and_fold_engine::substitute_and_fold):
18621 Call simple_dce_from_worklist instead of poping
18623 Don't update the stat on removal statements.
18625 2023-05-07 Andrew Pinski <apinski@marvell.com>
18628 * config/aarch64/aarch64-builtins.cc (aarch64_simd_switcher::aarch64_simd_switcher):
18629 Change argument type to aarch64_feature_flags.
18630 * config/aarch64/aarch64-protos.h (aarch64_simd_switcher): Change
18631 constructor argument type to aarch64_feature_flags.
18632 Change m_old_asm_isa_flags to be aarch64_feature_flags.
18634 2023-05-07 Jiufu Guo <guojiufu@linux.ibm.com>
18636 * config/rs6000/rs6000.cc (rs6000_emit_set_long_const): Generate
18637 more parallel code if can_create_pseudo_p.
18639 2023-05-07 Roger Sayle <roger@nextmovesoftware.com>
18642 * lower-subreg.cc (resolve_simple_move): Don't emit a clobber
18643 immediately before moving a multi-word register by parts.
18645 2023-05-06 Jeff Law <jlaw@ventanamicro>
18647 * config/riscv/riscv-v.cc (riscv_vector_preferred_simd_mode): Delete.
18649 2023-05-06 Michael Collison <collison@rivosinc.com>
18651 * tree-vect-slp.cc (can_duplicate_and_interleave_p):
18652 Check that GET_MODE_NUNITS is a multiple of 2.
18654 2023-05-06 Michael Collison <collison@rivosinc.com>
18656 * config/riscv/riscv.cc
18657 (riscv_estimated_poly_value): Implement
18658 TARGET_ESTIMATED_POLY_VALUE.
18659 (riscv_preferred_simd_mode): Implement
18660 TARGET_VECTORIZE_PREFERRED_SIMD_MODE.
18661 (riscv_get_mask_mode): Implement TARGET_VECTORIZE_GET_MASK_MODE.
18662 (riscv_empty_mask_is_expensive): Implement
18663 TARGET_VECTORIZE_EMPTY_MASK_IS_EXPENSIVE.
18664 (riscv_vectorize_create_costs): Implement
18665 TARGET_VECTORIZE_CREATE_COSTS.
18666 (riscv_support_vector_misalignment): Implement
18667 TARGET_VECTORIZE_SUPPORT_VECTOR_MISALIGNMENT.
18668 (TARGET_ESTIMATED_POLY_VALUE): Register target macro.
18669 (TARGET_VECTORIZE_GET_MASK_MODE): Ditto.
18670 (TARGET_VECTORIZE_EMPTY_MASK_IS_EXPENSIVE): Ditto.
18671 (TARGET_VECTORIZE_SUPPORT_VECTOR_MISALIGNMENT): Ditto.
18673 2023-05-06 Jeff Law <jlaw@ventanamicro>
18675 * config/riscv/riscv-v.cc (autovec_use_vlmax_p): Remove
18676 duplicate definition.
18678 2023-05-06 Michael Collison <collison@rivosinc.com>
18680 * config/riscv/riscv-v.cc (autovec_use_vlmax_p): New function.
18681 (riscv_vector_preferred_simd_mode): Ditto.
18682 (get_mask_policy_no_pred): Ditto.
18683 (get_tail_policy_no_pred): Ditto.
18684 (riscv_vector_mask_mode_p): Ditto.
18685 (riscv_vector_get_mask_mode): Ditto.
18687 2023-05-06 Michael Collison <collison@rivosinc.com>
18689 * config/riscv/riscv-vector-builtins.cc (get_tail_policy_for_pred):
18690 Remove static declaration to to make externally visible.
18691 (get_mask_policy_for_pred): Ditto.
18692 * config/riscv/riscv-vector-builtins.h (get_tail_policy_for_pred):
18693 New external declaration.
18694 (get_mask_policy_for_pred): Ditto.
18696 2023-05-06 Michael Collison <collison@rivosinc.com>
18698 * config/riscv/riscv-protos.h (riscv_vector_mask_mode_p): New.
18699 (riscv_vector_get_mask_mode): Ditto.
18700 (get_mask_policy_no_pred): Ditto.
18701 (get_tail_policy_no_pred): Ditto.
18703 2023-05-06 Xi Ruoyao <xry111@xry111.site>
18705 * config/loongarch/loongarch.h (struct machine_function): Add
18706 reg_is_wrapped_separately array for register wrapping
18708 * config/loongarch/loongarch.cc
18709 (loongarch_get_separate_components): New function.
18710 (loongarch_components_for_bb): Likewise.
18711 (loongarch_disqualify_components): Likewise.
18712 (loongarch_process_components): Likewise.
18713 (loongarch_emit_prologue_components): Likewise.
18714 (loongarch_emit_epilogue_components): Likewise.
18715 (loongarch_set_handled_components): Likewise.
18716 (TARGET_SHRINK_WRAP_GET_SEPARATE_COMPONENTS): Define.
18717 (TARGET_SHRINK_WRAP_COMPONENTS_FOR_BB): Likewise.
18718 (TARGET_SHRINK_WRAP_DISQUALIFY_COMPONENTS): Likewise.
18719 (TARGET_SHRINK_WRAP_EMIT_PROLOGUE_COMPONENTS): Likewise.
18720 (TARGET_SHRINK_WRAP_EMIT_EPILOGUE_COMPONENTS): Likewise.
18721 (TARGET_SHRINK_WRAP_SET_HANDLED_COMPONENTS): Likewise.
18722 (loongarch_for_each_saved_reg): Skip registers that are wrapped
18725 2023-05-06 Xi Ruoyao <xry111@xry111.site>
18728 * Makefile.in (s-macro_list): Pass -nostdinc to
18731 2023-05-06 Juzhe-Zhong <juzhe.zhong@rivai.ai>
18733 * config/riscv/riscv-protos.h (preferred_simd_mode): New function.
18734 * config/riscv/riscv-v.cc (autovec_use_vlmax_p): Ditto.
18735 (preferred_simd_mode): Ditto.
18736 * config/riscv/riscv.cc (riscv_get_arg_info): Handle RVV type in function arg.
18737 (riscv_convert_vector_bits): Adjust for RVV auto-vectorization.
18738 (riscv_preferred_simd_mode): New function.
18739 (TARGET_VECTORIZE_PREFERRED_SIMD_MODE): New target hook support.
18740 * config/riscv/vector.md: Add autovec.md.
18741 * config/riscv/autovec.md: New file.
18743 2023-05-06 Jakub Jelinek <jakub@redhat.com>
18745 * real.h (dconst_pi): Define.
18746 (dconst_e_ptr): Formatting fix.
18747 (dconst_pi_ptr): Declare.
18748 * real.cc (dconst_pi_ptr): New function.
18749 * gimple-range-op.cc (cfn_sincos::fold_range): Intersect the generic
18750 boundaries range with range computed from sin/cos of the particular
18751 bounds if the argument range is shorter than 2*pi.
18752 (cfn_sincos::op1_range): Take bulps into account when determining
18753 which result ranges are always invalid or behave like known NAN.
18755 2023-05-06 Aldy Hernandez <aldyh@redhat.com>
18757 * gimple-range-cache.cc (sbr_sparse_bitmap::set_bb_range): Do not
18758 pass type to vrange_storage::equal_p.
18759 * value-range-storage.cc (vrange_storage::equal_p): Remove type.
18760 (irange_storage::equal_p): Same.
18761 (frange_storage::equal_p): Same.
18762 * value-range-storage.h (class frange_storage): Same.
18764 2023-05-06 Juzhe-Zhong <juzhe.zhong@rivai.ai>
18767 * config/riscv/riscv-vsetvl.cc (local_eliminate_vsetvl_insn): Remove it.
18768 (pass_vsetvl::local_eliminate_vsetvl_insn): New function.
18770 2023-05-06 liuhongt <hongtao.liu@intel.com>
18772 * combine.cc (maybe_swap_commutative_operands): Canonicalize
18773 vec_merge when mask is constant.
18774 * doc/md.texi: Document vec_merge canonicalization.
18776 2023-05-06 Jakub Jelinek <jakub@redhat.com>
18778 * value-range.h (frange_arithmetic): Declare.
18779 * range-op-float.cc (frange_arithmetic): No longer static.
18780 * gimple-range-op.cc (frange_mpfr_arg1): New function.
18781 (cfn_sqrt::fold_range): Intersect the generic boundaries range
18782 with range computed from sqrt of the particular bounds.
18783 (cfn_sqrt::op1_range): Intersect the generic boundaries range
18784 with range computed from squared particular bounds.
18786 2023-05-06 Jakub Jelinek <jakub@redhat.com>
18788 * Makefile.in (check_p_numbers): Rename to one_to_9999, move
18789 earlier with helper variables also renamed.
18790 (MATCH_SPLUT_SEQ): Use $(wordlist 1,$(NUM_MATCH_SPLITS),$(one_to_9999))
18791 instead of $(shell seq 1 $(NUM_MATCH_SPLITS)).
18792 (check_p_subdirs): Use $(one_to_9999) instead of $(check_p_numbers).
18794 2023-05-06 Hans-Peter Nilsson <hp@axis.com>
18796 * config/cris/cris.md (splitop): Add PLUS.
18797 * config/cris/cris.cc (cris_split_constant): Also handle
18798 PLUS when a split into two insns may be useful.
18800 2023-05-05 Hans-Peter Nilsson <hp@axis.com>
18802 * config/cris/cris.md (movandsplit1): New define_peephole2.
18804 2023-05-05 Hans-Peter Nilsson <hp@axis.com>
18806 * config/cris/cris.md (lsrandsplit1): New define_peephole2.
18808 2023-05-05 Hans-Peter Nilsson <hp@axis.com>
18810 * doc/md.texi (define_peephole2): Document order of scanning.
18812 2023-05-05 Pan Li <pan2.li@intel.com>
18813 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
18815 * config/riscv/vector.md: Allow const as the operand of RVV
18816 indexed load/store.
18818 2023-05-05 Pan Li <pan2.li@intel.com>
18820 * config/riscv/riscv.h (VECTOR_STORE_FLAG_VALUE): Add new macro
18821 consumed by simplify_rtx.
18823 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18825 * config/arm/arm-mve-builtins-base.cc (vrshrq, vshrq): New.
18826 * config/arm/arm-mve-builtins-base.def (vrshrq, vshrq): New.
18827 * config/arm/arm-mve-builtins-base.h (vrshrq, vshrq): New.
18828 * config/arm/arm_mve.h (vshrq): Remove.
18830 (vrshrq_m): Remove.
18832 (vrshrq_x): Remove.
18834 (vshrq_n_s8): Remove.
18835 (vshrq_n_s16): Remove.
18836 (vshrq_n_s32): Remove.
18837 (vshrq_n_u8): Remove.
18838 (vshrq_n_u16): Remove.
18839 (vshrq_n_u32): Remove.
18840 (vrshrq_n_u8): Remove.
18841 (vrshrq_n_s8): Remove.
18842 (vrshrq_n_u16): Remove.
18843 (vrshrq_n_s16): Remove.
18844 (vrshrq_n_u32): Remove.
18845 (vrshrq_n_s32): Remove.
18846 (vrshrq_m_n_s8): Remove.
18847 (vrshrq_m_n_s32): Remove.
18848 (vrshrq_m_n_s16): Remove.
18849 (vrshrq_m_n_u8): Remove.
18850 (vrshrq_m_n_u32): Remove.
18851 (vrshrq_m_n_u16): Remove.
18852 (vshrq_m_n_s8): Remove.
18853 (vshrq_m_n_s32): Remove.
18854 (vshrq_m_n_s16): Remove.
18855 (vshrq_m_n_u8): Remove.
18856 (vshrq_m_n_u32): Remove.
18857 (vshrq_m_n_u16): Remove.
18858 (vrshrq_x_n_s8): Remove.
18859 (vrshrq_x_n_s16): Remove.
18860 (vrshrq_x_n_s32): Remove.
18861 (vrshrq_x_n_u8): Remove.
18862 (vrshrq_x_n_u16): Remove.
18863 (vrshrq_x_n_u32): Remove.
18864 (vshrq_x_n_s8): Remove.
18865 (vshrq_x_n_s16): Remove.
18866 (vshrq_x_n_s32): Remove.
18867 (vshrq_x_n_u8): Remove.
18868 (vshrq_x_n_u16): Remove.
18869 (vshrq_x_n_u32): Remove.
18870 (__arm_vshrq_n_s8): Remove.
18871 (__arm_vshrq_n_s16): Remove.
18872 (__arm_vshrq_n_s32): Remove.
18873 (__arm_vshrq_n_u8): Remove.
18874 (__arm_vshrq_n_u16): Remove.
18875 (__arm_vshrq_n_u32): Remove.
18876 (__arm_vrshrq_n_u8): Remove.
18877 (__arm_vrshrq_n_s8): Remove.
18878 (__arm_vrshrq_n_u16): Remove.
18879 (__arm_vrshrq_n_s16): Remove.
18880 (__arm_vrshrq_n_u32): Remove.
18881 (__arm_vrshrq_n_s32): Remove.
18882 (__arm_vrshrq_m_n_s8): Remove.
18883 (__arm_vrshrq_m_n_s32): Remove.
18884 (__arm_vrshrq_m_n_s16): Remove.
18885 (__arm_vrshrq_m_n_u8): Remove.
18886 (__arm_vrshrq_m_n_u32): Remove.
18887 (__arm_vrshrq_m_n_u16): Remove.
18888 (__arm_vshrq_m_n_s8): Remove.
18889 (__arm_vshrq_m_n_s32): Remove.
18890 (__arm_vshrq_m_n_s16): Remove.
18891 (__arm_vshrq_m_n_u8): Remove.
18892 (__arm_vshrq_m_n_u32): Remove.
18893 (__arm_vshrq_m_n_u16): Remove.
18894 (__arm_vrshrq_x_n_s8): Remove.
18895 (__arm_vrshrq_x_n_s16): Remove.
18896 (__arm_vrshrq_x_n_s32): Remove.
18897 (__arm_vrshrq_x_n_u8): Remove.
18898 (__arm_vrshrq_x_n_u16): Remove.
18899 (__arm_vrshrq_x_n_u32): Remove.
18900 (__arm_vshrq_x_n_s8): Remove.
18901 (__arm_vshrq_x_n_s16): Remove.
18902 (__arm_vshrq_x_n_s32): Remove.
18903 (__arm_vshrq_x_n_u8): Remove.
18904 (__arm_vshrq_x_n_u16): Remove.
18905 (__arm_vshrq_x_n_u32): Remove.
18906 (__arm_vshrq): Remove.
18907 (__arm_vrshrq): Remove.
18908 (__arm_vrshrq_m): Remove.
18909 (__arm_vshrq_m): Remove.
18910 (__arm_vrshrq_x): Remove.
18911 (__arm_vshrq_x): Remove.
18913 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18915 * config/arm/iterators.md (MVE_VSHRQ_M_N, MVE_VSHRQ_N): New.
18916 (mve_insn): Add vrshr, vshr.
18917 * config/arm/mve.md (mve_vshrq_n_<supf><mode>)
18918 (mve_vrshrq_n_<supf><mode>): Merge into ...
18919 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
18920 (mve_vrshrq_m_n_<supf><mode>, mve_vshrq_m_n_<supf><mode>): Merge
18922 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
18924 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18926 * config/arm/arm-mve-builtins-shapes.cc (binary_rshift): New.
18927 * config/arm/arm-mve-builtins-shapes.h (binary_rshift): New.
18929 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18931 * config/arm/arm-mve-builtins-base.cc (FUNCTION_ONLY_N_NO_U_F): New.
18932 (vqshrunbq, vqshruntq, vqrshrunbq, vqrshruntq): New.
18933 * config/arm/arm-mve-builtins-base.def (vqshrunbq, vqshruntq)
18934 (vqrshrunbq, vqrshruntq): New.
18935 * config/arm/arm-mve-builtins-base.h (vqshrunbq, vqshruntq)
18936 (vqrshrunbq, vqrshruntq): New.
18937 * config/arm/arm-mve-builtins.cc
18938 (function_instance::has_inactive_argument): Handle vqshrunbq,
18939 vqshruntq, vqrshrunbq, vqrshruntq.
18940 * config/arm/arm_mve.h (vqrshrunbq): Remove.
18941 (vqrshruntq): Remove.
18942 (vqrshrunbq_m): Remove.
18943 (vqrshruntq_m): Remove.
18944 (vqrshrunbq_n_s16): Remove.
18945 (vqrshrunbq_n_s32): Remove.
18946 (vqrshruntq_n_s16): Remove.
18947 (vqrshruntq_n_s32): Remove.
18948 (vqrshrunbq_m_n_s32): Remove.
18949 (vqrshrunbq_m_n_s16): Remove.
18950 (vqrshruntq_m_n_s32): Remove.
18951 (vqrshruntq_m_n_s16): Remove.
18952 (__arm_vqrshrunbq_n_s16): Remove.
18953 (__arm_vqrshrunbq_n_s32): Remove.
18954 (__arm_vqrshruntq_n_s16): Remove.
18955 (__arm_vqrshruntq_n_s32): Remove.
18956 (__arm_vqrshrunbq_m_n_s32): Remove.
18957 (__arm_vqrshrunbq_m_n_s16): Remove.
18958 (__arm_vqrshruntq_m_n_s32): Remove.
18959 (__arm_vqrshruntq_m_n_s16): Remove.
18960 (__arm_vqrshrunbq): Remove.
18961 (__arm_vqrshruntq): Remove.
18962 (__arm_vqrshrunbq_m): Remove.
18963 (__arm_vqrshruntq_m): Remove.
18964 (vqshrunbq): Remove.
18965 (vqshruntq): Remove.
18966 (vqshrunbq_m): Remove.
18967 (vqshruntq_m): Remove.
18968 (vqshrunbq_n_s16): Remove.
18969 (vqshruntq_n_s16): Remove.
18970 (vqshrunbq_n_s32): Remove.
18971 (vqshruntq_n_s32): Remove.
18972 (vqshrunbq_m_n_s32): Remove.
18973 (vqshrunbq_m_n_s16): Remove.
18974 (vqshruntq_m_n_s32): Remove.
18975 (vqshruntq_m_n_s16): Remove.
18976 (__arm_vqshrunbq_n_s16): Remove.
18977 (__arm_vqshruntq_n_s16): Remove.
18978 (__arm_vqshrunbq_n_s32): Remove.
18979 (__arm_vqshruntq_n_s32): Remove.
18980 (__arm_vqshrunbq_m_n_s32): Remove.
18981 (__arm_vqshrunbq_m_n_s16): Remove.
18982 (__arm_vqshruntq_m_n_s32): Remove.
18983 (__arm_vqshruntq_m_n_s16): Remove.
18984 (__arm_vqshrunbq): Remove.
18985 (__arm_vqshruntq): Remove.
18986 (__arm_vqshrunbq_m): Remove.
18987 (__arm_vqshruntq_m): Remove.
18989 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18991 * config/arm/iterators.md (MVE_SHRN_N): Add VQRSHRUNBQ,
18992 VQRSHRUNTQ, VQSHRUNBQ, VQSHRUNTQ.
18993 (MVE_SHRN_M_N): Likewise.
18994 (mve_insn): Add vqrshrunb, vqrshrunt, vqshrunb, vqshrunt.
18995 (isu): Add VQRSHRUNBQ, VQRSHRUNTQ, VQSHRUNBQ, VQSHRUNTQ.
18997 * config/arm/mve.md (mve_vqrshrunbq_n_s<mode>): Remove.
18998 (mve_vqrshruntq_n_s<mode>): Remove.
18999 (mve_vqshrunbq_n_s<mode>): Remove.
19000 (mve_vqshruntq_n_s<mode>): Remove.
19001 (mve_vqrshrunbq_m_n_s<mode>): Remove.
19002 (mve_vqrshruntq_m_n_s<mode>): Remove.
19003 (mve_vqshrunbq_m_n_s<mode>): Remove.
19004 (mve_vqshruntq_m_n_s<mode>): Remove.
19006 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19008 * config/arm/arm-mve-builtins-shapes.cc
19009 (binary_rshift_narrow_unsigned): New.
19010 * config/arm/arm-mve-builtins-shapes.h
19011 (binary_rshift_narrow_unsigned): New.
19013 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19015 * config/arm/arm-mve-builtins-base.cc (FUNCTION_ONLY_N_NO_F): New.
19016 (vshrnbq, vshrntq, vrshrnbq, vrshrntq, vqshrnbq, vqshrntq)
19017 (vqrshrnbq, vqrshrntq): New.
19018 * config/arm/arm-mve-builtins-base.def (vshrnbq, vshrntq)
19019 (vrshrnbq, vrshrntq, vqshrnbq, vqshrntq, vqrshrnbq, vqrshrntq):
19021 * config/arm/arm-mve-builtins-base.h (vshrnbq, vshrntq, vrshrnbq)
19022 (vrshrntq, vqshrnbq, vqshrntq, vqrshrnbq, vqrshrntq): New.
19023 * config/arm/arm-mve-builtins.cc
19024 (function_instance::has_inactive_argument): Handle vshrnbq,
19025 vshrntq, vrshrnbq, vrshrntq, vqshrnbq, vqshrntq, vqrshrnbq,
19027 * config/arm/arm_mve.h (vshrnbq): Remove.
19029 (vshrnbq_m): Remove.
19030 (vshrntq_m): Remove.
19031 (vshrnbq_n_s16): Remove.
19032 (vshrntq_n_s16): Remove.
19033 (vshrnbq_n_u16): Remove.
19034 (vshrntq_n_u16): Remove.
19035 (vshrnbq_n_s32): Remove.
19036 (vshrntq_n_s32): Remove.
19037 (vshrnbq_n_u32): Remove.
19038 (vshrntq_n_u32): Remove.
19039 (vshrnbq_m_n_s32): Remove.
19040 (vshrnbq_m_n_s16): Remove.
19041 (vshrnbq_m_n_u32): Remove.
19042 (vshrnbq_m_n_u16): Remove.
19043 (vshrntq_m_n_s32): Remove.
19044 (vshrntq_m_n_s16): Remove.
19045 (vshrntq_m_n_u32): Remove.
19046 (vshrntq_m_n_u16): Remove.
19047 (__arm_vshrnbq_n_s16): Remove.
19048 (__arm_vshrntq_n_s16): Remove.
19049 (__arm_vshrnbq_n_u16): Remove.
19050 (__arm_vshrntq_n_u16): Remove.
19051 (__arm_vshrnbq_n_s32): Remove.
19052 (__arm_vshrntq_n_s32): Remove.
19053 (__arm_vshrnbq_n_u32): Remove.
19054 (__arm_vshrntq_n_u32): Remove.
19055 (__arm_vshrnbq_m_n_s32): Remove.
19056 (__arm_vshrnbq_m_n_s16): Remove.
19057 (__arm_vshrnbq_m_n_u32): Remove.
19058 (__arm_vshrnbq_m_n_u16): Remove.
19059 (__arm_vshrntq_m_n_s32): Remove.
19060 (__arm_vshrntq_m_n_s16): Remove.
19061 (__arm_vshrntq_m_n_u32): Remove.
19062 (__arm_vshrntq_m_n_u16): Remove.
19063 (__arm_vshrnbq): Remove.
19064 (__arm_vshrntq): Remove.
19065 (__arm_vshrnbq_m): Remove.
19066 (__arm_vshrntq_m): Remove.
19067 (vrshrnbq): Remove.
19068 (vrshrntq): Remove.
19069 (vrshrnbq_m): Remove.
19070 (vrshrntq_m): Remove.
19071 (vrshrnbq_n_s16): Remove.
19072 (vrshrntq_n_s16): Remove.
19073 (vrshrnbq_n_u16): Remove.
19074 (vrshrntq_n_u16): Remove.
19075 (vrshrnbq_n_s32): Remove.
19076 (vrshrntq_n_s32): Remove.
19077 (vrshrnbq_n_u32): Remove.
19078 (vrshrntq_n_u32): Remove.
19079 (vrshrnbq_m_n_s32): Remove.
19080 (vrshrnbq_m_n_s16): Remove.
19081 (vrshrnbq_m_n_u32): Remove.
19082 (vrshrnbq_m_n_u16): Remove.
19083 (vrshrntq_m_n_s32): Remove.
19084 (vrshrntq_m_n_s16): Remove.
19085 (vrshrntq_m_n_u32): Remove.
19086 (vrshrntq_m_n_u16): Remove.
19087 (__arm_vrshrnbq_n_s16): Remove.
19088 (__arm_vrshrntq_n_s16): Remove.
19089 (__arm_vrshrnbq_n_u16): Remove.
19090 (__arm_vrshrntq_n_u16): Remove.
19091 (__arm_vrshrnbq_n_s32): Remove.
19092 (__arm_vrshrntq_n_s32): Remove.
19093 (__arm_vrshrnbq_n_u32): Remove.
19094 (__arm_vrshrntq_n_u32): Remove.
19095 (__arm_vrshrnbq_m_n_s32): Remove.
19096 (__arm_vrshrnbq_m_n_s16): Remove.
19097 (__arm_vrshrnbq_m_n_u32): Remove.
19098 (__arm_vrshrnbq_m_n_u16): Remove.
19099 (__arm_vrshrntq_m_n_s32): Remove.
19100 (__arm_vrshrntq_m_n_s16): Remove.
19101 (__arm_vrshrntq_m_n_u32): Remove.
19102 (__arm_vrshrntq_m_n_u16): Remove.
19103 (__arm_vrshrnbq): Remove.
19104 (__arm_vrshrntq): Remove.
19105 (__arm_vrshrnbq_m): Remove.
19106 (__arm_vrshrntq_m): Remove.
19107 (vqshrnbq): Remove.
19108 (vqshrntq): Remove.
19109 (vqshrnbq_m): Remove.
19110 (vqshrntq_m): Remove.
19111 (vqshrnbq_n_s16): Remove.
19112 (vqshrntq_n_s16): Remove.
19113 (vqshrnbq_n_u16): Remove.
19114 (vqshrntq_n_u16): Remove.
19115 (vqshrnbq_n_s32): Remove.
19116 (vqshrntq_n_s32): Remove.
19117 (vqshrnbq_n_u32): Remove.
19118 (vqshrntq_n_u32): Remove.
19119 (vqshrnbq_m_n_s32): Remove.
19120 (vqshrnbq_m_n_s16): Remove.
19121 (vqshrnbq_m_n_u32): Remove.
19122 (vqshrnbq_m_n_u16): Remove.
19123 (vqshrntq_m_n_s32): Remove.
19124 (vqshrntq_m_n_s16): Remove.
19125 (vqshrntq_m_n_u32): Remove.
19126 (vqshrntq_m_n_u16): Remove.
19127 (__arm_vqshrnbq_n_s16): Remove.
19128 (__arm_vqshrntq_n_s16): Remove.
19129 (__arm_vqshrnbq_n_u16): Remove.
19130 (__arm_vqshrntq_n_u16): Remove.
19131 (__arm_vqshrnbq_n_s32): Remove.
19132 (__arm_vqshrntq_n_s32): Remove.
19133 (__arm_vqshrnbq_n_u32): Remove.
19134 (__arm_vqshrntq_n_u32): Remove.
19135 (__arm_vqshrnbq_m_n_s32): Remove.
19136 (__arm_vqshrnbq_m_n_s16): Remove.
19137 (__arm_vqshrnbq_m_n_u32): Remove.
19138 (__arm_vqshrnbq_m_n_u16): Remove.
19139 (__arm_vqshrntq_m_n_s32): Remove.
19140 (__arm_vqshrntq_m_n_s16): Remove.
19141 (__arm_vqshrntq_m_n_u32): Remove.
19142 (__arm_vqshrntq_m_n_u16): Remove.
19143 (__arm_vqshrnbq): Remove.
19144 (__arm_vqshrntq): Remove.
19145 (__arm_vqshrnbq_m): Remove.
19146 (__arm_vqshrntq_m): Remove.
19147 (vqrshrnbq): Remove.
19148 (vqrshrntq): Remove.
19149 (vqrshrnbq_m): Remove.
19150 (vqrshrntq_m): Remove.
19151 (vqrshrnbq_n_s16): Remove.
19152 (vqrshrnbq_n_u16): Remove.
19153 (vqrshrnbq_n_s32): Remove.
19154 (vqrshrnbq_n_u32): Remove.
19155 (vqrshrntq_n_s16): Remove.
19156 (vqrshrntq_n_u16): Remove.
19157 (vqrshrntq_n_s32): Remove.
19158 (vqrshrntq_n_u32): Remove.
19159 (vqrshrnbq_m_n_s32): Remove.
19160 (vqrshrnbq_m_n_s16): Remove.
19161 (vqrshrnbq_m_n_u32): Remove.
19162 (vqrshrnbq_m_n_u16): Remove.
19163 (vqrshrntq_m_n_s32): Remove.
19164 (vqrshrntq_m_n_s16): Remove.
19165 (vqrshrntq_m_n_u32): Remove.
19166 (vqrshrntq_m_n_u16): Remove.
19167 (__arm_vqrshrnbq_n_s16): Remove.
19168 (__arm_vqrshrnbq_n_u16): Remove.
19169 (__arm_vqrshrnbq_n_s32): Remove.
19170 (__arm_vqrshrnbq_n_u32): Remove.
19171 (__arm_vqrshrntq_n_s16): Remove.
19172 (__arm_vqrshrntq_n_u16): Remove.
19173 (__arm_vqrshrntq_n_s32): Remove.
19174 (__arm_vqrshrntq_n_u32): Remove.
19175 (__arm_vqrshrnbq_m_n_s32): Remove.
19176 (__arm_vqrshrnbq_m_n_s16): Remove.
19177 (__arm_vqrshrnbq_m_n_u32): Remove.
19178 (__arm_vqrshrnbq_m_n_u16): Remove.
19179 (__arm_vqrshrntq_m_n_s32): Remove.
19180 (__arm_vqrshrntq_m_n_s16): Remove.
19181 (__arm_vqrshrntq_m_n_u32): Remove.
19182 (__arm_vqrshrntq_m_n_u16): Remove.
19183 (__arm_vqrshrnbq): Remove.
19184 (__arm_vqrshrntq): Remove.
19185 (__arm_vqrshrnbq_m): Remove.
19186 (__arm_vqrshrntq_m): Remove.
19188 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19190 * config/arm/iterators.md (MVE_SHRN_N, MVE_SHRN_M_N): New.
19191 (mve_insn): Add vqrshrnb, vqrshrnt, vqshrnb, vqshrnt, vrshrnb,
19192 vrshrnt, vshrnb, vshrnt.
19194 * config/arm/mve.md (mve_vqrshrnbq_n_<supf><mode>)
19195 (mve_vqrshrntq_n_<supf><mode>, mve_vqshrnbq_n_<supf><mode>)
19196 (mve_vqshrntq_n_<supf><mode>, mve_vrshrnbq_n_<supf><mode>)
19197 (mve_vrshrntq_n_<supf><mode>, mve_vshrnbq_n_<supf><mode>)
19198 (mve_vshrntq_n_<supf><mode>): Merge into ...
19199 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
19200 (mve_vqrshrnbq_m_n_<supf><mode>, mve_vqrshrntq_m_n_<supf><mode>)
19201 (mve_vqshrnbq_m_n_<supf><mode>, mve_vqshrntq_m_n_<supf><mode>)
19202 (mve_vrshrnbq_m_n_<supf><mode>, mve_vrshrntq_m_n_<supf><mode>)
19203 (mve_vshrnbq_m_n_<supf><mode>, mve_vshrntq_m_n_<supf><mode>):
19205 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
19207 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19209 * config/arm/arm-mve-builtins-shapes.cc (binary_rshift_narrow):
19211 * config/arm/arm-mve-builtins-shapes.h (binary_rshift_narrow): New.
19213 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19215 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITH_RTX_M_NO_F): New.
19216 (vmaxq, vminq): New.
19217 * config/arm/arm-mve-builtins-base.def (vmaxq, vminq): New.
19218 * config/arm/arm-mve-builtins-base.h (vmaxq, vminq): New.
19219 * config/arm/arm_mve.h (vminq): Remove.
19225 (vminq_u8): Remove.
19226 (vmaxq_u8): Remove.
19227 (vminq_s8): Remove.
19228 (vmaxq_s8): Remove.
19229 (vminq_u16): Remove.
19230 (vmaxq_u16): Remove.
19231 (vminq_s16): Remove.
19232 (vmaxq_s16): Remove.
19233 (vminq_u32): Remove.
19234 (vmaxq_u32): Remove.
19235 (vminq_s32): Remove.
19236 (vmaxq_s32): Remove.
19237 (vmaxq_m_s8): Remove.
19238 (vmaxq_m_s32): Remove.
19239 (vmaxq_m_s16): Remove.
19240 (vmaxq_m_u8): Remove.
19241 (vmaxq_m_u32): Remove.
19242 (vmaxq_m_u16): Remove.
19243 (vminq_m_s8): Remove.
19244 (vminq_m_s32): Remove.
19245 (vminq_m_s16): Remove.
19246 (vminq_m_u8): Remove.
19247 (vminq_m_u32): Remove.
19248 (vminq_m_u16): Remove.
19249 (vminq_x_s8): Remove.
19250 (vminq_x_s16): Remove.
19251 (vminq_x_s32): Remove.
19252 (vminq_x_u8): Remove.
19253 (vminq_x_u16): Remove.
19254 (vminq_x_u32): Remove.
19255 (vmaxq_x_s8): Remove.
19256 (vmaxq_x_s16): Remove.
19257 (vmaxq_x_s32): Remove.
19258 (vmaxq_x_u8): Remove.
19259 (vmaxq_x_u16): Remove.
19260 (vmaxq_x_u32): Remove.
19261 (__arm_vminq_u8): Remove.
19262 (__arm_vmaxq_u8): Remove.
19263 (__arm_vminq_s8): Remove.
19264 (__arm_vmaxq_s8): Remove.
19265 (__arm_vminq_u16): Remove.
19266 (__arm_vmaxq_u16): Remove.
19267 (__arm_vminq_s16): Remove.
19268 (__arm_vmaxq_s16): Remove.
19269 (__arm_vminq_u32): Remove.
19270 (__arm_vmaxq_u32): Remove.
19271 (__arm_vminq_s32): Remove.
19272 (__arm_vmaxq_s32): Remove.
19273 (__arm_vmaxq_m_s8): Remove.
19274 (__arm_vmaxq_m_s32): Remove.
19275 (__arm_vmaxq_m_s16): Remove.
19276 (__arm_vmaxq_m_u8): Remove.
19277 (__arm_vmaxq_m_u32): Remove.
19278 (__arm_vmaxq_m_u16): Remove.
19279 (__arm_vminq_m_s8): Remove.
19280 (__arm_vminq_m_s32): Remove.
19281 (__arm_vminq_m_s16): Remove.
19282 (__arm_vminq_m_u8): Remove.
19283 (__arm_vminq_m_u32): Remove.
19284 (__arm_vminq_m_u16): Remove.
19285 (__arm_vminq_x_s8): Remove.
19286 (__arm_vminq_x_s16): Remove.
19287 (__arm_vminq_x_s32): Remove.
19288 (__arm_vminq_x_u8): Remove.
19289 (__arm_vminq_x_u16): Remove.
19290 (__arm_vminq_x_u32): Remove.
19291 (__arm_vmaxq_x_s8): Remove.
19292 (__arm_vmaxq_x_s16): Remove.
19293 (__arm_vmaxq_x_s32): Remove.
19294 (__arm_vmaxq_x_u8): Remove.
19295 (__arm_vmaxq_x_u16): Remove.
19296 (__arm_vmaxq_x_u32): Remove.
19297 (__arm_vminq): Remove.
19298 (__arm_vmaxq): Remove.
19299 (__arm_vmaxq_m): Remove.
19300 (__arm_vminq_m): Remove.
19301 (__arm_vminq_x): Remove.
19302 (__arm_vmaxq_x): Remove.
19304 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19306 * config/arm/iterators.md (MAX_MIN_SU): New.
19307 (max_min_su_str): New.
19308 (max_min_supf): New.
19309 * config/arm/mve.md (mve_vmaxq_s<mode>, mve_vmaxq_u<mode>)
19310 (mve_vminq_s<mode>, mve_vminq_u<mode>): Merge into ...
19311 (mve_<max_min_su_str>q_<max_min_supf><mode>): ... this.
19313 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19315 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITH_M_N_R): New.
19316 (vqshlq, vshlq): New.
19317 * config/arm/arm-mve-builtins-base.def (vqshlq, vshlq): New.
19318 * config/arm/arm-mve-builtins-base.h (vqshlq, vshlq): New.
19319 * config/arm/arm_mve.h (vshlq): Remove.
19322 (vshlq_m_r): Remove.
19324 (vshlq_m_n): Remove.
19326 (vshlq_x_n): Remove.
19327 (vshlq_s8): Remove.
19328 (vshlq_s16): Remove.
19329 (vshlq_s32): Remove.
19330 (vshlq_u8): Remove.
19331 (vshlq_u16): Remove.
19332 (vshlq_u32): Remove.
19333 (vshlq_r_u8): Remove.
19334 (vshlq_n_u8): Remove.
19335 (vshlq_r_s8): Remove.
19336 (vshlq_n_s8): Remove.
19337 (vshlq_r_u16): Remove.
19338 (vshlq_n_u16): Remove.
19339 (vshlq_r_s16): Remove.
19340 (vshlq_n_s16): Remove.
19341 (vshlq_r_u32): Remove.
19342 (vshlq_n_u32): Remove.
19343 (vshlq_r_s32): Remove.
19344 (vshlq_n_s32): Remove.
19345 (vshlq_m_r_u8): Remove.
19346 (vshlq_m_r_s8): Remove.
19347 (vshlq_m_r_u16): Remove.
19348 (vshlq_m_r_s16): Remove.
19349 (vshlq_m_r_u32): Remove.
19350 (vshlq_m_r_s32): Remove.
19351 (vshlq_m_u8): Remove.
19352 (vshlq_m_s8): Remove.
19353 (vshlq_m_u16): Remove.
19354 (vshlq_m_s16): Remove.
19355 (vshlq_m_u32): Remove.
19356 (vshlq_m_s32): Remove.
19357 (vshlq_m_n_s8): Remove.
19358 (vshlq_m_n_s32): Remove.
19359 (vshlq_m_n_s16): Remove.
19360 (vshlq_m_n_u8): Remove.
19361 (vshlq_m_n_u32): Remove.
19362 (vshlq_m_n_u16): Remove.
19363 (vshlq_x_s8): Remove.
19364 (vshlq_x_s16): Remove.
19365 (vshlq_x_s32): Remove.
19366 (vshlq_x_u8): Remove.
19367 (vshlq_x_u16): Remove.
19368 (vshlq_x_u32): Remove.
19369 (vshlq_x_n_s8): Remove.
19370 (vshlq_x_n_s16): Remove.
19371 (vshlq_x_n_s32): Remove.
19372 (vshlq_x_n_u8): Remove.
19373 (vshlq_x_n_u16): Remove.
19374 (vshlq_x_n_u32): Remove.
19375 (__arm_vshlq_s8): Remove.
19376 (__arm_vshlq_s16): Remove.
19377 (__arm_vshlq_s32): Remove.
19378 (__arm_vshlq_u8): Remove.
19379 (__arm_vshlq_u16): Remove.
19380 (__arm_vshlq_u32): Remove.
19381 (__arm_vshlq_r_u8): Remove.
19382 (__arm_vshlq_n_u8): Remove.
19383 (__arm_vshlq_r_s8): Remove.
19384 (__arm_vshlq_n_s8): Remove.
19385 (__arm_vshlq_r_u16): Remove.
19386 (__arm_vshlq_n_u16): Remove.
19387 (__arm_vshlq_r_s16): Remove.
19388 (__arm_vshlq_n_s16): Remove.
19389 (__arm_vshlq_r_u32): Remove.
19390 (__arm_vshlq_n_u32): Remove.
19391 (__arm_vshlq_r_s32): Remove.
19392 (__arm_vshlq_n_s32): Remove.
19393 (__arm_vshlq_m_r_u8): Remove.
19394 (__arm_vshlq_m_r_s8): Remove.
19395 (__arm_vshlq_m_r_u16): Remove.
19396 (__arm_vshlq_m_r_s16): Remove.
19397 (__arm_vshlq_m_r_u32): Remove.
19398 (__arm_vshlq_m_r_s32): Remove.
19399 (__arm_vshlq_m_u8): Remove.
19400 (__arm_vshlq_m_s8): Remove.
19401 (__arm_vshlq_m_u16): Remove.
19402 (__arm_vshlq_m_s16): Remove.
19403 (__arm_vshlq_m_u32): Remove.
19404 (__arm_vshlq_m_s32): Remove.
19405 (__arm_vshlq_m_n_s8): Remove.
19406 (__arm_vshlq_m_n_s32): Remove.
19407 (__arm_vshlq_m_n_s16): Remove.
19408 (__arm_vshlq_m_n_u8): Remove.
19409 (__arm_vshlq_m_n_u32): Remove.
19410 (__arm_vshlq_m_n_u16): Remove.
19411 (__arm_vshlq_x_s8): Remove.
19412 (__arm_vshlq_x_s16): Remove.
19413 (__arm_vshlq_x_s32): Remove.
19414 (__arm_vshlq_x_u8): Remove.
19415 (__arm_vshlq_x_u16): Remove.
19416 (__arm_vshlq_x_u32): Remove.
19417 (__arm_vshlq_x_n_s8): Remove.
19418 (__arm_vshlq_x_n_s16): Remove.
19419 (__arm_vshlq_x_n_s32): Remove.
19420 (__arm_vshlq_x_n_u8): Remove.
19421 (__arm_vshlq_x_n_u16): Remove.
19422 (__arm_vshlq_x_n_u32): Remove.
19423 (__arm_vshlq): Remove.
19424 (__arm_vshlq_r): Remove.
19425 (__arm_vshlq_n): Remove.
19426 (__arm_vshlq_m_r): Remove.
19427 (__arm_vshlq_m): Remove.
19428 (__arm_vshlq_m_n): Remove.
19429 (__arm_vshlq_x): Remove.
19430 (__arm_vshlq_x_n): Remove.
19432 (vqshlq_r): Remove.
19433 (vqshlq_n): Remove.
19434 (vqshlq_m_r): Remove.
19435 (vqshlq_m_n): Remove.
19436 (vqshlq_m): Remove.
19437 (vqshlq_u8): Remove.
19438 (vqshlq_r_u8): Remove.
19439 (vqshlq_n_u8): Remove.
19440 (vqshlq_s8): Remove.
19441 (vqshlq_r_s8): Remove.
19442 (vqshlq_n_s8): Remove.
19443 (vqshlq_u16): Remove.
19444 (vqshlq_r_u16): Remove.
19445 (vqshlq_n_u16): Remove.
19446 (vqshlq_s16): Remove.
19447 (vqshlq_r_s16): Remove.
19448 (vqshlq_n_s16): Remove.
19449 (vqshlq_u32): Remove.
19450 (vqshlq_r_u32): Remove.
19451 (vqshlq_n_u32): Remove.
19452 (vqshlq_s32): Remove.
19453 (vqshlq_r_s32): Remove.
19454 (vqshlq_n_s32): Remove.
19455 (vqshlq_m_r_u8): Remove.
19456 (vqshlq_m_r_s8): Remove.
19457 (vqshlq_m_r_u16): Remove.
19458 (vqshlq_m_r_s16): Remove.
19459 (vqshlq_m_r_u32): Remove.
19460 (vqshlq_m_r_s32): Remove.
19461 (vqshlq_m_n_s8): Remove.
19462 (vqshlq_m_n_s32): Remove.
19463 (vqshlq_m_n_s16): Remove.
19464 (vqshlq_m_n_u8): Remove.
19465 (vqshlq_m_n_u32): Remove.
19466 (vqshlq_m_n_u16): Remove.
19467 (vqshlq_m_s8): Remove.
19468 (vqshlq_m_s32): Remove.
19469 (vqshlq_m_s16): Remove.
19470 (vqshlq_m_u8): Remove.
19471 (vqshlq_m_u32): Remove.
19472 (vqshlq_m_u16): Remove.
19473 (__arm_vqshlq_u8): Remove.
19474 (__arm_vqshlq_r_u8): Remove.
19475 (__arm_vqshlq_n_u8): Remove.
19476 (__arm_vqshlq_s8): Remove.
19477 (__arm_vqshlq_r_s8): Remove.
19478 (__arm_vqshlq_n_s8): Remove.
19479 (__arm_vqshlq_u16): Remove.
19480 (__arm_vqshlq_r_u16): Remove.
19481 (__arm_vqshlq_n_u16): Remove.
19482 (__arm_vqshlq_s16): Remove.
19483 (__arm_vqshlq_r_s16): Remove.
19484 (__arm_vqshlq_n_s16): Remove.
19485 (__arm_vqshlq_u32): Remove.
19486 (__arm_vqshlq_r_u32): Remove.
19487 (__arm_vqshlq_n_u32): Remove.
19488 (__arm_vqshlq_s32): Remove.
19489 (__arm_vqshlq_r_s32): Remove.
19490 (__arm_vqshlq_n_s32): Remove.
19491 (__arm_vqshlq_m_r_u8): Remove.
19492 (__arm_vqshlq_m_r_s8): Remove.
19493 (__arm_vqshlq_m_r_u16): Remove.
19494 (__arm_vqshlq_m_r_s16): Remove.
19495 (__arm_vqshlq_m_r_u32): Remove.
19496 (__arm_vqshlq_m_r_s32): Remove.
19497 (__arm_vqshlq_m_n_s8): Remove.
19498 (__arm_vqshlq_m_n_s32): Remove.
19499 (__arm_vqshlq_m_n_s16): Remove.
19500 (__arm_vqshlq_m_n_u8): Remove.
19501 (__arm_vqshlq_m_n_u32): Remove.
19502 (__arm_vqshlq_m_n_u16): Remove.
19503 (__arm_vqshlq_m_s8): Remove.
19504 (__arm_vqshlq_m_s32): Remove.
19505 (__arm_vqshlq_m_s16): Remove.
19506 (__arm_vqshlq_m_u8): Remove.
19507 (__arm_vqshlq_m_u32): Remove.
19508 (__arm_vqshlq_m_u16): Remove.
19509 (__arm_vqshlq): Remove.
19510 (__arm_vqshlq_r): Remove.
19511 (__arm_vqshlq_n): Remove.
19512 (__arm_vqshlq_m_r): Remove.
19513 (__arm_vqshlq_m_n): Remove.
19514 (__arm_vqshlq_m): Remove.
19516 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19518 * config/arm/arm-mve-builtins-functions.h (class
19519 unspec_mve_function_exact_insn_vshl): New.
19521 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19523 * config/arm/arm-mve-builtins-shapes.cc (binary_lshift_r): New.
19524 * config/arm/arm-mve-builtins-shapes.h (binary_lshift_r): New.
19526 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19528 * config/arm/arm-mve-builtins.cc (has_inactive_argument)
19529 (finish_opt_n_resolution): Handle MODE_r.
19530 * config/arm/arm-mve-builtins.def (r): New mode.
19532 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19534 * config/arm/arm-mve-builtins-shapes.cc (binary_lshift): New.
19535 * config/arm/arm-mve-builtins-shapes.h (binary_lshift): New.
19537 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19539 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITHOUT_N): New.
19541 * config/arm/arm-mve-builtins-base.def (vabdq): New.
19542 * config/arm/arm-mve-builtins-base.h (vabdq): New.
19543 * config/arm/arm_mve.h (vabdq): Remove.
19546 (vabdq_u8): Remove.
19547 (vabdq_s8): Remove.
19548 (vabdq_u16): Remove.
19549 (vabdq_s16): Remove.
19550 (vabdq_u32): Remove.
19551 (vabdq_s32): Remove.
19552 (vabdq_f16): Remove.
19553 (vabdq_f32): Remove.
19554 (vabdq_m_s8): Remove.
19555 (vabdq_m_s32): Remove.
19556 (vabdq_m_s16): Remove.
19557 (vabdq_m_u8): Remove.
19558 (vabdq_m_u32): Remove.
19559 (vabdq_m_u16): Remove.
19560 (vabdq_m_f32): Remove.
19561 (vabdq_m_f16): Remove.
19562 (vabdq_x_s8): Remove.
19563 (vabdq_x_s16): Remove.
19564 (vabdq_x_s32): Remove.
19565 (vabdq_x_u8): Remove.
19566 (vabdq_x_u16): Remove.
19567 (vabdq_x_u32): Remove.
19568 (vabdq_x_f16): Remove.
19569 (vabdq_x_f32): Remove.
19570 (__arm_vabdq_u8): Remove.
19571 (__arm_vabdq_s8): Remove.
19572 (__arm_vabdq_u16): Remove.
19573 (__arm_vabdq_s16): Remove.
19574 (__arm_vabdq_u32): Remove.
19575 (__arm_vabdq_s32): Remove.
19576 (__arm_vabdq_m_s8): Remove.
19577 (__arm_vabdq_m_s32): Remove.
19578 (__arm_vabdq_m_s16): Remove.
19579 (__arm_vabdq_m_u8): Remove.
19580 (__arm_vabdq_m_u32): Remove.
19581 (__arm_vabdq_m_u16): Remove.
19582 (__arm_vabdq_x_s8): Remove.
19583 (__arm_vabdq_x_s16): Remove.
19584 (__arm_vabdq_x_s32): Remove.
19585 (__arm_vabdq_x_u8): Remove.
19586 (__arm_vabdq_x_u16): Remove.
19587 (__arm_vabdq_x_u32): Remove.
19588 (__arm_vabdq_f16): Remove.
19589 (__arm_vabdq_f32): Remove.
19590 (__arm_vabdq_m_f32): Remove.
19591 (__arm_vabdq_m_f16): Remove.
19592 (__arm_vabdq_x_f16): Remove.
19593 (__arm_vabdq_x_f32): Remove.
19594 (__arm_vabdq): Remove.
19595 (__arm_vabdq_m): Remove.
19596 (__arm_vabdq_x): Remove.
19598 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19600 * config/arm/iterators.md (MVE_FP_M_BINARY): Add vabdq.
19601 (MVE_FP_VABDQ_ONLY): New.
19602 (mve_insn): Add vabd.
19603 * config/arm/mve.md (mve_vabdq_f<mode>): Move into ...
19604 (@mve_<mve_insn>q_f<mode>): ... this.
19605 (mve_vabdq_m_f<mode>): Remove.
19607 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19609 * config/arm/arm-mve-builtins-base.cc (vqrdmulhq): New.
19610 * config/arm/arm-mve-builtins-base.def (vqrdmulhq): New.
19611 * config/arm/arm-mve-builtins-base.h (vqrdmulhq): New.
19612 * config/arm/arm_mve.h (vqrdmulhq): Remove.
19613 (vqrdmulhq_m): Remove.
19614 (vqrdmulhq_s8): Remove.
19615 (vqrdmulhq_n_s8): Remove.
19616 (vqrdmulhq_s16): Remove.
19617 (vqrdmulhq_n_s16): Remove.
19618 (vqrdmulhq_s32): Remove.
19619 (vqrdmulhq_n_s32): Remove.
19620 (vqrdmulhq_m_n_s8): Remove.
19621 (vqrdmulhq_m_n_s32): Remove.
19622 (vqrdmulhq_m_n_s16): Remove.
19623 (vqrdmulhq_m_s8): Remove.
19624 (vqrdmulhq_m_s32): Remove.
19625 (vqrdmulhq_m_s16): Remove.
19626 (__arm_vqrdmulhq_s8): Remove.
19627 (__arm_vqrdmulhq_n_s8): Remove.
19628 (__arm_vqrdmulhq_s16): Remove.
19629 (__arm_vqrdmulhq_n_s16): Remove.
19630 (__arm_vqrdmulhq_s32): Remove.
19631 (__arm_vqrdmulhq_n_s32): Remove.
19632 (__arm_vqrdmulhq_m_n_s8): Remove.
19633 (__arm_vqrdmulhq_m_n_s32): Remove.
19634 (__arm_vqrdmulhq_m_n_s16): Remove.
19635 (__arm_vqrdmulhq_m_s8): Remove.
19636 (__arm_vqrdmulhq_m_s32): Remove.
19637 (__arm_vqrdmulhq_m_s16): Remove.
19638 (__arm_vqrdmulhq): Remove.
19639 (__arm_vqrdmulhq_m): Remove.
19641 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19643 * config/arm/iterators.md (MVE_SHIFT_M_R, MVE_SHIFT_M_N)
19644 (MVE_SHIFT_N, MVE_SHIFT_R): New.
19645 (mve_insn): Add vqshl, vshl.
19646 * config/arm/mve.md (mve_vqshlq_n_<supf><mode>)
19647 (mve_vshlq_n_<supf><mode>): Merge into ...
19648 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
19649 (mve_vqshlq_r_<supf><mode>, mve_vshlq_r_<supf><mode>): Merge into
19651 (@mve_<mve_insn>q_r_<supf><mode>): ... this.
19652 (mve_vqshlq_m_r_<supf><mode>, mve_vshlq_m_r_<supf><mode>): Merge
19654 (@mve_<mve_insn>q_m_r_<supf><mode>): ... this.
19655 (mve_vqshlq_m_n_<supf><mode>, mve_vshlq_m_n_<supf><mode>): Merge
19657 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
19658 * config/arm/vec-common.md (mve_vshlq_<supf><mode>): Transform
19660 (@mve_<mve_insn>q_<supf><mode>): ... this.
19662 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19664 * config/arm/arm-mve-builtins-base.cc (vqrshlq, vrshlq): New.
19665 * config/arm/arm-mve-builtins-base.def (vqrshlq, vrshlq): New.
19666 * config/arm/arm-mve-builtins-base.h (vqrshlq, vrshlq): New.
19667 * config/arm/arm-mve-builtins.cc (has_inactive_argument): Handle
19669 * config/arm/arm_mve.h (vrshlq): Remove.
19670 (vrshlq_m_n): Remove.
19671 (vrshlq_m): Remove.
19672 (vrshlq_x): Remove.
19673 (vrshlq_u8): Remove.
19674 (vrshlq_n_u8): Remove.
19675 (vrshlq_s8): Remove.
19676 (vrshlq_n_s8): Remove.
19677 (vrshlq_u16): Remove.
19678 (vrshlq_n_u16): Remove.
19679 (vrshlq_s16): Remove.
19680 (vrshlq_n_s16): Remove.
19681 (vrshlq_u32): Remove.
19682 (vrshlq_n_u32): Remove.
19683 (vrshlq_s32): Remove.
19684 (vrshlq_n_s32): Remove.
19685 (vrshlq_m_n_u8): Remove.
19686 (vrshlq_m_n_s8): Remove.
19687 (vrshlq_m_n_u16): Remove.
19688 (vrshlq_m_n_s16): Remove.
19689 (vrshlq_m_n_u32): Remove.
19690 (vrshlq_m_n_s32): Remove.
19691 (vrshlq_m_s8): Remove.
19692 (vrshlq_m_s32): Remove.
19693 (vrshlq_m_s16): Remove.
19694 (vrshlq_m_u8): Remove.
19695 (vrshlq_m_u32): Remove.
19696 (vrshlq_m_u16): Remove.
19697 (vrshlq_x_s8): Remove.
19698 (vrshlq_x_s16): Remove.
19699 (vrshlq_x_s32): Remove.
19700 (vrshlq_x_u8): Remove.
19701 (vrshlq_x_u16): Remove.
19702 (vrshlq_x_u32): Remove.
19703 (__arm_vrshlq_u8): Remove.
19704 (__arm_vrshlq_n_u8): Remove.
19705 (__arm_vrshlq_s8): Remove.
19706 (__arm_vrshlq_n_s8): Remove.
19707 (__arm_vrshlq_u16): Remove.
19708 (__arm_vrshlq_n_u16): Remove.
19709 (__arm_vrshlq_s16): Remove.
19710 (__arm_vrshlq_n_s16): Remove.
19711 (__arm_vrshlq_u32): Remove.
19712 (__arm_vrshlq_n_u32): Remove.
19713 (__arm_vrshlq_s32): Remove.
19714 (__arm_vrshlq_n_s32): Remove.
19715 (__arm_vrshlq_m_n_u8): Remove.
19716 (__arm_vrshlq_m_n_s8): Remove.
19717 (__arm_vrshlq_m_n_u16): Remove.
19718 (__arm_vrshlq_m_n_s16): Remove.
19719 (__arm_vrshlq_m_n_u32): Remove.
19720 (__arm_vrshlq_m_n_s32): Remove.
19721 (__arm_vrshlq_m_s8): Remove.
19722 (__arm_vrshlq_m_s32): Remove.
19723 (__arm_vrshlq_m_s16): Remove.
19724 (__arm_vrshlq_m_u8): Remove.
19725 (__arm_vrshlq_m_u32): Remove.
19726 (__arm_vrshlq_m_u16): Remove.
19727 (__arm_vrshlq_x_s8): Remove.
19728 (__arm_vrshlq_x_s16): Remove.
19729 (__arm_vrshlq_x_s32): Remove.
19730 (__arm_vrshlq_x_u8): Remove.
19731 (__arm_vrshlq_x_u16): Remove.
19732 (__arm_vrshlq_x_u32): Remove.
19733 (__arm_vrshlq): Remove.
19734 (__arm_vrshlq_m_n): Remove.
19735 (__arm_vrshlq_m): Remove.
19736 (__arm_vrshlq_x): Remove.
19738 (vqrshlq_m_n): Remove.
19739 (vqrshlq_m): Remove.
19740 (vqrshlq_u8): Remove.
19741 (vqrshlq_n_u8): Remove.
19742 (vqrshlq_s8): Remove.
19743 (vqrshlq_n_s8): Remove.
19744 (vqrshlq_u16): Remove.
19745 (vqrshlq_n_u16): Remove.
19746 (vqrshlq_s16): Remove.
19747 (vqrshlq_n_s16): Remove.
19748 (vqrshlq_u32): Remove.
19749 (vqrshlq_n_u32): Remove.
19750 (vqrshlq_s32): Remove.
19751 (vqrshlq_n_s32): Remove.
19752 (vqrshlq_m_n_u8): Remove.
19753 (vqrshlq_m_n_s8): Remove.
19754 (vqrshlq_m_n_u16): Remove.
19755 (vqrshlq_m_n_s16): Remove.
19756 (vqrshlq_m_n_u32): Remove.
19757 (vqrshlq_m_n_s32): Remove.
19758 (vqrshlq_m_s8): Remove.
19759 (vqrshlq_m_s32): Remove.
19760 (vqrshlq_m_s16): Remove.
19761 (vqrshlq_m_u8): Remove.
19762 (vqrshlq_m_u32): Remove.
19763 (vqrshlq_m_u16): Remove.
19764 (__arm_vqrshlq_u8): Remove.
19765 (__arm_vqrshlq_n_u8): Remove.
19766 (__arm_vqrshlq_s8): Remove.
19767 (__arm_vqrshlq_n_s8): Remove.
19768 (__arm_vqrshlq_u16): Remove.
19769 (__arm_vqrshlq_n_u16): Remove.
19770 (__arm_vqrshlq_s16): Remove.
19771 (__arm_vqrshlq_n_s16): Remove.
19772 (__arm_vqrshlq_u32): Remove.
19773 (__arm_vqrshlq_n_u32): Remove.
19774 (__arm_vqrshlq_s32): Remove.
19775 (__arm_vqrshlq_n_s32): Remove.
19776 (__arm_vqrshlq_m_n_u8): Remove.
19777 (__arm_vqrshlq_m_n_s8): Remove.
19778 (__arm_vqrshlq_m_n_u16): Remove.
19779 (__arm_vqrshlq_m_n_s16): Remove.
19780 (__arm_vqrshlq_m_n_u32): Remove.
19781 (__arm_vqrshlq_m_n_s32): Remove.
19782 (__arm_vqrshlq_m_s8): Remove.
19783 (__arm_vqrshlq_m_s32): Remove.
19784 (__arm_vqrshlq_m_s16): Remove.
19785 (__arm_vqrshlq_m_u8): Remove.
19786 (__arm_vqrshlq_m_u32): Remove.
19787 (__arm_vqrshlq_m_u16): Remove.
19788 (__arm_vqrshlq): Remove.
19789 (__arm_vqrshlq_m_n): Remove.
19790 (__arm_vqrshlq_m): Remove.
19792 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19794 * config/arm/iterators.md (MVE_RSHIFT_M_N, MVE_RSHIFT_N): New.
19795 (mve_insn): Add vqrshl, vrshl.
19796 * config/arm/mve.md (mve_vqrshlq_n_<supf><mode>)
19797 (mve_vrshlq_n_<supf><mode>): Merge into ...
19798 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
19799 (mve_vqrshlq_m_n_<supf><mode>, mve_vrshlq_m_n_<supf><mode>): Merge
19801 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
19803 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
19805 * config/arm/arm-mve-builtins-shapes.cc (binary_round_lshift): New.
19806 * config/arm/arm-mve-builtins-shapes.h (binary_round_lshift): New.
19808 2023-05-05 Juzhe-Zhong <juzhe.zhong@rivai.ai>
19811 * config/riscv/riscv-vsetvl.cc (avl_info::multiple_source_equal_p): Add
19812 denegrate PHI optmization.
19814 2023-05-05 Uros Bizjak <ubizjak@gmail.com>
19816 * config/i386/predicates.md (register_no_SP_operand):
19817 Rename from index_register_operand.
19818 (call_register_operand): Update for rename.
19819 * config/i386/i386.md (*lea<mode>_general_[1234]): Update for rename.
19821 2023-05-05 Tamar Christina <tamar.christina@arm.com>
19824 * Makefile.in (NUM_MATCH_SPLITS, MATCH_SPLITS_SEQ,
19825 GIMPLE_MATCH_PD_SEQ_SRC, GIMPLE_MATCH_PD_SEQ_O,
19826 GENERIC_MATCH_PD_SEQ_SRC, GENERIC_MATCH_PD_SEQ_O): New.
19827 (OBJS, MOSTLYCLEANFILES, .PRECIOUS): Use them.
19828 (s-match): Split into s-generic-match and s-gimple-match.
19829 * configure.ac (with-matchpd-partitions,
19830 DEFAULT_MATCHPD_PARTITIONS): New.
19831 * configure: Regenerate.
19833 2023-05-05 Tamar Christina <tamar.christina@arm.com>
19836 * genmatch.cc (emit_func, SIZED_BASED_CHUNKS, get_out_file): New.
19837 (decision_tree::gen): Accept list of files instead of single and update
19838 to write function definition to header and main file.
19839 (write_predicate): Likewise.
19840 (write_header): Emit pragmas and new includes.
19841 (main): Create file buffers and cleanup.
19842 (showUsage, write_header_includes): New.
19844 2023-05-05 Tamar Christina <tamar.christina@arm.com>
19847 * Makefile.in (OBJS): Add gimple-match-exports.o.
19848 * genmatch.cc (decision_tree::gen): Export gimple_gimplify helpers.
19849 * gimple-match-head.cc (gimple_simplify, gimple_resimplify1,
19850 gimple_resimplify2, gimple_resimplify3, gimple_resimplify4,
19851 gimple_resimplify5, constant_for_folding, convert_conditional_op,
19852 maybe_resimplify_conditional_op, gimple_match_op::resimplify,
19853 maybe_build_generic_op, build_call_internal, maybe_push_res_to_seq,
19854 do_valueize, try_conditional_simplification, gimple_extract,
19855 gimple_extract_op, canonicalize_code, commutative_binary_op_p,
19856 commutative_ternary_op_p, first_commutative_argument,
19857 associative_binary_op_p, directly_supported_p,
19858 get_conditional_internal_fn): Moved to gimple-match-exports.cc
19859 * gimple-match-exports.cc: New file.
19861 2023-05-05 Tamar Christina <tamar.christina@arm.com>
19864 * genmatch.cc (decision_tree::gen, write_predicate): Generate new
19866 (dt_simplify::gen_1): Use it.
19868 2023-05-05 Tamar Christina <tamar.christina@arm.com>
19871 * genmatch.cc (output_line_directive): Only emit commented directive
19874 2023-05-05 Tamar Christina <tamar.christina@arm.com>
19877 * genmatch.cc (dt_simplify::gen_1): Only emit labels if used.
19879 2023-05-05 Tobias Burnus <tobias@codesourcery.com>
19881 * config/gcn/gcn.cc (gcn_vectorize_builtin_vectorized_function): Remove
19882 unused in_mode/in_n variables.
19884 2023-05-05 Richard Biener <rguenther@suse.de>
19886 PR tree-optimization/109735
19887 * tree-vect-stmts.cc (vectorizable_operation): Perform
19888 conversion for POINTER_DIFF_EXPR unconditionally.
19890 2023-05-05 Uros Bizjak <ubizjak@gmail.com>
19892 * config/i386/mmx.md (mulv2si3): New expander.
19893 (*mulv2si3): New insn pattern.
19895 2023-05-05 Tobias Burnus <tobias@codesourcery.com>
19896 Thomas Schwinge <thomas@codesourcery.com>
19899 * config/nvptx/mkoffload.cc (process): Emit dummy procedure
19900 alongside reverse-offload function table to prevent NULL values
19901 of the function addresses.
19903 2023-05-05 Jakub Jelinek <jakub@redhat.com>
19905 * builtins.cc (do_mpfr_ckconv, do_mpc_ckconv): Fix comment typo,
19907 * fold-const-call.cc (do_mpfr_ckconv, do_mpc_ckconv): Likewise.
19909 2023-05-05 Andrew Pinski <apinski@marvell.com>
19911 PR tree-optimization/109732
19912 * tree-ssa-phiopt.cc (match_simplify_replacement): Fix the selection
19913 of the argtrue/argfalse.
19915 2023-05-05 Andrew Pinski <apinski@marvell.com>
19917 PR tree-optimization/109722
19918 * match.pd: Extend the `ABS<a> == 0` pattern
19919 to cover `ABSU<a> == 0` too.
19921 2023-05-04 Uros Bizjak <ubizjak@gmail.com>
19924 * config/i386/predicates.md (index_reg_operand): New predicate.
19925 * config/i386/i386.md (ashift to lea spliter): Use
19926 general_reg_operand and index_reg_operand predicates.
19928 2023-05-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19930 * config/aarch64/aarch64-simd.md (aarch64_<sur><addsub>hn2<mode>_insn_le):
19931 Rename and reimplement with RTL codes to...
19932 (aarch64_<optab>hn2<mode>_insn_le): .. This.
19933 (aarch64_r<optab>hn2<mode>_insn_le): New pattern.
19934 (aarch64_<sur><addsub>hn2<mode>_insn_be): Rename and reimplement with RTL
19936 (aarch64_<optab>hn2<mode>_insn_be): ... This.
19937 (aarch64_r<optab>hn2<mode>_insn_be): New pattern.
19938 (aarch64_<sur><addsub>hn2<mode>): Rename and adjust expander to...
19939 (aarch64_<optab>hn2<mode>): ... This.
19940 (aarch64_r<optab>hn2<mode>): New expander.
19941 * config/aarch64/iterators.md (UNSPEC_ADDHN, UNSPEC_RADDHN,
19942 UNSPEC_SUBHN, UNSPEC_RSUBHN): Delete unspecs.
19943 (ADDSUBHN): Delete.
19944 (sur): Remove handling of the above.
19945 (addsub): Likewise.
19947 2023-05-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19949 * config/aarch64/aarch64-simd.md (aarch64_<sur><addsub>hn<mode>_insn_le):
19951 (aarch64_<optab>hn<mode>_insn<vczle><vczbe>): New define_insn.
19952 (aarch64_<sur><addsub>hn<mode>_insn_be): Delete.
19953 (aarch64_r<optab>hn<mode>_insn<vczle><vczbe>): New define_insn.
19954 (aarch64_<sur><addsub>hn<mode>): Delete.
19955 (aarch64_<optab>hn<mode>): New define_expand.
19956 (aarch64_r<optab>hn<mode>): Likewise.
19957 * config/aarch64/predicates.md (aarch64_simd_raddsubhn_imm_vec):
19960 2023-05-04 Andrew Pinski <apinski@marvell.com>
19962 * tree-ssa-phiopt.cc (replace_phi_edge_with_variable): Handle
19963 diamond form bb with forwarder only empty blocks better.
19965 2023-05-04 Andrew Pinski <apinski@marvell.com>
19967 * tree-ssa-threadupdate.cc (copy_phi_arg_into_existing_phi): Move to ...
19968 * tree-cfg.cc (copy_phi_arg_into_existing_phi): Here and remove static.
19969 (gimple_duplicate_sese_tail): Use copy_phi_arg_into_existing_phi instead
19970 of an inline version of it.
19971 * tree-cfgcleanup.cc (remove_forwarder_block): Likewise.
19972 * tree-cfg.h (copy_phi_arg_into_existing_phi): New declaration.
19974 2023-05-04 Andrew Pinski <apinski@marvell.com>
19976 * tree-ssa-phiopt.cc (replace_phi_edge_with_variable): Change
19977 the default argument value for dce_ssa_names to nullptr.
19978 Check to make sure dce_ssa_names is a non-nullptr before
19979 calling simple_dce_from_worklist.
19981 2023-05-04 Uros Bizjak <ubizjak@gmail.com>
19983 * config/i386/predicates.md (index_register_operand): Reject
19984 arg_pointer_rtx, frame_pointer_rtx, stack_pointer_rtx and
19985 VIRTUAL_REGISTER_P operands. Allow subregs of memory before reload.
19986 (call_register_no_elim_operand): Rewrite as ...
19987 (call_register_operand): ... this.
19988 (call_insn_operand): Use call_register_operand predicate.
19990 2023-05-04 Richard Biener <rguenther@suse.de>
19992 PR tree-optimization/109721
19993 * tree-vect-stmts.cc (vectorizable_operation): Make sure
19994 to test word_mode for all !target_support_p operations.
19996 2023-05-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19999 * config/aarch64/aarch64-simd.md (aarch64_<su>aba<mode>): Rename to...
20000 (aarch64_<su>aba<mode><vczle><vczbe>): ... This.
20001 (aarch64_mla<mode>): Rename to...
20002 (aarch64_mla<mode><vczle><vczbe>): ... This.
20003 (*aarch64_mla_elt<mode>): Rename to...
20004 (*aarch64_mla_elt<mode><vczle><vczbe>): ... This.
20005 (*aarch64_mla_elt_<vswap_width_name><mode>): Rename to...
20006 (*aarch64_mla_elt_<vswap_width_name><mode><vczle><vczbe>): ... This.
20007 (aarch64_mla_n<mode>): Rename to...
20008 (aarch64_mla_n<mode><vczle><vczbe>): ... This.
20009 (aarch64_mls<mode>): Rename to...
20010 (aarch64_mls<mode><vczle><vczbe>): ... This.
20011 (*aarch64_mls_elt<mode>): Rename to...
20012 (*aarch64_mls_elt<mode><vczle><vczbe>): ... This.
20013 (*aarch64_mls_elt_<vswap_width_name><mode>): Rename to...
20014 (*aarch64_mls_elt_<vswap_width_name><mode><vczle><vczbe>): ... This.
20015 (aarch64_mls_n<mode>): Rename to...
20016 (aarch64_mls_n<mode><vczle><vczbe>): ... This.
20017 (fma<mode>4): Rename to...
20018 (fma<mode>4<vczle><vczbe>): ... This.
20019 (*aarch64_fma4_elt<mode>): Rename to...
20020 (*aarch64_fma4_elt<mode><vczle><vczbe>): ... This.
20021 (*aarch64_fma4_elt_<vswap_width_name><mode>): Rename to...
20022 (*aarch64_fma4_elt_<vswap_width_name><mode><vczle><vczbe>): ... This.
20023 (*aarch64_fma4_elt_from_dup<mode>): Rename to...
20024 (*aarch64_fma4_elt_from_dup<mode><vczle><vczbe>): ... This.
20025 (fnma<mode>4): Rename to...
20026 (fnma<mode>4<vczle><vczbe>): ... This.
20027 (*aarch64_fnma4_elt<mode>): Rename to...
20028 (*aarch64_fnma4_elt<mode><vczle><vczbe>): ... This.
20029 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Rename to...
20030 (*aarch64_fnma4_elt_<vswap_width_name><mode><vczle><vczbe>): ... This.
20031 (*aarch64_fnma4_elt_from_dup<mode>): Rename to...
20032 (*aarch64_fnma4_elt_from_dup<mode><vczle><vczbe>): ... This.
20033 (aarch64_simd_bsl<mode>_internal): Rename to...
20034 (aarch64_simd_bsl<mode>_internal<vczle><vczbe>): ... This.
20035 (*aarch64_simd_bsl<mode>_alt): Rename to...
20036 (*aarch64_simd_bsl<mode>_alt<vczle><vczbe>): ... This.
20038 2023-05-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20041 * config/aarch64/aarch64-simd.md (aarch64_<su>abd<mode>): Rename to...
20042 (aarch64_<su>abd<mode><vczle><vczbe>): ... This.
20043 (fabd<mode>3): Rename to...
20044 (fabd<mode>3<vczle><vczbe>): ... This.
20045 (aarch64_<optab>p<mode>): Rename to...
20046 (aarch64_<optab>p<mode><vczle><vczbe>): ... This.
20047 (aarch64_faddp<mode>): Rename to...
20048 (aarch64_faddp<mode><vczle><vczbe>): ... This.
20050 2023-05-04 Martin Liska <mliska@suse.cz>
20052 * gcov.cc (GCOV_JSON_FORMAT_VERSION): New definition.
20053 (print_version): Use it.
20054 (generate_results): Likewise.
20056 2023-05-04 Richard Biener <rguenther@suse.de>
20058 * tree-cfg.h (last_stmt): Rename to ...
20059 (last_nondebug_stmt): ... this.
20060 * tree-cfg.cc (last_stmt): Rename to ...
20061 (last_nondebug_stmt): ... this.
20062 (assign_discriminators): Adjust.
20063 (group_case_labels_stmt): Likewise.
20064 (gimple_can_duplicate_bb_p): Likewise.
20065 (execute_fixup_cfg): Likewise.
20066 * auto-profile.cc (afdo_propagate_circuit): Likewise.
20067 * gimple-range.cc (gimple_ranger::range_on_exit): Likewise.
20068 * omp-expand.cc (workshare_safe_to_combine_p): Likewise.
20069 (determine_parallel_type): Likewise.
20070 (adjust_context_and_scope): Likewise.
20071 (expand_task_call): Likewise.
20072 (remove_exit_barrier): Likewise.
20073 (expand_omp_taskreg): Likewise.
20074 (expand_omp_for_init_counts): Likewise.
20075 (expand_omp_for_init_vars): Likewise.
20076 (expand_omp_for_static_chunk): Likewise.
20077 (expand_omp_simd): Likewise.
20078 (expand_oacc_for): Likewise.
20079 (expand_omp_for): Likewise.
20080 (expand_omp_sections): Likewise.
20081 (expand_omp_atomic_fetch_op): Likewise.
20082 (expand_omp_atomic_cas): Likewise.
20083 (expand_omp_atomic): Likewise.
20084 (expand_omp_target): Likewise.
20085 (expand_omp): Likewise.
20086 (omp_make_gimple_edges): Likewise.
20087 * trans-mem.cc (tm_region_init): Likewise.
20088 * tree-inline.cc (redirect_all_calls): Likewise.
20089 * tree-parloops.cc (gen_parallel_loop): Likewise.
20090 * tree-ssa-loop-ch.cc (do_while_loop_p): Likewise.
20091 * tree-ssa-loop-ivcanon.cc (canonicalize_loop_induction_variables):
20093 * tree-ssa-loop-ivopts.cc (stmt_after_ip_normal_pos): Likewise.
20094 (may_eliminate_iv): Likewise.
20095 * tree-ssa-loop-manip.cc (standard_iv_increment_position): Likewise.
20096 * tree-ssa-loop-niter.cc (do_warn_aggressive_loop_optimizations):
20098 (estimate_numbers_of_iterations): Likewise.
20099 * tree-ssa-loop-split.cc (compute_added_num_insns): Likewise.
20100 * tree-ssa-loop-unswitch.cc (get_predicates_for_bb): Likewise.
20101 (set_predicates_for_bb): Likewise.
20102 (init_loop_unswitch_info): Likewise.
20103 (hoist_guard): Likewise.
20104 * tree-ssa-phiopt.cc (match_simplify_replacement): Likewise.
20105 (minmax_replacement): Likewise.
20106 * tree-ssa-reassoc.cc (update_range_test): Likewise.
20107 (optimize_range_tests_to_bit_test): Likewise.
20108 (optimize_range_tests_var_bound): Likewise.
20109 (optimize_range_tests): Likewise.
20110 (no_side_effect_bb): Likewise.
20111 (suitable_cond_bb): Likewise.
20112 (maybe_optimize_range_tests): Likewise.
20113 (reassociate_bb): Likewise.
20114 * tree-vrp.cc (rvrp_folder::pre_fold_bb): Likewise.
20116 2023-05-04 Jakub Jelinek <jakub@redhat.com>
20119 * config/i386/i386-features.cc (timode_scalar_chain::convert_insn):
20120 If src is REG, change its mode to V1TImode and call fix_debug_reg_uses
20121 for it only if it still has TImode. Don't decide whether to call
20122 fix_debug_reg_uses based on whether SRC is ever set or not.
20124 2023-05-04 Hans-Peter Nilsson <hp@axis.com>
20126 * config/cris/cris.cc (cris_split_constant): New function.
20127 * config/cris/cris.md (splitop): New iterator.
20128 (opsplit1): New define_peephole2.
20129 * config/cris/cris-protos.h (cris_split_constant): Declare.
20130 (cris_splittable_constant_p): New macro.
20132 2023-05-04 Hans-Peter Nilsson <hp@axis.com>
20134 * config/cris/cris.cc (TARGET_SPILL_CLASS): Define
20137 2023-05-04 Hans-Peter Nilsson <hp@axis.com>
20139 * config/cris/cris.cc (cris_side_effect_mode_ok): Use
20140 lra_in_progress, not reload_in_progress.
20141 * config/cris/cris.md ("movdi", "*addi_reload"): Ditto.
20142 * config/cris/constraints.md ("Q"): Ditto.
20144 2023-05-03 Andrew Pinski <apinski@marvell.com>
20146 * tree-ssa-dce.cc (simple_dce_from_worklist): Record
20147 stats on removed number of statements and phis.
20149 2023-05-03 Aldy Hernandez <aldyh@redhat.com>
20151 PR tree-optimization/109711
20152 * value-range.cc (irange::verify_range): Allow types of
20155 2023-05-03 Alexander Monakov <amonakov@ispras.ru>
20158 * calls.cc (can_implement_as_sibling_call_p): Reject calls
20159 to __sanitizer_cov_trace_pc.
20161 2023-05-03 Richard Sandiford <richard.sandiford@arm.com>
20164 * config/aarch64/aarch64.cc (aarch64_function_arg_alignment): Add
20165 a new ABI break parameter for GCC 14. Set it to the alignment
20166 of enums that have an underlying type. Take the true alignment
20167 of such enums from the TYPE_ALIGN of the underlying type's
20169 (aarch64_function_arg_boundary): Update accordingly.
20170 (aarch64_layout_arg, aarch64_gimplify_va_arg_expr): Likewise.
20171 Warn about ABI differences.
20173 2023-05-03 Richard Sandiford <richard.sandiford@arm.com>
20176 * config/aarch64/aarch64.cc (aarch64_function_arg_alignment): Rename
20177 ABI break variables to abi_break_gcc_9 and abi_break_gcc_13.
20178 (aarch64_layout_arg, aarch64_function_arg_boundary): Likewise.
20179 (aarch64_gimplify_va_arg_expr): Likewise.
20181 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20183 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITH_M_N_NO_F)
20184 (FUNCTION_WITHOUT_N_NO_F, FUNCTION_WITH_M_N_NO_U_F): New.
20185 (vhaddq, vhsubq, vmulhq, vqaddq, vqsubq, vqdmulhq, vrhaddq)
20187 * config/arm/arm-mve-builtins-base.def (vhaddq, vhsubq, vmulhq)
20188 (vqaddq, vqsubq, vqdmulhq, vrhaddq, vrmulhq): New.
20189 * config/arm/arm-mve-builtins-base.h (vhaddq, vhsubq, vmulhq)
20190 (vqaddq, vqsubq, vqdmulhq, vrhaddq, vrmulhq): New.
20191 * config/arm/arm_mve.h (vhsubq): Remove.
20193 (vhaddq_m): Remove.
20194 (vhsubq_m): Remove.
20195 (vhaddq_x): Remove.
20196 (vhsubq_x): Remove.
20197 (vhsubq_u8): Remove.
20198 (vhsubq_n_u8): Remove.
20199 (vhaddq_u8): Remove.
20200 (vhaddq_n_u8): Remove.
20201 (vhsubq_s8): Remove.
20202 (vhsubq_n_s8): Remove.
20203 (vhaddq_s8): Remove.
20204 (vhaddq_n_s8): Remove.
20205 (vhsubq_u16): Remove.
20206 (vhsubq_n_u16): Remove.
20207 (vhaddq_u16): Remove.
20208 (vhaddq_n_u16): Remove.
20209 (vhsubq_s16): Remove.
20210 (vhsubq_n_s16): Remove.
20211 (vhaddq_s16): Remove.
20212 (vhaddq_n_s16): Remove.
20213 (vhsubq_u32): Remove.
20214 (vhsubq_n_u32): Remove.
20215 (vhaddq_u32): Remove.
20216 (vhaddq_n_u32): Remove.
20217 (vhsubq_s32): Remove.
20218 (vhsubq_n_s32): Remove.
20219 (vhaddq_s32): Remove.
20220 (vhaddq_n_s32): Remove.
20221 (vhaddq_m_n_s8): Remove.
20222 (vhaddq_m_n_s32): Remove.
20223 (vhaddq_m_n_s16): Remove.
20224 (vhaddq_m_n_u8): Remove.
20225 (vhaddq_m_n_u32): Remove.
20226 (vhaddq_m_n_u16): Remove.
20227 (vhaddq_m_s8): Remove.
20228 (vhaddq_m_s32): Remove.
20229 (vhaddq_m_s16): Remove.
20230 (vhaddq_m_u8): Remove.
20231 (vhaddq_m_u32): Remove.
20232 (vhaddq_m_u16): Remove.
20233 (vhsubq_m_n_s8): Remove.
20234 (vhsubq_m_n_s32): Remove.
20235 (vhsubq_m_n_s16): Remove.
20236 (vhsubq_m_n_u8): Remove.
20237 (vhsubq_m_n_u32): Remove.
20238 (vhsubq_m_n_u16): Remove.
20239 (vhsubq_m_s8): Remove.
20240 (vhsubq_m_s32): Remove.
20241 (vhsubq_m_s16): Remove.
20242 (vhsubq_m_u8): Remove.
20243 (vhsubq_m_u32): Remove.
20244 (vhsubq_m_u16): Remove.
20245 (vhaddq_x_n_s8): Remove.
20246 (vhaddq_x_n_s16): Remove.
20247 (vhaddq_x_n_s32): Remove.
20248 (vhaddq_x_n_u8): Remove.
20249 (vhaddq_x_n_u16): Remove.
20250 (vhaddq_x_n_u32): Remove.
20251 (vhaddq_x_s8): Remove.
20252 (vhaddq_x_s16): Remove.
20253 (vhaddq_x_s32): Remove.
20254 (vhaddq_x_u8): Remove.
20255 (vhaddq_x_u16): Remove.
20256 (vhaddq_x_u32): Remove.
20257 (vhsubq_x_n_s8): Remove.
20258 (vhsubq_x_n_s16): Remove.
20259 (vhsubq_x_n_s32): Remove.
20260 (vhsubq_x_n_u8): Remove.
20261 (vhsubq_x_n_u16): Remove.
20262 (vhsubq_x_n_u32): Remove.
20263 (vhsubq_x_s8): Remove.
20264 (vhsubq_x_s16): Remove.
20265 (vhsubq_x_s32): Remove.
20266 (vhsubq_x_u8): Remove.
20267 (vhsubq_x_u16): Remove.
20268 (vhsubq_x_u32): Remove.
20269 (__arm_vhsubq_u8): Remove.
20270 (__arm_vhsubq_n_u8): Remove.
20271 (__arm_vhaddq_u8): Remove.
20272 (__arm_vhaddq_n_u8): Remove.
20273 (__arm_vhsubq_s8): Remove.
20274 (__arm_vhsubq_n_s8): Remove.
20275 (__arm_vhaddq_s8): Remove.
20276 (__arm_vhaddq_n_s8): Remove.
20277 (__arm_vhsubq_u16): Remove.
20278 (__arm_vhsubq_n_u16): Remove.
20279 (__arm_vhaddq_u16): Remove.
20280 (__arm_vhaddq_n_u16): Remove.
20281 (__arm_vhsubq_s16): Remove.
20282 (__arm_vhsubq_n_s16): Remove.
20283 (__arm_vhaddq_s16): Remove.
20284 (__arm_vhaddq_n_s16): Remove.
20285 (__arm_vhsubq_u32): Remove.
20286 (__arm_vhsubq_n_u32): Remove.
20287 (__arm_vhaddq_u32): Remove.
20288 (__arm_vhaddq_n_u32): Remove.
20289 (__arm_vhsubq_s32): Remove.
20290 (__arm_vhsubq_n_s32): Remove.
20291 (__arm_vhaddq_s32): Remove.
20292 (__arm_vhaddq_n_s32): Remove.
20293 (__arm_vhaddq_m_n_s8): Remove.
20294 (__arm_vhaddq_m_n_s32): Remove.
20295 (__arm_vhaddq_m_n_s16): Remove.
20296 (__arm_vhaddq_m_n_u8): Remove.
20297 (__arm_vhaddq_m_n_u32): Remove.
20298 (__arm_vhaddq_m_n_u16): Remove.
20299 (__arm_vhaddq_m_s8): Remove.
20300 (__arm_vhaddq_m_s32): Remove.
20301 (__arm_vhaddq_m_s16): Remove.
20302 (__arm_vhaddq_m_u8): Remove.
20303 (__arm_vhaddq_m_u32): Remove.
20304 (__arm_vhaddq_m_u16): Remove.
20305 (__arm_vhsubq_m_n_s8): Remove.
20306 (__arm_vhsubq_m_n_s32): Remove.
20307 (__arm_vhsubq_m_n_s16): Remove.
20308 (__arm_vhsubq_m_n_u8): Remove.
20309 (__arm_vhsubq_m_n_u32): Remove.
20310 (__arm_vhsubq_m_n_u16): Remove.
20311 (__arm_vhsubq_m_s8): Remove.
20312 (__arm_vhsubq_m_s32): Remove.
20313 (__arm_vhsubq_m_s16): Remove.
20314 (__arm_vhsubq_m_u8): Remove.
20315 (__arm_vhsubq_m_u32): Remove.
20316 (__arm_vhsubq_m_u16): Remove.
20317 (__arm_vhaddq_x_n_s8): Remove.
20318 (__arm_vhaddq_x_n_s16): Remove.
20319 (__arm_vhaddq_x_n_s32): Remove.
20320 (__arm_vhaddq_x_n_u8): Remove.
20321 (__arm_vhaddq_x_n_u16): Remove.
20322 (__arm_vhaddq_x_n_u32): Remove.
20323 (__arm_vhaddq_x_s8): Remove.
20324 (__arm_vhaddq_x_s16): Remove.
20325 (__arm_vhaddq_x_s32): Remove.
20326 (__arm_vhaddq_x_u8): Remove.
20327 (__arm_vhaddq_x_u16): Remove.
20328 (__arm_vhaddq_x_u32): Remove.
20329 (__arm_vhsubq_x_n_s8): Remove.
20330 (__arm_vhsubq_x_n_s16): Remove.
20331 (__arm_vhsubq_x_n_s32): Remove.
20332 (__arm_vhsubq_x_n_u8): Remove.
20333 (__arm_vhsubq_x_n_u16): Remove.
20334 (__arm_vhsubq_x_n_u32): Remove.
20335 (__arm_vhsubq_x_s8): Remove.
20336 (__arm_vhsubq_x_s16): Remove.
20337 (__arm_vhsubq_x_s32): Remove.
20338 (__arm_vhsubq_x_u8): Remove.
20339 (__arm_vhsubq_x_u16): Remove.
20340 (__arm_vhsubq_x_u32): Remove.
20341 (__arm_vhsubq): Remove.
20342 (__arm_vhaddq): Remove.
20343 (__arm_vhaddq_m): Remove.
20344 (__arm_vhsubq_m): Remove.
20345 (__arm_vhaddq_x): Remove.
20346 (__arm_vhsubq_x): Remove.
20348 (vmulhq_m): Remove.
20349 (vmulhq_x): Remove.
20350 (vmulhq_u8): Remove.
20351 (vmulhq_s8): Remove.
20352 (vmulhq_u16): Remove.
20353 (vmulhq_s16): Remove.
20354 (vmulhq_u32): Remove.
20355 (vmulhq_s32): Remove.
20356 (vmulhq_m_s8): Remove.
20357 (vmulhq_m_s32): Remove.
20358 (vmulhq_m_s16): Remove.
20359 (vmulhq_m_u8): Remove.
20360 (vmulhq_m_u32): Remove.
20361 (vmulhq_m_u16): Remove.
20362 (vmulhq_x_s8): Remove.
20363 (vmulhq_x_s16): Remove.
20364 (vmulhq_x_s32): Remove.
20365 (vmulhq_x_u8): Remove.
20366 (vmulhq_x_u16): Remove.
20367 (vmulhq_x_u32): Remove.
20368 (__arm_vmulhq_u8): Remove.
20369 (__arm_vmulhq_s8): Remove.
20370 (__arm_vmulhq_u16): Remove.
20371 (__arm_vmulhq_s16): Remove.
20372 (__arm_vmulhq_u32): Remove.
20373 (__arm_vmulhq_s32): Remove.
20374 (__arm_vmulhq_m_s8): Remove.
20375 (__arm_vmulhq_m_s32): Remove.
20376 (__arm_vmulhq_m_s16): Remove.
20377 (__arm_vmulhq_m_u8): Remove.
20378 (__arm_vmulhq_m_u32): Remove.
20379 (__arm_vmulhq_m_u16): Remove.
20380 (__arm_vmulhq_x_s8): Remove.
20381 (__arm_vmulhq_x_s16): Remove.
20382 (__arm_vmulhq_x_s32): Remove.
20383 (__arm_vmulhq_x_u8): Remove.
20384 (__arm_vmulhq_x_u16): Remove.
20385 (__arm_vmulhq_x_u32): Remove.
20386 (__arm_vmulhq): Remove.
20387 (__arm_vmulhq_m): Remove.
20388 (__arm_vmulhq_x): Remove.
20391 (vqaddq_m): Remove.
20392 (vqsubq_m): Remove.
20393 (vqsubq_u8): Remove.
20394 (vqsubq_n_u8): Remove.
20395 (vqaddq_u8): Remove.
20396 (vqaddq_n_u8): Remove.
20397 (vqsubq_s8): Remove.
20398 (vqsubq_n_s8): Remove.
20399 (vqaddq_s8): Remove.
20400 (vqaddq_n_s8): Remove.
20401 (vqsubq_u16): Remove.
20402 (vqsubq_n_u16): Remove.
20403 (vqaddq_u16): Remove.
20404 (vqaddq_n_u16): Remove.
20405 (vqsubq_s16): Remove.
20406 (vqsubq_n_s16): Remove.
20407 (vqaddq_s16): Remove.
20408 (vqaddq_n_s16): Remove.
20409 (vqsubq_u32): Remove.
20410 (vqsubq_n_u32): Remove.
20411 (vqaddq_u32): Remove.
20412 (vqaddq_n_u32): Remove.
20413 (vqsubq_s32): Remove.
20414 (vqsubq_n_s32): Remove.
20415 (vqaddq_s32): Remove.
20416 (vqaddq_n_s32): Remove.
20417 (vqaddq_m_n_s8): Remove.
20418 (vqaddq_m_n_s32): Remove.
20419 (vqaddq_m_n_s16): Remove.
20420 (vqaddq_m_n_u8): Remove.
20421 (vqaddq_m_n_u32): Remove.
20422 (vqaddq_m_n_u16): Remove.
20423 (vqaddq_m_s8): Remove.
20424 (vqaddq_m_s32): Remove.
20425 (vqaddq_m_s16): Remove.
20426 (vqaddq_m_u8): Remove.
20427 (vqaddq_m_u32): Remove.
20428 (vqaddq_m_u16): Remove.
20429 (vqsubq_m_n_s8): Remove.
20430 (vqsubq_m_n_s32): Remove.
20431 (vqsubq_m_n_s16): Remove.
20432 (vqsubq_m_n_u8): Remove.
20433 (vqsubq_m_n_u32): Remove.
20434 (vqsubq_m_n_u16): Remove.
20435 (vqsubq_m_s8): Remove.
20436 (vqsubq_m_s32): Remove.
20437 (vqsubq_m_s16): Remove.
20438 (vqsubq_m_u8): Remove.
20439 (vqsubq_m_u32): Remove.
20440 (vqsubq_m_u16): Remove.
20441 (__arm_vqsubq_u8): Remove.
20442 (__arm_vqsubq_n_u8): Remove.
20443 (__arm_vqaddq_u8): Remove.
20444 (__arm_vqaddq_n_u8): Remove.
20445 (__arm_vqsubq_s8): Remove.
20446 (__arm_vqsubq_n_s8): Remove.
20447 (__arm_vqaddq_s8): Remove.
20448 (__arm_vqaddq_n_s8): Remove.
20449 (__arm_vqsubq_u16): Remove.
20450 (__arm_vqsubq_n_u16): Remove.
20451 (__arm_vqaddq_u16): Remove.
20452 (__arm_vqaddq_n_u16): Remove.
20453 (__arm_vqsubq_s16): Remove.
20454 (__arm_vqsubq_n_s16): Remove.
20455 (__arm_vqaddq_s16): Remove.
20456 (__arm_vqaddq_n_s16): Remove.
20457 (__arm_vqsubq_u32): Remove.
20458 (__arm_vqsubq_n_u32): Remove.
20459 (__arm_vqaddq_u32): Remove.
20460 (__arm_vqaddq_n_u32): Remove.
20461 (__arm_vqsubq_s32): Remove.
20462 (__arm_vqsubq_n_s32): Remove.
20463 (__arm_vqaddq_s32): Remove.
20464 (__arm_vqaddq_n_s32): Remove.
20465 (__arm_vqaddq_m_n_s8): Remove.
20466 (__arm_vqaddq_m_n_s32): Remove.
20467 (__arm_vqaddq_m_n_s16): Remove.
20468 (__arm_vqaddq_m_n_u8): Remove.
20469 (__arm_vqaddq_m_n_u32): Remove.
20470 (__arm_vqaddq_m_n_u16): Remove.
20471 (__arm_vqaddq_m_s8): Remove.
20472 (__arm_vqaddq_m_s32): Remove.
20473 (__arm_vqaddq_m_s16): Remove.
20474 (__arm_vqaddq_m_u8): Remove.
20475 (__arm_vqaddq_m_u32): Remove.
20476 (__arm_vqaddq_m_u16): Remove.
20477 (__arm_vqsubq_m_n_s8): Remove.
20478 (__arm_vqsubq_m_n_s32): Remove.
20479 (__arm_vqsubq_m_n_s16): Remove.
20480 (__arm_vqsubq_m_n_u8): Remove.
20481 (__arm_vqsubq_m_n_u32): Remove.
20482 (__arm_vqsubq_m_n_u16): Remove.
20483 (__arm_vqsubq_m_s8): Remove.
20484 (__arm_vqsubq_m_s32): Remove.
20485 (__arm_vqsubq_m_s16): Remove.
20486 (__arm_vqsubq_m_u8): Remove.
20487 (__arm_vqsubq_m_u32): Remove.
20488 (__arm_vqsubq_m_u16): Remove.
20489 (__arm_vqsubq): Remove.
20490 (__arm_vqaddq): Remove.
20491 (__arm_vqaddq_m): Remove.
20492 (__arm_vqsubq_m): Remove.
20493 (vqdmulhq): Remove.
20494 (vqdmulhq_m): Remove.
20495 (vqdmulhq_s8): Remove.
20496 (vqdmulhq_n_s8): Remove.
20497 (vqdmulhq_s16): Remove.
20498 (vqdmulhq_n_s16): Remove.
20499 (vqdmulhq_s32): Remove.
20500 (vqdmulhq_n_s32): Remove.
20501 (vqdmulhq_m_n_s8): Remove.
20502 (vqdmulhq_m_n_s32): Remove.
20503 (vqdmulhq_m_n_s16): Remove.
20504 (vqdmulhq_m_s8): Remove.
20505 (vqdmulhq_m_s32): Remove.
20506 (vqdmulhq_m_s16): Remove.
20507 (__arm_vqdmulhq_s8): Remove.
20508 (__arm_vqdmulhq_n_s8): Remove.
20509 (__arm_vqdmulhq_s16): Remove.
20510 (__arm_vqdmulhq_n_s16): Remove.
20511 (__arm_vqdmulhq_s32): Remove.
20512 (__arm_vqdmulhq_n_s32): Remove.
20513 (__arm_vqdmulhq_m_n_s8): Remove.
20514 (__arm_vqdmulhq_m_n_s32): Remove.
20515 (__arm_vqdmulhq_m_n_s16): Remove.
20516 (__arm_vqdmulhq_m_s8): Remove.
20517 (__arm_vqdmulhq_m_s32): Remove.
20518 (__arm_vqdmulhq_m_s16): Remove.
20519 (__arm_vqdmulhq): Remove.
20520 (__arm_vqdmulhq_m): Remove.
20522 (vrhaddq_m): Remove.
20523 (vrhaddq_x): Remove.
20524 (vrhaddq_u8): Remove.
20525 (vrhaddq_s8): Remove.
20526 (vrhaddq_u16): Remove.
20527 (vrhaddq_s16): Remove.
20528 (vrhaddq_u32): Remove.
20529 (vrhaddq_s32): Remove.
20530 (vrhaddq_m_s8): Remove.
20531 (vrhaddq_m_s32): Remove.
20532 (vrhaddq_m_s16): Remove.
20533 (vrhaddq_m_u8): Remove.
20534 (vrhaddq_m_u32): Remove.
20535 (vrhaddq_m_u16): Remove.
20536 (vrhaddq_x_s8): Remove.
20537 (vrhaddq_x_s16): Remove.
20538 (vrhaddq_x_s32): Remove.
20539 (vrhaddq_x_u8): Remove.
20540 (vrhaddq_x_u16): Remove.
20541 (vrhaddq_x_u32): Remove.
20542 (__arm_vrhaddq_u8): Remove.
20543 (__arm_vrhaddq_s8): Remove.
20544 (__arm_vrhaddq_u16): Remove.
20545 (__arm_vrhaddq_s16): Remove.
20546 (__arm_vrhaddq_u32): Remove.
20547 (__arm_vrhaddq_s32): Remove.
20548 (__arm_vrhaddq_m_s8): Remove.
20549 (__arm_vrhaddq_m_s32): Remove.
20550 (__arm_vrhaddq_m_s16): Remove.
20551 (__arm_vrhaddq_m_u8): Remove.
20552 (__arm_vrhaddq_m_u32): Remove.
20553 (__arm_vrhaddq_m_u16): Remove.
20554 (__arm_vrhaddq_x_s8): Remove.
20555 (__arm_vrhaddq_x_s16): Remove.
20556 (__arm_vrhaddq_x_s32): Remove.
20557 (__arm_vrhaddq_x_u8): Remove.
20558 (__arm_vrhaddq_x_u16): Remove.
20559 (__arm_vrhaddq_x_u32): Remove.
20560 (__arm_vrhaddq): Remove.
20561 (__arm_vrhaddq_m): Remove.
20562 (__arm_vrhaddq_x): Remove.
20564 (vrmulhq_m): Remove.
20565 (vrmulhq_x): Remove.
20566 (vrmulhq_u8): Remove.
20567 (vrmulhq_s8): Remove.
20568 (vrmulhq_u16): Remove.
20569 (vrmulhq_s16): Remove.
20570 (vrmulhq_u32): Remove.
20571 (vrmulhq_s32): Remove.
20572 (vrmulhq_m_s8): Remove.
20573 (vrmulhq_m_s32): Remove.
20574 (vrmulhq_m_s16): Remove.
20575 (vrmulhq_m_u8): Remove.
20576 (vrmulhq_m_u32): Remove.
20577 (vrmulhq_m_u16): Remove.
20578 (vrmulhq_x_s8): Remove.
20579 (vrmulhq_x_s16): Remove.
20580 (vrmulhq_x_s32): Remove.
20581 (vrmulhq_x_u8): Remove.
20582 (vrmulhq_x_u16): Remove.
20583 (vrmulhq_x_u32): Remove.
20584 (__arm_vrmulhq_u8): Remove.
20585 (__arm_vrmulhq_s8): Remove.
20586 (__arm_vrmulhq_u16): Remove.
20587 (__arm_vrmulhq_s16): Remove.
20588 (__arm_vrmulhq_u32): Remove.
20589 (__arm_vrmulhq_s32): Remove.
20590 (__arm_vrmulhq_m_s8): Remove.
20591 (__arm_vrmulhq_m_s32): Remove.
20592 (__arm_vrmulhq_m_s16): Remove.
20593 (__arm_vrmulhq_m_u8): Remove.
20594 (__arm_vrmulhq_m_u32): Remove.
20595 (__arm_vrmulhq_m_u16): Remove.
20596 (__arm_vrmulhq_x_s8): Remove.
20597 (__arm_vrmulhq_x_s16): Remove.
20598 (__arm_vrmulhq_x_s32): Remove.
20599 (__arm_vrmulhq_x_u8): Remove.
20600 (__arm_vrmulhq_x_u16): Remove.
20601 (__arm_vrmulhq_x_u32): Remove.
20602 (__arm_vrmulhq): Remove.
20603 (__arm_vrmulhq_m): Remove.
20604 (__arm_vrmulhq_x): Remove.
20606 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20608 * config/arm/iterators.md (MVE_INT_SU_BINARY): New.
20609 (mve_insn): Add vabdq, vhaddq, vhsubq, vmulhq, vqaddq, vqdmulhq,
20610 vqrdmulhq, vqrshlq, vqshlq, vqsubq, vrhaddq, vrmulhq, vrshlq.
20611 (supf): Add VQDMULHQ_S, VQRDMULHQ_S.
20612 * config/arm/mve.md (mve_vabdq_<supf><mode>)
20613 (@mve_vhaddq_<supf><mode>, mve_vhsubq_<supf><mode>)
20614 (mve_vmulhq_<supf><mode>, mve_vqaddq_<supf><mode>)
20615 (mve_vqdmulhq_s<mode>, mve_vqrdmulhq_s<mode>)
20616 (mve_vqrshlq_<supf><mode>, mve_vqshlq_<supf><mode>)
20617 (mve_vqsubq_<supf><mode>, @mve_vrhaddq_<supf><mode>)
20618 (mve_vrmulhq_<supf><mode>, mve_vrshlq_<supf><mode>): Merge into
20620 (@mve_<mve_insn>q_<supf><mode>): ... this.
20621 * config/arm/vec-common.md (avg<mode>3_floor, uavg<mode>3_floor)
20622 (avg<mode>3_ceil, uavg<mode>3_ceil): Use gen_mve_q instead of
20623 gen_mve_vhaddq / gen_mve_vrhaddq.
20625 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20627 * config/arm/iterators.md (MVE_INT_SU_M_N_BINARY): New.
20628 (mve_insn): Add vhaddq, vhsubq, vmlaq, vmlasq, vqaddq, vqdmlahq,
20629 vqdmlashq, vqdmulhq, vqrdmlahq, vqrdmlashq, vqrdmulhq, vqsubq.
20630 (supf): Add VQDMLAHQ_M_N_S, VQDMLASHQ_M_N_S, VQRDMLAHQ_M_N_S,
20631 VQRDMLASHQ_M_N_S, VQDMULHQ_M_N_S, VQRDMULHQ_M_N_S.
20632 * config/arm/mve.md (mve_vhaddq_m_n_<supf><mode>)
20633 (mve_vhsubq_m_n_<supf><mode>, mve_vmlaq_m_n_<supf><mode>)
20634 (mve_vmlasq_m_n_<supf><mode>, mve_vqaddq_m_n_<supf><mode>)
20635 (mve_vqdmlahq_m_n_s<mode>, mve_vqdmlashq_m_n_s<mode>)
20636 (mve_vqrdmlahq_m_n_s<mode>, mve_vqrdmlashq_m_n_s<mode>)
20637 (mve_vqsubq_m_n_<supf><mode>, mve_vqdmulhq_m_n_s<mode>)
20638 (mve_vqrdmulhq_m_n_s<mode>): Merge into ...
20639 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
20641 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20643 * config/arm/iterators.md (MVE_INT_SU_N_BINARY): New.
20644 (mve_insn): Add vhaddq, vhsubq, vqaddq, vqdmulhq, vqrdmulhq,
20646 (supf): Add VQDMULHQ_N_S, VQRDMULHQ_N_S.
20647 * config/arm/mve.md (mve_vhaddq_n_<supf><mode>)
20648 (mve_vhsubq_n_<supf><mode>, mve_vqaddq_n_<supf><mode>)
20649 (mve_vqdmulhq_n_s<mode>, mve_vqrdmulhq_n_s<mode>)
20650 (mve_vqsubq_n_<supf><mode>): Merge into ...
20651 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
20653 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20655 * config/arm/iterators.md (MVE_INT_SU_M_BINARY): New.
20656 (mve_insn): Add vabdq, vhaddq, vhsubq, vmaxq, vminq, vmulhq,
20657 vqaddq, vqdmladhq, vqdmladhxq, vqdmlsdhq, vqdmlsdhxq, vqdmulhq,
20658 vqrdmladhq, vqrdmladhxq, vqrdmlsdhq, vqrdmlsdhxq, vqrdmulhq,
20659 vqrshlq, vqshlq, vqsubq, vrhaddq, vrmulhq, vrshlq, vshlq.
20660 (supf): Add VQDMLADHQ_M_S, VQDMLADHXQ_M_S, VQDMLSDHQ_M_S,
20661 VQDMLSDHXQ_M_S, VQDMULHQ_M_S, VQRDMLADHQ_M_S, VQRDMLADHXQ_M_S,
20662 VQRDMLSDHQ_M_S, VQRDMLSDHXQ_M_S, VQRDMULHQ_M_S.
20663 * config/arm/mve.md (@mve_<mve_insn>q_m_<supf><mode>): New.
20664 (mve_vshlq_m_<supf><mode>): Merged into
20665 @mve_<mve_insn>q_m_<supf><mode>.
20666 (mve_vabdq_m_<supf><mode>): Likewise.
20667 (mve_vhaddq_m_<supf><mode>): Likewise.
20668 (mve_vhsubq_m_<supf><mode>): Likewise.
20669 (mve_vmaxq_m_<supf><mode>): Likewise.
20670 (mve_vminq_m_<supf><mode>): Likewise.
20671 (mve_vmulhq_m_<supf><mode>): Likewise.
20672 (mve_vqaddq_m_<supf><mode>): Likewise.
20673 (mve_vqrshlq_m_<supf><mode>): Likewise.
20674 (mve_vqshlq_m_<supf><mode>): Likewise.
20675 (mve_vqsubq_m_<supf><mode>): Likewise.
20676 (mve_vrhaddq_m_<supf><mode>): Likewise.
20677 (mve_vrmulhq_m_<supf><mode>): Likewise.
20678 (mve_vrshlq_m_<supf><mode>): Likewise.
20679 (mve_vqdmladhq_m_s<mode>): Likewise.
20680 (mve_vqdmladhxq_m_s<mode>): Likewise.
20681 (mve_vqdmlsdhq_m_s<mode>): Likewise.
20682 (mve_vqdmlsdhxq_m_s<mode>): Likewise.
20683 (mve_vqdmulhq_m_s<mode>): Likewise.
20684 (mve_vqrdmladhq_m_s<mode>): Likewise.
20685 (mve_vqrdmladhxq_m_s<mode>): Likewise.
20686 (mve_vqrdmlsdhq_m_s<mode>): Likewise.
20687 (mve_vqrdmlsdhxq_m_s<mode>): Likewise.
20688 (mve_vqrdmulhq_m_s<mode>): Likewise.
20690 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20692 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITHOUT_M_N): New. (vcreateq): New.
20693 * config/arm/arm-mve-builtins-base.def (vcreateq): New.
20694 * config/arm/arm-mve-builtins-base.h (vcreateq): New.
20695 * config/arm/arm_mve.h (vcreateq_f16): Remove.
20696 (vcreateq_f32): Remove.
20697 (vcreateq_u8): Remove.
20698 (vcreateq_u16): Remove.
20699 (vcreateq_u32): Remove.
20700 (vcreateq_u64): Remove.
20701 (vcreateq_s8): Remove.
20702 (vcreateq_s16): Remove.
20703 (vcreateq_s32): Remove.
20704 (vcreateq_s64): Remove.
20705 (__arm_vcreateq_u8): Remove.
20706 (__arm_vcreateq_u16): Remove.
20707 (__arm_vcreateq_u32): Remove.
20708 (__arm_vcreateq_u64): Remove.
20709 (__arm_vcreateq_s8): Remove.
20710 (__arm_vcreateq_s16): Remove.
20711 (__arm_vcreateq_s32): Remove.
20712 (__arm_vcreateq_s64): Remove.
20713 (__arm_vcreateq_f16): Remove.
20714 (__arm_vcreateq_f32): Remove.
20716 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20718 * config/arm/iterators.md (MVE_FP_CREATE_ONLY): New.
20719 (mve_insn): Add VCREATEQ_S, VCREATEQ_U, VCREATEQ_F.
20720 * config/arm/mve.md (mve_vcreateq_f<mode>): Rename into ...
20721 (@mve_<mve_insn>q_f<mode>): ... this.
20722 (mve_vcreateq_<supf><mode>): Rename into ...
20723 (@mve_<mve_insn>q_<supf><mode>): ... this.
20725 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20727 * config/arm/arm-mve-builtins-shapes.cc (create): New.
20728 * config/arm/arm-mve-builtins-shapes.h: (create): New.
20730 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20732 * config/arm/arm-mve-builtins-functions.h (class
20733 unspec_mve_function_exact_insn): New.
20735 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20737 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITH_RTX_M_N_NO_N_F): New.
20739 * config/arm/arm-mve-builtins-base.def (vorrq): New.
20740 * config/arm/arm-mve-builtins-base.h (vorrq): New.
20741 * config/arm/arm-mve-builtins.cc
20742 (function_instance::has_inactive_argument): Handle vorrq.
20743 * config/arm/arm_mve.h (vorrq): Remove.
20744 (vorrq_m_n): Remove.
20747 (vorrq_u8): Remove.
20748 (vorrq_s8): Remove.
20749 (vorrq_u16): Remove.
20750 (vorrq_s16): Remove.
20751 (vorrq_u32): Remove.
20752 (vorrq_s32): Remove.
20753 (vorrq_n_u16): Remove.
20754 (vorrq_f16): Remove.
20755 (vorrq_n_s16): Remove.
20756 (vorrq_n_u32): Remove.
20757 (vorrq_f32): Remove.
20758 (vorrq_n_s32): Remove.
20759 (vorrq_m_n_s16): Remove.
20760 (vorrq_m_n_u16): Remove.
20761 (vorrq_m_n_s32): Remove.
20762 (vorrq_m_n_u32): Remove.
20763 (vorrq_m_s8): Remove.
20764 (vorrq_m_s32): Remove.
20765 (vorrq_m_s16): Remove.
20766 (vorrq_m_u8): Remove.
20767 (vorrq_m_u32): Remove.
20768 (vorrq_m_u16): Remove.
20769 (vorrq_m_f32): Remove.
20770 (vorrq_m_f16): Remove.
20771 (vorrq_x_s8): Remove.
20772 (vorrq_x_s16): Remove.
20773 (vorrq_x_s32): Remove.
20774 (vorrq_x_u8): Remove.
20775 (vorrq_x_u16): Remove.
20776 (vorrq_x_u32): Remove.
20777 (vorrq_x_f16): Remove.
20778 (vorrq_x_f32): Remove.
20779 (__arm_vorrq_u8): Remove.
20780 (__arm_vorrq_s8): Remove.
20781 (__arm_vorrq_u16): Remove.
20782 (__arm_vorrq_s16): Remove.
20783 (__arm_vorrq_u32): Remove.
20784 (__arm_vorrq_s32): Remove.
20785 (__arm_vorrq_n_u16): Remove.
20786 (__arm_vorrq_n_s16): Remove.
20787 (__arm_vorrq_n_u32): Remove.
20788 (__arm_vorrq_n_s32): Remove.
20789 (__arm_vorrq_m_n_s16): Remove.
20790 (__arm_vorrq_m_n_u16): Remove.
20791 (__arm_vorrq_m_n_s32): Remove.
20792 (__arm_vorrq_m_n_u32): Remove.
20793 (__arm_vorrq_m_s8): Remove.
20794 (__arm_vorrq_m_s32): Remove.
20795 (__arm_vorrq_m_s16): Remove.
20796 (__arm_vorrq_m_u8): Remove.
20797 (__arm_vorrq_m_u32): Remove.
20798 (__arm_vorrq_m_u16): Remove.
20799 (__arm_vorrq_x_s8): Remove.
20800 (__arm_vorrq_x_s16): Remove.
20801 (__arm_vorrq_x_s32): Remove.
20802 (__arm_vorrq_x_u8): Remove.
20803 (__arm_vorrq_x_u16): Remove.
20804 (__arm_vorrq_x_u32): Remove.
20805 (__arm_vorrq_f16): Remove.
20806 (__arm_vorrq_f32): Remove.
20807 (__arm_vorrq_m_f32): Remove.
20808 (__arm_vorrq_m_f16): Remove.
20809 (__arm_vorrq_x_f16): Remove.
20810 (__arm_vorrq_x_f32): Remove.
20811 (__arm_vorrq): Remove.
20812 (__arm_vorrq_m_n): Remove.
20813 (__arm_vorrq_m): Remove.
20814 (__arm_vorrq_x): Remove.
20816 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20818 * config/arm/arm-mve-builtins-shapes.cc (binary_orrq): New.
20819 * config/arm/arm-mve-builtins-shapes.h (binary_orrq): New.
20820 * config/arm/arm-mve-builtins.cc (preds_m_or_none): Remove static.
20821 * config/arm/arm-mve-builtins.h (preds_m_or_none): Declare.
20823 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20825 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITH_RTX_M): New.
20826 (vandq,veorq): New.
20827 * config/arm/arm-mve-builtins-base.def (vandq, veorq): New.
20828 * config/arm/arm-mve-builtins-base.h (vandq, veorq): New.
20829 * config/arm/arm_mve.h (vandq): Remove.
20832 (vandq_u8): Remove.
20833 (vandq_s8): Remove.
20834 (vandq_u16): Remove.
20835 (vandq_s16): Remove.
20836 (vandq_u32): Remove.
20837 (vandq_s32): Remove.
20838 (vandq_f16): Remove.
20839 (vandq_f32): Remove.
20840 (vandq_m_s8): Remove.
20841 (vandq_m_s32): Remove.
20842 (vandq_m_s16): Remove.
20843 (vandq_m_u8): Remove.
20844 (vandq_m_u32): Remove.
20845 (vandq_m_u16): Remove.
20846 (vandq_m_f32): Remove.
20847 (vandq_m_f16): Remove.
20848 (vandq_x_s8): Remove.
20849 (vandq_x_s16): Remove.
20850 (vandq_x_s32): Remove.
20851 (vandq_x_u8): Remove.
20852 (vandq_x_u16): Remove.
20853 (vandq_x_u32): Remove.
20854 (vandq_x_f16): Remove.
20855 (vandq_x_f32): Remove.
20856 (__arm_vandq_u8): Remove.
20857 (__arm_vandq_s8): Remove.
20858 (__arm_vandq_u16): Remove.
20859 (__arm_vandq_s16): Remove.
20860 (__arm_vandq_u32): Remove.
20861 (__arm_vandq_s32): Remove.
20862 (__arm_vandq_m_s8): Remove.
20863 (__arm_vandq_m_s32): Remove.
20864 (__arm_vandq_m_s16): Remove.
20865 (__arm_vandq_m_u8): Remove.
20866 (__arm_vandq_m_u32): Remove.
20867 (__arm_vandq_m_u16): Remove.
20868 (__arm_vandq_x_s8): Remove.
20869 (__arm_vandq_x_s16): Remove.
20870 (__arm_vandq_x_s32): Remove.
20871 (__arm_vandq_x_u8): Remove.
20872 (__arm_vandq_x_u16): Remove.
20873 (__arm_vandq_x_u32): Remove.
20874 (__arm_vandq_f16): Remove.
20875 (__arm_vandq_f32): Remove.
20876 (__arm_vandq_m_f32): Remove.
20877 (__arm_vandq_m_f16): Remove.
20878 (__arm_vandq_x_f16): Remove.
20879 (__arm_vandq_x_f32): Remove.
20880 (__arm_vandq): Remove.
20881 (__arm_vandq_m): Remove.
20882 (__arm_vandq_x): Remove.
20885 (veorq_u8): Remove.
20886 (veorq_s8): Remove.
20887 (veorq_u16): Remove.
20888 (veorq_s16): Remove.
20889 (veorq_u32): Remove.
20890 (veorq_s32): Remove.
20891 (veorq_f16): Remove.
20892 (veorq_f32): Remove.
20893 (veorq_m_s8): Remove.
20894 (veorq_m_s32): Remove.
20895 (veorq_m_s16): Remove.
20896 (veorq_m_u8): Remove.
20897 (veorq_m_u32): Remove.
20898 (veorq_m_u16): Remove.
20899 (veorq_m_f32): Remove.
20900 (veorq_m_f16): Remove.
20901 (veorq_x_s8): Remove.
20902 (veorq_x_s16): Remove.
20903 (veorq_x_s32): Remove.
20904 (veorq_x_u8): Remove.
20905 (veorq_x_u16): Remove.
20906 (veorq_x_u32): Remove.
20907 (veorq_x_f16): Remove.
20908 (veorq_x_f32): Remove.
20909 (__arm_veorq_u8): Remove.
20910 (__arm_veorq_s8): Remove.
20911 (__arm_veorq_u16): Remove.
20912 (__arm_veorq_s16): Remove.
20913 (__arm_veorq_u32): Remove.
20914 (__arm_veorq_s32): Remove.
20915 (__arm_veorq_m_s8): Remove.
20916 (__arm_veorq_m_s32): Remove.
20917 (__arm_veorq_m_s16): Remove.
20918 (__arm_veorq_m_u8): Remove.
20919 (__arm_veorq_m_u32): Remove.
20920 (__arm_veorq_m_u16): Remove.
20921 (__arm_veorq_x_s8): Remove.
20922 (__arm_veorq_x_s16): Remove.
20923 (__arm_veorq_x_s32): Remove.
20924 (__arm_veorq_x_u8): Remove.
20925 (__arm_veorq_x_u16): Remove.
20926 (__arm_veorq_x_u32): Remove.
20927 (__arm_veorq_f16): Remove.
20928 (__arm_veorq_f32): Remove.
20929 (__arm_veorq_m_f32): Remove.
20930 (__arm_veorq_m_f16): Remove.
20931 (__arm_veorq_x_f16): Remove.
20932 (__arm_veorq_x_f32): Remove.
20933 (__arm_veorq): Remove.
20934 (__arm_veorq_m): Remove.
20935 (__arm_veorq_x): Remove.
20937 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20939 * config/arm/iterators.md (MVE_INT_M_BINARY_LOGIC)
20940 (MVE_FP_M_BINARY_LOGIC): New.
20941 (MVE_INT_M_N_BINARY_LOGIC): New.
20942 (MVE_INT_N_BINARY_LOGIC): New.
20943 (mve_insn): Add vand, veor, vorr, vbic.
20944 * config/arm/mve.md (mve_vandq_m_<supf><mode>)
20945 (mve_veorq_m_<supf><mode>, mve_vorrq_m_<supf><mode>)
20946 (mve_vbicq_m_<supf><mode>): Merge into ...
20947 (@mve_<mve_insn>q_m_<supf><mode>): ... this.
20948 (mve_vandq_m_f<mode>, mve_veorq_m_f<mode>, mve_vorrq_m_f<mode>)
20949 (mve_vbicq_m_f<mode>): Merge into ...
20950 (@mve_<mve_insn>q_m_f<mode>): ... this.
20951 (mve_vorrq_n_<supf><mode>)
20952 (mve_vbicq_n_<supf><mode>): Merge into ...
20953 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
20954 (mve_vorrq_m_n_<supf><mode>, mve_vbicq_m_n_<supf><mode>): Merge
20956 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
20958 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20960 * config/arm/arm-mve-builtins-shapes.cc (binary): New.
20961 * config/arm/arm-mve-builtins-shapes.h (binary): New.
20963 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20965 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITH_RTX_M_N):
20967 (vaddq, vmulq, vsubq): New.
20968 * config/arm/arm-mve-builtins-base.def (vaddq, vmulq, vsubq): New.
20969 * config/arm/arm-mve-builtins-base.h (vaddq, vmulq, vsubq): New.
20970 * config/arm/arm_mve.h (vaddq): Remove.
20973 (vaddq_n_u8): Remove.
20974 (vaddq_n_s8): Remove.
20975 (vaddq_n_u16): Remove.
20976 (vaddq_n_s16): Remove.
20977 (vaddq_n_u32): Remove.
20978 (vaddq_n_s32): Remove.
20979 (vaddq_n_f16): Remove.
20980 (vaddq_n_f32): Remove.
20981 (vaddq_m_n_s8): Remove.
20982 (vaddq_m_n_s32): Remove.
20983 (vaddq_m_n_s16): Remove.
20984 (vaddq_m_n_u8): Remove.
20985 (vaddq_m_n_u32): Remove.
20986 (vaddq_m_n_u16): Remove.
20987 (vaddq_m_s8): Remove.
20988 (vaddq_m_s32): Remove.
20989 (vaddq_m_s16): Remove.
20990 (vaddq_m_u8): Remove.
20991 (vaddq_m_u32): Remove.
20992 (vaddq_m_u16): Remove.
20993 (vaddq_m_f32): Remove.
20994 (vaddq_m_f16): Remove.
20995 (vaddq_m_n_f32): Remove.
20996 (vaddq_m_n_f16): Remove.
20997 (vaddq_s8): Remove.
20998 (vaddq_s16): Remove.
20999 (vaddq_s32): Remove.
21000 (vaddq_u8): Remove.
21001 (vaddq_u16): Remove.
21002 (vaddq_u32): Remove.
21003 (vaddq_f16): Remove.
21004 (vaddq_f32): Remove.
21005 (vaddq_x_s8): Remove.
21006 (vaddq_x_s16): Remove.
21007 (vaddq_x_s32): Remove.
21008 (vaddq_x_n_s8): Remove.
21009 (vaddq_x_n_s16): Remove.
21010 (vaddq_x_n_s32): Remove.
21011 (vaddq_x_u8): Remove.
21012 (vaddq_x_u16): Remove.
21013 (vaddq_x_u32): Remove.
21014 (vaddq_x_n_u8): Remove.
21015 (vaddq_x_n_u16): Remove.
21016 (vaddq_x_n_u32): Remove.
21017 (vaddq_x_f16): Remove.
21018 (vaddq_x_f32): Remove.
21019 (vaddq_x_n_f16): Remove.
21020 (vaddq_x_n_f32): Remove.
21021 (__arm_vaddq_n_u8): Remove.
21022 (__arm_vaddq_n_s8): Remove.
21023 (__arm_vaddq_n_u16): Remove.
21024 (__arm_vaddq_n_s16): Remove.
21025 (__arm_vaddq_n_u32): Remove.
21026 (__arm_vaddq_n_s32): Remove.
21027 (__arm_vaddq_m_n_s8): Remove.
21028 (__arm_vaddq_m_n_s32): Remove.
21029 (__arm_vaddq_m_n_s16): Remove.
21030 (__arm_vaddq_m_n_u8): Remove.
21031 (__arm_vaddq_m_n_u32): Remove.
21032 (__arm_vaddq_m_n_u16): Remove.
21033 (__arm_vaddq_m_s8): Remove.
21034 (__arm_vaddq_m_s32): Remove.
21035 (__arm_vaddq_m_s16): Remove.
21036 (__arm_vaddq_m_u8): Remove.
21037 (__arm_vaddq_m_u32): Remove.
21038 (__arm_vaddq_m_u16): Remove.
21039 (__arm_vaddq_s8): Remove.
21040 (__arm_vaddq_s16): Remove.
21041 (__arm_vaddq_s32): Remove.
21042 (__arm_vaddq_u8): Remove.
21043 (__arm_vaddq_u16): Remove.
21044 (__arm_vaddq_u32): Remove.
21045 (__arm_vaddq_x_s8): Remove.
21046 (__arm_vaddq_x_s16): Remove.
21047 (__arm_vaddq_x_s32): Remove.
21048 (__arm_vaddq_x_n_s8): Remove.
21049 (__arm_vaddq_x_n_s16): Remove.
21050 (__arm_vaddq_x_n_s32): Remove.
21051 (__arm_vaddq_x_u8): Remove.
21052 (__arm_vaddq_x_u16): Remove.
21053 (__arm_vaddq_x_u32): Remove.
21054 (__arm_vaddq_x_n_u8): Remove.
21055 (__arm_vaddq_x_n_u16): Remove.
21056 (__arm_vaddq_x_n_u32): Remove.
21057 (__arm_vaddq_n_f16): Remove.
21058 (__arm_vaddq_n_f32): Remove.
21059 (__arm_vaddq_m_f32): Remove.
21060 (__arm_vaddq_m_f16): Remove.
21061 (__arm_vaddq_m_n_f32): Remove.
21062 (__arm_vaddq_m_n_f16): Remove.
21063 (__arm_vaddq_f16): Remove.
21064 (__arm_vaddq_f32): Remove.
21065 (__arm_vaddq_x_f16): Remove.
21066 (__arm_vaddq_x_f32): Remove.
21067 (__arm_vaddq_x_n_f16): Remove.
21068 (__arm_vaddq_x_n_f32): Remove.
21069 (__arm_vaddq): Remove.
21070 (__arm_vaddq_m): Remove.
21071 (__arm_vaddq_x): Remove.
21075 (vmulq_u8): Remove.
21076 (vmulq_n_u8): Remove.
21077 (vmulq_s8): Remove.
21078 (vmulq_n_s8): Remove.
21079 (vmulq_u16): Remove.
21080 (vmulq_n_u16): Remove.
21081 (vmulq_s16): Remove.
21082 (vmulq_n_s16): Remove.
21083 (vmulq_u32): Remove.
21084 (vmulq_n_u32): Remove.
21085 (vmulq_s32): Remove.
21086 (vmulq_n_s32): Remove.
21087 (vmulq_n_f16): Remove.
21088 (vmulq_f16): Remove.
21089 (vmulq_n_f32): Remove.
21090 (vmulq_f32): Remove.
21091 (vmulq_m_n_s8): Remove.
21092 (vmulq_m_n_s32): Remove.
21093 (vmulq_m_n_s16): Remove.
21094 (vmulq_m_n_u8): Remove.
21095 (vmulq_m_n_u32): Remove.
21096 (vmulq_m_n_u16): Remove.
21097 (vmulq_m_s8): Remove.
21098 (vmulq_m_s32): Remove.
21099 (vmulq_m_s16): Remove.
21100 (vmulq_m_u8): Remove.
21101 (vmulq_m_u32): Remove.
21102 (vmulq_m_u16): Remove.
21103 (vmulq_m_f32): Remove.
21104 (vmulq_m_f16): Remove.
21105 (vmulq_m_n_f32): Remove.
21106 (vmulq_m_n_f16): Remove.
21107 (vmulq_x_s8): Remove.
21108 (vmulq_x_s16): Remove.
21109 (vmulq_x_s32): Remove.
21110 (vmulq_x_n_s8): Remove.
21111 (vmulq_x_n_s16): Remove.
21112 (vmulq_x_n_s32): Remove.
21113 (vmulq_x_u8): Remove.
21114 (vmulq_x_u16): Remove.
21115 (vmulq_x_u32): Remove.
21116 (vmulq_x_n_u8): Remove.
21117 (vmulq_x_n_u16): Remove.
21118 (vmulq_x_n_u32): Remove.
21119 (vmulq_x_f16): Remove.
21120 (vmulq_x_f32): Remove.
21121 (vmulq_x_n_f16): Remove.
21122 (vmulq_x_n_f32): Remove.
21123 (__arm_vmulq_u8): Remove.
21124 (__arm_vmulq_n_u8): Remove.
21125 (__arm_vmulq_s8): Remove.
21126 (__arm_vmulq_n_s8): Remove.
21127 (__arm_vmulq_u16): Remove.
21128 (__arm_vmulq_n_u16): Remove.
21129 (__arm_vmulq_s16): Remove.
21130 (__arm_vmulq_n_s16): Remove.
21131 (__arm_vmulq_u32): Remove.
21132 (__arm_vmulq_n_u32): Remove.
21133 (__arm_vmulq_s32): Remove.
21134 (__arm_vmulq_n_s32): Remove.
21135 (__arm_vmulq_m_n_s8): Remove.
21136 (__arm_vmulq_m_n_s32): Remove.
21137 (__arm_vmulq_m_n_s16): Remove.
21138 (__arm_vmulq_m_n_u8): Remove.
21139 (__arm_vmulq_m_n_u32): Remove.
21140 (__arm_vmulq_m_n_u16): Remove.
21141 (__arm_vmulq_m_s8): Remove.
21142 (__arm_vmulq_m_s32): Remove.
21143 (__arm_vmulq_m_s16): Remove.
21144 (__arm_vmulq_m_u8): Remove.
21145 (__arm_vmulq_m_u32): Remove.
21146 (__arm_vmulq_m_u16): Remove.
21147 (__arm_vmulq_x_s8): Remove.
21148 (__arm_vmulq_x_s16): Remove.
21149 (__arm_vmulq_x_s32): Remove.
21150 (__arm_vmulq_x_n_s8): Remove.
21151 (__arm_vmulq_x_n_s16): Remove.
21152 (__arm_vmulq_x_n_s32): Remove.
21153 (__arm_vmulq_x_u8): Remove.
21154 (__arm_vmulq_x_u16): Remove.
21155 (__arm_vmulq_x_u32): Remove.
21156 (__arm_vmulq_x_n_u8): Remove.
21157 (__arm_vmulq_x_n_u16): Remove.
21158 (__arm_vmulq_x_n_u32): Remove.
21159 (__arm_vmulq_n_f16): Remove.
21160 (__arm_vmulq_f16): Remove.
21161 (__arm_vmulq_n_f32): Remove.
21162 (__arm_vmulq_f32): Remove.
21163 (__arm_vmulq_m_f32): Remove.
21164 (__arm_vmulq_m_f16): Remove.
21165 (__arm_vmulq_m_n_f32): Remove.
21166 (__arm_vmulq_m_n_f16): Remove.
21167 (__arm_vmulq_x_f16): Remove.
21168 (__arm_vmulq_x_f32): Remove.
21169 (__arm_vmulq_x_n_f16): Remove.
21170 (__arm_vmulq_x_n_f32): Remove.
21171 (__arm_vmulq): Remove.
21172 (__arm_vmulq_m): Remove.
21173 (__arm_vmulq_x): Remove.
21177 (vsubq_n_f16): Remove.
21178 (vsubq_n_f32): Remove.
21179 (vsubq_u8): Remove.
21180 (vsubq_n_u8): Remove.
21181 (vsubq_s8): Remove.
21182 (vsubq_n_s8): Remove.
21183 (vsubq_u16): Remove.
21184 (vsubq_n_u16): Remove.
21185 (vsubq_s16): Remove.
21186 (vsubq_n_s16): Remove.
21187 (vsubq_u32): Remove.
21188 (vsubq_n_u32): Remove.
21189 (vsubq_s32): Remove.
21190 (vsubq_n_s32): Remove.
21191 (vsubq_f16): Remove.
21192 (vsubq_f32): Remove.
21193 (vsubq_m_s8): Remove.
21194 (vsubq_m_u8): Remove.
21195 (vsubq_m_s16): Remove.
21196 (vsubq_m_u16): Remove.
21197 (vsubq_m_s32): Remove.
21198 (vsubq_m_u32): Remove.
21199 (vsubq_m_n_s8): Remove.
21200 (vsubq_m_n_s32): Remove.
21201 (vsubq_m_n_s16): Remove.
21202 (vsubq_m_n_u8): Remove.
21203 (vsubq_m_n_u32): Remove.
21204 (vsubq_m_n_u16): Remove.
21205 (vsubq_m_f32): Remove.
21206 (vsubq_m_f16): Remove.
21207 (vsubq_m_n_f32): Remove.
21208 (vsubq_m_n_f16): Remove.
21209 (vsubq_x_s8): Remove.
21210 (vsubq_x_s16): Remove.
21211 (vsubq_x_s32): Remove.
21212 (vsubq_x_n_s8): Remove.
21213 (vsubq_x_n_s16): Remove.
21214 (vsubq_x_n_s32): Remove.
21215 (vsubq_x_u8): Remove.
21216 (vsubq_x_u16): Remove.
21217 (vsubq_x_u32): Remove.
21218 (vsubq_x_n_u8): Remove.
21219 (vsubq_x_n_u16): Remove.
21220 (vsubq_x_n_u32): Remove.
21221 (vsubq_x_f16): Remove.
21222 (vsubq_x_f32): Remove.
21223 (vsubq_x_n_f16): Remove.
21224 (vsubq_x_n_f32): Remove.
21225 (__arm_vsubq_u8): Remove.
21226 (__arm_vsubq_n_u8): Remove.
21227 (__arm_vsubq_s8): Remove.
21228 (__arm_vsubq_n_s8): Remove.
21229 (__arm_vsubq_u16): Remove.
21230 (__arm_vsubq_n_u16): Remove.
21231 (__arm_vsubq_s16): Remove.
21232 (__arm_vsubq_n_s16): Remove.
21233 (__arm_vsubq_u32): Remove.
21234 (__arm_vsubq_n_u32): Remove.
21235 (__arm_vsubq_s32): Remove.
21236 (__arm_vsubq_n_s32): Remove.
21237 (__arm_vsubq_m_s8): Remove.
21238 (__arm_vsubq_m_u8): Remove.
21239 (__arm_vsubq_m_s16): Remove.
21240 (__arm_vsubq_m_u16): Remove.
21241 (__arm_vsubq_m_s32): Remove.
21242 (__arm_vsubq_m_u32): Remove.
21243 (__arm_vsubq_m_n_s8): Remove.
21244 (__arm_vsubq_m_n_s32): Remove.
21245 (__arm_vsubq_m_n_s16): Remove.
21246 (__arm_vsubq_m_n_u8): Remove.
21247 (__arm_vsubq_m_n_u32): Remove.
21248 (__arm_vsubq_m_n_u16): Remove.
21249 (__arm_vsubq_x_s8): Remove.
21250 (__arm_vsubq_x_s16): Remove.
21251 (__arm_vsubq_x_s32): Remove.
21252 (__arm_vsubq_x_n_s8): Remove.
21253 (__arm_vsubq_x_n_s16): Remove.
21254 (__arm_vsubq_x_n_s32): Remove.
21255 (__arm_vsubq_x_u8): Remove.
21256 (__arm_vsubq_x_u16): Remove.
21257 (__arm_vsubq_x_u32): Remove.
21258 (__arm_vsubq_x_n_u8): Remove.
21259 (__arm_vsubq_x_n_u16): Remove.
21260 (__arm_vsubq_x_n_u32): Remove.
21261 (__arm_vsubq_n_f16): Remove.
21262 (__arm_vsubq_n_f32): Remove.
21263 (__arm_vsubq_f16): Remove.
21264 (__arm_vsubq_f32): Remove.
21265 (__arm_vsubq_m_f32): Remove.
21266 (__arm_vsubq_m_f16): Remove.
21267 (__arm_vsubq_m_n_f32): Remove.
21268 (__arm_vsubq_m_n_f16): Remove.
21269 (__arm_vsubq_x_f16): Remove.
21270 (__arm_vsubq_x_f32): Remove.
21271 (__arm_vsubq_x_n_f16): Remove.
21272 (__arm_vsubq_x_n_f32): Remove.
21273 (__arm_vsubq): Remove.
21274 (__arm_vsubq_m): Remove.
21275 (__arm_vsubq_x): Remove.
21276 * config/arm/arm_mve_builtins.def (vsubq_u, vsubq_s, vsubq_f):
21278 (vmulq_u, vmulq_s, vmulq_f): Remove.
21279 * config/arm/mve.md (mve_vsubq_<supf><mode>): Remove.
21280 (mve_vmulq_<supf><mode>): Remove.
21282 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
21284 * config/arm/iterators.md (MVE_INT_BINARY_RTX, MVE_INT_M_BINARY)
21285 (MVE_INT_M_N_BINARY, MVE_INT_N_BINARY, MVE_FP_M_BINARY)
21286 (MVE_FP_M_N_BINARY, MVE_FP_N_BINARY, mve_addsubmul, mve_insn): New
21288 * config/arm/mve.md
21289 (mve_vsubq_n_f<mode>, mve_vaddq_n_f<mode>, mve_vmulq_n_f<mode>):
21291 (@mve_<mve_insn>q_n_f<mode>): ... this.
21292 (mve_vaddq_n_<supf><mode>, mve_vmulq_n_<supf><mode>)
21293 (mve_vsubq_n_<supf><mode>): Factorize into ...
21294 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
21295 (mve_vaddq<mode>, mve_vmulq<mode>, mve_vsubq<mode>): Factorize
21297 (mve_<mve_addsubmul>q<mode>): ... this.
21298 (mve_vaddq_f<mode>, mve_vmulq_f<mode>, mve_vsubq_f<mode>):
21300 (mve_<mve_addsubmul>q_f<mode>): ... this.
21301 (mve_vaddq_m_<supf><mode>, mve_vmulq_m_<supf><mode>)
21302 (mve_vsubq_m_<supf><mode>): Factorize into ...
21303 (@mve_<mve_insn>q_m_<supf><mode>): ... this,
21304 (mve_vaddq_m_n_<supf><mode>, mve_vmulq_m_n_<supf><mode>)
21305 (mve_vsubq_m_n_<supf><mode>): Factorize into ...
21306 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
21307 (mve_vaddq_m_f<mode>, mve_vmulq_m_f<mode>, mve_vsubq_m_f<mode>):
21309 (@mve_<mve_insn>q_m_f<mode>): ... this.
21310 (mve_vaddq_m_n_f<mode>, mve_vmulq_m_n_f<mode>)
21311 (mve_vsubq_m_n_f<mode>): Factorize into ...
21312 (@mve_<mve_insn>q_m_n_f<mode>): ... this.
21314 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
21316 * config/arm/arm-mve-builtins-functions.h (class
21317 unspec_based_mve_function_base): New.
21318 (class unspec_based_mve_function_exact_insn): New.
21320 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
21322 * config/arm/arm-mve-builtins-shapes.cc (binary_opt_n): New.
21323 * config/arm/arm-mve-builtins-shapes.h (binary_opt_n): New.
21325 2023-05-03 Murray Steele <murray.steele@arm.com>
21326 Christophe Lyon <christophe.lyon@arm.com>
21328 * config/arm/arm-mve-builtins-base.cc (class
21329 vuninitializedq_impl): New.
21330 * config/arm/arm-mve-builtins-base.def (vuninitializedq): New.
21331 * config/arm/arm-mve-builtins-base.h (vuninitializedq): New
21333 * config/arm/arm-mve-builtins-shapes.cc (inherent): New.
21334 * config/arm/arm-mve-builtins-shapes.h (inherent): New
21336 * config/arm/arm_mve_types.h (__arm_vuninitializedq): Move to ...
21337 * config/arm/arm_mve.h (__arm_vuninitializedq): ... here.
21338 (__arm_vuninitializedq_u8): Remove.
21339 (__arm_vuninitializedq_u16): Remove.
21340 (__arm_vuninitializedq_u32): Remove.
21341 (__arm_vuninitializedq_u64): Remove.
21342 (__arm_vuninitializedq_s8): Remove.
21343 (__arm_vuninitializedq_s16): Remove.
21344 (__arm_vuninitializedq_s32): Remove.
21345 (__arm_vuninitializedq_s64): Remove.
21346 (__arm_vuninitializedq_f16): Remove.
21347 (__arm_vuninitializedq_f32): Remove.
21349 2023-05-03 Murray Steele <murray.steele@arm.com>
21350 Christophe Lyon <christophe.lyon@arm.com>
21352 * config/arm/arm-mve-builtins-base.cc (vreinterpretq_impl): New class.
21353 * config/arm/arm-mve-builtins-base.def: Define vreinterpretq.
21354 * config/arm/arm-mve-builtins-base.h (vreinterpretq): New declaration.
21355 * config/arm/arm-mve-builtins-shapes.cc (parse_element_type): New function.
21356 (parse_type): Likewise.
21357 (parse_signature): Likewise.
21358 (build_one): Likewise.
21359 (build_all): Likewise.
21360 (overloaded_base): New struct.
21361 (unary_convert_def): Likewise.
21362 * config/arm/arm-mve-builtins-shapes.h (unary_convert): Declare.
21363 * config/arm/arm-mve-builtins.cc (TYPES_reinterpret_signed1): New
21365 (TYPES_reinterpret_unsigned1): Likewise.
21366 (TYPES_reinterpret_integer): Likewise.
21367 (TYPES_reinterpret_integer1): Likewise.
21368 (TYPES_reinterpret_float1): Likewise.
21369 (TYPES_reinterpret_float): Likewise.
21370 (reinterpret_integer): New.
21371 (reinterpret_float): New.
21372 (handle_arm_mve_h): Register builtins.
21373 * config/arm/arm_mve.h (vreinterpretq_s16): Remove.
21374 (vreinterpretq_s32): Likewise.
21375 (vreinterpretq_s64): Likewise.
21376 (vreinterpretq_s8): Likewise.
21377 (vreinterpretq_u16): Likewise.
21378 (vreinterpretq_u32): Likewise.
21379 (vreinterpretq_u64): Likewise.
21380 (vreinterpretq_u8): Likewise.
21381 (vreinterpretq_f16): Likewise.
21382 (vreinterpretq_f32): Likewise.
21383 (vreinterpretq_s16_s32): Likewise.
21384 (vreinterpretq_s16_s64): Likewise.
21385 (vreinterpretq_s16_s8): Likewise.
21386 (vreinterpretq_s16_u16): Likewise.
21387 (vreinterpretq_s16_u32): Likewise.
21388 (vreinterpretq_s16_u64): Likewise.
21389 (vreinterpretq_s16_u8): Likewise.
21390 (vreinterpretq_s32_s16): Likewise.
21391 (vreinterpretq_s32_s64): Likewise.
21392 (vreinterpretq_s32_s8): Likewise.
21393 (vreinterpretq_s32_u16): Likewise.
21394 (vreinterpretq_s32_u32): Likewise.
21395 (vreinterpretq_s32_u64): Likewise.
21396 (vreinterpretq_s32_u8): Likewise.
21397 (vreinterpretq_s64_s16): Likewise.
21398 (vreinterpretq_s64_s32): Likewise.
21399 (vreinterpretq_s64_s8): Likewise.
21400 (vreinterpretq_s64_u16): Likewise.
21401 (vreinterpretq_s64_u32): Likewise.
21402 (vreinterpretq_s64_u64): Likewise.
21403 (vreinterpretq_s64_u8): Likewise.
21404 (vreinterpretq_s8_s16): Likewise.
21405 (vreinterpretq_s8_s32): Likewise.
21406 (vreinterpretq_s8_s64): Likewise.
21407 (vreinterpretq_s8_u16): Likewise.
21408 (vreinterpretq_s8_u32): Likewise.
21409 (vreinterpretq_s8_u64): Likewise.
21410 (vreinterpretq_s8_u8): Likewise.
21411 (vreinterpretq_u16_s16): Likewise.
21412 (vreinterpretq_u16_s32): Likewise.
21413 (vreinterpretq_u16_s64): Likewise.
21414 (vreinterpretq_u16_s8): Likewise.
21415 (vreinterpretq_u16_u32): Likewise.
21416 (vreinterpretq_u16_u64): Likewise.
21417 (vreinterpretq_u16_u8): Likewise.
21418 (vreinterpretq_u32_s16): Likewise.
21419 (vreinterpretq_u32_s32): Likewise.
21420 (vreinterpretq_u32_s64): Likewise.
21421 (vreinterpretq_u32_s8): Likewise.
21422 (vreinterpretq_u32_u16): Likewise.
21423 (vreinterpretq_u32_u64): Likewise.
21424 (vreinterpretq_u32_u8): Likewise.
21425 (vreinterpretq_u64_s16): Likewise.
21426 (vreinterpretq_u64_s32): Likewise.
21427 (vreinterpretq_u64_s64): Likewise.
21428 (vreinterpretq_u64_s8): Likewise.
21429 (vreinterpretq_u64_u16): Likewise.
21430 (vreinterpretq_u64_u32): Likewise.
21431 (vreinterpretq_u64_u8): Likewise.
21432 (vreinterpretq_u8_s16): Likewise.
21433 (vreinterpretq_u8_s32): Likewise.
21434 (vreinterpretq_u8_s64): Likewise.
21435 (vreinterpretq_u8_s8): Likewise.
21436 (vreinterpretq_u8_u16): Likewise.
21437 (vreinterpretq_u8_u32): Likewise.
21438 (vreinterpretq_u8_u64): Likewise.
21439 (vreinterpretq_s32_f16): Likewise.
21440 (vreinterpretq_s32_f32): Likewise.
21441 (vreinterpretq_u16_f16): Likewise.
21442 (vreinterpretq_u16_f32): Likewise.
21443 (vreinterpretq_u32_f16): Likewise.
21444 (vreinterpretq_u32_f32): Likewise.
21445 (vreinterpretq_u64_f16): Likewise.
21446 (vreinterpretq_u64_f32): Likewise.
21447 (vreinterpretq_u8_f16): Likewise.
21448 (vreinterpretq_u8_f32): Likewise.
21449 (vreinterpretq_f16_f32): Likewise.
21450 (vreinterpretq_f16_s16): Likewise.
21451 (vreinterpretq_f16_s32): Likewise.
21452 (vreinterpretq_f16_s64): Likewise.
21453 (vreinterpretq_f16_s8): Likewise.
21454 (vreinterpretq_f16_u16): Likewise.
21455 (vreinterpretq_f16_u32): Likewise.
21456 (vreinterpretq_f16_u64): Likewise.
21457 (vreinterpretq_f16_u8): Likewise.
21458 (vreinterpretq_f32_f16): Likewise.
21459 (vreinterpretq_f32_s16): Likewise.
21460 (vreinterpretq_f32_s32): Likewise.
21461 (vreinterpretq_f32_s64): Likewise.
21462 (vreinterpretq_f32_s8): Likewise.
21463 (vreinterpretq_f32_u16): Likewise.
21464 (vreinterpretq_f32_u32): Likewise.
21465 (vreinterpretq_f32_u64): Likewise.
21466 (vreinterpretq_f32_u8): Likewise.
21467 (vreinterpretq_s16_f16): Likewise.
21468 (vreinterpretq_s16_f32): Likewise.
21469 (vreinterpretq_s64_f16): Likewise.
21470 (vreinterpretq_s64_f32): Likewise.
21471 (vreinterpretq_s8_f16): Likewise.
21472 (vreinterpretq_s8_f32): Likewise.
21473 (__arm_vreinterpretq_f16): Likewise.
21474 (__arm_vreinterpretq_f32): Likewise.
21475 (__arm_vreinterpretq_s16): Likewise.
21476 (__arm_vreinterpretq_s32): Likewise.
21477 (__arm_vreinterpretq_s64): Likewise.
21478 (__arm_vreinterpretq_s8): Likewise.
21479 (__arm_vreinterpretq_u16): Likewise.
21480 (__arm_vreinterpretq_u32): Likewise.
21481 (__arm_vreinterpretq_u64): Likewise.
21482 (__arm_vreinterpretq_u8): Likewise.
21483 * config/arm/arm_mve_types.h (__arm_vreinterpretq_s16_s32): Remove.
21484 (__arm_vreinterpretq_s16_s64): Likewise.
21485 (__arm_vreinterpretq_s16_s8): Likewise.
21486 (__arm_vreinterpretq_s16_u16): Likewise.
21487 (__arm_vreinterpretq_s16_u32): Likewise.
21488 (__arm_vreinterpretq_s16_u64): Likewise.
21489 (__arm_vreinterpretq_s16_u8): Likewise.
21490 (__arm_vreinterpretq_s32_s16): Likewise.
21491 (__arm_vreinterpretq_s32_s64): Likewise.
21492 (__arm_vreinterpretq_s32_s8): Likewise.
21493 (__arm_vreinterpretq_s32_u16): Likewise.
21494 (__arm_vreinterpretq_s32_u32): Likewise.
21495 (__arm_vreinterpretq_s32_u64): Likewise.
21496 (__arm_vreinterpretq_s32_u8): Likewise.
21497 (__arm_vreinterpretq_s64_s16): Likewise.
21498 (__arm_vreinterpretq_s64_s32): Likewise.
21499 (__arm_vreinterpretq_s64_s8): Likewise.
21500 (__arm_vreinterpretq_s64_u16): Likewise.
21501 (__arm_vreinterpretq_s64_u32): Likewise.
21502 (__arm_vreinterpretq_s64_u64): Likewise.
21503 (__arm_vreinterpretq_s64_u8): Likewise.
21504 (__arm_vreinterpretq_s8_s16): Likewise.
21505 (__arm_vreinterpretq_s8_s32): Likewise.
21506 (__arm_vreinterpretq_s8_s64): Likewise.
21507 (__arm_vreinterpretq_s8_u16): Likewise.
21508 (__arm_vreinterpretq_s8_u32): Likewise.
21509 (__arm_vreinterpretq_s8_u64): Likewise.
21510 (__arm_vreinterpretq_s8_u8): Likewise.
21511 (__arm_vreinterpretq_u16_s16): Likewise.
21512 (__arm_vreinterpretq_u16_s32): Likewise.
21513 (__arm_vreinterpretq_u16_s64): Likewise.
21514 (__arm_vreinterpretq_u16_s8): Likewise.
21515 (__arm_vreinterpretq_u16_u32): Likewise.
21516 (__arm_vreinterpretq_u16_u64): Likewise.
21517 (__arm_vreinterpretq_u16_u8): Likewise.
21518 (__arm_vreinterpretq_u32_s16): Likewise.
21519 (__arm_vreinterpretq_u32_s32): Likewise.
21520 (__arm_vreinterpretq_u32_s64): Likewise.
21521 (__arm_vreinterpretq_u32_s8): Likewise.
21522 (__arm_vreinterpretq_u32_u16): Likewise.
21523 (__arm_vreinterpretq_u32_u64): Likewise.
21524 (__arm_vreinterpretq_u32_u8): Likewise.
21525 (__arm_vreinterpretq_u64_s16): Likewise.
21526 (__arm_vreinterpretq_u64_s32): Likewise.
21527 (__arm_vreinterpretq_u64_s64): Likewise.
21528 (__arm_vreinterpretq_u64_s8): Likewise.
21529 (__arm_vreinterpretq_u64_u16): Likewise.
21530 (__arm_vreinterpretq_u64_u32): Likewise.
21531 (__arm_vreinterpretq_u64_u8): Likewise.
21532 (__arm_vreinterpretq_u8_s16): Likewise.
21533 (__arm_vreinterpretq_u8_s32): Likewise.
21534 (__arm_vreinterpretq_u8_s64): Likewise.
21535 (__arm_vreinterpretq_u8_s8): Likewise.
21536 (__arm_vreinterpretq_u8_u16): Likewise.
21537 (__arm_vreinterpretq_u8_u32): Likewise.
21538 (__arm_vreinterpretq_u8_u64): Likewise.
21539 (__arm_vreinterpretq_s32_f16): Likewise.
21540 (__arm_vreinterpretq_s32_f32): Likewise.
21541 (__arm_vreinterpretq_s16_f16): Likewise.
21542 (__arm_vreinterpretq_s16_f32): Likewise.
21543 (__arm_vreinterpretq_s64_f16): Likewise.
21544 (__arm_vreinterpretq_s64_f32): Likewise.
21545 (__arm_vreinterpretq_s8_f16): Likewise.
21546 (__arm_vreinterpretq_s8_f32): Likewise.
21547 (__arm_vreinterpretq_u16_f16): Likewise.
21548 (__arm_vreinterpretq_u16_f32): Likewise.
21549 (__arm_vreinterpretq_u32_f16): Likewise.
21550 (__arm_vreinterpretq_u32_f32): Likewise.
21551 (__arm_vreinterpretq_u64_f16): Likewise.
21552 (__arm_vreinterpretq_u64_f32): Likewise.
21553 (__arm_vreinterpretq_u8_f16): Likewise.
21554 (__arm_vreinterpretq_u8_f32): Likewise.
21555 (__arm_vreinterpretq_f16_f32): Likewise.
21556 (__arm_vreinterpretq_f16_s16): Likewise.
21557 (__arm_vreinterpretq_f16_s32): Likewise.
21558 (__arm_vreinterpretq_f16_s64): Likewise.
21559 (__arm_vreinterpretq_f16_s8): Likewise.
21560 (__arm_vreinterpretq_f16_u16): Likewise.
21561 (__arm_vreinterpretq_f16_u32): Likewise.
21562 (__arm_vreinterpretq_f16_u64): Likewise.
21563 (__arm_vreinterpretq_f16_u8): Likewise.
21564 (__arm_vreinterpretq_f32_f16): Likewise.
21565 (__arm_vreinterpretq_f32_s16): Likewise.
21566 (__arm_vreinterpretq_f32_s32): Likewise.
21567 (__arm_vreinterpretq_f32_s64): Likewise.
21568 (__arm_vreinterpretq_f32_s8): Likewise.
21569 (__arm_vreinterpretq_f32_u16): Likewise.
21570 (__arm_vreinterpretq_f32_u32): Likewise.
21571 (__arm_vreinterpretq_f32_u64): Likewise.
21572 (__arm_vreinterpretq_f32_u8): Likewise.
21573 (__arm_vreinterpretq_s16): Likewise.
21574 (__arm_vreinterpretq_s32): Likewise.
21575 (__arm_vreinterpretq_s64): Likewise.
21576 (__arm_vreinterpretq_s8): Likewise.
21577 (__arm_vreinterpretq_u16): Likewise.
21578 (__arm_vreinterpretq_u32): Likewise.
21579 (__arm_vreinterpretq_u64): Likewise.
21580 (__arm_vreinterpretq_u8): Likewise.
21581 (__arm_vreinterpretq_f16): Likewise.
21582 (__arm_vreinterpretq_f32): Likewise.
21583 * config/arm/mve.md (@arm_mve_reinterpret<mode>): New pattern.
21584 * config/arm/unspecs.md: (REINTERPRET): New unspec.
21586 2023-05-03 Murray Steele <murray.steele@arm.com>
21587 Christophe Lyon <christophe.lyon@arm.com>
21588 Christophe Lyon <christophe.lyon@arm.com
21590 * config.gcc: Add arm-mve-builtins-base.o and
21591 arm-mve-builtins-shapes.o to extra_objs.
21592 * config/arm/arm-builtins.cc (arm_builtin_decl): Handle MVE builtin
21594 (arm_expand_builtin): Likewise
21595 (arm_check_builtin_call): Likewise
21596 (arm_describe_resolver): Likewise.
21597 * config/arm/arm-builtins.h (enum resolver_ident): Add
21599 * config/arm/arm-c.cc (arm_pragma_arm): Handle new pragma.
21600 (arm_resolve_overloaded_builtin): Handle MVE builtins.
21601 (arm_register_target_pragmas): Register arm_check_builtin_call.
21602 * config/arm/arm-mve-builtins.cc (class registered_function): New
21604 (struct registered_function_hasher): New struct.
21605 (pred_suffixes): New table.
21606 (mode_suffixes): New table.
21607 (type_suffix_info): New table.
21608 (TYPES_float16): New.
21609 (TYPES_all_float): New.
21610 (TYPES_integer_8): New.
21611 (TYPES_integer_8_16): New.
21612 (TYPES_integer_16_32): New.
21613 (TYPES_integer_32): New.
21614 (TYPES_signed_16_32): New.
21615 (TYPES_signed_32): New.
21616 (TYPES_all_signed): New.
21617 (TYPES_all_unsigned): New.
21618 (TYPES_all_integer): New.
21619 (TYPES_all_integer_with_64): New.
21620 (DEF_VECTOR_TYPE): New.
21621 (DEF_DOUBLE_TYPE): New.
21622 (DEF_MVE_TYPES_ARRAY): New.
21623 (all_integer): New.
21624 (all_integer_with_64): New.
21628 (all_unsigned): New.
21630 (integer_8_16): New.
21631 (integer_16_32): New.
21633 (signed_16_32): New.
21635 (register_vector_type): Use void_type_node for mve.fp-only types when
21636 mve.fp is not enabled.
21637 (register_builtin_tuple_types): Likewise.
21638 (handle_arm_mve_h): New function..
21639 (matches_type_p): Likewise..
21640 (report_out_of_range): Likewise.
21641 (report_not_enum): Likewise.
21642 (report_missing_float): Likewise.
21643 (report_non_ice): Likewise.
21644 (check_requires_float): Likewise.
21645 (function_instance::hash): Likewise
21646 (function_instance::call_properties): Likewise.
21647 (function_instance::reads_global_state_p): Likewise.
21648 (function_instance::modifies_global_state_p): Likewise.
21649 (function_instance::could_trap_p): Likewise.
21650 (function_instance::has_inactive_argument): Likewise.
21651 (registered_function_hasher::hash): Likewise.
21652 (registered_function_hasher::equal): Likewise.
21653 (function_builder::function_builder): Likewise.
21654 (function_builder::~function_builder): Likewise.
21655 (function_builder::append_name): Likewise.
21656 (function_builder::finish_name): Likewise.
21657 (function_builder::get_name): Likewise.
21658 (add_attribute): Likewise.
21659 (function_builder::get_attributes): Likewise.
21660 (function_builder::add_function): Likewise.
21661 (function_builder::add_unique_function): Likewise.
21662 (function_builder::add_overloaded_function): Likewise.
21663 (function_builder::add_overloaded_functions): Likewise.
21664 (function_builder::register_function_group): Likewise.
21665 (function_call_info::function_call_info): Likewise.
21666 (function_resolver::function_resolver): Likewise.
21667 (function_resolver::get_vector_type): Likewise.
21668 (function_resolver::get_scalar_type_name): Likewise.
21669 (function_resolver::get_argument_type): Likewise.
21670 (function_resolver::scalar_argument_p): Likewise.
21671 (function_resolver::report_no_such_form): Likewise.
21672 (function_resolver::lookup_form): Likewise.
21673 (function_resolver::resolve_to): Likewise.
21674 (function_resolver::infer_vector_or_tuple_type): Likewise.
21675 (function_resolver::infer_vector_type): Likewise.
21676 (function_resolver::require_vector_or_scalar_type): Likewise.
21677 (function_resolver::require_vector_type): Likewise.
21678 (function_resolver::require_matching_vector_type): Likewise.
21679 (function_resolver::require_derived_vector_type): Likewise.
21680 (function_resolver::require_derived_scalar_type): Likewise.
21681 (function_resolver::require_integer_immediate): Likewise.
21682 (function_resolver::require_scalar_type): Likewise.
21683 (function_resolver::check_num_arguments): Likewise.
21684 (function_resolver::check_gp_argument): Likewise.
21685 (function_resolver::finish_opt_n_resolution): Likewise.
21686 (function_resolver::resolve_unary): Likewise.
21687 (function_resolver::resolve_unary_n): Likewise.
21688 (function_resolver::resolve_uniform): Likewise.
21689 (function_resolver::resolve_uniform_opt_n): Likewise.
21690 (function_resolver::resolve): Likewise.
21691 (function_checker::function_checker): Likewise.
21692 (function_checker::argument_exists_p): Likewise.
21693 (function_checker::require_immediate): Likewise.
21694 (function_checker::require_immediate_enum): Likewise.
21695 (function_checker::require_immediate_range): Likewise.
21696 (function_checker::check): Likewise.
21697 (gimple_folder::gimple_folder): Likewise.
21698 (gimple_folder::fold): Likewise.
21699 (function_expander::function_expander): Likewise.
21700 (function_expander::direct_optab_handler): Likewise.
21701 (function_expander::get_fallback_value): Likewise.
21702 (function_expander::get_reg_target): Likewise.
21703 (function_expander::add_output_operand): Likewise.
21704 (function_expander::add_input_operand): Likewise.
21705 (function_expander::add_integer_operand): Likewise.
21706 (function_expander::generate_insn): Likewise.
21707 (function_expander::use_exact_insn): Likewise.
21708 (function_expander::use_unpred_insn): Likewise.
21709 (function_expander::use_pred_x_insn): Likewise.
21710 (function_expander::use_cond_insn): Likewise.
21711 (function_expander::map_to_rtx_codes): Likewise.
21712 (function_expander::expand): Likewise.
21713 (resolve_overloaded_builtin): Likewise.
21714 (check_builtin_call): Likewise.
21715 (gimple_fold_builtin): Likewise.
21716 (expand_builtin): Likewise.
21717 (gt_ggc_mx): Likewise.
21718 (gt_pch_nx): Likewise.
21719 (gt_pch_nx): Likewise.
21720 * config/arm/arm-mve-builtins.def(s8): Define new type suffix.
21731 (offset): New mode.
21732 * config/arm/arm-mve-builtins.h (MAX_TUPLE_SIZE): New constant.
21733 (CP_READ_FPCR): Likewise.
21734 (CP_RAISE_FP_EXCEPTIONS): Likewise.
21735 (CP_READ_MEMORY): Likewise.
21736 (CP_WRITE_MEMORY): Likewise.
21737 (enum units_index): New enum.
21738 (enum predication_index): New.
21739 (enum type_class_index): New.
21740 (enum mode_suffix_index): New enum.
21741 (enum type_suffix_index): New.
21742 (struct mode_suffix_info): New struct.
21743 (struct type_suffix_info): New.
21744 (struct function_group_info): Likewise.
21745 (class function_instance): Likewise.
21746 (class registered_function): Likewise.
21747 (class function_builder): Likewise.
21748 (class function_call_info): Likewise.
21749 (class function_resolver): Likewise.
21750 (class function_checker): Likewise.
21751 (class gimple_folder): Likewise.
21752 (class function_expander): Likewise.
21753 (get_mve_pred16_t): Likewise.
21754 (find_mode_suffix): New function.
21755 (class function_base): Likewise.
21756 (class function_shape): Likewise.
21757 (function_instance::operator==): New function.
21758 (function_instance::operator!=): Likewise.
21759 (function_instance::vectors_per_tuple): Likewise.
21760 (function_instance::mode_suffix): Likewise.
21761 (function_instance::type_suffix): Likewise.
21762 (function_instance::scalar_type): Likewise.
21763 (function_instance::vector_type): Likewise.
21764 (function_instance::tuple_type): Likewise.
21765 (function_instance::vector_mode): Likewise.
21766 (function_call_info::function_returns_void_p): Likewise.
21767 (function_base::call_properties): Likewise.
21768 * config/arm/arm-protos.h (enum arm_builtin_class): Add
21770 (handle_arm_mve_h): New.
21771 (resolve_overloaded_builtin): New.
21772 (check_builtin_call): New.
21773 (gimple_fold_builtin): New.
21774 (expand_builtin): New.
21775 * config/arm/arm.cc (TARGET_GIMPLE_FOLD_BUILTIN): Define as
21776 arm_gimple_fold_builtin.
21777 (arm_gimple_fold_builtin): New function.
21778 * config/arm/arm_mve.h: Use new arm_mve.h pragma.
21779 * config/arm/predicates.md (arm_any_register_operand): New predicate.
21780 * config/arm/t-arm: (arm-mve-builtins.o): Add includes.
21781 (arm-mve-builtins-shapes.o): New target.
21782 (arm-mve-builtins-base.o): New target.
21783 * config/arm/arm-mve-builtins-base.cc: New file.
21784 * config/arm/arm-mve-builtins-base.def: New file.
21785 * config/arm/arm-mve-builtins-base.h: New file.
21786 * config/arm/arm-mve-builtins-functions.h: New file.
21787 * config/arm/arm-mve-builtins-shapes.cc: New file.
21788 * config/arm/arm-mve-builtins-shapes.h: New file.
21790 2023-05-03 Murray Steele <murray.steele@arm.com>
21791 Christophe Lyon <christophe.lyon@arm.com>
21792 Christophe Lyon <christophe.lyon@arm.com>
21794 * config/arm/arm-builtins.cc (arm_general_add_builtin_function):
21796 (arm_init_builtin): Use arm_general_add_builtin_function instead
21797 of arm_add_builtin_function.
21798 (arm_init_acle_builtins): Likewise.
21799 (arm_init_mve_builtins): Likewise.
21800 (arm_init_crypto_builtins): Likewise.
21801 (arm_init_builtins): Likewise.
21802 (arm_general_builtin_decl): New function.
21803 (arm_builtin_decl): Defer to numberspace-specialized functions.
21804 (arm_expand_builtin_args): Rename into arm_general_expand_builtin_args.
21805 (arm_expand_builtin_1): Rename into arm_general_expand_builtin_1 and ...
21806 (arm_general_expand_builtin_1): ... specialize for general builtins.
21807 (arm_expand_acle_builtin): Use arm_general_expand_builtin
21808 instead of arm_expand_builtin.
21809 (arm_expand_mve_builtin): Likewise.
21810 (arm_expand_neon_builtin): Likewise.
21811 (arm_expand_vfp_builtin): Likewise.
21812 (arm_general_expand_builtin): New function.
21813 (arm_expand_builtin): Specialize for general builtins.
21814 (arm_general_check_builtin_call): New function.
21815 (arm_check_builtin_call): Specialize for general builtins.
21816 (arm_describe_resolver): Validate numberspace.
21817 (arm_cde_end_args): Likewise.
21818 * config/arm/arm-protos.h (enum arm_builtin_class): New enum.
21819 (ARM_BUILTIN_SHIFT, ARM_BUILTIN_CLASS): New constants.
21821 2023-05-03 Martin Liska <mliska@suse.cz>
21824 * config/riscv/sync.md: Add gcc_unreachable to a switch.
21826 2023-05-03 Richard Biener <rguenther@suse.de>
21828 * tree-ssa-loop-split.cc (split_at_bb_p): Avoid last_stmt.
21829 (patch_loop_exit): Likewise.
21830 (connect_loops): Likewise.
21831 (split_loop): Likewise.
21832 (control_dep_semi_invariant_p): Likewise.
21833 (do_split_loop_on_cond): Likewise.
21834 (split_loop_on_cond): Likewise.
21835 * tree-ssa-loop-unswitch.cc (find_unswitching_predicates_for_bb):
21837 (simplify_loop_version): Likewise.
21838 (evaluate_bbs): Likewise.
21839 (find_loop_guard): Likewise.
21840 (clean_up_after_unswitching): Likewise.
21841 * tree-ssa-math-opts.cc (maybe_optimize_guarding_check):
21843 (optimize_spaceship): Take a gcond * argument, avoid
21845 (math_opts_dom_walker::after_dom_children): Adjust call to
21846 optimize_spaceship.
21847 * tree-vrp.cc (maybe_set_nonzero_bits): Avoid last_stmt.
21848 * value-pointer-equiv.cc (pointer_equiv_analyzer::visit_edge):
21851 2023-05-03 Andreas Schwab <schwab@suse.de>
21853 * config/riscv/linux.h (LIB_SPEC): Don't redefine.
21855 2023-05-03 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
21857 * config/riscv/riscv-vector-builtins-bases.cc (fold_fault_load):
21859 (class vlseg): New class.
21860 (class vsseg): Ditto.
21861 (class vlsseg): Ditto.
21862 (class vssseg): Ditto.
21863 (class seg_indexed_load): Ditto.
21864 (class seg_indexed_store): Ditto.
21865 (class vlsegff): Ditto.
21867 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
21868 * config/riscv/riscv-vector-builtins-functions.def (vlseg):
21878 * config/riscv/riscv-vector-builtins-shapes.cc (struct
21879 seg_loadstore_def): Ditto.
21880 (struct seg_indexed_loadstore_def): Ditto.
21881 (struct seg_fault_load_def): Ditto.
21883 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
21884 * config/riscv/riscv-vector-builtins.cc
21885 (function_builder::append_nf): New function.
21886 * config/riscv/riscv-vector-builtins.def (vfloat32m1x2_t):
21887 Change ptr from double into float.
21888 (vfloat32m1x3_t): Ditto.
21889 (vfloat32m1x4_t): Ditto.
21890 (vfloat32m1x5_t): Ditto.
21891 (vfloat32m1x6_t): Ditto.
21892 (vfloat32m1x7_t): Ditto.
21893 (vfloat32m1x8_t): Ditto.
21894 (vfloat32m2x2_t): Ditto.
21895 (vfloat32m2x3_t): Ditto.
21896 (vfloat32m2x4_t): Ditto.
21897 (vfloat32m4x2_t): Ditto.
21898 * config/riscv/riscv-vector-builtins.h: Add segment intrinsics.
21899 * config/riscv/riscv-vsetvl.cc (fault_first_load_p): Adapt for
21901 * config/riscv/riscv.md: Add segment instructions.
21902 * config/riscv/vector-iterators.md: Support segment intrinsics.
21903 * config/riscv/vector.md (@pred_unit_strided_load<mode>): New
21905 (@pred_unit_strided_store<mode>): Ditto.
21906 (@pred_strided_load<mode>): Ditto.
21907 (@pred_strided_store<mode>): Ditto.
21908 (@pred_fault_load<mode>): Ditto.
21909 (@pred_indexed_<order>load<V1T:mode><V1I:mode>): Ditto.
21910 (@pred_indexed_<order>load<V2T:mode><V2I:mode>): Ditto.
21911 (@pred_indexed_<order>load<V4T:mode><V4I:mode>): Ditto.
21912 (@pred_indexed_<order>load<V8T:mode><V8I:mode>): Ditto.
21913 (@pred_indexed_<order>load<V16T:mode><V16I:mode>): Ditto.
21914 (@pred_indexed_<order>load<V32T:mode><V32I:mode>): Ditto.
21915 (@pred_indexed_<order>load<V64T:mode><V64I:mode>): Ditto.
21916 (@pred_indexed_<order>store<V1T:mode><V1I:mode>): Ditto.
21917 (@pred_indexed_<order>store<V2T:mode><V2I:mode>): Ditto.
21918 (@pred_indexed_<order>store<V4T:mode><V4I:mode>): Ditto.
21919 (@pred_indexed_<order>store<V8T:mode><V8I:mode>): Ditto.
21920 (@pred_indexed_<order>store<V16T:mode><V16I:mode>): Ditto.
21921 (@pred_indexed_<order>store<V32T:mode><V32I:mode>): Ditto.
21922 (@pred_indexed_<order>store<V64T:mode><V64I:mode>): Ditto.
21924 2023-05-03 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
21926 * config/riscv/genrvv-type-indexer.cc (valid_type): Adapt for
21927 tuple type support.
21929 (floattype): Ditto.
21931 * config/riscv/riscv-vector-builtins-bases.cc: Ditto.
21932 * config/riscv/riscv-vector-builtins-functions.def (vset): Add
21934 (vget): Add tuple type vget.
21935 * config/riscv/riscv-vector-builtins-types.def
21936 (DEF_RVV_TUPLE_OPS): New macro.
21937 (vint8mf8x2_t): Ditto.
21938 (vuint8mf8x2_t): Ditto.
21939 (vint8mf8x3_t): Ditto.
21940 (vuint8mf8x3_t): Ditto.
21941 (vint8mf8x4_t): Ditto.
21942 (vuint8mf8x4_t): Ditto.
21943 (vint8mf8x5_t): Ditto.
21944 (vuint8mf8x5_t): Ditto.
21945 (vint8mf8x6_t): Ditto.
21946 (vuint8mf8x6_t): Ditto.
21947 (vint8mf8x7_t): Ditto.
21948 (vuint8mf8x7_t): Ditto.
21949 (vint8mf8x8_t): Ditto.
21950 (vuint8mf8x8_t): Ditto.
21951 (vint8mf4x2_t): Ditto.
21952 (vuint8mf4x2_t): Ditto.
21953 (vint8mf4x3_t): Ditto.
21954 (vuint8mf4x3_t): Ditto.
21955 (vint8mf4x4_t): Ditto.
21956 (vuint8mf4x4_t): Ditto.
21957 (vint8mf4x5_t): Ditto.
21958 (vuint8mf4x5_t): Ditto.
21959 (vint8mf4x6_t): Ditto.
21960 (vuint8mf4x6_t): Ditto.
21961 (vint8mf4x7_t): Ditto.
21962 (vuint8mf4x7_t): Ditto.
21963 (vint8mf4x8_t): Ditto.
21964 (vuint8mf4x8_t): Ditto.
21965 (vint8mf2x2_t): Ditto.
21966 (vuint8mf2x2_t): Ditto.
21967 (vint8mf2x3_t): Ditto.
21968 (vuint8mf2x3_t): Ditto.
21969 (vint8mf2x4_t): Ditto.
21970 (vuint8mf2x4_t): Ditto.
21971 (vint8mf2x5_t): Ditto.
21972 (vuint8mf2x5_t): Ditto.
21973 (vint8mf2x6_t): Ditto.
21974 (vuint8mf2x6_t): Ditto.
21975 (vint8mf2x7_t): Ditto.
21976 (vuint8mf2x7_t): Ditto.
21977 (vint8mf2x8_t): Ditto.
21978 (vuint8mf2x8_t): Ditto.
21979 (vint8m1x2_t): Ditto.
21980 (vuint8m1x2_t): Ditto.
21981 (vint8m1x3_t): Ditto.
21982 (vuint8m1x3_t): Ditto.
21983 (vint8m1x4_t): Ditto.
21984 (vuint8m1x4_t): Ditto.
21985 (vint8m1x5_t): Ditto.
21986 (vuint8m1x5_t): Ditto.
21987 (vint8m1x6_t): Ditto.
21988 (vuint8m1x6_t): Ditto.
21989 (vint8m1x7_t): Ditto.
21990 (vuint8m1x7_t): Ditto.
21991 (vint8m1x8_t): Ditto.
21992 (vuint8m1x8_t): Ditto.
21993 (vint8m2x2_t): Ditto.
21994 (vuint8m2x2_t): Ditto.
21995 (vint8m2x3_t): Ditto.
21996 (vuint8m2x3_t): Ditto.
21997 (vint8m2x4_t): Ditto.
21998 (vuint8m2x4_t): Ditto.
21999 (vint8m4x2_t): Ditto.
22000 (vuint8m4x2_t): Ditto.
22001 (vint16mf4x2_t): Ditto.
22002 (vuint16mf4x2_t): Ditto.
22003 (vint16mf4x3_t): Ditto.
22004 (vuint16mf4x3_t): Ditto.
22005 (vint16mf4x4_t): Ditto.
22006 (vuint16mf4x4_t): Ditto.
22007 (vint16mf4x5_t): Ditto.
22008 (vuint16mf4x5_t): Ditto.
22009 (vint16mf4x6_t): Ditto.
22010 (vuint16mf4x6_t): Ditto.
22011 (vint16mf4x7_t): Ditto.
22012 (vuint16mf4x7_t): Ditto.
22013 (vint16mf4x8_t): Ditto.
22014 (vuint16mf4x8_t): Ditto.
22015 (vint16mf2x2_t): Ditto.
22016 (vuint16mf2x2_t): Ditto.
22017 (vint16mf2x3_t): Ditto.
22018 (vuint16mf2x3_t): Ditto.
22019 (vint16mf2x4_t): Ditto.
22020 (vuint16mf2x4_t): Ditto.
22021 (vint16mf2x5_t): Ditto.
22022 (vuint16mf2x5_t): Ditto.
22023 (vint16mf2x6_t): Ditto.
22024 (vuint16mf2x6_t): Ditto.
22025 (vint16mf2x7_t): Ditto.
22026 (vuint16mf2x7_t): Ditto.
22027 (vint16mf2x8_t): Ditto.
22028 (vuint16mf2x8_t): Ditto.
22029 (vint16m1x2_t): Ditto.
22030 (vuint16m1x2_t): Ditto.
22031 (vint16m1x3_t): Ditto.
22032 (vuint16m1x3_t): Ditto.
22033 (vint16m1x4_t): Ditto.
22034 (vuint16m1x4_t): Ditto.
22035 (vint16m1x5_t): Ditto.
22036 (vuint16m1x5_t): Ditto.
22037 (vint16m1x6_t): Ditto.
22038 (vuint16m1x6_t): Ditto.
22039 (vint16m1x7_t): Ditto.
22040 (vuint16m1x7_t): Ditto.
22041 (vint16m1x8_t): Ditto.
22042 (vuint16m1x8_t): Ditto.
22043 (vint16m2x2_t): Ditto.
22044 (vuint16m2x2_t): Ditto.
22045 (vint16m2x3_t): Ditto.
22046 (vuint16m2x3_t): Ditto.
22047 (vint16m2x4_t): Ditto.
22048 (vuint16m2x4_t): Ditto.
22049 (vint16m4x2_t): Ditto.
22050 (vuint16m4x2_t): Ditto.
22051 (vint32mf2x2_t): Ditto.
22052 (vuint32mf2x2_t): Ditto.
22053 (vint32mf2x3_t): Ditto.
22054 (vuint32mf2x3_t): Ditto.
22055 (vint32mf2x4_t): Ditto.
22056 (vuint32mf2x4_t): Ditto.
22057 (vint32mf2x5_t): Ditto.
22058 (vuint32mf2x5_t): Ditto.
22059 (vint32mf2x6_t): Ditto.
22060 (vuint32mf2x6_t): Ditto.
22061 (vint32mf2x7_t): Ditto.
22062 (vuint32mf2x7_t): Ditto.
22063 (vint32mf2x8_t): Ditto.
22064 (vuint32mf2x8_t): Ditto.
22065 (vint32m1x2_t): Ditto.
22066 (vuint32m1x2_t): Ditto.
22067 (vint32m1x3_t): Ditto.
22068 (vuint32m1x3_t): Ditto.
22069 (vint32m1x4_t): Ditto.
22070 (vuint32m1x4_t): Ditto.
22071 (vint32m1x5_t): Ditto.
22072 (vuint32m1x5_t): Ditto.
22073 (vint32m1x6_t): Ditto.
22074 (vuint32m1x6_t): Ditto.
22075 (vint32m1x7_t): Ditto.
22076 (vuint32m1x7_t): Ditto.
22077 (vint32m1x8_t): Ditto.
22078 (vuint32m1x8_t): Ditto.
22079 (vint32m2x2_t): Ditto.
22080 (vuint32m2x2_t): Ditto.
22081 (vint32m2x3_t): Ditto.
22082 (vuint32m2x3_t): Ditto.
22083 (vint32m2x4_t): Ditto.
22084 (vuint32m2x4_t): Ditto.
22085 (vint32m4x2_t): Ditto.
22086 (vuint32m4x2_t): Ditto.
22087 (vint64m1x2_t): Ditto.
22088 (vuint64m1x2_t): Ditto.
22089 (vint64m1x3_t): Ditto.
22090 (vuint64m1x3_t): Ditto.
22091 (vint64m1x4_t): Ditto.
22092 (vuint64m1x4_t): Ditto.
22093 (vint64m1x5_t): Ditto.
22094 (vuint64m1x5_t): Ditto.
22095 (vint64m1x6_t): Ditto.
22096 (vuint64m1x6_t): Ditto.
22097 (vint64m1x7_t): Ditto.
22098 (vuint64m1x7_t): Ditto.
22099 (vint64m1x8_t): Ditto.
22100 (vuint64m1x8_t): Ditto.
22101 (vint64m2x2_t): Ditto.
22102 (vuint64m2x2_t): Ditto.
22103 (vint64m2x3_t): Ditto.
22104 (vuint64m2x3_t): Ditto.
22105 (vint64m2x4_t): Ditto.
22106 (vuint64m2x4_t): Ditto.
22107 (vint64m4x2_t): Ditto.
22108 (vuint64m4x2_t): Ditto.
22109 (vfloat32mf2x2_t): Ditto.
22110 (vfloat32mf2x3_t): Ditto.
22111 (vfloat32mf2x4_t): Ditto.
22112 (vfloat32mf2x5_t): Ditto.
22113 (vfloat32mf2x6_t): Ditto.
22114 (vfloat32mf2x7_t): Ditto.
22115 (vfloat32mf2x8_t): Ditto.
22116 (vfloat32m1x2_t): Ditto.
22117 (vfloat32m1x3_t): Ditto.
22118 (vfloat32m1x4_t): Ditto.
22119 (vfloat32m1x5_t): Ditto.
22120 (vfloat32m1x6_t): Ditto.
22121 (vfloat32m1x7_t): Ditto.
22122 (vfloat32m1x8_t): Ditto.
22123 (vfloat32m2x2_t): Ditto.
22124 (vfloat32m2x3_t): Ditto.
22125 (vfloat32m2x4_t): Ditto.
22126 (vfloat32m4x2_t): Ditto.
22127 (vfloat64m1x2_t): Ditto.
22128 (vfloat64m1x3_t): Ditto.
22129 (vfloat64m1x4_t): Ditto.
22130 (vfloat64m1x5_t): Ditto.
22131 (vfloat64m1x6_t): Ditto.
22132 (vfloat64m1x7_t): Ditto.
22133 (vfloat64m1x8_t): Ditto.
22134 (vfloat64m2x2_t): Ditto.
22135 (vfloat64m2x3_t): Ditto.
22136 (vfloat64m2x4_t): Ditto.
22137 (vfloat64m4x2_t): Ditto.
22138 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_TUPLE_OPS):
22140 (DEF_RVV_TYPE_INDEX): Ditto.
22141 (rvv_arg_type_info::get_tuple_subpart_type): New function.
22142 (DEF_RVV_TUPLE_TYPE): New macro.
22143 * config/riscv/riscv-vector-builtins.def (DEF_RVV_TYPE_INDEX):
22144 Adapt for tuple vget/vset support.
22145 (vint8mf4_t): Ditto.
22146 (vuint8mf4_t): Ditto.
22147 (vint8mf2_t): Ditto.
22148 (vuint8mf2_t): Ditto.
22149 (vint8m1_t): Ditto.
22150 (vuint8m1_t): Ditto.
22151 (vint8m2_t): Ditto.
22152 (vuint8m2_t): Ditto.
22153 (vint8m4_t): Ditto.
22154 (vuint8m4_t): Ditto.
22155 (vint8m8_t): Ditto.
22156 (vuint8m8_t): Ditto.
22157 (vint16mf4_t): Ditto.
22158 (vuint16mf4_t): Ditto.
22159 (vint16mf2_t): Ditto.
22160 (vuint16mf2_t): Ditto.
22161 (vint16m1_t): Ditto.
22162 (vuint16m1_t): Ditto.
22163 (vint16m2_t): Ditto.
22164 (vuint16m2_t): Ditto.
22165 (vint16m4_t): Ditto.
22166 (vuint16m4_t): Ditto.
22167 (vint16m8_t): Ditto.
22168 (vuint16m8_t): Ditto.
22169 (vint32mf2_t): Ditto.
22170 (vuint32mf2_t): Ditto.
22171 (vint32m1_t): Ditto.
22172 (vuint32m1_t): Ditto.
22173 (vint32m2_t): Ditto.
22174 (vuint32m2_t): Ditto.
22175 (vint32m4_t): Ditto.
22176 (vuint32m4_t): Ditto.
22177 (vint32m8_t): Ditto.
22178 (vuint32m8_t): Ditto.
22179 (vint64m1_t): Ditto.
22180 (vuint64m1_t): Ditto.
22181 (vint64m2_t): Ditto.
22182 (vuint64m2_t): Ditto.
22183 (vint64m4_t): Ditto.
22184 (vuint64m4_t): Ditto.
22185 (vint64m8_t): Ditto.
22186 (vuint64m8_t): Ditto.
22187 (vfloat32mf2_t): Ditto.
22188 (vfloat32m1_t): Ditto.
22189 (vfloat32m2_t): Ditto.
22190 (vfloat32m4_t): Ditto.
22191 (vfloat32m8_t): Ditto.
22192 (vfloat64m1_t): Ditto.
22193 (vfloat64m2_t): Ditto.
22194 (vfloat64m4_t): Ditto.
22195 (vfloat64m8_t): Ditto.
22196 (tuple_subpart): Add tuple subpart base type.
22197 * config/riscv/riscv-vector-builtins.h (struct
22198 rvv_arg_type_info): Ditto.
22199 (tuple_type_field): New function.
22201 2023-05-03 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
22203 * config/riscv/riscv-modes.def (RVV_TUPLE_MODES): New macro.
22204 (RVV_TUPLE_PARTIAL_MODES): Ditto.
22205 * config/riscv/riscv-protos.h (riscv_v_ext_tuple_mode_p): New
22208 (get_subpart_mode): Ditto.
22209 (get_tuple_mode): Ditto.
22210 (expand_tuple_move): Ditto.
22211 * config/riscv/riscv-v.cc (ENTRY): New macro.
22212 (TUPLE_ENTRY): Ditto.
22213 (get_nf): New function.
22214 (get_subpart_mode): Ditto.
22215 (get_tuple_mode): Ditto.
22216 (expand_tuple_move): Ditto.
22217 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_TUPLE_TYPE):
22219 (register_tuple_type): New function
22220 * config/riscv/riscv-vector-builtins.def (DEF_RVV_TUPLE_TYPE):
22222 (vint8mf8x2_t): New macro.
22223 (vuint8mf8x2_t): Ditto.
22224 (vint8mf8x3_t): Ditto.
22225 (vuint8mf8x3_t): Ditto.
22226 (vint8mf8x4_t): Ditto.
22227 (vuint8mf8x4_t): Ditto.
22228 (vint8mf8x5_t): Ditto.
22229 (vuint8mf8x5_t): Ditto.
22230 (vint8mf8x6_t): Ditto.
22231 (vuint8mf8x6_t): Ditto.
22232 (vint8mf8x7_t): Ditto.
22233 (vuint8mf8x7_t): Ditto.
22234 (vint8mf8x8_t): Ditto.
22235 (vuint8mf8x8_t): Ditto.
22236 (vint8mf4x2_t): Ditto.
22237 (vuint8mf4x2_t): Ditto.
22238 (vint8mf4x3_t): Ditto.
22239 (vuint8mf4x3_t): Ditto.
22240 (vint8mf4x4_t): Ditto.
22241 (vuint8mf4x4_t): Ditto.
22242 (vint8mf4x5_t): Ditto.
22243 (vuint8mf4x5_t): Ditto.
22244 (vint8mf4x6_t): Ditto.
22245 (vuint8mf4x6_t): Ditto.
22246 (vint8mf4x7_t): Ditto.
22247 (vuint8mf4x7_t): Ditto.
22248 (vint8mf4x8_t): Ditto.
22249 (vuint8mf4x8_t): Ditto.
22250 (vint8mf2x2_t): Ditto.
22251 (vuint8mf2x2_t): Ditto.
22252 (vint8mf2x3_t): Ditto.
22253 (vuint8mf2x3_t): Ditto.
22254 (vint8mf2x4_t): Ditto.
22255 (vuint8mf2x4_t): Ditto.
22256 (vint8mf2x5_t): Ditto.
22257 (vuint8mf2x5_t): Ditto.
22258 (vint8mf2x6_t): Ditto.
22259 (vuint8mf2x6_t): Ditto.
22260 (vint8mf2x7_t): Ditto.
22261 (vuint8mf2x7_t): Ditto.
22262 (vint8mf2x8_t): Ditto.
22263 (vuint8mf2x8_t): Ditto.
22264 (vint8m1x2_t): Ditto.
22265 (vuint8m1x2_t): Ditto.
22266 (vint8m1x3_t): Ditto.
22267 (vuint8m1x3_t): Ditto.
22268 (vint8m1x4_t): Ditto.
22269 (vuint8m1x4_t): Ditto.
22270 (vint8m1x5_t): Ditto.
22271 (vuint8m1x5_t): Ditto.
22272 (vint8m1x6_t): Ditto.
22273 (vuint8m1x6_t): Ditto.
22274 (vint8m1x7_t): Ditto.
22275 (vuint8m1x7_t): Ditto.
22276 (vint8m1x8_t): Ditto.
22277 (vuint8m1x8_t): Ditto.
22278 (vint8m2x2_t): Ditto.
22279 (vuint8m2x2_t): Ditto.
22280 (vint8m2x3_t): Ditto.
22281 (vuint8m2x3_t): Ditto.
22282 (vint8m2x4_t): Ditto.
22283 (vuint8m2x4_t): Ditto.
22284 (vint8m4x2_t): Ditto.
22285 (vuint8m4x2_t): Ditto.
22286 (vint16mf4x2_t): Ditto.
22287 (vuint16mf4x2_t): Ditto.
22288 (vint16mf4x3_t): Ditto.
22289 (vuint16mf4x3_t): Ditto.
22290 (vint16mf4x4_t): Ditto.
22291 (vuint16mf4x4_t): Ditto.
22292 (vint16mf4x5_t): Ditto.
22293 (vuint16mf4x5_t): Ditto.
22294 (vint16mf4x6_t): Ditto.
22295 (vuint16mf4x6_t): Ditto.
22296 (vint16mf4x7_t): Ditto.
22297 (vuint16mf4x7_t): Ditto.
22298 (vint16mf4x8_t): Ditto.
22299 (vuint16mf4x8_t): Ditto.
22300 (vint16mf2x2_t): Ditto.
22301 (vuint16mf2x2_t): Ditto.
22302 (vint16mf2x3_t): Ditto.
22303 (vuint16mf2x3_t): Ditto.
22304 (vint16mf2x4_t): Ditto.
22305 (vuint16mf2x4_t): Ditto.
22306 (vint16mf2x5_t): Ditto.
22307 (vuint16mf2x5_t): Ditto.
22308 (vint16mf2x6_t): Ditto.
22309 (vuint16mf2x6_t): Ditto.
22310 (vint16mf2x7_t): Ditto.
22311 (vuint16mf2x7_t): Ditto.
22312 (vint16mf2x8_t): Ditto.
22313 (vuint16mf2x8_t): Ditto.
22314 (vint16m1x2_t): Ditto.
22315 (vuint16m1x2_t): Ditto.
22316 (vint16m1x3_t): Ditto.
22317 (vuint16m1x3_t): Ditto.
22318 (vint16m1x4_t): Ditto.
22319 (vuint16m1x4_t): Ditto.
22320 (vint16m1x5_t): Ditto.
22321 (vuint16m1x5_t): Ditto.
22322 (vint16m1x6_t): Ditto.
22323 (vuint16m1x6_t): Ditto.
22324 (vint16m1x7_t): Ditto.
22325 (vuint16m1x7_t): Ditto.
22326 (vint16m1x8_t): Ditto.
22327 (vuint16m1x8_t): Ditto.
22328 (vint16m2x2_t): Ditto.
22329 (vuint16m2x2_t): Ditto.
22330 (vint16m2x3_t): Ditto.
22331 (vuint16m2x3_t): Ditto.
22332 (vint16m2x4_t): Ditto.
22333 (vuint16m2x4_t): Ditto.
22334 (vint16m4x2_t): Ditto.
22335 (vuint16m4x2_t): Ditto.
22336 (vint32mf2x2_t): Ditto.
22337 (vuint32mf2x2_t): Ditto.
22338 (vint32mf2x3_t): Ditto.
22339 (vuint32mf2x3_t): Ditto.
22340 (vint32mf2x4_t): Ditto.
22341 (vuint32mf2x4_t): Ditto.
22342 (vint32mf2x5_t): Ditto.
22343 (vuint32mf2x5_t): Ditto.
22344 (vint32mf2x6_t): Ditto.
22345 (vuint32mf2x6_t): Ditto.
22346 (vint32mf2x7_t): Ditto.
22347 (vuint32mf2x7_t): Ditto.
22348 (vint32mf2x8_t): Ditto.
22349 (vuint32mf2x8_t): Ditto.
22350 (vint32m1x2_t): Ditto.
22351 (vuint32m1x2_t): Ditto.
22352 (vint32m1x3_t): Ditto.
22353 (vuint32m1x3_t): Ditto.
22354 (vint32m1x4_t): Ditto.
22355 (vuint32m1x4_t): Ditto.
22356 (vint32m1x5_t): Ditto.
22357 (vuint32m1x5_t): Ditto.
22358 (vint32m1x6_t): Ditto.
22359 (vuint32m1x6_t): Ditto.
22360 (vint32m1x7_t): Ditto.
22361 (vuint32m1x7_t): Ditto.
22362 (vint32m1x8_t): Ditto.
22363 (vuint32m1x8_t): Ditto.
22364 (vint32m2x2_t): Ditto.
22365 (vuint32m2x2_t): Ditto.
22366 (vint32m2x3_t): Ditto.
22367 (vuint32m2x3_t): Ditto.
22368 (vint32m2x4_t): Ditto.
22369 (vuint32m2x4_t): Ditto.
22370 (vint32m4x2_t): Ditto.
22371 (vuint32m4x2_t): Ditto.
22372 (vint64m1x2_t): Ditto.
22373 (vuint64m1x2_t): Ditto.
22374 (vint64m1x3_t): Ditto.
22375 (vuint64m1x3_t): Ditto.
22376 (vint64m1x4_t): Ditto.
22377 (vuint64m1x4_t): Ditto.
22378 (vint64m1x5_t): Ditto.
22379 (vuint64m1x5_t): Ditto.
22380 (vint64m1x6_t): Ditto.
22381 (vuint64m1x6_t): Ditto.
22382 (vint64m1x7_t): Ditto.
22383 (vuint64m1x7_t): Ditto.
22384 (vint64m1x8_t): Ditto.
22385 (vuint64m1x8_t): Ditto.
22386 (vint64m2x2_t): Ditto.
22387 (vuint64m2x2_t): Ditto.
22388 (vint64m2x3_t): Ditto.
22389 (vuint64m2x3_t): Ditto.
22390 (vint64m2x4_t): Ditto.
22391 (vuint64m2x4_t): Ditto.
22392 (vint64m4x2_t): Ditto.
22393 (vuint64m4x2_t): Ditto.
22394 (vfloat32mf2x2_t): Ditto.
22395 (vfloat32mf2x3_t): Ditto.
22396 (vfloat32mf2x4_t): Ditto.
22397 (vfloat32mf2x5_t): Ditto.
22398 (vfloat32mf2x6_t): Ditto.
22399 (vfloat32mf2x7_t): Ditto.
22400 (vfloat32mf2x8_t): Ditto.
22401 (vfloat32m1x2_t): Ditto.
22402 (vfloat32m1x3_t): Ditto.
22403 (vfloat32m1x4_t): Ditto.
22404 (vfloat32m1x5_t): Ditto.
22405 (vfloat32m1x6_t): Ditto.
22406 (vfloat32m1x7_t): Ditto.
22407 (vfloat32m1x8_t): Ditto.
22408 (vfloat32m2x2_t): Ditto.
22409 (vfloat32m2x3_t): Ditto.
22410 (vfloat32m2x4_t): Ditto.
22411 (vfloat32m4x2_t): Ditto.
22412 (vfloat64m1x2_t): Ditto.
22413 (vfloat64m1x3_t): Ditto.
22414 (vfloat64m1x4_t): Ditto.
22415 (vfloat64m1x5_t): Ditto.
22416 (vfloat64m1x6_t): Ditto.
22417 (vfloat64m1x7_t): Ditto.
22418 (vfloat64m1x8_t): Ditto.
22419 (vfloat64m2x2_t): Ditto.
22420 (vfloat64m2x3_t): Ditto.
22421 (vfloat64m2x4_t): Ditto.
22422 (vfloat64m4x2_t): Ditto.
22423 * config/riscv/riscv-vector-builtins.h (DEF_RVV_TUPLE_TYPE):
22425 * config/riscv/riscv-vector-switch.def (TUPLE_ENTRY): Ditto.
22426 * config/riscv/riscv.cc (riscv_v_ext_tuple_mode_p): New
22428 (TUPLE_ENTRY): Ditto.
22429 (riscv_v_ext_mode_p): New function.
22430 (riscv_v_adjust_nunits): Add tuple mode adjustment.
22431 (riscv_classify_address): Ditto.
22432 (riscv_binary_cost): Ditto.
22433 (riscv_rtx_costs): Ditto.
22434 (riscv_secondary_memory_needed): Ditto.
22435 (riscv_hard_regno_nregs): Ditto.
22436 (riscv_hard_regno_mode_ok): Ditto.
22437 (riscv_vector_mode_supported_p): Ditto.
22438 (riscv_regmode_natural_size): Ditto.
22439 (riscv_array_mode): New function.
22440 (TARGET_ARRAY_MODE): New target hook.
22441 * config/riscv/riscv.md: Add tuple modes.
22442 * config/riscv/vector-iterators.md: Ditto.
22443 * config/riscv/vector.md (mov<mode>): Add tuple modes data
22445 (*mov<VT:mode>_<P:mode>): Ditto.
22447 2023-05-03 Richard Biener <rguenther@suse.de>
22449 * cse.cc (cse_insn): Track an equivalence to the destination
22450 separately and delay using src_related for it.
22452 2023-05-03 Richard Biener <rguenther@suse.de>
22454 * cse.cc (HASH): Turn into inline function and mix
22455 in another HASH_SHIFT bits.
22456 (SAFE_HASH): Likewise.
22458 2023-05-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22461 * config/aarch64/aarch64-simd.md (aarch64_<sur>h<addsub><mode>): Rename to...
22462 (aarch64_<sur>h<addsub><mode><vczle><vczbe>): ... This.
22464 2023-05-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22467 * config/aarch64/aarch64-simd.md (add<mode>3): Rename to...
22468 (add<mode>3<vczle><vczbe>): ... This.
22469 (sub<mode>3): Rename to...
22470 (sub<mode>3<vczle><vczbe>): ... This.
22471 (mul<mode>3): Rename to...
22472 (mul<mode>3<vczle><vczbe>): ... This.
22473 (*div<mode>3): Rename to...
22474 (*div<mode>3<vczle><vczbe>): ... This.
22475 (neg<mode>2): Rename to...
22476 (neg<mode>2<vczle><vczbe>): ... This.
22477 (abs<mode>2): Rename to...
22478 (abs<mode>2<vczle><vczbe>): ... This.
22479 (<frint_pattern><mode>2): Rename to...
22480 (<frint_pattern><mode>2<vczle><vczbe>): ... This.
22481 (<fmaxmin><mode>3): Rename to...
22482 (<fmaxmin><mode>3<vczle><vczbe>): ... This.
22483 (*sqrt<mode>2): Rename to...
22484 (*sqrt<mode>2<vczle><vczbe>): ... This.
22486 2023-05-03 Kito Cheng <kito.cheng@sifive.com>
22488 * doc/md.texi (RISC-V): Add vr, vm, vd constarint.
22490 2023-05-03 Martin Liska <mliska@suse.cz>
22492 PR tree-optimization/109693
22493 * value-range-storage.cc (vrange_allocator::vrange_allocator):
22494 Remove unused field.
22495 * value-range-storage.h: Likewise.
22497 2023-05-02 Andrew Pinski <apinski@marvell.com>
22499 * tree-ssa-phiopt.cc (move_stmt): New function.
22500 (match_simplify_replacement): Use move_stmt instead
22501 of the inlined version.
22503 2023-05-02 Andrew Pinski <apinski@marvell.com>
22505 * match.pd (a != 0 ? CLRSB(a) : CST -> CLRSB(a)): New
22508 2023-05-02 Andrew Pinski <apinski@marvell.com>
22510 PR tree-optimization/109702
22511 * match.pd: Fix "a != 0 ? FUNC(a) : CST" patterns
22512 for FUNC of POPCOUNT BSWAP FFS PARITY CLZ and CTZ.
22514 2023-05-02 Andrew Pinski <apinski@marvell.com>
22517 * config/aarch64/aarch64.md (*cmov<mode>_insn_m1): New
22518 insn_and_split pattern.
22520 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
22522 * config/riscv/sync.md (atomic_load<mode>): Implement atomic
22525 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
22527 * config/riscv/sync.md (mem_thread_fence_1): Change fence
22528 depending on the given memory model.
22530 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
22532 * config/riscv/riscv-protos.h (riscv_union_memmodels): Expose
22533 riscv_union_memmodels function to sync.md.
22534 * config/riscv/riscv.cc (riscv_union_memmodels): Add function to
22535 get the union of two memmodels in sync.md.
22536 (riscv_print_operand): Add %I and %J flags that output the
22537 optimal LR/SC flag bits for a given memory model.
22538 * config/riscv/sync.md: Remove static .aqrl bits on LR op/.rl
22539 bits on SC op and replace with optimized %I, %J flags.
22541 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
22543 * config/riscv/riscv.cc
22544 (riscv_memmodel_needs_amo_release): Change function name.
22545 (riscv_print_operand): Remove unneeded %F case.
22546 * config/riscv/sync.md: Remove unneeded fences.
22548 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
22551 * config/riscv/sync.md (atomic_store<mode>): Use simple store
22552 instruction in combination with fence(s).
22554 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
22556 * config/riscv/riscv.cc (riscv_print_operand): Change behavior
22557 of %A to include release bits.
22559 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
22561 * config/riscv/sync.md (atomic_cas_value_strong<mode>): Change
22562 FENCE/LR.aq/SC.aq into sequentially consistent LR.aqrl/SC.rl
22565 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
22567 * config/riscv/sync.md: Change LR.aq/SC.rl pairs into
22568 sequentially consistent LR.aqrl/SC.rl pairs.
22570 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
22572 * config/riscv/riscv.cc: Remove MEMMODEL_SYNC_* cases and
22573 sanitize memmodel input with memmodel_base.
22575 2023-05-02 Yanzhang Wang <yanzhang.wang@intel.com>
22576 Pan Li <pan2.li@intel.com>
22579 * config/riscv/vector-iterators.md: Support VNx2HI and VNX4DI when MIN_VLEN >= 128.
22581 2023-05-02 Romain Naour <romain.naour@gmail.com>
22583 * config/riscv/genrvv-type-indexer.cc: Use log2 from the C header, without
22586 2023-05-02 Martin Liska <mliska@suse.cz>
22588 * doc/invoke.texi: Update documentation based on param.opt file.
22590 2023-05-02 Richard Biener <rguenther@suse.de>
22592 PR tree-optimization/109672
22593 * tree-vect-stmts.cc (vectorizable_operation): For plus,
22594 minus and negate always check the vector mode is word mode.
22596 2023-05-01 Andrew Pinski <apinski@marvell.com>
22598 * tree-ssa-phiopt.cc: Update comment about
22599 how the transformation are implemented.
22601 2023-05-01 Jeff Law <jlaw@ventanamicro>
22603 * config/stormy16/stormy16.cc (TARGET_LRA_P): Remove defintion.
22605 2023-05-01 Jeff Law <jlaw@ventanamicro>
22607 * config/cris/cris.cc (TARGET_LRA_P): Remove.
22608 * config/epiphany/epiphany.cc (TARGET_LRA_P): Remove.
22609 * config/iq2000/iq2000.cc (TARGET_LRA_P): Remove.
22610 * config/m32r/m32r.cc (TARGET_LRA_P): Remove.
22611 * config/microblaze/microblaze.cc (TARGET_LRA_P): Remove.
22612 * config/mmix/mmix.cc (TARGET_LRA_P): Remove.
22614 2023-05-01 Rasmus Villemoes <rasmus.villemoes@prevas.dk>
22616 * print-tree.h (PRINT_DECL_REMAP_DEBUG): New flag.
22617 * print-tree.cc (print_decl_identifier): Implement it.
22618 * toplev.cc (output_stack_usage_1): Use it.
22620 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
22622 * value-range.h (class int_range): Remove gt_ggc_mx and gt_pch_nx
22625 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
22627 * value-range.h (irange::set_nonzero): Inline.
22629 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
22631 * gimple-range-op.cc (cfn_ffs::fold_range): Use the correct
22633 * gimple-ssa-warn-alloca.cc (alloca_call_type): Use <2> for
22634 invalid_range, as it is an inverse range.
22635 * tree-vrp.cc (find_case_label_range): Avoid trees.
22636 * value-range.cc (irange::irange_set): Delete.
22637 (irange::irange_set_1bit_anti_range): Delete.
22638 (irange::irange_set_anti_range): Delete.
22639 (irange::set): Cleanup.
22640 * value-range.h (class irange): Remove irange_set,
22641 irange_set_anti_range, irange_set_1bit_anti_range.
22642 (irange::set_undefined): Remove set to m_type.
22644 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
22646 * range-op.cc (update_known_bitmask): Adjust for irange containing
22647 wide_ints internally.
22648 * tree-ssanames.cc (set_nonzero_bits): Same.
22649 * tree-ssanames.h (set_nonzero_bits): Same.
22650 * value-range-storage.cc (irange_storage::set_irange): Same.
22651 (irange_storage::get_irange): Same.
22652 * value-range.cc (irange::operator=): Same.
22653 (irange::irange_set): Same.
22654 (irange::irange_set_1bit_anti_range): Same.
22655 (irange::irange_set_anti_range): Same.
22656 (irange::set): Same.
22657 (irange::verify_range): Same.
22658 (irange::contains_p): Same.
22659 (irange::irange_single_pair_union): Same.
22660 (irange::union_): Same.
22661 (irange::irange_contains_p): Same.
22662 (irange::intersect): Same.
22663 (irange::invert): Same.
22664 (irange::set_range_from_nonzero_bits): Same.
22665 (irange::set_nonzero_bits): Same.
22666 (mask_to_wi): Same.
22667 (irange::intersect_nonzero_bits): Same.
22668 (irange::union_nonzero_bits): Same.
22671 (tree_range): Same.
22672 (range_tests_strict_enum): Same.
22673 (range_tests_misc): Same.
22674 (range_tests_nonzero_bits): Same.
22675 * value-range.h (irange::type): Same.
22676 (irange::varying_compatible_p): Same.
22677 (irange::irange): Same.
22678 (int_range::int_range): Same.
22679 (irange::set_undefined): Same.
22680 (irange::set_varying): Same.
22681 (irange::lower_bound): Same.
22682 (irange::upper_bound): Same.
22684 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
22686 * gimple-range-fold.cc (tree_lower_bound): Delete.
22687 (tree_upper_bound): Delete.
22688 (vrp_val_max): Delete.
22689 (vrp_val_min): Delete.
22690 (fold_using_range::range_of_ssa_name_with_loop_info): Call
22691 range_of_var_in_loop.
22692 * vr-values.cc (valid_value_p): Delete.
22693 (fix_overflow): Delete.
22694 (get_scev_info): New.
22695 (bounds_of_var_in_loop): Refactor into...
22696 (induction_variable_may_overflow_p): ...this,
22697 (range_from_loop_direction): ...and this,
22698 (range_of_var_in_loop): ...and this.
22699 * vr-values.h (bounds_of_var_in_loop): Delete.
22700 (range_of_var_in_loop): New.
22702 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
22704 * gimple-range-fold.cc (adjust_pointer_diff_expr): Rewrite with
22706 (vrp_val_max): New.
22707 (vrp_val_min): New.
22708 * gimple-range-op.cc (cfn_strlen::fold_range): Use irange_val_*.
22709 * range-op.cc (max_limit): Same.
22711 (plus_minus_ranges): Same.
22712 (operator_rshift::op1_range): Same.
22713 (operator_cast::inside_domain_p): Same.
22714 * value-range.cc (vrp_val_is_max): Delete.
22715 (vrp_val_is_min): Delete.
22716 (range_tests_misc): Use irange_val_*.
22717 * value-range.h (vrp_val_is_min): Delete.
22718 (vrp_val_is_max): Delete.
22719 (vrp_val_max): Delete.
22720 (irange_val_min): New.
22721 (vrp_val_min): Delete.
22722 (irange_val_max): New.
22723 * vr-values.cc (check_for_binary_op_overflow): Use irange_val_*.
22725 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
22727 * fold-const.cc (expr_not_equal_to): Convert to irange wide_int API.
22728 * gimple-fold.cc (size_must_be_zero_p): Same.
22729 * gimple-loop-versioning.cc
22730 (loop_versioning::prune_loop_conditions): Same.
22731 * gimple-range-edge.cc (gcond_edge_range): Same.
22732 (gimple_outgoing_range::calc_switch_ranges): Same.
22733 * gimple-range-fold.cc (adjust_imagpart_expr): Same.
22734 (adjust_realpart_expr): Same.
22735 (fold_using_range::range_of_address): Same.
22736 (fold_using_range::relation_fold_and_or): Same.
22737 * gimple-range-gori.cc (gori_compute::gori_compute): Same.
22738 (range_is_either_true_or_false): Same.
22739 * gimple-range-op.cc (cfn_toupper_tolower::get_letter_range): Same.
22740 (cfn_clz::fold_range): Same.
22741 (cfn_ctz::fold_range): Same.
22742 * gimple-range-tests.cc (class test_expr_eval): Same.
22743 * gimple-ssa-warn-alloca.cc (alloca_call_type): Same.
22744 * ipa-cp.cc (ipa_value_range_from_jfunc): Same.
22745 (propagate_vr_across_jump_function): Same.
22746 (decide_whether_version_node): Same.
22747 * ipa-prop.cc (ipa_get_value_range): Same.
22748 * ipa-prop.h (ipa_range_set_and_normalize): Same.
22749 * range-op.cc (get_shift_range): Same.
22750 (value_range_from_overflowed_bounds): Same.
22751 (value_range_with_overflow): Same.
22752 (create_possibly_reversed_range): Same.
22753 (equal_op1_op2_relation): Same.
22754 (not_equal_op1_op2_relation): Same.
22755 (lt_op1_op2_relation): Same.
22756 (le_op1_op2_relation): Same.
22757 (gt_op1_op2_relation): Same.
22758 (ge_op1_op2_relation): Same.
22759 (operator_mult::op1_range): Same.
22760 (operator_exact_divide::op1_range): Same.
22761 (operator_lshift::op1_range): Same.
22762 (operator_rshift::op1_range): Same.
22763 (operator_cast::op1_range): Same.
22764 (operator_logical_and::fold_range): Same.
22765 (set_nonzero_range_from_mask): Same.
22766 (operator_bitwise_or::op1_range): Same.
22767 (operator_bitwise_xor::op1_range): Same.
22768 (operator_addr_expr::fold_range): Same.
22769 (pointer_plus_operator::wi_fold): Same.
22770 (pointer_or_operator::op1_range): Same.
22777 (range_op_cast_tests): Same.
22778 (range_op_lshift_tests): Same.
22779 (range_op_rshift_tests): Same.
22780 (range_op_bitwise_and_tests): Same.
22781 (range_relational_tests): Same.
22782 * range.cc (range_zero): Same.
22783 (range_nonzero): Same.
22784 * range.h (range_true): Same.
22785 (range_false): Same.
22786 (range_true_and_false): Same.
22787 * tree-data-ref.cc (split_constant_offset_1): Same.
22788 * tree-ssa-loop-ch.cc (entry_loop_condition_is_static): Same.
22789 * tree-ssa-loop-unswitch.cc (struct unswitch_predicate): Same.
22790 (find_unswitching_predicates_for_bb): Same.
22791 * tree-ssa-phiopt.cc (value_replacement): Same.
22792 * tree-ssa-threadbackward.cc
22793 (back_threader::find_taken_edge_cond): Same.
22794 * tree-ssanames.cc (ssa_name_has_boolean_range): Same.
22795 * tree-vrp.cc (find_case_label_range): Same.
22796 * value-query.cc (range_query::get_tree_range): Same.
22797 * value-range.cc (irange::set_nonnegative): Same.
22798 (frange::contains_p): Same.
22799 (frange::singleton_p): Same.
22800 (frange::internal_singleton_p): Same.
22801 (irange::irange_set): Same.
22802 (irange::irange_set_1bit_anti_range): Same.
22803 (irange::irange_set_anti_range): Same.
22804 (irange::set): Same.
22805 (irange::operator==): Same.
22806 (irange::singleton_p): Same.
22807 (irange::contains_p): Same.
22808 (irange::set_range_from_nonzero_bits): Same.
22809 (DEFINE_INT_RANGE_INSTANCE): Same.
22819 (range_uint128): New.
22820 (range_uchar): New.
22822 (build_range3): Convert to irange wide_int API.
22823 (range_tests_irange3): Same.
22824 (range_tests_int_range_max): Same.
22825 (range_tests_strict_enum): Same.
22826 (range_tests_misc): Same.
22827 (range_tests_nonzero_bits): Same.
22828 (range_tests_nan): Same.
22829 (range_tests_signed_zeros): Same.
22830 * value-range.h (Value_Range::Value_Range): Same.
22831 (irange::set): Same.
22832 (irange::nonzero_p): Same.
22833 (irange::contains_p): Same.
22834 (range_includes_zero_p): Same.
22835 (irange::set_nonzero): Same.
22836 (irange::set_zero): Same.
22837 (contains_zero_p): Same.
22838 (frange::contains_p): Same.
22840 (simplify_using_ranges::op_with_boolean_value_range_p): Same.
22841 (bounds_of_var_in_loop): Same.
22842 (simplify_using_ranges::legacy_fold_cond_overflow): Same.
22844 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
22846 * value-range.cc (irange::irange_union): Rename to...
22847 (irange::union_): ...this.
22848 (irange::irange_intersect): Rename to...
22849 (irange::intersect): ...this.
22850 * value-range.h (irange::union_): Delete.
22851 (irange::intersect): Delete.
22853 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
22855 * vr-values.cc (bounds_of_var_in_loop): Convert to irange API.
22857 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
22859 * vr-values.cc (check_for_binary_op_overflow): Tidy up by using
22861 (compare_ranges): Delete.
22862 (compare_range_with_value): Delete.
22863 (bounds_of_var_in_loop): Tidy up by using ranger API.
22864 (simplify_using_ranges::fold_cond_with_ops): Cleanup and rename
22865 from vrp_evaluate_conditional_warnv_with_ops_using_ranges.
22866 (simplify_using_ranges::legacy_fold_cond_overflow): Remove
22867 strict_overflow_p and only_ranges.
22868 (simplify_using_ranges::legacy_fold_cond): Adjust call to
22869 legacy_fold_cond_overflow.
22870 (simplify_using_ranges::simplify_abs_using_ranges): Adjust for
22872 (range_fits_type_p): Rename value_range to irange.
22873 * vr-values.h (range_fits_type_p): Adjust prototype.
22875 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
22877 * value-range.cc (irange::irange_set_anti_range): Remove uses of
22878 tree_lower_bound and tree_upper_bound.
22879 (irange::verify_range): Same.
22880 (irange::operator==): Same.
22881 (irange::singleton_p): Same.
22882 * value-range.h (irange::tree_lower_bound): Delete.
22883 (irange::tree_upper_bound): Delete.
22884 (irange::lower_bound): Delete.
22885 (irange::upper_bound): Delete.
22886 (irange::zero_p): Remove uses of tree_lower_bound and
22889 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
22891 * tree-ssa-loop-niter.cc (refine_value_range_using_guard): Remove
22893 (determine_value_range): Same.
22894 (record_nonwrapping_iv): Same.
22895 (infer_loop_bounds_from_signedness): Same.
22896 (scev_var_range_cant_overflow): Same.
22897 * tree-vrp.cc (operand_less_p): Delete.
22898 * tree-vrp.h (operand_less_p): Delete.
22899 * value-range.cc (get_legacy_range): Remove uses of deprecated API.
22900 (irange::value_inside_range): Delete.
22901 * value-range.h (vrange::kind): Delete.
22902 (irange::num_pairs): Remove check of m_kind.
22903 (irange::min): Delete.
22904 (irange::max): Delete.
22906 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
22908 * gimple-fold.cc (maybe_fold_comparisons_from_match_pd): Adjust
22909 for vrange_storage.
22910 * gimple-range-cache.cc (sbr_vector::sbr_vector): Same.
22911 (sbr_vector::grow): Same.
22912 (sbr_vector::set_bb_range): Same.
22913 (sbr_vector::get_bb_range): Same.
22914 (sbr_sparse_bitmap::sbr_sparse_bitmap): Same.
22915 (sbr_sparse_bitmap::set_bb_range): Same.
22916 (sbr_sparse_bitmap::get_bb_range): Same.
22917 (block_range_cache::block_range_cache): Same.
22918 (ssa_global_cache::ssa_global_cache): Same.
22919 (ssa_global_cache::get_global_range): Same.
22920 (ssa_global_cache::set_global_range): Same.
22921 * gimple-range-cache.h: Same.
22922 * gimple-range-edge.cc
22923 (gimple_outgoing_range::gimple_outgoing_range): Same.
22924 (gimple_outgoing_range::switch_edge_range): Same.
22925 (gimple_outgoing_range::calc_switch_ranges): Same.
22926 * gimple-range-edge.h: Same.
22927 * gimple-range-infer.cc
22928 (infer_range_manager::infer_range_manager): Same.
22929 (infer_range_manager::get_nonzero): Same.
22930 (infer_range_manager::maybe_adjust_range): Same.
22931 (infer_range_manager::add_range): Same.
22932 * gimple-range-infer.h: Rename obstack_vrange_allocator to
22934 * tree-core.h (struct irange_storage_slot): Remove.
22935 (struct tree_ssa_name): Remove irange_info and frange_info. Make
22936 range_info a pointer to vrange_storage.
22937 * tree-ssanames.cc (range_info_fits_p): Adjust for vrange_storage.
22938 (range_info_alloc): Same.
22939 (range_info_free): Same.
22940 (range_info_get_range): Same.
22941 (range_info_set_range): Same.
22942 (get_nonzero_bits): Same.
22943 * value-query.cc (get_ssa_name_range_info): Same.
22944 * value-range-storage.cc (class vrange_internal_alloc): New.
22945 (class vrange_obstack_alloc): New.
22946 (class vrange_ggc_alloc): New.
22947 (vrange_allocator::vrange_allocator): New.
22948 (vrange_allocator::~vrange_allocator): New.
22949 (vrange_storage::alloc_slot): New.
22950 (vrange_allocator::alloc): New.
22951 (vrange_allocator::free): New.
22952 (vrange_allocator::clone): New.
22953 (vrange_allocator::clone_varying): New.
22954 (vrange_allocator::clone_undefined): New.
22955 (vrange_storage::alloc): New.
22956 (vrange_storage::set_vrange): Remove slot argument.
22957 (vrange_storage::get_vrange): Same.
22958 (vrange_storage::fits_p): Same.
22959 (vrange_storage::equal_p): New.
22960 (irange_storage::write_lengths_address): New.
22961 (irange_storage::lengths_address): New.
22962 (irange_storage_slot::alloc_slot): Remove.
22963 (irange_storage::alloc): New.
22964 (irange_storage_slot::irange_storage_slot): Remove.
22965 (irange_storage::irange_storage): New.
22966 (write_wide_int): New.
22967 (irange_storage_slot::set_irange): Remove.
22968 (irange_storage::set_irange): New.
22969 (read_wide_int): New.
22970 (irange_storage_slot::get_irange): Remove.
22971 (irange_storage::get_irange): New.
22972 (irange_storage_slot::size): Remove.
22973 (irange_storage::equal_p): New.
22974 (irange_storage_slot::num_wide_ints_needed): Remove.
22975 (irange_storage::size): New.
22976 (irange_storage_slot::fits_p): Remove.
22977 (irange_storage::fits_p): New.
22978 (irange_storage_slot::dump): Remove.
22979 (irange_storage::dump): New.
22980 (frange_storage_slot::alloc_slot): Remove.
22981 (frange_storage::alloc): New.
22982 (frange_storage_slot::set_frange): Remove.
22983 (frange_storage::set_frange): New.
22984 (frange_storage_slot::get_frange): Remove.
22985 (frange_storage::get_frange): New.
22986 (frange_storage_slot::fits_p): Remove.
22987 (frange_storage::equal_p): New.
22988 (frange_storage::fits_p): New.
22989 (ggc_vrange_allocator): New.
22990 (ggc_alloc_vrange_storage): New.
22991 * value-range-storage.h (class vrange_storage): Rewrite.
22992 (class irange_storage): Rewrite.
22993 (class frange_storage): Rewrite.
22994 (class obstack_vrange_allocator): Remove.
22995 (class ggc_vrange_allocator): Remove.
22996 (vrange_allocator::alloc_vrange): Remove.
22997 (vrange_allocator::alloc_irange): Remove.
22998 (vrange_allocator::alloc_frange): Remove.
22999 (ggc_alloc_vrange_storage): New.
23000 * value-range.h (class irange): Rename vrange_allocator to
23002 (class frange): Same.
23004 2023-04-30 Roger Sayle <roger@nextmovesoftware.com>
23006 * config/stormy16/stormy16.md (neghi2): Rewrite pattern using
23007 inc to avoid clobbering the carry flag.
23009 2023-04-30 Andrew Pinski <apinski@marvell.com>
23011 * match.pd: Add patterns for "a != 0 ? FUNC(a) : CST"
23012 for FUNC of POPCOUNT BSWAP FFS PARITY CLZ and CTZ.
23014 2023-04-30 Andrew Pinski <apinski@marvell.com>
23016 * tree-ssa-phiopt.cc (empty_bb_or_one_feeding_into_p):
23017 Allow some builtin/internal function calls which
23018 are known not to trap/throw.
23019 (phiopt_worker::match_simplify_replacement):
23020 Use name instead of getting the lhs again.
23022 2023-04-30 Joakim Nohlgård <joakim@nohlgard.se>
23024 * configure: Regenerate.
23025 * configure.ac: Use ld -r in the check for HAVE_LD_RO_RW_SECTION_MIXING
23027 2023-04-29 Hans-Peter Nilsson <hp@axis.com>
23029 * reload1.cc (emit_insn_if_valid_for_reload_1): Rename from
23030 emit_insn_if_valid_for_reload.
23031 (emit_insn_if_valid_for_reload): Call new helper, and if a SET fails
23032 to be recognized, also try emitting a parallel that clobbers
23033 TARGET_FLAGS_REGNUM, as applicable.
23035 2023-04-29 Roger Sayle <roger@nextmovesoftware.com>
23037 * config/stormy16/stormy16.md (neghi2): Convert from a define_expand
23039 (*rotatehi_1): New define_insn for efficient 2 insn sequence.
23040 (*rotatehi_8, *rotaterthi_8): New define_insn to emit a swpb.
23042 2023-04-29 Roger Sayle <roger@nextmovesoftware.com>
23044 * config/stormy16/stormy16.md (any_lshift): New code iterator.
23045 (any_or_plus): Likewise.
23046 (any_rotate): Likewise.
23047 (*<any_lshift>_and_internal): New define_insn_and_split to
23048 recognize a logical shift followed by an AND, and split it
23049 again after reload.
23050 (*swpn): New define_insn matching xstormy16's swpn.
23051 (*swpn_zext): New define_insn recognizing swpn followed by
23052 zero_extendqihi2, i.e. with the high byte set to zero.
23053 (*swpn_sext): Likewise, for swpn followed by cbw.
23054 (*swpn_sext_2): Likewise, for an alternate RTL form.
23055 (*swpn_zext_ior): A pre-reload splitter so that an swpn+zext+ior
23056 sequence is split in the correct place to recognize the *swpn_zext
23057 followed by any_or_plus (ior, xor or plus) instruction.
23059 2023-04-29 Mikael Pettersson <mikpelinux@gmail.com>
23062 * config.gcc (vax-*-linux*): Add glibc-stdint.h.
23063 (lm32-*-uclinux*): Likewise.
23065 2023-04-29 Fei Gao <gaofei@eswincomputing.com>
23067 * config/riscv/riscv.cc (riscv_avoid_save_libcall): helper function
23068 for riscv_use_save_libcall.
23069 (riscv_use_save_libcall): call riscv_avoid_save_libcall.
23070 (riscv_compute_frame_info): restructure to decouple stack allocation
23071 for rv32e w/o save-restore.
23073 2023-04-28 Eugene Rozenfeld <erozen@microsoft.com>
23075 * doc/install.texi: Fix documentation typo
23077 2023-04-28 Matevos Mehrabyan <matevosmehrabyan@gmail.com>
23079 * config/riscv/iterators.md (only_div, paired_mod): New iterators.
23080 (u): Add div/udiv cases.
23081 * config/riscv/riscv-protos.h (riscv_use_divmod_expander): Prototype.
23082 * config/riscv/riscv.cc (struct riscv_tune_param): Add field for
23084 (rocket_tune_info, sifive_7_tune_info): Initialize new field.
23085 (thead_c906_tune_info): Likewise.
23086 (optimize_size_tune_info): Likewise.
23087 (riscv_use_divmod_expander): New function.
23088 * config/riscv/riscv.md (<u>divmod<mode>4): New expander.
23090 2023-04-28 Karen Sargsyan <karen1999411@gmail.com>
23092 * config/riscv/bitmanip.md: Added clmulr instruction.
23093 * config/riscv/riscv-builtins.cc (AVAIL): Add new.
23094 * config/riscv/riscv.md: (UNSPEC_CLMULR): Add new unspec type.
23096 * config/riscv/riscv-cmo.def: Added built-in function for clmulr.
23097 * config/riscv/crypto.md: Move clmul[h] instructions to bitmanip.md.
23098 * config/riscv/riscv-scalar-crypto.def: Move clmul[h] built-in
23099 functions to riscv-cmo.def.
23100 * config/riscv/generic.md: Add clmul to list of instructions
23101 using the generic_imul reservation.
23103 2023-04-28 Jivan Hakobyan <jivanhakobyan9@gmail.com>
23105 * config/riscv/bitmanip.md: Added expanders for minu/maxu instructions
23107 2023-04-28 Andrew Pinski <apinski@marvell.com>
23109 PR tree-optimization/100958
23110 * tree-ssa-phiopt.cc (two_value_replacement): Remove.
23111 (pass_phiopt::execute): Don't call two_value_replacement.
23112 * match.pd (a !=/== CST1 ? CST2 : CST3): Add pattern to
23113 handle what two_value_replacement did.
23115 2023-04-28 Andrew Pinski <apinski@marvell.com>
23117 * match.pd: Add patterns for
23118 "(A CMP B) ? MIN/MAX<A, C> : MIN/MAX <B, C>".
23120 2023-04-28 Andrew Pinski <apinski@marvell.com>
23122 * match.pd: Factor out the deciding the min/max from
23123 the "(cond (cmp (convert1? x) c1) (convert2? x) c2)"
23125 * fold-const.cc (minmax_from_comparison): this new function.
23126 * fold-const.h (minmax_from_comparison): New prototype.
23128 2023-04-28 Roger Sayle <roger@nextmovesoftware.com>
23130 PR rtl-optimization/109476
23131 * lower-subreg.cc: Include explow.h for force_reg.
23132 (find_decomposable_shift_zext): Pass an additional SPEED_P argument.
23133 If decomposing a suitable LSHIFTRT and we're not splitting
23134 ZERO_EXTEND (based on the current SPEED_P), then use a ZERO_EXTEND
23135 instead of setting a high part SUBREG to zero, which helps combine.
23136 (decompose_multiword_subregs): Update call to resolve_shift_zext.
23138 2023-04-28 Richard Biener <rguenther@suse.de>
23140 * tree-vect-data-refs.cc (vect_analyze_data_refs): Always
23142 * tree-vect-stmts.cc (vect_model_store_cost): Pass in the
23143 gather-scatter info and cost emulated scatters accordingly.
23144 (get_load_store_type): Support emulated scatters.
23145 (vectorizable_store): Likewise. Emulate them by extracting
23146 scalar offsets and data, doing scalar stores.
23148 2023-04-28 Richard Biener <rguenther@suse.de>
23150 * config/i386/i386.cc (ix86_vector_costs::add_stmt_cost):
23151 Tame down element extracts and scalar loads for gather/scatter
23152 similar to elementwise strided accesses.
23154 2023-04-28 Pan Li <pan2.li@intel.com>
23155 kito-cheng <kito.cheng@sifive.com>
23157 * config/riscv/vector.md: Add new define split to perform
23158 the simplification.
23160 2023-04-28 Richard Biener <rguenther@suse.de>
23163 * ipa-param-manipulation.cc
23164 (ipa_param_body_adjustments::modify_expression): Allow
23165 conversion of a register to a non-register type. Elide
23166 conversions inside BIT_FIELD_REFs.
23168 2023-04-28 Richard Biener <rguenther@suse.de>
23170 PR tree-optimization/109644
23171 * tree-cfg.cc (verify_types_in_gimple_reference): Check
23172 register constraints on the outermost VIEW_CONVERT_EXPR
23173 only. Do not allow register or invariant bases on
23174 multi-level or possibly variable index handled components.
23176 2023-04-28 Richard Biener <rguenther@suse.de>
23178 * gimplify.cc (gimplify_compound_lval): When there's a
23179 non-register type produced by one of the handled component
23180 operations make sure we get a non-register base.
23182 2023-04-28 Richard Biener <rguenther@suse.de>
23184 PR tree-optimization/108752
23185 * tree-vect-generic.cc (build_replicated_const): Rename
23186 to build_replicated_int_cst and move to tree.{h,cc}.
23187 (do_plus_minus): Adjust.
23188 (do_negate): Likewise.
23189 * tree-vect-stmts.cc (vectorizable_operation): Emit emulated
23190 arithmetic vector operations in lowered form.
23191 * tree.h (build_replicated_int_cst): Declare.
23192 * tree.cc (build_replicated_int_cst): Moved from
23193 tree-vect-generic.cc build_replicated_const.
23195 2023-04-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23198 * config/aarch64/aarch64-simd.md (aarch64_rbit<mode>): Rename to...
23199 (aarch64_rbit<mode><vczle><vczbe>): ... This.
23200 (neg<mode>2): Rename to...
23201 (neg<mode>2<vczle><vczbe>): ... This.
23202 (abs<mode>2): Rename to...
23203 (abs<mode>2<vczle><vczbe>): ... This.
23204 (aarch64_abs<mode>): Rename to...
23205 (aarch64_abs<mode><vczle><vczbe>): ... This.
23206 (one_cmpl<mode>2): Rename to...
23207 (one_cmpl<mode>2<vczle><vczbe>): ... This.
23208 (clrsb<mode>2): Rename to...
23209 (clrsb<mode>2<vczle><vczbe>): ... This.
23210 (clz<mode>2): Rename to...
23211 (clz<mode>2<vczle><vczbe>): ... This.
23212 (popcount<mode>2): Rename to...
23213 (popcount<mode>2<vczle><vczbe>): ... This.
23215 2023-04-28 Jakub Jelinek <jakub@redhat.com>
23217 * gimple-range-op.cc (class cfn_sqrt): New type.
23218 (op_cfn_sqrt): New variable.
23219 (gimple_range_op_handler::maybe_builtin_call): Handle
23220 CASE_CFN_SQRT{,_FN}.
23222 2023-04-28 Aldy Hernandez <aldyh@redhat.com>
23223 Jakub Jelinek <jakub@redhat.com>
23225 * value-range.h (frange_nextafter): Declare.
23226 * gimple-range-op.cc (class cfn_sincos): New.
23227 (op_cfn_sin, op_cfn_cos): New variables.
23228 (gimple_range_op_handler::maybe_builtin_call): Handle
23229 CASE_CFN_{SIN,COS}{,_FN}.
23231 2023-04-28 Jakub Jelinek <jakub@redhat.com>
23233 * target.def (libm_function_max_error): New target hook.
23234 * doc/tm.texi.in (TARGET_LIBM_FUNCTION_MAX_ERROR): Add.
23235 * doc/tm.texi: Regenerated.
23236 * targhooks.h (default_libm_function_max_error,
23237 glibc_linux_libm_function_max_error): Declare.
23238 * targhooks.cc: Include case-cfn-macros.h.
23239 (default_libm_function_max_error,
23240 glibc_linux_libm_function_max_error): New functions.
23241 * config/linux.h (TARGET_LIBM_FUNCTION_MAX_ERROR): Redefine.
23242 * config/linux-protos.h (linux_libm_function_max_error): Declare.
23243 * config/linux.cc: Include target.h and targhooks.h.
23244 (linux_libm_function_max_error): New function.
23245 * config/arc/arc.cc: Include targhooks.h and case-cfn-macros.h.
23246 (arc_libm_function_max_error): New function.
23247 (TARGET_LIBM_FUNCTION_MAX_ERROR): Redefine.
23248 * config/i386/i386.cc (ix86_libc_has_fast_function): Formatting fix.
23249 (ix86_libm_function_max_error): New function.
23250 (TARGET_LIBM_FUNCTION_MAX_ERROR): Redefine.
23251 * config/rs6000/rs6000-protos.h
23252 (rs6000_linux_libm_function_max_error): Declare.
23253 * config/rs6000/rs6000-linux.cc: Include target.h, targhooks.h, tree.h
23254 and case-cfn-macros.h.
23255 (rs6000_linux_libm_function_max_error): New function.
23256 * config/rs6000/linux.h (TARGET_LIBM_FUNCTION_MAX_ERROR): Redefine.
23257 * config/rs6000/linux64.h (TARGET_LIBM_FUNCTION_MAX_ERROR): Redefine.
23258 * config/or1k/or1k.cc: Include targhooks.h and case-cfn-macros.h.
23259 (or1k_libm_function_max_error): New function.
23260 (TARGET_LIBM_FUNCTION_MAX_ERROR): Redefine.
23262 2023-04-28 Alexandre Oliva <oliva@adacore.com>
23264 * gimple-harden-conditionals.cc (insert_edge_check_and_trap):
23265 Move detach value calls...
23266 (pass_harden_conditional_branches::execute): ... here.
23267 (pass_harden_compares::execute): Detach values before
23270 2023-04-27 Andrew Stubbs <ams@codesourcery.com>
23272 * config/gcn/gcn-valu.md (cmul<conj_op><mode>3): Use gcn_gen_undef.
23273 (cml<addsub_as><mode>4): Likewise.
23274 (vec_addsub<mode>3): Likewise.
23275 (cadd<rot><mode>3): Likewise.
23276 (vec_fmaddsub<mode>4): Likewise.
23277 (vec_fmsubadd<mode>4): Likewise, and use sub for the odd lanes.
23279 2023-04-27 Andrew Pinski <apinski@marvell.com>
23281 * tree-ssa-phiopt.cc (phiopt_early_allow): Allow for
23282 up to 2 min/max expressions in the sequence/match code.
23284 2023-04-27 Andrew Pinski <apinski@marvell.com>
23286 * rtlanal.cc (may_trap_p_1): Treat SMIN/SMAX similar as
23288 * tree-eh.cc (operation_could_trap_helper_p): Treate
23289 MIN_EXPR/MAX_EXPR similar as other comparisons.
23291 2023-04-27 Andrew Pinski <apinski@marvell.com>
23293 * tree-ssa-phiopt.cc (cond_store_replacement): Remove
23295 (cond_if_else_store_replacement): Likewise.
23296 (get_non_trapping): Likewise.
23297 (store_elim_worker): Move into ...
23298 (pass_cselim::execute): This.
23300 2023-04-27 Andrew Pinski <apinski@marvell.com>
23302 * tree-ssa-phiopt.cc (two_value_replacement): Remove
23304 (match_simplify_replacement): Likewise.
23305 (factor_out_conditional_conversion): Likewise.
23306 (value_replacement): Likewise.
23307 (minmax_replacement): Likewise.
23308 (spaceship_replacement): Likewise.
23309 (cond_removal_in_builtin_zero_pattern): Likewise.
23310 (hoist_adjacent_loads): Likewise.
23311 (tree_ssa_phiopt_worker): Move into ...
23312 (pass_phiopt::execute): this.
23314 2023-04-27 Andrew Pinski <apinski@marvell.com>
23316 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker): Remove
23317 do_store_elim argument and split that part out to ...
23318 (store_elim_worker): This new function.
23319 (pass_cselim::execute): Call store_elim_worker.
23320 (pass_phiopt::execute): Update call to tree_ssa_phiopt_worker.
23322 2023-04-27 Jan Hubicka <jh@suse.cz>
23324 * cfgloopmanip.h (unloop_loops): Export.
23325 * tree-ssa-loop-ch.cc (ch_base::copy_headers): Unloop loops
23326 that no longer loop.
23327 * tree-ssa-loop-ivcanon.cc (unloop_loops): Export; do not free
23328 vectors of loops to unloop.
23329 (canonicalize_induction_variables): Free vectors here.
23330 (tree_unroll_loops_completely): Free vectors here.
23332 2023-04-27 Richard Biener <rguenther@suse.de>
23334 PR tree-optimization/109170
23335 * gimple-range-op.cc (gimple_range_op_handler::maybe_builtin_call):
23336 Handle __builtin_expect and similar via cfn_pass_through_arg1
23337 and inspecting the calls fnspec.
23338 * builtins.cc (builtin_fnspec): Handle BUILT_IN_EXPECT
23339 and BUILT_IN_EXPECT_WITH_PROBABILITY.
23341 2023-04-27 Alexandre Oliva <oliva@adacore.com>
23343 * genmultilib: Use CONFIG_SHELL to run sub-scripts.
23345 2023-04-27 Aldy Hernandez <aldyh@redhat.com>
23347 PR tree-optimization/109639
23348 * ipa-cp.cc (ipa_value_range_from_jfunc): Normalize range.
23349 (propagate_vr_across_jump_function): Same.
23350 * ipa-fnsummary.cc (evaluate_conditions_for_known_args): Same.
23351 * ipa-prop.h (ipa_range_set_and_normalize): New.
23352 * value-range.cc (irange::set): Assert min and max are INTEGER_CST.
23354 2023-04-27 Richard Biener <rguenther@suse.de>
23356 * match.pd (BIT_FIELD_REF CONSTRUCTOR@0 @1 @2): Do not
23357 create a CTOR operand in the result when simplifying GIMPLE.
23359 2023-04-27 Richard Biener <rguenther@suse.de>
23361 * gimplify.cc (gimplify_compound_lval): When the base
23362 gimplified to a register make sure to split up chains
23365 2023-04-27 Richard Biener <rguenther@suse.de>
23368 * ipa-param-manipulation.h
23369 (ipa_param_body_adjustments::modify_expression): Add extra_stmts
23371 * ipa-param-manipulation.cc
23372 (ipa_param_body_adjustments::modify_expression): Likewise.
23373 When we need a conversion and the replacement is a register
23374 split the conversion out.
23375 (ipa_param_body_adjustments::modify_assignment): Pass
23376 extra_stmts to RHS modify_expression.
23378 2023-04-27 Jonathan Wakely <jwakely@redhat.com>
23380 * doc/extend.texi (Zero Length): Describe example.
23382 2023-04-27 Richard Biener <rguenther@suse.de>
23384 PR tree-optimization/109594
23385 * tree-ssa.cc (non_rewritable_mem_ref_base): Constrain
23386 what we rewrite to a register based on the above.
23388 2023-04-26 Patrick O'Neill <patrick@rivosinc.com>
23390 * config/riscv/riscv.cc: Fix whitespace.
23391 * config/riscv/sync.md: Fix whitespace.
23393 2023-04-26 Andrew MacLeod <amacleod@redhat.com>
23395 PR tree-optimization/108697
23396 * gimple-range-cache.cc (ssa_global_cache::clear_range): Do
23397 not clear the vector on an out of range query.
23398 (ssa_cache::dump): Use dump_range_query instead of get_range.
23399 (ssa_cache::dump_range_query): New.
23400 (ssa_lazy_cache::dump_range_query): New.
23401 (ssa_lazy_cache::set_range): New.
23402 * gimple-range-cache.h (ssa_cache::dump_range_query): New.
23403 (class ssa_lazy_cache): New.
23404 (ssa_lazy_cache::ssa_lazy_cache): New.
23405 (ssa_lazy_cache::~ssa_lazy_cache): New.
23406 (ssa_lazy_cache::get_range): New.
23407 (ssa_lazy_cache::clear_range): New.
23408 (ssa_lazy_cache::clear): New.
23409 (ssa_lazy_cache::dump): New.
23410 * gimple-range-path.cc (path_range_query::path_range_query): Do
23411 not allocate a ssa_cache object nor has_cache bitmap.
23412 (path_range_query::~path_range_query): Do not free objects.
23413 (path_range_query::clear_cache): Remove.
23414 (path_range_query::get_cache): Adjust.
23415 (path_range_query::set_cache): Remove.
23416 (path_range_query::dump): Don't call through a pointer.
23417 (path_range_query::internal_range_of_expr): Set cache directly.
23418 (path_range_query::reset_path): Clear cache directly.
23419 (path_range_query::ssa_range_in_phi): Fold with globals only.
23420 (path_range_query::compute_ranges_in_phis): Simply set range.
23421 (path_range_query::compute_ranges_in_block): Call cache directly.
23422 * gimple-range-path.h (class path_range_query): Replace bitmap
23423 and cache pointer with lazy cache object.
23424 * gimple-range.h (class assume_query): Use ssa_lazy_cache.
23426 2023-04-26 Andrew MacLeod <amacleod@redhat.com>
23428 * gimple-range-cache.cc (ssa_cache::ssa_cache): Rename.
23429 (ssa_cache::~ssa_cache): Rename.
23430 (ssa_cache::has_range): New.
23431 (ssa_cache::get_range): Rename.
23432 (ssa_cache::set_range): Rename.
23433 (ssa_cache::clear_range): Rename.
23434 (ssa_cache::clear): Rename.
23435 (ssa_cache::dump): Rename and use get_range.
23436 (ranger_cache::get_global_range): Use get_range and set_range.
23437 (ranger_cache::range_of_def): Use get_range.
23438 * gimple-range-cache.h (class ssa_cache): Rename class and methods.
23439 (class ranger_cache): Use ssa_cache.
23440 * gimple-range-path.cc (path_range_query::path_range_query): Use
23442 (path_range_query::get_cache): Use get_range.
23443 (path_range_query::set_cache): Use set_range.
23444 * gimple-range-path.h (class path_range_query): Use ssa_cache.
23445 * gimple-range.cc (assume_query::assume_range_p): Use get_range.
23446 (assume_query::range_of_expr): Use get_range.
23447 (assume_query::assume_query): Use set_range.
23448 (assume_query::calculate_op): Use get_range and set_range.
23449 * gimple-range.h (class assume_query): Use ssa_cache.
23451 2023-04-26 Andrew MacLeod <amacleod@redhat.com>
23453 * gimple-range-cache.cc (sbr_vector::sbr_vector): Add parameter
23454 and local to optionally zero memory.
23455 (br_vector::grow): Only zero memory if flag is set.
23456 (class sbr_lazy_vector): New.
23457 (sbr_lazy_vector::sbr_lazy_vector): New.
23458 (sbr_lazy_vector::set_bb_range): New.
23459 (sbr_lazy_vector::get_bb_range): New.
23460 (sbr_lazy_vector::bb_range_p): New.
23461 (block_range_cache::set_bb_range): Check flags and Use sbr_lazy_vector.
23462 * gimple-range-gori.cc (gori_map::calculate_gori): Use
23463 param_vrp_switch_limit.
23464 (gori_compute::gori_compute): Use param_vrp_switch_limit.
23465 * params.opt (vrp_sparse_threshold): Rename from evrp_sparse_threshold.
23466 (vrp_switch_limit): Rename from evrp_switch_limit.
23467 (vrp_vector_threshold): New.
23469 2023-04-26 Andrew MacLeod <amacleod@redhat.com>
23471 * value-relation.cc (dom_oracle::query_relation): Check early for lack
23473 * value-relation.h (equiv_oracle::has_equiv_p): New.
23475 2023-04-26 Andrew MacLeod <amacleod@redhat.com>
23477 PR tree-optimization/109417
23478 * gimple-range-gori.cc (range_def_chain::register_dependency):
23479 Save the ssa version number, not the pointer.
23480 (gori_compute::may_recompute_p): No need to check if a dependency
23481 is in the free list.
23482 * gimple-range-gori.h (class range_def_chain): Change ssa1 and ssa2
23483 fields to be unsigned int instead of trees.
23484 (ange_def_chain::depend1): Adjust.
23485 (ange_def_chain::depend2): Adjust.
23486 * gimple-range.h: Include "ssa.h" to inline ssa_name().
23488 2023-04-26 David Edelsohn <dje.gcc@gmail.com>
23490 * config/rs6000/aix72.h (TARGET_DEFAULT): Use ISA_2_6_MASKS_SERVER.
23491 * config/rs6000/aix73.h (TARGET_DEFAULT): Use ISA_2_7_MASKS_SERVER.
23492 (PROCESSOR_DEFAULT): Use PROCESSOR_POWER8.
23494 2023-04-26 Patrick O'Neill <patrick@rivosinc.com>
23497 * config/riscv/riscv-protos.h: Add helper function stubs.
23498 * config/riscv/riscv.cc: Add helper functions for subword masking.
23499 * config/riscv/riscv.opt: Add command-line flags -minline-atomics and
23500 -mno-inline-atomics.
23501 * config/riscv/sync.md: Add masking logic and inline asm for fetch_and_op,
23502 fetch_and_nand, CAS, and exchange ops.
23503 * doc/invoke.texi: Add blurb regarding new command-line flags
23504 -minline-atomics and -mno-inline-atomics.
23506 2023-04-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23508 * config/aarch64/aarch64-simd.md (aarch64_rshrn2<mode>_insn_le):
23509 Reimplement using standard RTL codes instead of unspec.
23510 (aarch64_rshrn2<mode>_insn_be): Likewise.
23511 (aarch64_rshrn2<mode>): Adjust for the above.
23512 * config/aarch64/aarch64.md (UNSPEC_RSHRN): Delete.
23514 2023-04-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23516 * config/aarch64/aarch64-simd.md (aarch64_rshrn<mode>_insn_le): Reimplement
23517 with standard RTL codes instead of an UNSPEC.
23518 (aarch64_rshrn<mode>_insn_be): Likewise.
23519 (aarch64_rshrn<mode>): Adjust for the above.
23520 * config/aarch64/predicates.md (aarch64_simd_rshrn_imm_vec): Define.
23522 2023-04-26 Pan Li <pan2.li@intel.com>
23523 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
23525 * config/riscv/riscv.cc (riscv_classify_address): Allow
23526 const0_rtx for the RVV load/store.
23528 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23530 * range-op.cc (range_op_cast_tests): Remove legacy support.
23531 * value-range-storage.h (vrange_allocator::alloc_irange): Same.
23532 * value-range.cc (irange::operator=): Same.
23533 (get_legacy_range): Same.
23534 (irange::copy_legacy_to_multi_range): Delete.
23535 (irange::copy_to_legacy): Delete.
23536 (irange::irange_set_anti_range): Delete.
23537 (irange::set): Remove legacy support.
23538 (irange::verify_range): Same.
23539 (irange::legacy_lower_bound): Delete.
23540 (irange::legacy_upper_bound): Delete.
23541 (irange::legacy_equal_p): Delete.
23542 (irange::operator==): Remove legacy support.
23543 (irange::singleton_p): Same.
23544 (irange::value_inside_range): Same.
23545 (irange::contains_p): Same.
23546 (intersect_ranges): Delete.
23547 (irange::legacy_intersect): Delete.
23548 (union_ranges): Delete.
23549 (irange::legacy_union): Delete.
23550 (irange::legacy_verbose_union_): Delete.
23551 (irange::legacy_verbose_intersect): Delete.
23552 (irange::irange_union): Remove legacy support.
23553 (irange::irange_intersect): Same.
23554 (irange::intersect): Same.
23555 (irange::invert): Same.
23556 (ranges_from_anti_range): Delete.
23557 (gt_pch_nx): Adjust for legacy removal.
23559 (range_tests_legacy): Delete.
23560 (range_tests_misc): Adjust for legacy removal.
23561 (range_tests): Same.
23562 * value-range.h (class irange): Same.
23563 (irange::legacy_mode_p): Delete.
23564 (ranges_from_anti_range): Delete.
23565 (irange::nonzero_p): Adjust for legacy removal.
23566 (irange::lower_bound): Same.
23567 (irange::upper_bound): Same.
23568 (irange::union_): Same.
23569 (irange::intersect): Same.
23570 (irange::set_nonzero): Same.
23571 (irange::set_zero): Same.
23572 * vr-values.cc (simplify_using_ranges::legacy_fold_cond_overflow): Same.
23574 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23576 * value-range.cc (irange::copy_legacy_to_multi_range): Rewrite use
23577 of range_has_numeric_bounds_p with irange API.
23578 (range_has_numeric_bounds_p): Delete.
23579 * value-range.h (range_has_numeric_bounds_p): Delete.
23581 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23583 * tree-data-ref.cc (compute_distributive_range): Replace uses of
23584 range_int_cst_p with irange API.
23585 * tree-ssa-strlen.cc (get_range_strlen_dynamic): Same.
23586 * tree-vrp.h (range_int_cst_p): Delete.
23587 * vr-values.cc (check_for_binary_op_overflow): Replace usees of
23588 range_int_cst_p with irange API.
23589 (vr_set_zero_nonzero_bits): Same.
23590 (range_fits_type_p): Same.
23591 (simplify_using_ranges::simplify_casted_cond): Same.
23592 * tree-vrp.cc (range_int_cst_p): Remove.
23594 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23596 * tree-ssa-strlen.cc (compare_nonzero_chars): Convert to wide_ints.
23598 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23600 * builtins.cc (expand_builtin_strnlen): Rewrite deprecated irange
23601 API uses to new API.
23602 * gimple-predicate-analysis.cc (find_var_cmp_const): Same.
23603 * internal-fn.cc (get_min_precision): Same.
23605 * tree-affine.cc (expr_to_aff_combination): Same.
23606 * tree-data-ref.cc (dr_step_indicator): Same.
23607 * tree-dfa.cc (get_ref_base_and_extent): Same.
23608 * tree-scalar-evolution.cc (iv_can_overflow_p): Same.
23609 * tree-ssa-phiopt.cc (two_value_replacement): Same.
23610 * tree-ssa-pre.cc (insert_into_preds_of_block): Same.
23611 * tree-ssa-reassoc.cc (optimize_range_tests_to_bit_test): Same.
23612 * tree-ssa-strlen.cc (compare_nonzero_chars): Same.
23613 * tree-switch-conversion.cc (bit_test_cluster::emit): Same.
23614 * tree-vect-patterns.cc (vect_recog_divmod_pattern): Same.
23615 * tree.cc (get_range_pos_neg): Same.
23617 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23619 * ipa-prop.cc (ipa_print_node_jump_functions_for_edge): Use
23620 vrange::dump instead of ad-hoc dumper.
23621 * tree-ssa-strlen.cc (dump_strlen_info): Same.
23622 * value-range-pretty-print.cc (visit): Pass TDF_NOUID to
23625 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23627 * range-op.cc (operator_cast::op1_range): Use
23628 create_possibly_reversed_range.
23629 (operator_bitwise_and::simple_op1_range_solver): Same.
23630 * value-range.cc (swap_out_of_order_endpoints): Delete.
23631 (irange::set): Remove call to swap_out_of_order_endpoints.
23633 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23635 * builtins.cc (determine_block_size): Convert use of legacy API to
23637 * gimple-array-bounds.cc (check_out_of_bounds_and_warn): Same.
23638 (array_bounds_checker::check_array_ref): Same.
23639 * gimple-ssa-warn-restrict.cc
23640 (builtin_memref::extend_offset_range): Same.
23641 * ipa-cp.cc (ipcp_store_vr_results): Same.
23642 * ipa-fnsummary.cc (set_switch_stmt_execution_predicate): Same.
23643 * ipa-prop.cc (struct ipa_vr_ggc_hash_traits): Same.
23644 (ipa_write_jump_function): Same.
23645 * pointer-query.cc (get_size_range): Same.
23646 * tree-data-ref.cc (split_constant_offset): Same.
23647 * tree-ssa-strlen.cc (get_range): Same.
23648 (maybe_diag_stxncpy_trunc): Same.
23649 (strlen_pass::get_len_or_size): Same.
23650 (strlen_pass::count_nonzero_bytes_addr): Same.
23651 * tree-vect-patterns.cc (vect_get_range_info): Same.
23652 * value-range.cc (irange::maybe_anti_range): Remove.
23653 (get_legacy_range): New.
23654 (irange::copy_to_legacy): Use get_legacy_range.
23655 (ranges_from_anti_range): Same.
23656 * value-range.h (class irange): Remove maybe_anti_range.
23657 (get_legacy_range): New.
23658 * vr-values.cc (check_for_binary_op_overflow): Convert use of
23659 legacy API to get_legacy_range.
23660 (compare_ranges): Same.
23661 (compare_range_with_value): Same.
23662 (bounds_of_var_in_loop): Same.
23663 (find_case_label_ranges): Same.
23664 (simplify_using_ranges::simplify_switch_using_ranges): Same.
23666 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23668 * value-range-pretty-print.cc (vrange_printer::visit): Remove
23670 * value-range.cc (irange::constant_p): Remove.
23671 (irange::get_nonzero_bits_from_range): Remove constant_p use.
23672 * value-range.h (class irange): Remove constant_p.
23673 (irange::num_pairs): Remove constant_p use.
23675 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23677 * value-range.cc (irange::copy_legacy_to_multi_range): Remove
23679 (irange::set): Same.
23680 (irange::legacy_lower_bound): Same.
23681 (irange::legacy_upper_bound): Same.
23682 (irange::contains_p): Same.
23683 (range_tests_legacy): Same.
23684 (irange::normalize_addresses): Remove.
23685 (irange::normalize_symbolics): Remove.
23686 (irange::symbolic_p): Remove.
23687 * value-range.h (class irange): Remove symbolic_p,
23688 normalize_symbolics, and normalize_addresses.
23689 * vr-values.cc (simplify_using_ranges::two_valued_val_range_p):
23690 Remove symbolics support.
23692 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23694 * value-range.cc (irange::may_contain_p): Remove.
23695 * value-range.h (range_includes_zero_p): Rewrite may_contain_p
23696 usage with contains_p.
23697 * vr-values.cc (compare_range_with_value): Same.
23699 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23701 * tree-vrp.cc (supported_types_p): Remove.
23702 (defined_ranges_p): Remove.
23703 (range_fold_binary_expr): Remove.
23704 (range_fold_unary_expr): Remove.
23705 * tree-vrp.h (range_fold_unary_expr): Remove.
23706 (range_fold_binary_expr): Remove.
23708 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23710 * ipa-cp.cc (ipa_vr_operation_and_type_effects): Convert to ranger API.
23711 (ipa_value_range_from_jfunc): Same.
23712 (propagate_vr_across_jump_function): Same.
23713 * ipa-fnsummary.cc (evaluate_conditions_for_known_args): Same.
23714 * ipa-prop.cc (ipa_compute_jump_functions_for_edge): Same.
23715 * vr-values.cc (bounds_of_var_in_loop): Same.
23717 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23719 * gimple-array-bounds.cc (array_bounds_checker::get_value_range):
23720 Add irange argument.
23721 (check_out_of_bounds_and_warn): Remove check for vr.
23722 (array_bounds_checker::check_array_ref): Remove pointer qualifier
23723 for vr and adjust accordingly.
23724 * gimple-array-bounds.h (get_value_range): Add irange argument.
23725 * value-query.cc (class equiv_allocator): Delete.
23726 (range_query::get_value_range): Delete.
23727 (range_query::range_query): Remove allocator access.
23728 (range_query::~range_query): Same.
23729 * value-query.h (get_value_range): Delete.
23731 (simplify_using_ranges::op_with_boolean_value_range_p): Remove
23732 call to get_value_range.
23733 (check_for_binary_op_overflow): Same.
23734 (simplify_using_ranges::legacy_fold_cond_overflow): Same.
23735 (simplify_using_ranges::simplify_abs_using_ranges): Same.
23736 (simplify_using_ranges::simplify_cond_using_ranges_1): Same.
23737 (simplify_using_ranges::simplify_casted_cond): Same.
23738 (simplify_using_ranges::simplify_switch_using_ranges): Same.
23739 (simplify_using_ranges::two_valued_val_range_p): Same.
23741 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23744 (simplify_using_ranges::vrp_evaluate_conditional_warnv_with_ops):
23746 (simplify_using_ranges::legacy_fold_cond_overflow): ...this.
23747 (simplify_using_ranges::vrp_visit_cond_stmt): Rename to...
23748 (simplify_using_ranges::legacy_fold_cond): ...this.
23749 (simplify_using_ranges::fold_cond): Rename
23750 vrp_evaluate_conditional_warnv_with_ops to
23751 legacy_fold_cond_overflow.
23752 * vr-values.h (class vr_values): Replace vrp_visit_cond_stmt and
23753 vrp_evaluate_conditional_warnv_with_ops with legacy_fold_cond and
23754 legacy_fold_cond_overflow respectively.
23756 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
23758 * vr-values.cc (get_vr_for_comparison): Remove.
23759 (compare_name_with_value): Same.
23760 (vrp_evaluate_conditional_warnv_with_ops): Remove calls to
23761 compare_name_with_value.
23762 * vr-values.h: Remove compare_name_with_value.
23763 Remove get_vr_for_comparison.
23765 2023-04-26 Roger Sayle <roger@nextmovesoftware.com>
23767 * config/stormy16/stormy16.md (bswaphi2): New define_insn.
23768 (bswapsi2): New define_insn.
23769 (swaphi): New define_insn to exchange two registers (swpw).
23770 (define_peephole2): Recognize exchange of registers as swaphi.
23772 2023-04-26 Richard Biener <rguenther@suse.de>
23774 * gimple-range-path.cc (path_range_query::compute_outgoing_relations):
23776 * ipa-pure-const.cc (pass_nothrow::execute): Likewise.
23777 * predict.cc (apply_return_prediction): Likewise.
23778 * sese.cc (set_ifsese_condition): Likewise. Simplify.
23779 * tree-cfg.cc (assert_unreachable_fallthru_edge_p): Avoid last_stmt.
23780 (make_edges_bb): Likewise.
23781 (make_cond_expr_edges): Likewise.
23782 (end_recording_case_labels): Likewise.
23783 (make_gimple_asm_edges): Likewise.
23784 (cleanup_dead_labels): Likewise.
23785 (group_case_labels): Likewise.
23786 (gimple_can_merge_blocks_p): Likewise.
23787 (gimple_merge_blocks): Likewise.
23788 (find_taken_edge): Likewise. Also handle empty fallthru blocks.
23789 (gimple_duplicate_sese_tail): Avoid last_stmt.
23790 (find_loop_dist_alias): Likewise.
23791 (gimple_block_ends_with_condjump_p): Likewise.
23792 (gimple_purge_dead_eh_edges): Likewise.
23793 (gimple_purge_dead_abnormal_call_edges): Likewise.
23794 (pass_warn_function_return::execute): Likewise.
23795 (execute_fixup_cfg): Likewise.
23796 * tree-eh.cc (redirect_eh_edge_1): Likewise.
23797 (pass_lower_resx::execute): Likewise.
23798 (pass_lower_eh_dispatch::execute): Likewise.
23799 (cleanup_empty_eh): Likewise.
23800 * tree-if-conv.cc (if_convertible_bb_p): Likewise.
23801 (predicate_bbs): Likewise.
23802 (ifcvt_split_critical_edges): Likewise.
23803 * tree-loop-distribution.cc (create_edge_for_control_dependence):
23805 (loop_distribution::transform_reduction_loop): Likewise.
23806 * tree-parloops.cc (transform_to_exit_first_loop_alt): Likewise.
23807 (try_transform_to_exit_first_loop_alt): Likewise.
23808 (transform_to_exit_first_loop): Likewise.
23809 (create_parallel_loop): Likewise.
23810 * tree-scalar-evolution.cc (get_loop_exit_condition): Likewise.
23811 * tree-ssa-dce.cc (mark_last_stmt_necessary): Likewise.
23812 (eliminate_unnecessary_stmts): Likewise.
23814 (dom_opt_dom_walker::set_global_ranges_from_unreachable_edges):
23816 * tree-ssa-ifcombine.cc (ifcombine_ifandif): Likewise.
23817 (pass_tree_ifcombine::execute): Likewise.
23818 * tree-ssa-loop-ch.cc (entry_loop_condition_is_static): Likewise.
23819 (should_duplicate_loop_header_p): Likewise.
23820 * tree-ssa-loop-ivcanon.cc (create_canonical_iv): Likewise.
23821 (tree_estimate_loop_size): Likewise.
23822 (try_unroll_loop_completely): Likewise.
23823 * tree-ssa-loop-ivopts.cc (tree_ssa_iv_optimize_loop): Likewise.
23824 * tree-ssa-loop-manip.cc (ip_normal_pos): Likewise.
23825 (canonicalize_loop_ivs): Likewise.
23826 * tree-ssa-loop-niter.cc (determine_value_range): Likewise.
23827 (bound_difference): Likewise.
23828 (number_of_iterations_popcount): Likewise.
23829 (number_of_iterations_cltz): Likewise.
23830 (number_of_iterations_cltz_complement): Likewise.
23831 (simplify_using_initial_conditions): Likewise.
23832 (number_of_iterations_exit_assumptions): Likewise.
23833 (loop_niter_by_eval): Likewise.
23834 (estimate_numbers_of_iterations): Likewise.
23836 2023-04-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
23838 * config/riscv/vector.md: Refine vmadc/vmsbc RA constraint.
23840 2023-04-26 Kewen Lin <linkw@linux.ibm.com>
23843 * config/rs6000/rs6000-builtins.def
23844 (__builtin_vsx_scalar_cmp_exp_qp_eq, __builtin_vsx_scalar_cmp_exp_qp_gt
23845 __builtin_vsx_scalar_cmp_exp_qp_lt,
23846 __builtin_vsx_scalar_cmp_exp_qp_unordered): Move from stanza ieee128-hw
23849 2023-04-26 Kewen Lin <linkw@linux.ibm.com>
23852 * config/rs6000/altivec.md (sldoi_to_mov<mode>): Replace predicate
23853 easy_vector_constant with const_vector_each_byte_same, add
23854 handlings in preparation for !easy_vector_constant, and update
23855 VECTOR_UNIT_ALTIVEC_OR_VSX_P with VECTOR_MEM_ALTIVEC_OR_VSX_P.
23856 * config/rs6000/predicates.md (const_vector_each_byte_same): New
23859 2023-04-26 Juzhe-Zhong <juzhe.zhong@rivai.ai>
23861 * config/riscv/vector.md (*pred_cmp<mode>_merge_tie_mask): New pattern.
23862 (*pred_ltge<mode>_merge_tie_mask): Ditto.
23863 (*pred_cmp<mode>_scalar_merge_tie_mask): Ditto.
23864 (*pred_eqne<mode>_scalar_merge_tie_mask): Ditto.
23865 (*pred_cmp<mode>_extended_scalar_merge_tie_mask): Ditto.
23866 (*pred_eqne<mode>_extended_scalar_merge_tie_mask): Ditto.
23867 (*pred_cmp<mode>_narrow_merge_tie_mask): Ditto.
23869 2023-04-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
23871 * config/riscv/vector.md: Fix redundant vmv1r.v.
23873 2023-04-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
23875 * config/riscv/vector.md: Fix RA constraint.
23877 2023-04-26 Pan Li <pan2.li@intel.com>
23880 * tree-ssa-sccvn.cc (vn_reference_eq): add type vector subparts
23881 check for vn_reference equal.
23883 2023-04-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
23885 * config/riscv/riscv-opts.h (enum riscv_autovec_preference_enum): Add enum for
23886 auto-vectorization preference.
23887 (enum riscv_autovec_lmul_enum): Add enum for choosing LMUL of RVV
23888 auto-vectorization.
23889 * config/riscv/riscv.opt: Add compile option for RVV auto-vectorization.
23891 2023-04-26 Jivan Hakobyan <jivanhakobyan9@gmail.com>
23893 * config/riscv/bitmanip.md: Updated predicates of bclri<mode>_nottwobits
23894 and bclridisi_nottwobits patterns.
23895 * config/riscv/predicates.md: (not_uimm_extra_bit_or_nottwobits): Adjust
23896 predicate to avoid splitting arith constants.
23897 (const_nottwobits_not_arith_operand): New predicate.
23899 2023-04-25 Hans-Peter Nilsson <hp@axis.com>
23901 * recog.cc (peep2_attempt, peep2_update_life): Correct
23902 head-comment description of parameter match_len.
23904 2023-04-25 Vineet Gupta <vineetg@rivosinc.com>
23906 * config/riscv/riscv.md: riscv_move_integer() drop in_splitter arg.
23907 riscv_split_symbol() drop in_splitter arg.
23908 * config/riscv/riscv.cc: riscv_move_integer() drop in_splitter arg.
23909 riscv_split_symbol() drop in_splitter arg.
23910 riscv_force_temporary() drop in_splitter arg.
23911 * config/riscv/riscv-protos.h: riscv_move_integer() drop in_splitter arg.
23912 riscv_split_symbol() drop in_splitter arg.
23914 2023-04-25 Eric Botcazou <ebotcazou@adacore.com>
23916 * tree-ssa.cc (insert_debug_temp_for_var_def): Do not create
23917 superfluous debug temporaries for single GIMPLE assignments.
23919 2023-04-25 Richard Biener <rguenther@suse.de>
23921 PR tree-optimization/109609
23922 * attr-fnspec.h (arg_max_access_size_given_by_arg_p):
23924 * tree-ssa-alias.cc (check_fnspec): Correctly interpret
23925 the size given by arg_max_access_size_given_by_arg_p as
23926 maximum, not exact, size.
23928 2023-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23931 * config/aarch64/aarch64-simd.md (orn<mode>3): Rename to...
23932 (orn<mode>3<vczle><vczbe>): ... This.
23933 (bic<mode>3): Rename to...
23934 (bic<mode>3<vczle><vczbe>): ... This.
23935 (<su><maxmin><mode>3): Rename to...
23936 (<su><maxmin><mode>3<vczle><vczbe>): ... This.
23938 2023-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23940 * config/aarch64/aarch64-simd.md (<su_optab>div<mode>3): New define_expand.
23941 * config/aarch64/iterators.md (VQDIV): New mode iterator.
23942 (vnx2di): New mode attribute.
23944 2023-04-25 Richard Biener <rguenther@suse.de>
23946 PR rtl-optimization/109585
23947 * tree-ssa-alias.cc (aliasing_component_refs_p): Fix typo.
23949 2023-04-25 Jakub Jelinek <jakub@redhat.com>
23952 * config/rs6000/rs6000.cc (rs6000_is_valid_rotate_dot_mask): For
23953 !TARGET_64BIT, don't return true if UINTVAL (mask) << (63 - nb)
23954 is larger than signed int maximum.
23956 2023-04-25 Martin Liska <mliska@suse.cz>
23958 * doc/gcov.texi: Document the new "calls" field and document
23959 the API bump. Mention also "block_ids" for lines.
23960 * gcov.cc (output_intermediate_json_line): Output info about
23961 calls and extend branches as well.
23962 (generate_results): Bump version to 2.
23963 (output_line_details): Use block ID instead of a non-sensual
23966 2023-04-25 Roger Sayle <roger@nextmovesoftware.com>
23968 * config/stormy16/stormy16.md (zero_extendqihi2): Restore/fix
23969 length attribute for the first (memory operand) alternative.
23971 2023-04-25 Victor Do Nascimento <victor.donascimento@arm.com>
23973 * config/aarch64/aarch64-simd.md(aarch64_simd_stp<mode>): New.
23974 * config/aarch64/constraints.md: Make "Umn" relaxed memory
23976 * config/aarch64/iterators.md(ldpstp_vel_sz): New.
23978 2023-04-25 Aldy Hernandez <aldyh@redhat.com>
23980 * value-range.cc (frange::set): Adjust constructor.
23981 * value-range.h (nan_state::nan_state): Replace default
23982 constructor with one taking an argument.
23984 2023-04-25 Aldy Hernandez <aldyh@redhat.com>
23986 * ipa-cp.cc (ipa_range_contains_p): New.
23987 (decide_whether_version_node): Use it.
23989 2023-04-24 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
23991 * tree-ssa-forwprop.cc (is_combined_permutation_identity): Try to
23992 simplify two successive VEC_PERM_EXPRs with same VLA mask,
23993 where mask chooses elements in reverse order.
23995 2023-04-24 Andrew Pinski <apinski@marvell.com>
23997 * tree-ssa-phiopt.cc (match_simplify_replacement): Add new arguments
23998 and support diamond shaped basic block form.
23999 (tree_ssa_phiopt_worker): Update call to match_simplify_replacement
24001 2023-04-24 Andrew Pinski <apinski@marvell.com>
24003 * tree-ssa-phiopt.cc (empty_bb_or_one_feeding_into_p):
24004 Instead of calling last_and_only_stmt, look for the last statement
24007 2023-04-24 Andrew Pinski <apinski@marvell.com>
24009 * tree-ssa-phiopt.cc (empty_bb_or_one_feeding_into_p):
24011 (match_simplify_replacement): Call
24012 empty_bb_or_one_feeding_into_p instead of doing it inline.
24014 2023-04-24 Andrew Pinski <apinski@marvell.com>
24016 PR tree-optimization/68894
24017 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker): Remove the
24018 continue for the do_hoist_loads diamond case.
24020 2023-04-24 Andrew Pinski <apinski@marvell.com>
24022 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker): Rearrange
24023 code for better code readability.
24025 2023-04-24 Andrew Pinski <apinski@marvell.com>
24027 PR tree-optimization/109604
24028 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker): Move the
24029 diamond form check from ...
24030 (minmax_replacement): Here.
24032 2023-04-24 Patrick Palka <ppalka@redhat.com>
24034 * tree.cc (strip_array_types): Don't define here.
24035 (is_typedef_decl): Don't define here.
24036 (typedef_variant_p): Don't define here.
24037 * tree.h (strip_array_types): Define here.
24038 (is_typedef_decl): Define here.
24039 (typedef_variant_p): Define here.
24041 2023-04-24 Frederik Harwath <frederik@codesourcery.com>
24043 * doc/generic.texi (OpenMP): Add != to allowed
24044 conditions and state that vars can be unsigned.
24045 * tree.def (OMP_FOR): Likewise.
24047 2023-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24049 * config/aarch64/aarch64-simd.md (mulv2di3): New expander.
24051 2023-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24053 * doc/install.texi: Consistently use Solaris rather than Solaris 2.
24054 Remove explicit Solaris 11 references.
24056 (Options specification, --with-gnu-as): as and gas always differ
24058 Remove /usr/ccs/bin reference.
24059 (Installing GCC: Binaries, Solaris (SPARC, Intel)): Remove.
24060 (i?86-*-solaris2*): Merge assembler, linker recommendations ...
24061 (*-*-solaris2*): ... here.
24062 Update bundled GCC versions.
24063 Don't refer to pre-built binaries.
24064 Remove /bin/sh warning.
24065 Update assembler, linker recommendations.
24066 Document GNAT bootstrap compiler.
24067 (sparc-sun-solaris2*): Remove non-UltraSPARC reference.
24068 (sparc64-*-solaris2*): Move content...
24069 (sparcv9-*-solaris2*): ...here.
24070 Add GDC for 64-bit bootstrap compilers.
24072 2023-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24075 * config/aarch64/aarch64-sve.md (<optab><mode>3): Handle TARGET_SVE2 MUL
24077 * config/aarch64/aarch64-sve2.md (*aarch64_mul_unpredicated_<mode>): New
24080 2023-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24082 * config/aarch64/aarch64-simd.md (aarch64_<sur>abal2<mode>): Rename to...
24083 (aarch64_<su>abal2<mode>_insn): ... This. Use RTL codes instead of unspec.
24084 (aarch64_<su>abal2<mode>): New define_expand.
24085 * config/aarch64/aarch64.cc (aarch64_abd_rtx_p): New function.
24086 (aarch64_rtx_costs): Handle ABD rtxes.
24087 * config/aarch64/aarch64.md (UNSPEC_SABAL2, UNSPEC_UABAL2): Delete.
24088 * config/aarch64/iterators.md (ABAL2): Delete.
24089 (sur): Remove handling of UNSPEC_UABAL2 and UNSPEC_SABAL2.
24091 2023-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24093 * config/aarch64/aarch64-simd.md (aarch64_<sur>abal<mode>): Rename to...
24094 (aarch64_<su>abal<mode>): ... This. Use RTL codes instead of unspec.
24095 (<sur>sadv16qi): Rename to...
24096 (<su>sadv16qi): ... This. Adjust for the above.
24097 * config/aarch64/aarch64-sve.md (<sur>sad<vsi2qi>): Rename to...
24098 (<su>sad<vsi2qi>): ... This. Adjust for the above.
24099 * config/aarch64/aarch64.md (UNSPEC_SABAL, UNSPEC_UABAL): Delete.
24100 * config/aarch64/iterators.md (ABAL): Delete.
24101 (sur): Remove handling of UNSPEC_SABAL and UNSPEC_UABAL.
24103 2023-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24105 * config/aarch64/aarch64-simd.md (aarch64_<sur>abdl2<mode>): Rename to...
24106 (aarch64_<su>abdl2<mode>_insn): ... This. Use RTL codes instead of unspec.
24107 (aarch64_<su>abdl2<mode>): New define_expand.
24108 * config/aarch64/aarch64.md (UNSPEC_SABDL2, UNSPEC_UABDL2): Delete.
24109 * config/aarch64/iterators.md (ABDL2): Delete.
24110 (sur): Remove handling of UNSPEC_SABDL2 and UNSPEC_UABDL2.
24112 2023-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24114 * config/aarch64/aarch64-simd.md (aarch64_<sur>abdl<mode>): Rename to...
24115 (aarch64_<su>abdl<mode>): ... This. Use standard RTL ops instead of
24117 * config/aarch64/aarch64.md (UNSPEC_SABDL, UNSPEC_UABDL): Delete.
24118 * config/aarch64/iterators.md (ABDL): Delete.
24119 (sur): Remove handling of UNSPEC_SABDL and UNSPEC_UABDL.
24121 2023-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24123 * config/aarch64/aarch64-simd.md
24124 (*aarch64_<su>addlv<VDQV_L:mode>_ze<GPI:mode>): New pattern.
24126 2023-04-24 Richard Biener <rguenther@suse.de>
24128 * gimple-ssa-split-paths.cc (is_feasible_trace): Avoid
24130 * graphite-scop-detection.cc (single_pred_cond_non_loop_exit):
24132 * ipa-fnsummary.cc (set_cond_stmt_execution_predicate): Likewise.
24133 (set_switch_stmt_execution_predicate): Likewise.
24134 (phi_result_unknown_predicate): Likewise.
24135 * ipa-prop.cc (compute_complex_ancestor_jump_func): Likewise.
24136 (ipa_analyze_indirect_call_uses): Likewise.
24137 * predict.cc (predict_iv_comparison): Likewise.
24138 (predict_extra_loop_exits): Likewise.
24139 (predict_loops): Likewise.
24140 (tree_predict_by_opcode): Likewise.
24141 * gimple-predicate-analysis.cc (predicate::init_from_control_deps):
24143 * gimple-pretty-print.cc (dump_implicit_edges): Likewise.
24144 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker): Likewise.
24145 (replace_phi_edge_with_variable): Likewise.
24146 (two_value_replacement): Likewise.
24147 (value_replacement): Likewise.
24148 (minmax_replacement): Likewise.
24149 (spaceship_replacement): Likewise.
24150 (cond_removal_in_builtin_zero_pattern): Likewise.
24151 * tree-ssa-reassoc.cc (maybe_optimize_range_tests): Likewise.
24152 * tree-ssa-sccvn.cc (vn_phi_eq): Likewise.
24153 (vn_phi_lookup): Likewise.
24154 (vn_phi_insert): Likewise.
24155 * tree-ssa-structalias.cc (compute_points_to_sets): Likewise.
24156 * tree-ssa-threadbackward.cc (back_threader::maybe_thread_block):
24158 (back_threader_profitability::possibly_profitable_path_p):
24160 * tree-ssa-threadedge.cc (jump_threader::thread_outgoing_edges):
24162 * tree-switch-conversion.cc (pass_convert_switch::execute):
24164 (pass_lower_switch<O0>::execute): Likewise.
24165 * tree-tailcall.cc (tree_optimize_tail_calls_1): Likewise.
24166 * tree-vect-loop-manip.cc (vect_loop_versioning): Likewise.
24167 * tree-vect-slp.cc (vect_slp_function): Likewise.
24168 * tree-vect-stmts.cc (cfun_returns): Likewise.
24169 * tree-vectorizer.cc (vect_loop_vectorized_call): Likewise.
24170 (vect_loop_dist_alias_call): Likewise.
24172 2023-04-24 Richard Biener <rguenther@suse.de>
24174 * cfgcleanup.cc (outgoing_edges_match): Use FORWARDER_BLOCK_P.
24176 2023-04-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
24178 * config/riscv/riscv-vsetvl.cc
24179 (vector_infos_manager::all_avail_in_compatible_p): New function.
24180 (pass_vsetvl::refine_vsetvls): Optimize vsetvls.
24181 * config/riscv/riscv-vsetvl.h: New function.
24183 2023-04-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
24185 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::pre_vsetvl): Add function
24186 comment for cleanup_insns.
24188 2023-04-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
24190 * config/riscv/vector-iterators.md: New unspec to refine fault first load pattern.
24191 * config/riscv/vector.md: Refine fault first load pattern to erase avl from instructions
24192 with the fault first load property.
24194 2023-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24196 * config/aarch64/aarch64-simd.md (aarch64_float_truncate_lo_): Rename to...
24197 (aarch64_float_truncate_lo_<mode><vczle><vczbe>): ... This.
24199 2023-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24202 * config/aarch64/aarch64-simd.md (aarch64_addp<mode>): Rename to...
24203 (aarch64_addp<mode><vczle><vczbe>): ... This.
24205 2023-04-23 Roger Sayle <roger@nextmovesoftware.com>
24207 * config/stormy16/stormy16.cc (xstormy16_rtx_costs): Rewrite to
24208 provide reasonable values for common arithmetic operations and
24209 immediate operands (in several machine modes).
24211 2023-04-23 Roger Sayle <roger@nextmovesoftware.com>
24213 * config/stormy16/stormy16.cc (xstormy16_print_operand): Add %h
24214 format specifier to output high_part register name of SImode reg.
24215 * config/stormy16/stormy16.md (extendhisi2): New define_insn.
24216 (zero_extendqihi2): Fix lengths, consistent formatting and add
24217 "and Rx,#255" alternative, for documentation purposes.
24218 (zero_extendhisi2): New define_insn.
24220 2023-04-23 Roger Sayle <roger@nextmovesoftware.com>
24222 * config/stormy16/stormy16.cc (xstormy16_output_shift): Implement
24223 SImode shifts by two by performing a single bit SImode shift twice.
24225 2023-04-23 Aldy Hernandez <aldyh@redhat.com>
24227 PR tree-optimization/109593
24228 * value-range.cc (frange::operator==): Handle NANs.
24230 2023-04-23 liuhongt <hongtao.liu@intel.com>
24232 PR rtl-optimization/108707
24233 * ira-costs.cc (scan_one_insn): Use NO_REGS instead of
24234 GENERAL_REGS when preferred reg_class is not known.
24236 2023-04-22 Andrew Pinski <apinski@marvell.com>
24238 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker):
24239 Change the code around slightly to move diamond
24240 handling for do_store_elim/do_hoist_loads out of
24243 2023-04-22 Andrew Pinski <apinski@marvell.com>
24245 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker):
24246 Remove check on empty_block_p.
24248 2023-04-22 Jakub Jelinek <jakub@redhat.com>
24250 PR bootstrap/109589
24251 * system.h (class auto_mpz): Workaround PR62101 bug in GCC 4.8 and 4.9.
24252 * realmpfr.h (class auto_mpfr): Likewise.
24254 2023-04-22 Jakub Jelinek <jakub@redhat.com>
24256 PR tree-optimization/109583
24257 * match.pd (fneg/fadd simplify): Don't call related_vector_mode
24258 if vec_mode is not VECTOR_MODE_P.
24260 2023-04-22 Jan Hubicka <hubicka@ucw.cz>
24261 Ondrej Kubanek <kubanek0ondrej@gmail.com>
24263 * cfgloopmanip.h (adjust_loop_info_after_peeling): Declare.
24264 * tree-ssa-loop-ch.cc (ch_base::copy_headers): Fix updating of
24265 loop profile and bounds after header duplication.
24266 * tree-ssa-loop-ivcanon.cc (adjust_loop_info_after_peeling):
24267 Break out from try_peel_loop; fix handling of 0 iterations.
24268 (try_peel_loop): Use adjust_loop_info_after_peeling.
24270 2023-04-21 Andrew MacLeod <amacleod@redhat.com>
24272 PR tree-optimization/109546
24273 * tree-vrp.cc (remove_unreachable::remove_and_update_globals): Do
24274 not fold conditions with ADDR_EXPR early.
24276 2023-04-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24278 * config/aarch64/aarch64.md (aarch64_umax<mode>3_insn): Delete.
24279 (umax<mode>3): Emit raw UMAX RTL instead of going through gen_ function
24281 (<optab><mode>3): New define_expand for MAXMIN_NOUMAX codes.
24282 (*aarch64_<optab><mode>3_zero): Define.
24283 (*aarch64_<optab><mode>3_cssc): Likewise.
24284 * config/aarch64/iterators.md (maxminand): New code attribute.
24286 2023-04-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24289 * config/aarch64/aarch64-opts.h (enum aarch64_tp_reg): Define.
24290 * config/aarch64/aarch64-protos.h (aarch64_output_load_tp):
24292 * config/aarch64/aarch64.cc (aarch64_tpidr_register): Declare.
24293 (aarch64_override_options_internal): Handle the above.
24294 (aarch64_output_load_tp): New function.
24295 * config/aarch64/aarch64.md (aarch64_load_tp_hard): Call
24296 aarch64_output_load_tp.
24297 * config/aarch64/aarch64.opt (aarch64_tp_reg): Define enum.
24298 (mtp=): New option.
24299 * doc/invoke.texi (AArch64 Options): Document -mtp=.
24301 2023-04-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24304 * config/aarch64/aarch64-simd.md (add_vec_concat_subst_le): Define.
24305 (add_vec_concat_subst_be): Likewise.
24308 (add<mode>3): Rename to...
24309 (add<mode>3<vczle><vczbe>): ... This.
24310 (sub<mode>3): Rename to...
24311 (sub<mode>3<vczle><vczbe>): ... This.
24312 (mul<mode>3): Rename to...
24313 (mul<mode>3<vczle><vczbe>): ... This.
24314 (and<mode>3): Rename to...
24315 (and<mode>3<vczle><vczbe>): ... This.
24316 (ior<mode>3): Rename to...
24317 (ior<mode>3<vczle><vczbe>): ... This.
24318 (xor<mode>3): Rename to...
24319 (xor<mode>3<vczle><vczbe>): ... This.
24320 * config/aarch64/iterators.md (VDZ): Define.
24322 2023-04-21 Patrick Palka <ppalka@redhat.com>
24324 * tree.cc (walk_tree_1): Avoid repeatedly dereferencing tp
24327 2023-04-21 Jan Hubicka <jh@suse.cz>
24329 * tree-ssa-loop-ch.cc (ch_base::copy_headers): Fix previous
24332 2023-04-21 Vineet Gupta <vineetg@rivosinc.com>
24334 * expmed.h (x_shift*_cost): convert to int [speed][mode][shift].
24335 (shift*_cost_ptr ()): Access x_shift*_cost array directly.
24337 2023-04-21 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
24339 * config/aarch64/aarch64.cc (aarch64_simd_dup_constant): Use
24340 force_reg instead of copy_to_mode_reg.
24341 (aarch64_expand_vector_init): Likewise.
24343 2023-04-21 Uroš Bizjak <ubizjak@gmail.com>
24345 * config/i386/i386.h (REG_OK_FOR_INDEX_P, REG_OK_FOR_BASE_P): Remove.
24346 (REG_OK_FOR_INDEX_NONSTRICT_P, REG_OK_FOR_BASE_NONSTRICT_P): Ditto.
24347 (REG_OK_FOR_INDEX_STRICT_P, REG_OK_FOR_BASE_STRICT_P): Ditto.
24348 (FIRST_INDEX_REG, LAST_INDEX_REG): New defines.
24349 (LEGACY_INDEX_REG_P, LEGACY_INDEX_REGNO_P): New macros.
24350 (INDEX_REG_P, INDEX_REGNO_P): Ditto.
24351 (REGNO_OK_FOR_INDEX_P): Use INDEX_REGNO_P predicates.
24352 (REGNO_OK_FOR_INDEX_NONSTRICT_P): New macro.
24353 (EG_OK_FOR_BASE_NONSTRICT_P): Ditto.
24354 * config/i386/predicates.md (index_register_operand):
24355 Use REGNO_OK_FOR_INDEX_P and REGNO_OK_FOR_INDEX_NONSTRICT_P macros.
24356 * config/i386/i386.cc (ix86_legitimate_address_p): Use
24357 REGNO_OK_FOR_BASE_P, REGNO_OK_FOR_BASE_NONSTRICT_P,
24358 REGNO_OK_FOR_INDEX_P and REGNO_OK_FOR_INDEX_NONSTRICT_P macros.
24360 2023-04-21 Jan Hubicka <hubicka@ucw.cz>
24361 Ondrej Kubanek <kubanek0ondrej@gmail.com>
24363 * tree-ssa-loop-ch.cc (ch_base::copy_headers): Update loop header and
24366 2023-04-21 Richard Biener <rguenther@suse.de>
24368 * is-a.h (safe_is_a): New.
24370 2023-04-21 Richard Biener <rguenther@suse.de>
24372 * gimple-iterator.h (gimple_stmt_iterator::operator*): Add.
24373 (gphi_iterator::operator*): Likewise.
24375 2023-04-21 Jan Hubicka <hubicka@ucw.cz>
24376 Michal Jires <michal@jires.eu>
24378 * ipa-inline.cc (class inline_badness): New class.
24379 (edge_heap_t, edge_heap_node_t): Use inline_badness for badness instead
24381 (update_edge_key): Update.
24382 (lookup_recursive_calls): Likewise.
24383 (recursive_inlining): Likewise.
24384 (add_new_edges_to_heap): Likewise.
24385 (inline_small_functions): Likewise.
24387 2023-04-21 Jan Hubicka <hubicka@ucw.cz>
24389 * ipa-devirt.cc (odr_types_equivalent_p): Cleanup warned checks.
24391 2023-04-21 Richard Biener <rguenther@suse.de>
24393 PR tree-optimization/109573
24394 * tree-vect-loop.cc (vectorizable_live_operation): Allow
24395 unhandled SSA copy as well. Demote assert to checking only.
24397 2023-04-21 Richard Biener <rguenther@suse.de>
24399 * df-core.cc (df_analyze): Compute RPO on the reverse graph
24400 for DF_BACKWARD problems.
24401 (loop_post_order_compute): Rename to ...
24402 (loop_rev_post_order_compute): ... this, compute a RPO.
24403 (loop_inverted_post_order_compute): Rename to ...
24404 (loop_inverted_rev_post_order_compute): ... this, compute a RPO.
24405 (df_analyze_loop): Use RPO on the forward graph for DF_FORWARD
24406 problems, RPO on the inverted graph for DF_BACKWARD.
24408 2023-04-21 Richard Biener <rguenther@suse.de>
24410 * cfganal.h (inverted_rev_post_order_compute): Rename
24412 (inverted_post_order_compute): ... this. Add struct function
24413 argument, change allocation to a C array.
24414 * cfganal.cc (inverted_rev_post_order_compute): Likewise.
24415 * lcm.cc (compute_antinout_edge): Adjust.
24416 * lra-lives.cc (lra_create_live_ranges_1): Likewise.
24417 * tree-ssa-dce.cc (remove_dead_stmt): Likewise.
24418 * tree-ssa-pre.cc (compute_antic): Likewise.
24420 2023-04-21 Richard Biener <rguenther@suse.de>
24422 * df.h (df_d::postorder_inverted): Change back to int *,
24424 * df-core.cc (rest_of_handle_df_finish): Adjust.
24425 (df_analyze_1): Likewise.
24426 (df_analyze): For DF_FORWARD problems use RPO on the forward
24428 (loop_inverted_post_order_compute): Adjust API.
24429 (df_analyze_loop): Adjust.
24430 (df_get_n_blocks): Likewise.
24431 (df_get_postorder): Likewise.
24433 2023-04-21 Juzhe-Zhong <juzhe.zhong@rivai.ai>
24436 * config/riscv/riscv-vsetvl.cc
24437 (vector_infos_manager::all_empty_predecessor_p): New function.
24438 (pass_vsetvl::backward_demand_fusion): Ditto.
24439 * config/riscv/riscv-vsetvl.h: Ditto.
24441 2023-04-21 Robin Dapp <rdapp@ventanamicro.com>
24444 * config/riscv/generic.md: Change standard names to insn names.
24446 2023-04-21 Richard Biener <rguenther@suse.de>
24448 * lcm.cc (compute_antinout_edge): Use RPO on the inverted graph.
24449 (compute_laterin): Use RPO.
24450 (compute_available): Likewise.
24452 2023-04-21 Peng Fan <fanpeng@loongson.cn>
24454 * config/loongarch/gnu-user.h (MUSL_DYNAMIC_LINKER): Redefine.
24456 2023-04-21 Juzhe-Zhong <juzhe.zhong@rivai.ai>
24459 * config/riscv/riscv-vsetvl.cc (local_eliminate_vsetvl_insn): New function.
24460 (vector_insn_info::skip_avl_compatible_p): Ditto.
24461 (vector_insn_info::merge): Remove default value.
24462 (pass_vsetvl::compute_local_backward_infos): Ditto.
24463 (pass_vsetvl::cleanup_insns): Add local vsetvl elimination.
24464 * config/riscv/riscv-vsetvl.h: Ditto.
24466 2023-04-20 Alejandro Colomar <alx.manpages@gmail.com>
24468 * doc/extend.texi (Common Function Attributes): Remove duplicate
24471 2023-04-20 Andrew MacLeod <amacleod@redhat.com>
24473 PR tree-optimization/109564
24474 * gimple-range-fold.cc (fold_using_range::range_of_phi): Do no ignore
24475 UNDEFINED range names when deciding if all PHI arguments are the same,
24477 2023-04-20 Jakub Jelinek <jakub@redhat.com>
24479 PR tree-optimization/109011
24480 * tree-vect-patterns.cc (vect_recog_ctz_ffs_pattern): Use
24481 .CTZ (X) = .POPCOUNT ((X - 1) & ~X) in preference to
24482 .CTZ (X) = PREC - .POPCOUNT (X | -X).
24484 2023-04-20 Vladimir N. Makarov <vmakarov@redhat.com>
24486 * lra-constraints.cc (match_reload): Exclude some hard regs for
24487 multi-reg inout reload pseudos used in asm in different mode.
24489 2023-04-20 Uros Bizjak <ubizjak@gmail.com>
24491 * config/arm/arm.cc (thumb1_legitimate_address_p):
24492 Use VIRTUAL_REGISTER_P predicate.
24493 (arm_eliminable_register): Ditto.
24494 * config/avr/avr.md (push<mode>_1): Ditto.
24495 * config/bfin/predicates.md (register_no_elim_operand): Ditto.
24496 * config/h8300/predicates.md (register_no_sp_elim_operand): Ditto.
24497 * config/i386/predicates.md (register_no_elim_operand): Ditto.
24498 * config/iq2000/predicates.md (call_insn_operand): Ditto.
24499 * config/microblaze/microblaze.h (CALL_INSN_OP): Ditto.
24501 2023-04-20 Uros Bizjak <ubizjak@gmail.com>
24504 * config/i386/predicates.md (extract_operator): New predicate.
24505 * config/i386/i386.md (any_extract): Remove code iterator.
24506 (*cmpqi_ext<mode>_1_mem_rex64): Use extract_operator predicate.
24507 (*cmpqi_ext<mode>_1): Ditto.
24508 (*cmpqi_ext<mode>_2): Ditto.
24509 (*cmpqi_ext<mode>_3_mem_rex64): Ditto.
24510 (*cmpqi_ext<mode>_3): Ditto.
24511 (*cmpqi_ext<mode>_4): Ditto.
24512 (*extzvqi_mem_rex64): Ditto.
24514 (*insvqi_2): Ditto.
24515 (*extendqi<SWI24:mode>_ext_1): Ditto.
24516 (*addqi_ext<mode>_0): Ditto.
24517 (*addqi_ext<mode>_1): Ditto.
24518 (*addqi_ext<mode>_2): Ditto.
24519 (*subqi_ext<mode>_0): Ditto.
24520 (*subqi_ext<mode>_2): Ditto.
24521 (*testqi_ext<mode>_1): Ditto.
24522 (*testqi_ext<mode>_2): Ditto.
24523 (*andqi_ext<mode>_0): Ditto.
24524 (*andqi_ext<mode>_1): Ditto.
24525 (*andqi_ext<mode>_1_cc): Ditto.
24526 (*andqi_ext<mode>_2): Ditto.
24527 (*<any_or:code>qi_ext<mode>_0): Ditto.
24528 (*<any_or:code>qi_ext<mode>_1): Ditto.
24529 (*<any_or:code>qi_ext<mode>_2): Ditto.
24530 (*xorqi_ext<mode>_1_cc): Ditto.
24531 (*negqi_ext<mode>_2): Ditto.
24532 (*ashlqi_ext<mode>_2): Ditto.
24533 (*<any_shiftrt:insn>qi_ext<mode>_2): Ditto.
24535 2023-04-20 Raphael Zinsly <rzinsly@ventanamicro.com>
24538 * config/riscv/bitmanip.md (clz, ctz, pcnt, min, max patterns): Use
24539 <bitmanip_insn> as the type to allow for fine grained control of
24540 scheduling these insns.
24541 * config/riscv/generic.md (generic_alu): Add bitmanip, clz, ctz, pcnt,
24543 * config/riscv/riscv.md (type attribute): Add types for clz, ctz,
24544 pcnt, signed and unsigned min/max.
24546 2023-04-20 Juzhe-Zhong <juzhe.zhong@rivai.ai>
24547 kito-cheng <kito.cheng@sifive.com>
24549 * config/riscv/riscv.h (enum reg_class): Fix RVV register order.
24551 2023-04-20 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
24552 kito-cheng <kito.cheng@sifive.com>
24555 * config/riscv/riscv-vsetvl.cc (count_regno_occurrences): New function.
24556 (pass_vsetvl::cleanup_insns): Fix bug.
24558 2023-04-20 Andrew Stubbs <ams@codesourcery.com>
24560 * config/gcn/gcn-valu.md (vnsi, VnSI): Add scalar modes.
24561 (ldexp<mode>3): Delete.
24562 (ldexp<mode>3<exec>): Change "B" to "A".
24564 2023-04-20 Jakub Jelinek <jakub@redhat.com>
24565 Jonathan Wakely <jwakely@redhat.com>
24567 * tree.h (built_in_function_equal_p): New helper function.
24568 (fndecl_built_in_p): Turn into variadic template to support
24569 1 or more built_in_function arguments.
24570 * builtins.cc (fold_builtin_expect): Use 3 argument fndecl_built_in_p.
24571 * gimplify.cc (goa_stabilize_expr): Likewise.
24572 * cgraphclones.cc (cgraph_node::create_clone): Likewise.
24573 * ipa-fnsummary.cc (compute_fn_summary): Likewise.
24574 * omp-low.cc (setjmp_or_longjmp_p): Likewise.
24575 * cgraph.cc (cgraph_edge::redirect_call_stmt_to_callee,
24576 cgraph_update_edges_for_call_stmt_node,
24577 cgraph_edge::verify_corresponds_to_fndecl,
24578 cgraph_node::verify_node): Likewise.
24579 * tree-stdarg.cc (optimize_va_list_gpr_fpr_size): Likewise.
24580 * gimple-ssa-warn-access.cc (matching_alloc_calls_p): Likewise.
24581 * ipa-prop.cc (try_make_edge_direct_virtual_call): Likewise.
24583 2023-04-20 Jakub Jelinek <jakub@redhat.com>
24585 PR tree-optimization/109011
24586 * tree-vect-patterns.cc (vect_recog_ctz_ffs_pattern): New function.
24587 (vect_recog_popcount_clz_ctz_ffs_pattern): Move vect_pattern_detected
24588 call later. Don't punt for IFN_CTZ or IFN_FFS if it doesn't have
24589 direct optab support, but has instead IFN_CLZ, IFN_POPCOUNT or
24590 for IFN_FFS IFN_CTZ support, use vect_recog_ctz_ffs_pattern for that
24592 (vect_vect_recog_func_ptrs): Add ctz_ffs entry.
24594 2023-04-20 Richard Biener <rguenther@suse.de>
24596 * df-core.cc (rest_of_handle_df_initialize): Remove
24597 computation of df->postorder, df->postorder_inverted and
24600 2023-04-20 Haochen Jiang <haochen.jiang@intel.com>
24602 * common/config/i386/i386-common.cc
24603 (OPTION_MASK_ISA2_AVX_UNSET): Add OPTION_MASK_ISA2_VAES_UNSET.
24604 (ix86_handle_option): Set AVX flag for VAES.
24605 * config/i386/i386-builtins.cc (ix86_init_mmx_sse_builtins):
24606 Add OPTION_MASK_ISA2_VAES_UNSET.
24607 (def_builtin): Share builtin between AES and VAES.
24608 * config/i386/i386-expand.cc (ix86_check_builtin_isa_match):
24610 * config/i386/i386.md (aes): New isa attribute.
24611 * config/i386/sse.md (aesenc): Add pattern for VAES with xmm.
24612 (aesenclast): Ditto.
24614 (aesdeclast): Ditto.
24615 * config/i386/vaesintrin.h: Remove redundant avx target push.
24616 * config/i386/wmmintrin.h (_mm_aesdec_si128): Change to macro.
24617 (_mm_aesdeclast_si128): Ditto.
24618 (_mm_aesenc_si128): Ditto.
24619 (_mm_aesenclast_si128): Ditto.
24621 2023-04-20 Hu, Lin1 <lin1.hu@intel.com>
24623 * config/i386/avx2intrin.h
24624 (_MM_REDUCE_OPERATOR_BASIC_EPI16): New macro.
24625 (_MM_REDUCE_OPERATOR_MAX_MIN_EP16): Ditto.
24626 (_MM256_REDUCE_OPERATOR_BASIC_EPI16): Ditto.
24627 (_MM256_REDUCE_OPERATOR_MAX_MIN_EP16): Ditto.
24628 (_MM_REDUCE_OPERATOR_BASIC_EPI8): Ditto.
24629 (_MM_REDUCE_OPERATOR_MAX_MIN_EP8): Ditto.
24630 (_MM256_REDUCE_OPERATOR_BASIC_EPI8): Ditto.
24631 (_MM256_REDUCE_OPERATOR_MAX_MIN_EP8): Ditto.
24632 (_mm_reduce_add_epi16): New instrinsics.
24633 (_mm_reduce_mul_epi16): Ditto.
24634 (_mm_reduce_and_epi16): Ditto.
24635 (_mm_reduce_or_epi16): Ditto.
24636 (_mm_reduce_max_epi16): Ditto.
24637 (_mm_reduce_max_epu16): Ditto.
24638 (_mm_reduce_min_epi16): Ditto.
24639 (_mm_reduce_min_epu16): Ditto.
24640 (_mm256_reduce_add_epi16): Ditto.
24641 (_mm256_reduce_mul_epi16): Ditto.
24642 (_mm256_reduce_and_epi16): Ditto.
24643 (_mm256_reduce_or_epi16): Ditto.
24644 (_mm256_reduce_max_epi16): Ditto.
24645 (_mm256_reduce_max_epu16): Ditto.
24646 (_mm256_reduce_min_epi16): Ditto.
24647 (_mm256_reduce_min_epu16): Ditto.
24648 (_mm_reduce_add_epi8): Ditto.
24649 (_mm_reduce_mul_epi8): Ditto.
24650 (_mm_reduce_and_epi8): Ditto.
24651 (_mm_reduce_or_epi8): Ditto.
24652 (_mm_reduce_max_epi8): Ditto.
24653 (_mm_reduce_max_epu8): Ditto.
24654 (_mm_reduce_min_epi8): Ditto.
24655 (_mm_reduce_min_epu8): Ditto.
24656 (_mm256_reduce_add_epi8): Ditto.
24657 (_mm256_reduce_mul_epi8): Ditto.
24658 (_mm256_reduce_and_epi8): Ditto.
24659 (_mm256_reduce_or_epi8): Ditto.
24660 (_mm256_reduce_max_epi8): Ditto.
24661 (_mm256_reduce_max_epu8): Ditto.
24662 (_mm256_reduce_min_epi8): Ditto.
24663 (_mm256_reduce_min_epu8): Ditto.
24664 * config/i386/avx512vlbwintrin.h:
24665 (_mm_mask_reduce_add_epi16): Ditto.
24666 (_mm_mask_reduce_mul_epi16): Ditto.
24667 (_mm_mask_reduce_and_epi16): Ditto.
24668 (_mm_mask_reduce_or_epi16): Ditto.
24669 (_mm_mask_reduce_max_epi16): Ditto.
24670 (_mm_mask_reduce_max_epu16): Ditto.
24671 (_mm_mask_reduce_min_epi16): Ditto.
24672 (_mm_mask_reduce_min_epu16): Ditto.
24673 (_mm256_mask_reduce_add_epi16): Ditto.
24674 (_mm256_mask_reduce_mul_epi16): Ditto.
24675 (_mm256_mask_reduce_and_epi16): Ditto.
24676 (_mm256_mask_reduce_or_epi16): Ditto.
24677 (_mm256_mask_reduce_max_epi16): Ditto.
24678 (_mm256_mask_reduce_max_epu16): Ditto.
24679 (_mm256_mask_reduce_min_epi16): Ditto.
24680 (_mm256_mask_reduce_min_epu16): Ditto.
24681 (_mm_mask_reduce_add_epi8): Ditto.
24682 (_mm_mask_reduce_mul_epi8): Ditto.
24683 (_mm_mask_reduce_and_epi8): Ditto.
24684 (_mm_mask_reduce_or_epi8): Ditto.
24685 (_mm_mask_reduce_max_epi8): Ditto.
24686 (_mm_mask_reduce_max_epu8): Ditto.
24687 (_mm_mask_reduce_min_epi8): Ditto.
24688 (_mm_mask_reduce_min_epu8): Ditto.
24689 (_mm256_mask_reduce_add_epi8): Ditto.
24690 (_mm256_mask_reduce_mul_epi8): Ditto.
24691 (_mm256_mask_reduce_and_epi8): Ditto.
24692 (_mm256_mask_reduce_or_epi8): Ditto.
24693 (_mm256_mask_reduce_max_epi8): Ditto.
24694 (_mm256_mask_reduce_max_epu8): Ditto.
24695 (_mm256_mask_reduce_min_epi8): Ditto.
24696 (_mm256_mask_reduce_min_epu8): Ditto.
24698 2023-04-20 Haochen Jiang <haochen.jiang@intel.com>
24700 * common/config/i386/i386-common.cc
24701 (OPTION_MASK_ISA_VPCLMULQDQ_SET):
24702 Add OPTION_MASK_ISA_PCLMUL_SET and OPTION_MASK_ISA_AVX_SET.
24703 (OPTION_MASK_ISA_AVX_UNSET):
24704 Add OPTION_MASK_ISA_VPCLMULQDQ_UNSET.
24705 (OPTION_MASK_ISA_PCLMUL_UNSET): Ditto.
24706 * config/i386/i386.md (vpclmulqdqvl): New.
24707 * config/i386/sse.md (pclmulqdq): Add evex encoding.
24708 * config/i386/vpclmulqdqintrin.h: Remove redudant avx target
24711 2023-04-20 Haochen Jiang <haochen.jiang@intel.com>
24713 * config/i386/avx512vlbwintrin.h
24714 (_mm_mask_blend_epi16): Remove __OPTIMIZE__ wrapper.
24715 (_mm_mask_blend_epi8): Ditto.
24716 (_mm256_mask_blend_epi16): Ditto.
24717 (_mm256_mask_blend_epi8): Ditto.
24718 * config/i386/avx512vlintrin.h
24719 (_mm256_mask_blend_pd): Ditto.
24720 (_mm256_mask_blend_ps): Ditto.
24721 (_mm256_mask_blend_epi64): Ditto.
24722 (_mm256_mask_blend_epi32): Ditto.
24723 (_mm_mask_blend_pd): Ditto.
24724 (_mm_mask_blend_ps): Ditto.
24725 (_mm_mask_blend_epi64): Ditto.
24726 (_mm_mask_blend_epi32): Ditto.
24727 * config/i386/sse.md (VF_AVX512BWHFBF16): Removed.
24728 (VF_AVX512HFBFVL): Move it before the first usage.
24729 (<avx512>_blendm<mode>): Change iterator from VF_AVX512BWHFBF16
24730 to VF_AVX512HFBFVL.
24732 2023-04-20 Haochen Jiang <haochen.jiang@intel.com>
24734 * common/config/i386/i386-common.cc
24735 (OPTION_MASK_ISA_AVX512VBMI2_SET): Change OPTION_MASK_ISA_AVX512F_SET
24736 to OPTION_MASK_ISA_AVX512BW_SET.
24737 (OPTION_MASK_ISA_AVX512F_UNSET):
24738 Remove OPTION_MASK_ISA_AVX512VBMI2_UNSET.
24739 (OPTION_MASK_ISA_AVX512BW_UNSET):
24740 Add OPTION_MASK_ISA_AVX512VBMI2_UNSET.
24741 * config/i386/avx512vbmi2intrin.h: Do not push avx512bw.
24742 * config/i386/avx512vbmi2vlintrin.h: Ditto.
24743 * config/i386/i386-builtin.def: Remove OPTION_MASK_ISA_AVX512BW.
24744 * config/i386/sse.md (VI12_AVX512VLBW): Removed.
24745 (VI12_VI48F_AVX512VLBW): Rename to VI12_VI48F_AVX512VL.
24746 (compress<mode>_mask): Change iterator from VI12_AVX512VLBW to
24748 (compressstore<mode>_mask): Ditto.
24749 (expand<mode>_mask): Ditto.
24750 (expand<mode>_maskz): Ditto.
24751 (*expand<mode>_mask): Change iterator from VI12_VI48F_AVX512VLBW to
24752 VI12_VI48F_AVX512VL.
24754 2023-04-20 Haochen Jiang <haochen.jiang@intel.com>
24756 * common/config/i386/i386-common.cc
24757 (OPTION_MASK_ISA_AVX512BITALG_SET):
24758 Change OPTION_MASK_ISA_AVX512F_SET
24759 to OPTION_MASK_ISA_AVX512BW_SET.
24760 (OPTION_MASK_ISA_AVX512F_UNSET):
24761 Remove OPTION_MASK_ISA_AVX512BITALG_SET.
24762 (OPTION_MASK_ISA_AVX512BW_UNSET):
24763 Add OPTION_MASK_ISA_AVX512BITALG_SET.
24764 * config/i386/avx512bitalgintrin.h: Do not push avx512bw.
24765 * config/i386/i386-builtin.def:
24766 Remove redundant OPTION_MASK_ISA_AVX512BW.
24767 * config/i386/sse.md (VI1_AVX512VLBW): Removed.
24768 (avx512vl_vpshufbitqmb<mode><mask_scalar_merge_name>):
24769 Change the iterator from VI1_AVX512VLBW to VI1_AVX512VL.
24771 2023-04-20 Haochen Jiang <haochen.jiang@intel.com>
24773 * config/i386/i386-expand.cc
24774 (ix86_check_builtin_isa_match): Correct wrong comments.
24775 Add a new macro SHARE_BUILTIN and refactor the current if
24778 2023-04-20 Mo, Zewei <zewei.mo@intel.com>
24780 * config/i386/cpuid.h: Open a new section for Extended Features
24781 Leaf (%eax == 7, %ecx == 0) and Extended Features Sub-leaf (%eax == 7,
24784 2023-04-20 Hu, Lin1 <lin1.hu@intel.com>
24786 * config/i386/sse.md: Modify insn vperm{i,f}
24789 2023-04-19 Max Filippov <jcmvbkbc@gmail.com>
24791 * config/xtensa/xtensa-opts.h: New header.
24792 * config/xtensa/xtensa.h (STRICT_ALIGNMENT): Redefine as
24793 xtensa_strict_align.
24794 * config/xtensa/xtensa.cc (xtensa_option_override): When
24795 -m[no-]strict-align is not specified in the command line set
24796 xtensa_strict_align to 0 if the hardware supports both unaligned
24797 loads and stores or to 1 otherwise.
24798 * config/xtensa/xtensa.opt (mstrict-align): New option.
24799 * doc/invoke.texi (Xtensa Options): Document -m[no-]strict-align.
24801 2023-04-19 Max Filippov <jcmvbkbc@gmail.com>
24803 * config/xtensa/xtensa-dynconfig.cc (xtensa_get_config_v4): New
24806 2023-04-19 Andrew Pinski <apinski@marvell.com>
24808 * config/i386/i386.md (*movsicc_noc_zext_1): New pattern.
24810 2023-04-19 Juzhe-Zhong <juzhe.zhong@rivai.ai>
24812 * config/riscv/riscv-modes.def (FLOAT_MODE): Add chunk 128 support.
24813 (VECTOR_BOOL_MODE): Ditto.
24814 (ADJUST_NUNITS): Ditto.
24815 (ADJUST_ALIGNMENT): Ditto.
24816 (ADJUST_BYTESIZE): Ditto.
24817 (ADJUST_PRECISION): Ditto.
24818 (RVV_MODES): Ditto.
24819 (VECTOR_MODE_WITH_PREFIX): Ditto.
24820 * config/riscv/riscv-v.cc (ENTRY): Ditto.
24821 (get_vlmul): Ditto.
24822 (get_ratio): Ditto.
24823 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_TYPE): Ditto.
24824 * config/riscv/riscv-vector-builtins.def (DEF_RVV_TYPE): Ditto.
24825 (vbool64_t): Ditto.
24826 (vbool32_t): Ditto.
24827 (vbool16_t): Ditto.
24832 (vint8mf8_t): Ditto.
24833 (vuint8mf8_t): Ditto.
24834 (vint8mf4_t): Ditto.
24835 (vuint8mf4_t): Ditto.
24836 (vint8mf2_t): Ditto.
24837 (vuint8mf2_t): Ditto.
24838 (vint8m1_t): Ditto.
24839 (vuint8m1_t): Ditto.
24840 (vint8m2_t): Ditto.
24841 (vuint8m2_t): Ditto.
24842 (vint8m4_t): Ditto.
24843 (vuint8m4_t): Ditto.
24844 (vint8m8_t): Ditto.
24845 (vuint8m8_t): Ditto.
24846 (vint16mf4_t): Ditto.
24847 (vuint16mf4_t): Ditto.
24848 (vint16mf2_t): Ditto.
24849 (vuint16mf2_t): Ditto.
24850 (vint16m1_t): Ditto.
24851 (vuint16m1_t): Ditto.
24852 (vint16m2_t): Ditto.
24853 (vuint16m2_t): Ditto.
24854 (vint16m4_t): Ditto.
24855 (vuint16m4_t): Ditto.
24856 (vint16m8_t): Ditto.
24857 (vuint16m8_t): Ditto.
24858 (vint32mf2_t): Ditto.
24859 (vuint32mf2_t): Ditto.
24860 (vint32m1_t): Ditto.
24861 (vuint32m1_t): Ditto.
24862 (vint32m2_t): Ditto.
24863 (vuint32m2_t): Ditto.
24864 (vint32m4_t): Ditto.
24865 (vuint32m4_t): Ditto.
24866 (vint32m8_t): Ditto.
24867 (vuint32m8_t): Ditto.
24868 (vint64m1_t): Ditto.
24869 (vuint64m1_t): Ditto.
24870 (vint64m2_t): Ditto.
24871 (vuint64m2_t): Ditto.
24872 (vint64m4_t): Ditto.
24873 (vuint64m4_t): Ditto.
24874 (vint64m8_t): Ditto.
24875 (vuint64m8_t): Ditto.
24876 (vfloat32mf2_t): Ditto.
24877 (vfloat32m1_t): Ditto.
24878 (vfloat32m2_t): Ditto.
24879 (vfloat32m4_t): Ditto.
24880 (vfloat32m8_t): Ditto.
24881 (vfloat64m1_t): Ditto.
24882 (vfloat64m2_t): Ditto.
24883 (vfloat64m4_t): Ditto.
24884 (vfloat64m8_t): Ditto.
24885 * config/riscv/riscv-vector-switch.def (ENTRY): Ditto.
24886 * config/riscv/riscv.cc (riscv_legitimize_poly_move): Ditto.
24887 (riscv_convert_vector_bits): Ditto.
24888 * config/riscv/riscv.md:
24889 * config/riscv/vector-iterators.md:
24890 * config/riscv/vector.md
24891 (@pred_indexed_<order>store<VNX32_QH:mode><VNX32_QHI:mode>): Ditto.
24892 (@pred_indexed_<order>store<VNX32_QHS:mode><VNX32_QHSI:mode>): Ditto.
24893 (@pred_indexed_<order>store<VNX64_Q:mode><VNX64_Q:mode>): Ditto.
24894 (@pred_indexed_<order>store<VNX64_QH:mode><VNX64_QHI:mode>): Ditto.
24895 (@pred_indexed_<order>store<VNX128_Q:mode><VNX128_Q:mode>): Ditto.
24896 (@pred_reduc_<reduc><mode><vlmul1_zve64>): Ditto.
24897 (@pred_widen_reduc_plus<v_su><mode><vwlmul1_zve64>): Ditto.
24898 (@pred_reduc_plus<order><mode><vlmul1_zve64>): Ditto.
24899 (@pred_widen_reduc_plus<order><mode><vwlmul1_zve64>): Ditto.
24901 2023-04-19 Pan Li <pan2.li@intel.com>
24903 * simplify-rtx.cc (simplify_context::simplify_binary_operation_1):
24904 Align IOR (A | (~A) -> -1) optimization MODE_CLASS condition to AND.
24906 2023-04-19 Uros Bizjak <ubizjak@gmail.com>
24910 * config/i386/i386.md (*cmpqi_ext<mode>_1_mem_rex64): New insn pattern.
24911 (*cmpqi_ext<mode>_1): Use nonimmediate_operand predicate
24912 for operand 0. Use any_extract code iterator.
24913 (*cmpqi_ext<mode>_1 peephole2): New peephole2 pattern.
24914 (*cmpqi_ext<mode>_2): Use any_extract code iterator.
24915 (*cmpqi_ext<mode>_3_mem_rex64): New insn pattern.
24916 (*cmpqi_ext<mode>_1): Use general_operand predicate
24917 for operand 1. Use any_extract code iterator.
24918 (*cmpqi_ext<mode>_3 peephole2): New peephole2 pattern.
24919 (*cmpqi_ext<mode>_4): Use any_extract code iterator.
24921 2023-04-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24923 * config/aarch64/aarch64-simd.md (aarch64_saddw2<mode>): Delete.
24924 (aarch64_uaddw2<mode>): Delete.
24925 (aarch64_ssubw2<mode>): Delete.
24926 (aarch64_usubw2<mode>): Delete.
24927 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>): New define_expand.
24929 2023-04-19 Richard Biener <rguenther@suse.de>
24931 * tree-ssa-structalias.cc (do_ds_constraint): Use
24932 solve_add_graph_edge.
24934 2023-04-19 Richard Biener <rguenther@suse.de>
24936 * tree-ssa-structalias.cc (solve_add_graph_edge): New function,
24938 (do_sd_constraint): ... here.
24940 2023-04-19 Richard Biener <rguenther@suse.de>
24942 * tree-cfg.cc (gimple_can_merge_blocks_p): Remove condition
24943 rejecting the merge when A contains only a non-local label.
24945 2023-04-19 Uros Bizjak <ubizjak@gmail.com>
24947 * rtl.h (VIRTUAL_REGISTER_P): New predicate.
24948 (VIRTUAL_REGISTER_NUM_P): Ditto.
24949 (REGNO_PTR_FRAME_P): Use VIRTUAL_REGISTER_NUM_P predicate.
24950 * expr.cc (force_operand): Use VIRTUAL_REGISTER_P predicate.
24951 * function.cc (instantiate_decl_rtl): Ditto.
24952 * rtlanal.cc (rtx_addr_can_trap_p_1): Ditto.
24953 (nonzero_address_p): Ditto.
24954 (refers_to_regno_p): Use VIRTUAL_REGISTER_NUM_P predicate.
24956 2023-04-19 Aldy Hernandez <aldyh@redhat.com>
24958 * value-range.h (Value_Range::Value_Range): Avoid pointer sharing.
24960 2023-04-19 Richard Biener <rguenther@suse.de>
24962 * system.h (auto_mpz::operator->()): New.
24963 * realmpfr.h (auto_mpfr::operator->()): New.
24964 * builtins.cc (do_mpfr_lgamma_r): Use auto_mpfr.
24965 * real.cc (real_from_string): Likewise.
24966 (dconst_e_ptr): Likewise.
24967 (dconst_sqrt2_ptr): Likewise.
24968 * tree-ssa-loop-niter.cc (refine_value_range_using_guard):
24970 (bound_difference_of_offsetted_base): Likewise.
24971 (number_of_iterations_ne): Likewise.
24972 (number_of_iterations_lt_to_ne): Likewise.
24973 * ubsan.cc: Include realmpfr.h.
24974 (ubsan_instrument_float_cast): Use auto_mpfr.
24976 2023-04-19 Richard Biener <rguenther@suse.de>
24978 * tree-ssa-structalias.cc (solve_graph): Remove self-copy
24979 edges, remove edges from escaped after special-casing them.
24981 2023-04-19 Richard Biener <rguenther@suse.de>
24983 * tree-ssa-structalias.cc (do_sd_constraint): Fixup escape
24986 2023-04-19 Richard Biener <rguenther@suse.de>
24988 * tree-ssa-structalias.cc (do_sd_constraint): Do not write
24989 to the LHS varinfo solution member.
24991 2023-04-19 Richard Biener <rguenther@suse.de>
24993 * tree-ssa-structalias.cc (topo_visit): Look at the real
24994 destination of edges.
24996 2023-04-19 Richard Biener <rguenther@suse.de>
24998 PR tree-optimization/44794
24999 * tree-ssa-loop-manip.cc (tree_transform_and_unroll_loop):
25000 If an epilogue loop is required set its iteration upper bound.
25002 2023-04-19 Xi Ruoyao <xry111@xry111.site>
25005 * config/loongarch/loongarch-protos.h
25006 (loongarch_expand_block_move): Add a parameter as alignment RTX.
25007 * config/loongarch/loongarch.h:
25008 (LARCH_MAX_MOVE_BYTES_PER_LOOP_ITER): Remove.
25009 (LARCH_MAX_MOVE_BYTES_STRAIGHT): Remove.
25010 (LARCH_MAX_MOVE_OPS_PER_LOOP_ITER): Define.
25011 (LARCH_MAX_MOVE_OPS_STRAIGHT): Define.
25012 (MOVE_RATIO): Use LARCH_MAX_MOVE_OPS_PER_LOOP_ITER instead of
25013 LARCH_MAX_MOVE_BYTES_PER_LOOP_ITER.
25014 * config/loongarch/loongarch.cc (loongarch_expand_block_move):
25015 Take the alignment from the parameter, but set it to
25016 UNITS_PER_WORD if !TARGET_STRICT_ALIGN. Limit the length of
25017 straight-line implementation with LARCH_MAX_MOVE_OPS_STRAIGHT
25018 instead of LARCH_MAX_MOVE_BYTES_STRAIGHT.
25019 (loongarch_block_move_straight): When there are left-over bytes,
25020 half the mode size instead of falling back to byte mode at once.
25021 (loongarch_block_move_loop): Limit the length of loop body with
25022 LARCH_MAX_MOVE_OPS_PER_LOOP_ITER instead of
25023 LARCH_MAX_MOVE_BYTES_PER_LOOP_ITER.
25024 * config/loongarch/loongarch.md (cpymemsi): Pass the alignment
25025 to loongarch_expand_block_move.
25027 2023-04-19 Xi Ruoyao <xry111@xry111.site>
25029 * config/loongarch/loongarch.cc
25030 (loongarch_setup_incoming_varargs): Don't save more GARs than
25031 cfun->va_list_gpr_size / UNITS_PER_WORD.
25033 2023-04-19 Richard Biener <rguenther@suse.de>
25035 * tree-ssa-loop-manip.cc (determine_exit_conditions): Fix
25036 no epilogue condition.
25038 2023-04-19 Richard Biener <rguenther@suse.de>
25040 * gimple.h (gimple_assign_load): Outline...
25041 * gimple.cc (gimple_assign_load): ... here. Avoid
25042 get_base_address and instead just strip the outermost
25043 handled component, treating a remaining handled component
25046 2023-04-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25048 * config/aarch64/aarch64-simd-builtins.def (neg): Delete builtins
25050 * config/aarch64/arm_fp16.h (vnegh_f16): Reimplement using normal negation.
25052 2023-04-19 Jakub Jelinek <jakub@redhat.com>
25054 PR tree-optimization/109011
25055 * tree-vect-patterns.cc (vect_recog_popcount_pattern): Rename to ...
25056 (vect_recog_popcount_clz_ctz_ffs_pattern): ... this. Handle also
25057 CLZ, CTZ and FFS. Remove vargs variable, use
25058 gimple_build_call_internal rather than gimple_build_call_internal_vec.
25059 (vect_vect_recog_func_ptrs): Adjust popcount entry.
25061 2023-04-19 Jakub Jelinek <jakub@redhat.com>
25064 * dse.cc (replace_read): If read_reg is a SUBREG of a word mode
25065 REG, for WORD_REGISTER_OPERATIONS copy SUBREG_REG of it into
25066 a new REG rather than the SUBREG.
25068 2023-04-19 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
25070 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_set_zero<mode>):
25073 2023-04-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25076 * config/aarch64/aarch64.cc (aarch64_rtx_costs): Merge ASHIFT and
25077 ROTATE, ROTATERT, LSHIFTRT, ASHIFTRT cases. Handle subregs in op1.
25079 2023-04-19 Richard Biener <rguenther@suse.de>
25081 PR rtl-optimization/109237
25082 * cse.cc (insn_live_p): Remove NEXT_INSN walk, instead check
25083 TREE_VISITED on INSN_VAR_LOCATION_DECL.
25084 (delete_trivially_dead_insns): Maintain TREE_VISITED on
25085 active debug bind INSN_VAR_LOCATION_DECL.
25087 2023-04-19 Richard Biener <rguenther@suse.de>
25089 PR rtl-optimization/109237
25090 * cfgcleanup.cc (bb_is_just_return): Walk insns backwards.
25092 2023-04-19 Christophe Lyon <christophe.lyon@arm.com>
25094 * doc/install.texi (enable-decimal-float): Add AArch64.
25096 2023-04-19 liuhongt <hongtao.liu@intel.com>
25098 PR rtl-optimization/109351
25099 * ira.cc (setup_class_subset_and_memory_move_costs): Check
25100 hard_regno_mode_ok before setting lowest memory move cost for
25101 the mode with different reg classes.
25103 2023-04-18 Jason Merrill <jason@redhat.com>
25105 * doc/invoke.texi: Remove stray @gol.
25107 2023-04-18 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
25109 * ifcvt.cc (cond_move_process_if_block): Consider the result of
25110 targetm.noce_conversion_profitable_p() when replacing the original
25111 sequence with the converted one.
25113 2023-04-18 Mark Harmstone <mark@harmstone.com>
25115 * common.opt (gcodeview): Add new option.
25116 * gcc.cc (driver_handle_option); Handle OPT_gcodeview.
25117 * opts.cc (command_handle_option): Similarly.
25118 * doc/invoke.texi: Add documentation for -gcodeview.
25120 2023-04-18 Andrew Pinski <apinski@marvell.com>
25122 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker): Remove declaration.
25123 (make_pass_phiopt): Make execute out of line.
25124 (tree_ssa_cs_elim): Move code into ...
25125 (pass_cselim::execute): here.
25127 2023-04-18 Sam James <sam@gentoo.org>
25129 * system.h: Drop unused INCLUDE_PTHREAD_H.
25131 2023-04-18 Kevin Lee <kevinl@rivosinc.com>
25133 * tree-vect-data-refs.cc (vect_grouped_store_supported): Add new
25136 2023-04-18 Sinan Lin <sinan.lin@linux.alibaba.com>
25138 * config/riscv/bitmanip.md (rotr<mode>3 expander): Enable for ZBKB.
25139 (bswapdi2, bswapsi2): Similarly.
25141 2023-04-18 Uros Bizjak <ubizjak@gmail.com>
25144 * config/i386/i386-builtin.def (__builtin_ia32_insertps128):
25145 Use CODE_FOR_sse4_1_insertps_v4sf.
25146 * config/i386/i386-expand.cc (expand_vec_perm_insertps): New.
25147 (expand_vec_perm_1): Call expand_vec_per_insertps.
25148 * config/i386/i386.md ("unspec"): Declare UNSPEC_INSERTPS here.
25149 * config/i386/mmx.md (mmxscalarmode): New mode attribute.
25150 (@sse4_1_insertps_<mode>): New insn pattern.
25151 * config/i386/sse.md (@sse4_1_insertps_<mode>): Macroize insn
25152 pattern from sse4_1_insertps using VI4F_128 mode iterator.
25154 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
25156 * value-range.cc (gt_ggc_mx): New.
25158 * value-range.h (class vrange): Add GTY marker.
25159 (class frange): Same.
25160 (gt_ggc_mx): Remove.
25161 (gt_pch_nx): Remove.
25163 2023-04-18 Victor L. Do Nascimento <victor.donascimento@arm.com>
25165 * lra-constraints.cc (constraint_unique): New.
25166 (process_address_1): Apply constraint_unique test.
25167 * recog.cc (constrain_operands): Allow relaxed memory
25170 2023-04-18 Kito Cheng <kito.cheng@sifive.com>
25172 * doc/extend.texi (Target Builtins): Add RISC-V Vector
25174 (RISC-V Vector Intrinsics): Document GCC implemented which
25175 version of RISC-V vector intrinsics and its reference.
25177 2023-04-18 Richard Biener <rguenther@suse.de>
25179 PR middle-end/108786
25180 * bitmap.h (bitmap_clear_first_set_bit): New.
25181 * bitmap.cc (bitmap_first_set_bit_worker): Rename from
25182 bitmap_first_set_bit and add optional clearing of the bit.
25183 (bitmap_first_set_bit): Wrap bitmap_first_set_bit_worker.
25184 (bitmap_clear_first_set_bit): Likewise.
25185 * df-core.cc (df_worklist_dataflow_doublequeue): Use
25186 bitmap_clear_first_set_bit.
25187 * graphite-scop-detection.cc (scop_detection::merge_sese):
25189 * sanopt.cc (sanitize_asan_mark_unpoison): Likewise.
25190 (sanitize_asan_mark_poison): Likewise.
25191 * tree-cfgcleanup.cc (cleanup_tree_cfg_noloop): Likewise.
25192 * tree-into-ssa.cc (rewrite_blocks): Likewise.
25193 * tree-ssa-dce.cc (simple_dce_from_worklist): Likewise.
25194 * tree-ssa-sccvn.cc (do_rpo_vn_1): Likewise.
25196 2023-04-18 Richard Biener <rguenther@suse.de>
25198 * tree-ssa-structalias.cc (dump_sa_stats): Split out from...
25199 (dump_sa_points_to_info): ... this function.
25200 (compute_points_to_sets): Guard large dumps with TDF_DETAILS,
25201 and call dump_sa_stats guarded with TDF_STATS.
25202 (ipa_pta_execute): Likewise.
25203 (compute_may_aliases): Guard dump_alias_info with
25204 TDF_DETAILS|TDF_ALIAS.
25206 2023-04-18 Andrew Pinski <apinski@marvell.com>
25208 * tree-ssa-phiopt.cc (gimple_simplify_phiopt): Dump
25209 the expression that is being tried when TDF_FOLDING
25211 (phiopt_worker::match_simplify_replacement): Dump
25212 the sequence which was created by gimple_simplify_phiopt
25213 when TDF_FOLDING is true.
25215 2023-04-18 Andrew Pinski <apinski@marvell.com>
25217 * tree-ssa-phiopt.cc (match_simplify_replacement):
25218 Simplify code that does the movement slightly.
25220 2023-04-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25222 * config/aarch64/aarch64.md (@aarch64_rev16<mode>): Change to
25224 (rev16<mode>2): Rename to...
25225 (aarch64_rev16<mode>2_alt1): ... This.
25226 (rev16<mode>2_alt): Rename to...
25227 (*aarch64_rev16<mode>2_alt2): ... This.
25229 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
25231 * emit-rtl.cc (init_emit_once): Initialize dconstm0.
25232 * gimple-range-op.cc (class cfn_signbit): Remove dconstm0
25234 * range-op-float.cc (zero_range): Use dconstm0.
25235 (zero_to_inf_range): Same.
25236 * real.h (dconstm0): New.
25237 * value-range.cc (frange::flush_denormals_to_zero): Use dconstm0.
25238 (frange::set_zero): Do not declare dconstm0.
25240 2023-04-18 Richard Biener <rguenther@suse.de>
25242 * system.h (class auto_mpz): New,
25243 * realmpfr.h (class auto_mpfr): Likewise.
25244 * fold-const-call.cc (do_mpfr_arg1): Use auto_mpfr.
25245 (do_mpfr_arg2): Likewise.
25246 * tree-ssa-loop-niter.cc (bound_difference): Use auto_mpz;
25248 2023-04-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25250 * config/aarch64/aarch64-builtins.cc (aarch64_init_simd_intrinsics): Take
25251 builtin flags from intrinsic data rather than hardcoded FLAG_AUTO_FP.
25253 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
25255 * value-range.cc (frange::operator==): Adjust for NAN.
25256 (range_tests_nan): Remove some NAN tests.
25258 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
25260 * inchash.cc (hash::add_real_value): New.
25261 * inchash.h (class hash): Add add_real_value.
25262 * value-range.cc (add_vrange): New.
25263 * value-range.h (inchash::add_vrange): New.
25265 2023-04-18 Richard Biener <rguenther@suse.de>
25267 PR tree-optimization/109539
25268 * gimple-ssa-warn-access.cc (pass_waccess::check_pointer_uses):
25269 Re-implement pointer relatedness for PHIs.
25271 2023-04-18 Andrew Stubbs <ams@codesourcery.com>
25273 * config/gcn/gcn-valu.md (SV_SFDF): New iterator.
25274 (SV_FP): New iterator.
25275 (scalar_mode, SCALAR_MODE): Add identity mappings for scalar modes.
25276 (recip<mode>2): Unify the two patterns using SV_FP.
25277 (div_scale<mode><exec_vcc>): New insn.
25278 (div_fmas<mode><exec>): New insn.
25279 (div_fixup<mode><exec>): New insn.
25280 (div<mode>3): Unify the two expanders and rewrite using hardfp.
25281 * config/gcn/gcn.cc (gcn_md_reorg): Support "vccwait" attribute.
25282 * config/gcn/gcn.md (unspec): Add UNSPEC_DIV_SCALE, UNSPEC_DIV_FMAS,
25283 and UNSPEC_DIV_FIXUP.
25284 (vccwait): New attribute.
25286 2023-04-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25288 * config/aarch64/aarch64.cc (aarch64_validate_mcpu): Add hint to use -march
25289 if the argument matches that.
25291 2023-04-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25293 * config/aarch64/atomics.md
25294 (*aarch64_atomic_load<ALLX:mode>_rcpc_zext):
25295 Use SD_HSDI for destination mode iterator.
25297 2023-04-18 Jin Ma <jinma@linux.alibaba.com>
25299 * common/config/riscv/riscv-common.cc (multi_letter_subset_rank): Swap the order
25300 of z-extensions and s-extensions.
25301 (riscv_subset_list::parse): Likewise.
25303 2023-04-18 Jakub Jelinek <jakub@redhat.com>
25305 PR tree-optimization/109240
25306 * match.pd (fneg/fadd): Rewrite such that it handles both plus as
25307 first vec_perm operand and minus as second using fneg/fadd and
25308 minus as first vec_perm operand and plus as second using fneg/fsub.
25310 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
25312 * data-streamer.cc (bp_pack_real_value): New.
25313 (bp_unpack_real_value): New.
25314 * data-streamer.h (bp_pack_real_value): New.
25315 (bp_unpack_real_value): New.
25316 * tree-streamer-in.cc (unpack_ts_real_cst_value_fields): Use
25317 bp_unpack_real_value.
25318 * tree-streamer-out.cc (pack_ts_real_cst_value_fields): Use
25319 bp_pack_real_value.
25321 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
25323 * wide-int.h (WIDE_INT_MAX_HWIS): New.
25324 (class fixed_wide_int_storage): Use it.
25325 (trailing_wide_ints <N>::set_precision): Use it.
25326 (trailing_wide_ints <N>::extra_size): Use it.
25328 2023-04-18 Xi Ruoyao <xry111@xry111.site>
25330 * config/loongarch/loongarch-protos.h
25331 (loongarch_addu16i_imm12_operand_p): New function prototype.
25332 (loongarch_split_plus_constant): Likewise.
25333 * config/loongarch/loongarch.cc
25334 (loongarch_addu16i_imm12_operand_p): New function.
25335 (loongarch_split_plus_constant): Likewise.
25336 * config/loongarch/loongarch.h (ADDU16I_OPERAND): New macro.
25337 (DUAL_IMM12_OPERAND): Likewise.
25338 (DUAL_ADDU16I_OPERAND): Likewise.
25339 * config/loongarch/constraints.md (La, Lb, Lc, Ld, Le): New
25341 * config/loongarch/predicates.md (const_dual_imm12_operand): New
25343 (const_addu16i_operand): Likewise.
25344 (const_addu16i_imm12_di_operand): Likewise.
25345 (const_addu16i_imm12_si_operand): Likewise.
25346 (plus_di_operand): Likewise.
25347 (plus_si_operand): Likewise.
25348 (plus_si_extend_operand): Likewise.
25349 * config/loongarch/loongarch.md (add<mode>3): Convert to
25350 define_insn_and_split. Use plus_<mode>_operand predicate
25351 instead of arith_operand. Add alternatives for La, Lb, Lc, Ld,
25352 and Le constraints.
25353 (*addsi3_extended): Convert to define_insn_and_split. Use
25354 plus_si_extend_operand instead of arith_operand. Add
25355 alternatives for La and Le alternatives.
25357 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
25359 * value-range.h (Value_Range::Value_Range): New.
25360 (Value_Range::contains_p): New.
25362 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
25364 * value-range.h (class vrange): Make m_discriminator const.
25365 (class irange): Make m_max_ranges const. Adjust constructors
25367 (class unsupported_range): Construct vrange appropriately.
25368 (class frange): Same.
25370 2023-04-18 Lulu Cheng <chenglulu@loongson.cn>
25372 * config/loongarch/loongarch.h (LOGICAL_OP_NON_SHORT_CIRCUIT): Remove the macro
25375 2023-04-18 Lulu Cheng <chenglulu@loongson.cn>
25377 * doc/extend.texi: Add section for LoongArch Base Built-in functions.
25379 2023-04-18 Fei Gao <gaofei@eswincomputing.com>
25381 * config/riscv/riscv.cc (riscv_first_stack_step): Make codes more
25383 (riscv_expand_epilogue): Likewise.
25385 2023-04-17 Fei Gao <gaofei@eswincomputing.com>
25387 * config/riscv/riscv.cc (riscv_expand_prologue): Consider save-restore in
25389 (riscv_expand_epilogue): Consider save-restore in stack deallocation.
25391 2023-04-17 Andrew Pinski <apinski@marvell.com>
25393 * tree-ssa-phiopt.cc (gate_hoist_loads): Remove
25396 2023-04-17 Aldy Hernandez <aldyh@redhat.com>
25398 * gimple-ssa-warn-alloca.cc (pass_walloca::execute): Do not export
25401 2023-04-17 Fei Gao <gaofei@eswincomputing.com>
25403 * config/riscv/riscv.cc (riscv_first_stack_step): Add a new function
25404 parameter remaining_size.
25405 (riscv_compute_frame_info): Adapt new riscv_first_stack_step interface.
25406 (riscv_expand_prologue): Likewise.
25407 (riscv_expand_epilogue): Likewise.
25409 2023-04-17 Feng Wang <wangfeng@eswincomputing.com>
25411 * config/riscv/bitmanip.md (rotrsi3_sext): Support generating
25412 roriw for constant counts.
25413 * rtl.h (reverse_rotate_by_imm_p): Add function declartion
25414 * simplify-rtx.cc (reverse_rotate_by_imm_p): New function.
25415 (simplify_context::simplify_binary_operation_1): Use it.
25416 * expmed.cc (expand_shift_1): Likewise.
25418 2023-04-17 Martin Jambor <mjambor@suse.cz>
25422 * cgraph.h (symtab_node::find_reference): Add parameter use_type.
25423 * ipa-prop.h (ipa_pass_through_data): New flag refdesc_decremented.
25424 (ipa_zap_jf_refdesc): New function.
25425 (ipa_get_jf_pass_through_refdesc_decremented): Likewise.
25426 (ipa_set_jf_pass_through_refdesc_decremented): Likewise.
25427 * ipa-cp.cc (ipcp_discover_new_direct_edges): Provide a value for
25428 the new parameter of find_reference.
25429 (adjust_references_in_caller): Likewise. Make sure the constant jump
25430 function is not used to decrement a refdec counter again. Only
25431 decrement refdesc counters when the pass_through jump function allows
25432 it. Added a detailed dump when decrementing refdesc counters.
25433 * ipa-prop.cc (ipa_print_node_jump_functions_for_edge): Dump new flag.
25434 (ipa_set_jf_simple_pass_through): Initialize the new flag.
25435 (ipa_set_jf_unary_pass_through): Likewise.
25436 (ipa_set_jf_arith_pass_through): Likewise.
25437 (remove_described_reference): Provide a value for the new parameter of
25439 (update_jump_functions_after_inlining): Zap refdesc of new jfunc if
25440 the previous pass_through had a flag mandating that we do so.
25441 (propagate_controlled_uses): Likewise. Only decrement refdesc
25442 counters when the pass_through jump function allows it.
25443 (ipa_edge_args_sum_t::duplicate): Provide a value for the new
25444 parameter of find_reference.
25445 (ipa_write_jump_function): Assert the new flag does not have to be
25447 * symtab.cc (symtab_node::find_reference): Add parameter use_type, use
25450 2023-04-17 Philipp Tomsich <philipp.tomsich@vrull.eu>
25451 Di Zhao <di.zhao@amperecomputing.com>
25453 * config/aarch64/aarch64-tuning-flags.def (AARCH64_EXTRA_TUNING_OPTION):
25454 Add AARCH64_EXTRA_TUNE_NO_LDP_COMBINE.
25455 * config/aarch64/aarch64.cc (aarch64_operands_ok_for_ldpstp):
25456 Check for the above tuning option when processing loads.
25458 2023-04-17 Richard Biener <rguenther@suse.de>
25460 PR tree-optimization/109524
25461 * tree-vrp.cc (remove_unreachable::m_list): Change to a
25462 vector of pairs of block indices.
25463 (remove_unreachable::maybe_register_block): Adjust.
25464 (remove_unreachable::remove_and_update_globals): Likewise.
25465 Deal with removed blocks.
25467 2023-04-16 Jeff Law <jlaw@ventanamicro>
25470 * config/riscv/riscv.cc (riscv_expand_conditional_move): For
25471 TARGET_SFB_ALU, force the true arm into a register.
25473 2023-04-15 John David Anglin <danglin@gcc.gnu.org>
25476 * config/pa/pa-protos.h (pa_function_arg_size): Update prototype.
25477 * config/pa/pa.cc (pa_function_arg): Return NULL_RTX if argument
25479 (pa_arg_partial_bytes): Don't call pa_function_arg_size twice.
25480 (pa_function_arg_size): Change return type to int. Return zero
25481 for arguments larger than 1 GB. Update comments.
25483 2023-04-15 Jakub Jelinek <jakub@redhat.com>
25485 PR tree-optimization/109154
25486 * tree-if-conv.cc (predicate_scalar_phi): For complex PHIs, emit just
25487 args_len - 1 COND_EXPRs rather than args_len. Formatting fix.
25489 2023-04-15 Jason Merrill <jason@redhat.com>
25492 * gimple-ssa-warn-access.cc (pass_waccess::check_dangling_stores):
25493 Overhaul lhs_ref.ref analysis.
25495 2023-04-14 Richard Biener <rguenther@suse.de>
25497 PR tree-optimization/109502
25498 * tree-vect-stmts.cc (vectorizable_assignment): Fix
25499 check for conversion between mask and non-mask types.
25501 2023-04-14 Jeff Law <jlaw@ventanamicro.com>
25502 Jakub Jelinek <jakub@redhat.com>
25506 * combine.cc (simplify_and_const_int_1): Compute nonzero_bits in
25507 word_mode rather than mode if WORD_REGISTER_OPERATIONS and mode is
25508 smaller than word_mode.
25509 * simplify-rtx.cc (simplify_context::simplify_binary_operation_1)
25510 <case AND>: Likewise.
25512 2023-04-14 Jakub Jelinek <jakub@redhat.com>
25514 * loop-iv.cc (iv_number_of_iterations): Use gen_int_mode instead
25517 2023-04-13 Andrew MacLeod <amacleod@redhat.com>
25519 PR tree-optimization/108139
25520 PR tree-optimization/109462
25521 * gimple-range-cache.cc (ranger_cache::fill_block_cache): Remove
25522 equivalency check for PHI nodes.
25523 * gimple-range-fold.cc (fold_using_range::range_of_phi): Ensure def
25524 does not dominate single-arg equivalency edges.
25526 2023-04-13 Richard Sandiford <richard.sandiford@arm.com>
25529 * config/aarch64/aarch64.cc (aarch64_function_arg_alignment): Do
25530 not trust TYPE_ALIGN for pointer types; use POINTER_SIZE instead.
25532 2023-04-13 Richard Biener <rguenther@suse.de>
25534 PR tree-optimization/109491
25535 * tree-ssa-sccvn.cc (expressions_equal_p): Restore the
25536 NULL operands test.
25538 2023-04-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
25541 * config/riscv/riscv-vector-builtins-types.def (vint8mf8_t): Fix predicate.
25542 (vint16mf4_t): Ditto.
25543 (vint32mf2_t): Ditto.
25544 (vint64m1_t): Ditto.
25545 (vint64m2_t): Ditto.
25546 (vint64m4_t): Ditto.
25547 (vint64m8_t): Ditto.
25548 (vuint8mf8_t): Ditto.
25549 (vuint16mf4_t): Ditto.
25550 (vuint32mf2_t): Ditto.
25551 (vuint64m1_t): Ditto.
25552 (vuint64m2_t): Ditto.
25553 (vuint64m4_t): Ditto.
25554 (vuint64m8_t): Ditto.
25555 (vfloat32mf2_t): Ditto.
25556 (vbool64_t): Ditto.
25557 * config/riscv/riscv-vector-builtins.cc (register_builtin_type): Add comments.
25558 (register_vector_type): Ditto.
25559 (check_required_extensions): Fix condition.
25560 * config/riscv/riscv-vector-builtins.h (RVV_REQUIRE_ZVE64): Remove it.
25561 (RVV_REQUIRE_ELEN_64): New define.
25562 (RVV_REQUIRE_MIN_VLEN_64): Ditto.
25563 * config/riscv/riscv-vector-switch.def (TARGET_VECTOR_FP32): Remove it.
25564 (TARGET_VECTOR_FP64): Ditto.
25565 (ENTRY): Fix predicate.
25566 * config/riscv/vector-iterators.md: Fix predicate.
25568 2023-04-12 Jakub Jelinek <jakub@redhat.com>
25570 PR tree-optimization/109410
25571 * tree-ssa-reassoc.cc (build_and_add_sum): Split edge from entry
25572 block if first statement of the function is a call to returns_twice
25575 2023-04-12 Jakub Jelinek <jakub@redhat.com>
25578 * config/i386/i386.cc: Include rtl-error.h.
25579 (ix86_print_operand): For z modifier warning, use warning_for_asm
25580 if this_is_asm_operands. For Z modifier errors, use %c and code
25581 instead of hardcoded Z.
25583 2023-04-12 Costas Argyris <costas.argyris@gmail.com>
25585 * config/i386/x-mingw32-utf8: Remove extrataneous $@
25587 2023-04-12 Andrew MacLeod <amacleod@redhat.com>
25589 PR tree-optimization/109462
25590 * gimple-range-cache.cc (ranger_cache::fill_block_cache): Don't
25591 check for equivalences if NAME is a phi node.
25593 2023-04-12 Richard Biener <rguenther@suse.de>
25595 PR tree-optimization/109473
25596 * tree-vect-loop.cc (vect_create_epilog_for_reduction):
25597 Convert scalar result to the computation type before performing
25598 the reduction adjustment.
25600 2023-04-12 Richard Biener <rguenther@suse.de>
25602 PR tree-optimization/109469
25603 * tree-vect-slp.cc (vect_slp_function): Skip region starts with
25604 a returns-twice call.
25606 2023-04-12 Richard Biener <rguenther@suse.de>
25608 PR tree-optimization/109434
25609 * tree-ssa-dse.cc (initialize_ao_ref_for_dse): Properly
25610 handle possibly throwing calls when processing the LHS
25611 and may-defs are not OK.
25613 2023-04-11 Lin Sinan <mynameisxiaou@gmail.com>
25615 * config/riscv/predicates.md (uimm_extra_bit_or_twobits): Adjust
25616 predicate to avoid splitting arith constants.
25618 2023-04-11 Yanzhang Wang <yanzhang.wang@intel.com>
25619 Pan Li <pan2.li@intel.com>
25620 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
25621 Kito Cheng <kito.cheng@sifive.com>
25624 * config/riscv/riscv-protos.h (emit_hard_vlmax_vsetvl): New.
25625 * config/riscv/riscv-v.cc (emit_hard_vlmax_vsetvl): New.
25626 (emit_vlmax_vsetvl): Use emit_hard_vlmax_vsetvl.
25627 * config/riscv/riscv.cc (vector_zero_call_used_regs): New.
25628 (riscv_zero_call_used_regs): New.
25629 (TARGET_ZERO_CALL_USED_REGS): New.
25631 2023-04-11 Martin Liska <mliska@suse.cz>
25634 * opts.cc (finish_options): Drop also
25635 x_flag_var_tracking_assignments.
25637 2023-04-11 Andre Vieira <andre.simoesdiasvieira@arm.com>
25639 PR tree-optimization/108888
25640 * tree-if-conv.cc (predicate_statements): Fix gimple call check.
25642 2023-04-11 Haochen Gui <guihaoc@gcc.gnu.org>
25645 * config/rs6000/vsx.md (vsx_sign_extend_qi_<mode>): Rename to...
25646 (vsx_sign_extend_v16qi_<mode>): ... this.
25647 (vsx_sign_extend_hi_<mode>): Rename to...
25648 (vsx_sign_extend_v8hi_<mode>): ... this.
25649 (vsx_sign_extend_si_v2di): Rename to...
25650 (vsx_sign_extend_v4si_v2di): ... this.
25651 (vsignextend_qi_<mode>): Remove.
25652 (vsignextend_hi_<mode>): Remove.
25653 (vsignextend_si_v2di): Remove.
25654 (vsignextend_v2di_v1ti): Remove.
25655 (*xxspltib_<mode>_split): Replace gen_vsx_sign_extend_qi_v2di with
25656 gen_vsx_sign_extend_v16qi_v2di and gen_vsx_sign_extend_qi_v4si
25657 with gen_vsx_sign_extend_v16qi_v4si.
25658 * config/rs6000/rs6000.md (split for DI constant generation):
25659 Replace gen_vsx_sign_extend_qi_si with gen_vsx_sign_extend_v16qi_si.
25660 (split for HSDI constant generation): Replace gen_vsx_sign_extend_qi_di
25661 with gen_vsx_sign_extend_v16qi_di and gen_vsx_sign_extend_qi_si
25662 with gen_vsx_sign_extend_v16qi_si.
25663 * config/rs6000/rs6000-builtins.def (__builtin_altivec_vsignextsb2d):
25664 Set bif-pattern to vsx_sign_extend_v16qi_v2di.
25665 (__builtin_altivec_vsignextsb2w): Set bif-pattern to
25666 vsx_sign_extend_v16qi_v4si.
25667 (__builtin_altivec_visgnextsh2d): Set bif-pattern to
25668 vsx_sign_extend_v8hi_v2di.
25669 (__builtin_altivec_vsignextsh2w): Set bif-pattern to
25670 vsx_sign_extend_v8hi_v4si.
25671 (__builtin_altivec_vsignextsw2d): Set bif-pattern to
25672 vsx_sign_extend_si_v2di.
25673 (__builtin_altivec_vsignext): Set bif-pattern to
25674 vsx_sign_extend_v2di_v1ti.
25675 * config/rs6000/rs6000-builtin.cc (lxvrse_expand_builtin): Replace
25676 gen_vsx_sign_extend_qi_v2di with gen_vsx_sign_extend_v16qi_v2di,
25677 gen_vsx_sign_extend_hi_v2di with gen_vsx_sign_extend_v8hi_v2di and
25678 gen_vsx_sign_extend_si_v2di with gen_vsx_sign_extend_v4si_v2di.
25680 2023-04-10 Michael Meissner <meissner@linux.ibm.com>
25683 * config/rs6000/vsx.md (vsx_fmav4sf4): Do not generate vmaddfp.
25684 (vsx_nfmsv4sf4): Do not generate vnmsubfp.
25686 2023-04-10 Haochen Jiang <haochen.jiang@intel.com>
25688 * config/i386/i386.h (PTA_GRANITERAPIDS): Add PTA_AMX_COMPLEX.
25690 2023-04-10 Haochen Jiang <haochen.jiang@intel.com>
25692 * common/config/i386/cpuinfo.h (get_available_features):
25693 Detect AMX-COMPLEX.
25694 * common/config/i386/i386-common.cc
25695 (OPTION_MASK_ISA2_AMX_COMPLEX_SET,
25696 OPTION_MASK_ISA2_AMX_COMPLEX_UNSET): New.
25697 (ix86_handle_option): Handle -mamx-complex.
25698 * common/config/i386/i386-cpuinfo.h (enum processor_features):
25699 Add FEATURE_AMX_COMPLEX.
25700 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
25702 * config.gcc: Add amxcomplexintrin.h.
25703 * config/i386/cpuid.h (bit_AMX_COMPLEX): New.
25704 * config/i386/i386-c.cc (ix86_target_macros_internal): Define
25706 * config/i386/i386-isa.def (AMX_COMPLEX): Add DEF_PTA(AMX_COMPLEX).
25707 * config/i386/i386-options.cc (ix86_valid_target_attribute_inner_p):
25708 Handle amx-complex.
25709 * config/i386/i386.opt: Add option -mamx-complex.
25710 * config/i386/immintrin.h: Include amxcomplexintrin.h.
25711 * doc/extend.texi: Document amx-complex.
25712 * doc/invoke.texi: Document -mamx-complex.
25713 * doc/sourcebuild.texi: Document target amx-complex.
25714 * config/i386/amxcomplexintrin.h: New file.
25716 2023-04-08 Jakub Jelinek <jakub@redhat.com>
25718 PR tree-optimization/109392
25719 * tree-vect-generic.cc (tree_vec_extract): Handle failure
25720 of maybe_push_res_to_seq better.
25722 2023-04-08 Jakub Jelinek <jakub@redhat.com>
25724 * Makefile.in (CORETYPES_H): Depend on align.h, poly-int.h and
25726 (SYSTEM_H): Depend on $(HASHTAB_H).
25727 * config/riscv/t-riscv (build/genrvv-type-indexer.o): Remove unused
25728 dependency on $(RTL_BASE_H), remove redundant dependency on
25731 2023-04-06 Richard Earnshaw <rearnsha@arm.com>
25734 * config/arm/arm.cc (arm_effective_regno): New function.
25735 (mve_vector_mem_operand): Use it.
25737 2023-04-06 Andrew MacLeod <amacleod@redhat.com>
25739 PR tree-optimization/109417
25740 * gimple-range-gori.cc (gori_compute::may_recompute_p): Check if
25741 dependency is in SSA_NAME_FREE_LIST.
25743 2023-04-06 Andrew Pinski <apinski@marvell.com>
25745 PR tree-optimization/109427
25746 * params.opt (-param=vect-induction-float=):
25747 Fix option attribute typo for IntegerRange.
25749 2023-04-05 Jeff Law <jlaw@ventanamicro>
25752 * combine.cc (combine_instructions): Force re-recognition when
25753 after restoring the body of an insn to its original form.
25755 2023-04-05 Martin Jambor <mjambor@suse.cz>
25758 * ipa-sra.cc (zap_useless_ipcp_results): New function.
25759 (process_isra_node_results): Call it.
25761 2023-04-05 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
25763 * config/riscv/vector.md: Fix incorrect operand order.
25765 2023-04-05 Juzhe-Zhong <juzhe.zhong@rivai.ai>
25767 * config/riscv/riscv-vsetvl.cc
25768 (pass_vsetvl::compute_local_backward_infos): Update user vsetvl in local
25771 2023-04-05 Li Xu <xuli1@eswincomputing.com>
25773 * config/riscv/riscv-vector-builtins.def: Fix typo.
25774 * config/riscv/riscv.cc (riscv_dwarf_poly_indeterminate_value): Ditto.
25775 * config/riscv/vector-iterators.md: Ditto.
25777 2023-04-04 Hans-Peter Nilsson <hp@axis.com>
25779 * doc/md.texi (Including Patterns): Fix page break.
25781 2023-04-04 Jakub Jelinek <jakub@redhat.com>
25783 PR tree-optimization/109386
25784 * range-op-float.cc (foperator_lt::op1_range, foperator_lt::op2_range,
25785 foperator_le::op1_range, foperator_le::op2_range,
25786 foperator_gt::op1_range, foperator_gt::op2_range,
25787 foperator_ge::op1_range, foperator_ge::op2_range): Make r varying for
25788 BRS_FALSE case even if the other op is maybe_isnan, not just
25790 (foperator_unordered_lt::op1_range, foperator_unordered_lt::op2_range,
25791 foperator_unordered_le::op1_range, foperator_unordered_le::op2_range,
25792 foperator_unordered_gt::op1_range, foperator_unordered_gt::op2_range,
25793 foperator_unordered_ge::op1_range, foperator_unordered_ge::op2_range):
25794 Make r varying for BRS_TRUE case even if the other op is maybe_isnan,
25795 not just known_isnan.
25797 2023-04-04 Marek Polacek <polacek@redhat.com>
25799 PR sanitizer/109107
25800 * fold-const.cc (fold_binary_loc): Use TYPE_OVERFLOW_SANITIZED
25802 * match.pd: Use TYPE_OVERFLOW_SANITIZED.
25804 2023-04-04 Stam Markianos-Wright <stam.markianos-wright@arm.com>
25806 * config/arm/mve.md (mve_vcvtq_n_to_f_<supf><mode>): Swap operands.
25807 (mve_vcreateq_f<mode>): Swap operands.
25809 2023-04-04 Andrew Stubbs <ams@codesourcery.com>
25811 * config/gcn/gcn-valu.md (one_cmpl<mode>2<exec>): New.
25813 2023-04-04 Jakub Jelinek <jakub@redhat.com>
25816 * common/config/riscv/riscv-common.cc (riscv_subset_list::parse):
25817 Reword diagnostics about zfinx conflict with f, formatting fixes.
25819 2023-04-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25821 * config/sol2.h (LIB_SPEC): Don't link with -lpthread.
25823 2023-04-04 Richard Biener <rguenther@suse.de>
25825 PR tree-optimization/109304
25826 * tree-profile.cc (tree_profiling): Use symtab node
25827 availability to decide whether to skip adjusting calls.
25828 Do not adjust calls to internal functions.
25830 2023-04-04 Kewen Lin <linkw@linux.ibm.com>
25833 * config/rs6000/rs6000.cc (rs6000_expand_vector_set_var_p9): Fix gen
25834 function for permutation control vector by considering big endianness.
25836 2023-04-04 Kewen Lin <linkw@linux.ibm.com>
25839 * config/rs6000/altivec.md (*p9v_parity<mode>2): Rename to ...
25840 (rs6000_vprtyb<mode>2): ... this.
25841 * config/rs6000/rs6000-builtins.def (VPRTYBD): Replace parityv2di2 with
25842 rs6000_vprtybv2di2.
25843 (VPRTYBW): Replace parityv4si2 with rs6000_vprtybv4si2.
25844 (VPRTYBQ): Replace parityv1ti2 with rs6000_vprtybv1ti2.
25845 * config/rs6000/vector.md (parity<mode>2 with VEC_IP): Expand with
25846 popcountv16qi2 and the corresponding rs6000_vprtyb<mode>2.
25848 2023-04-04 Hans-Peter Nilsson <hp@axis.com>
25849 Sandra Loosemore <sandra@codesourcery.com>
25851 * doc/md.texi (Insn Splitting): Tweak wording for readability.
25853 2023-04-03 Martin Jambor <mjambor@suse.cz>
25856 * ipa-prop.cc (determine_known_aggregate_parts): Check that the
25857 offset + size will be representable in unsigned int.
25859 2023-04-03 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25861 * configure.ac (ZSTD_LIB): Move before zstd.h check.
25862 Unset gcc_cv_header_zstd_h without libzstd.
25863 * configure: Regenerate.
25865 2023-04-03 Martin Liska <mliska@suse.cz>
25867 * doc/invoke.texi: Document new param.
25869 2023-04-03 Cupertino Miranda <cupertino.miranda@oracle.com>
25871 * doc/sourcebuild.texi (const_volatile_readonly_section): Document
25872 new check_effective_target function.
25874 2023-04-03 Li Xu <xuli1@eswincomputing.com>
25876 * config/riscv/riscv-vector-builtins.def (vuint32m8_t): Fix typo.
25877 (vfloat32m8_t): Likewise
25879 2023-04-03 liuhongt <hongtao.liu@intel.com>
25881 * doc/md.texi: Document signbitm2.
25883 2023-04-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
25884 kito-cheng <kito.cheng@sifive.com>
25886 * config/riscv/vector.md: Fix RA constraint.
25888 2023-04-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
25890 * config/riscv/riscv-protos.h (gen_avl_for_scalar_move): New function.
25891 * config/riscv/riscv-v.cc (gen_avl_for_scalar_move): New function.
25892 * config/riscv/vector.md: Fix scalar move bug.
25894 2023-04-01 Jakub Jelinek <jakub@redhat.com>
25896 * range-op-float.cc (foperator_equal::fold_range): If at least
25897 one of the op ranges is not singleton and neither is NaN and all
25898 4 bounds are zero, return [1, 1].
25899 (foperator_not_equal::fold_range): In the same case return [0, 0].
25901 2023-04-01 Jakub Jelinek <jakub@redhat.com>
25903 * range-op-float.cc (foperator_equal::fold_range): Perform the
25904 non-singleton handling regardless of maybe_isnan (op1, op2).
25905 (foperator_not_equal::fold_range): Likewise.
25906 (foperator_lt::fold_range, foperator_le::fold_range,
25907 foperator_gt::fold_range, foperator_ge::fold_range): Perform the
25908 real_* comparison check which results in range_false (type)
25909 even if maybe_isnan (op1, op2). Simplify.
25910 (foperator_ltgt): New class.
25911 (fop_ltgt): New variable.
25912 (floating_op_table::floating_op_table): Handle LTGT_EXPR using
25915 2023-04-01 Jakub Jelinek <jakub@redhat.com>
25918 * builtins.cc (apply_args_size): If targetm.calls.get_raw_arg_mode
25919 returns VOIDmode, handle it like if the register isn't used for
25920 passing arguments at all.
25921 (apply_result_size): If targetm.calls.get_raw_result_mode returns
25922 VOIDmode, handle it like if the register isn't used for returning
25924 * target.def (get_raw_result_mode, get_raw_arg_mode): Document what it
25925 means to return VOIDmode.
25926 * doc/tm.texi: Regenerated.
25927 * config/aarch64/aarch64.cc (aarch64_function_value_regno_p): Return
25928 TARGET_SVE for P0_REGNUM.
25929 (aarch64_function_arg_regno_p): Also return true for p0-p3.
25930 (aarch64_get_reg_raw_mode): Return VOIDmode for PR_REGNUM_P regs.
25932 2023-03-31 Vladimir N. Makarov <vmakarov@redhat.com>
25934 * lra-constraints.cc: (combine_reload_insn): New function.
25936 2023-03-31 Jakub Jelinek <jakub@redhat.com>
25938 PR tree-optimization/91645
25939 * range-op-float.cc (foperator_unordered_lt::fold_range,
25940 foperator_unordered_le::fold_range,
25941 foperator_unordered_gt::fold_range,
25942 foperator_unordered_ge::fold_range,
25943 foperator_unordered_equal::fold_range): Call the ordered
25944 fold_range on ranges with cleared NaNs.
25945 * value-query.cc (range_query::get_tree_range): Handle also
25946 COMPARISON_CLASS_P trees.
25948 2023-03-31 Kito Cheng <kito.cheng@sifive.com>
25949 Andrew Pinski <pinskia@gmail.com>
25952 * config/riscv/t-riscv: Add missing dependencies.
25954 2023-03-31 liuhongt <hongtao.liu@intel.com>
25956 * config/i386/i386.cc (inline_memory_move_cost): Return 100
25957 for MASK_REGS when MODE_SIZE > 8.
25959 2023-03-31 liuhongt <hongtao.liu@intel.com>
25962 * config/i386/i386-builtin.def (BDESC): Adjust icode name from
25963 ufloat/ufix to floatuns/fixuns.
25964 * config/i386/i386-expand.cc
25965 (ix86_expand_vector_convert_uns_vsivsf): Adjust comments.
25966 * config/i386/sse.md
25967 (ufloat<sseintvecmodelower><mode>2<mask_name><round_name>):
25969 (<mask_codefor>floatuns<sseintvecmodelower><mode>2<mask_name><round_name>):.. this.
25970 (<mask_codefor><avx512>_ufix_notrunc<sf2simodelower><mode><mask_name><round_name>):
25972 (<mask_codefor><avx512>_fixuns_notrunc<sf2simodelower><mode><mask_name><round_name>):
25974 (<fixsuffix>fix_truncv16sfv16si2<mask_name><round_saeonly_name>):
25976 (fix<fixunssuffix>_truncv16sfv16si2<mask_name><round_saeonly_name>):.. this.
25977 (ufloat<si2dfmodelower><mode>2<mask_name>): Renamed to ..
25978 (floatuns<si2dfmodelower><mode>2<mask_name>): .. this.
25979 (ufloatv2siv2df2<mask_name>): Renamed to ..
25980 (<mask_codefor>floatunsv2siv2df2<mask_name>): .. this.
25981 (ufix_notrunc<mode><si2dfmodelower>2<mask_name><round_name>):
25983 (fixuns_notrunc<mode><si2dfmodelower>2<mask_name><round_name>):
25985 (ufix_notruncv2dfv2si2): Renamed to ..
25986 (fixuns_notruncv2dfv2si2):.. this.
25987 (ufix_notruncv2dfv2si2_mask): Renamed to ..
25988 (fixuns_notruncv2dfv2si2_mask): .. this.
25989 (*ufix_notruncv2dfv2si2_mask_1): Renamed to ..
25990 (*fixuns_notruncv2dfv2si2_mask_1): .. this.
25991 (ufix_truncv2dfv2si2): Renamed to ..
25992 (*fixuns_truncv2dfv2si2): .. this.
25993 (ufix_truncv2dfv2si2_mask): Renamed to ..
25994 (fixuns_truncv2dfv2si2_mask): .. this.
25995 (*ufix_truncv2dfv2si2_mask_1): Renamed to ..
25996 (*fixuns_truncv2dfv2si2_mask_1): .. this.
25997 (ufix_truncv4dfv4si2<mask_name>): Renamed to ..
25998 (fixuns_truncv4dfv4si2<mask_name>): .. this.
25999 (ufix_notrunc<mode><sseintvecmodelower>2<mask_name><round_name>):
26001 (fixuns_notrunc<mode><sseintvecmodelower>2<mask_name><round_name>):
26003 (ufix_trunc<mode><sseintvecmodelower>2<mask_name>): Renamed to ..
26004 (<mask_codefor>fixuns_trunc<mode><sseintvecmodelower>2<mask_name>):
26007 2023-03-30 Andrew MacLeod <amacleod@redhat.com>
26009 PR tree-optimization/109154
26010 * gimple-range-gori.cc (gori_compute::may_recompute_p): Add depth limit.
26011 * gimple-range-gori.h (may_recompute_p): Add depth param.
26012 * params.opt (ranger-recompute-depth): New param.
26014 2023-03-30 Jason Merrill <jason@redhat.com>
26018 * cgraph.h: Move reset() from cgraph_node to symtab_node.
26019 * cgraphunit.cc (symtab_node::reset): Adjust. Also call
26020 remove_from_same_comdat_group.
26022 2023-03-30 Richard Biener <rguenther@suse.de>
26024 PR tree-optimization/107561
26025 * gimple-ssa-warn-access.cc (get_size_range): Add flags
26026 argument and pass it on.
26027 (check_access): When querying for the size range pass
26028 SR_ALLOW_ZERO when the known destination size is zero.
26030 2023-03-30 Richard Biener <rguenther@suse.de>
26032 PR tree-optimization/109342
26033 * tree-ssa-sccvn.cc (vn_nary_op_get_predicated_value): New
26034 overload for edge. When that edge is a backedge use
26035 dominated_by_p directly.
26037 2023-03-30 liuhongt <hongtao.liu@intel.com>
26039 * config/i386/i386-expand.cc (expand_vec_perm_blend): Generate
26040 vpblendd instead of vpblendw for V4SI under avx2.
26042 2023-03-29 Hans-Peter Nilsson <hp@axis.com>
26044 * config/cris/cris.cc (cris_rtx_costs) [CONST_INT]: Return 0
26045 for many quick operands, for register-sized modes.
26047 2023-03-29 Jiawei <jiawei@iscas.ac.cn>
26049 * common/config/riscv/riscv-common.cc (riscv_subset_list::parse):
26052 2023-03-29 Martin Liska <mliska@suse.cz>
26054 PR bootstrap/109310
26055 * configure.ac: Emit a warning for deprecated option
26056 --enable-link-mutex.
26057 * configure: Regenerate.
26059 2023-03-29 Richard Biener <rguenther@suse.de>
26061 PR tree-optimization/109331
26062 * tree-ssa-forwprop.cc (pass_forwprop::execute): When we
26063 discover a taken edge make sure to cleanup the CFG.
26065 2023-03-29 Richard Biener <rguenther@suse.de>
26067 PR tree-optimization/109327
26068 * tree-ssa-forwprop.cc (pass_forwprop::execute): Deal with
26069 already removed stmts when draining to_remove.
26071 2023-03-29 Richard Biener <rguenther@suse.de>
26074 * dwarf2out.cc (lookup_type_die): Reset TREE_ASM_WRITTEN
26075 so we can re-create the DIE for the type if required.
26077 2023-03-29 Jakub Jelinek <jakub@redhat.com>
26078 Richard Biener <rguenther@suse.de>
26080 PR tree-optimization/109301
26081 * tree-ssa-math-opts.cc (pass_data_cse_sincos): Change
26082 properties_provided from PROP_gimple_opt_math to 0.
26083 (pass_data_expand_powcabs): Change properties_provided from 0 to
26084 PROP_gimple_opt_math.
26086 2023-03-29 Richard Biener <rguenther@suse.de>
26088 PR tree-optimization/109154
26089 * tree-if-conv.cc (gen_phi_arg_condition): Handle single
26090 inverted condition specially by inverting at the caller.
26091 (gen_phi_arg_condition): Swap COND_EXPR arms if requested.
26093 2023-03-28 David Malcolm <dmalcolm@redhat.com>
26096 * diagnostic-show-locus.cc (column_range::column_range): Factor
26097 out assertion conditional into...
26098 (column_range::valid_p): ...this new function.
26099 (line_corrections::add_hint): Don't attempt to consolidate hints
26100 if it would lead to invalid column_range instances.
26102 2023-03-28 Kito Cheng <kito.cheng@sifive.com>
26105 * config/riscv/riscv-c.cc (riscv_ext_version_value): New.
26106 (riscv_cpu_cpp_builtins): Define __riscv_v_intrinsic and
26109 2023-03-28 Alexander Monakov <amonakov@ispras.ru>
26111 PR rtl-optimization/109187
26112 * haifa-sched.cc (autopref_rank_for_schedule): Avoid use of overflowing
26113 subtraction in three-way comparison.
26115 2023-03-28 Andrew MacLeod <amacleod@redhat.com>
26117 PR tree-optimization/109265
26118 PR tree-optimization/109274
26119 * gimple-range-gori.cc (gori_compute::compute_operand_range): Do
26120 not create a relation record is op1 and op2 are the same symbol.
26121 (gori_compute::compute_operand1_range): Pass op1 == op2 to the
26122 handler for this stmt, but create a new record only if this statement
26123 generates a relation based on the ranges.
26124 (gori_compute::compute_operand2_range): Ditto.
26125 * value-relation.h (value_relation::set_relation): Always create the
26126 record that is requested.
26128 2023-03-28 Richard Biener <rguenther@suse.de>
26130 PR tree-optimization/107087
26131 * tree-ssa-forwprop.cc (pass_forwprop::execute): Track
26132 executable regions to avoid useless work and to better
26133 propagate degenerate PHIs.
26135 2023-03-28 Costas Argyris <costas.argyris@gmail.com>
26137 * config/i386/x-mingw32-utf8: update comments.
26139 2023-03-28 Richard Sandiford <richard.sandiford@arm.com>
26142 * config/aarch64/aarch64-protos.h (aarch64_vector_load_decl): Declare.
26143 * config/aarch64/aarch64.h (machine_function::vector_load_decls): New
26145 * config/aarch64/aarch64-builtins.cc (aarch64_record_vector_load_arg):
26147 (aarch64_general_gimple_fold_builtin): Delay folding of vld1 until
26148 after inlining. Record which decls are loaded from. Fix handling
26149 of vops for loads and stores.
26150 * config/aarch64/aarch64.cc (aarch64_vector_load_decl): New function.
26151 (aarch64_accesses_vector_load_decl_p): Likewise.
26152 (aarch64_vector_costs::m_stores_to_vector_load_decl): New member
26154 (aarch64_vector_costs::add_stmt_cost): If the function has a vld1
26155 that loads from a decl, treat vector stores to those decls as
26157 (aarch64_vector_costs::finish_cost): ...and in that case,
26158 if the vector code does nothing more than a store, give the
26159 prologue a zero cost as well.
26161 2023-03-28 Richard Biener <rguenther@suse.de>
26164 PR tree-optimization/108129
26165 * genmatch.cc (lower_for): For (match ...) delay
26166 substituting into the match operator if possible.
26167 (dt_operand::gen_gimple_expr): For user_id look at the
26168 first substitute for determining how to access operands.
26169 (dt_operand::gen_generic_expr): Likewise.
26170 (dt_node::gen_kids): Properly sort user_ids according
26171 to their substitutes.
26172 (dt_node::gen_kids_1): Code-generate user_id matching.
26174 2023-03-28 Jakub Jelinek <jakub@redhat.com>
26175 Jonathan Wakely <jwakely@redhat.com>
26177 * gcov-tool.cc (do_merge, do_merge_stream, do_rewrite, do_overlap):
26178 Use subcommand rather than sub-command in function comments.
26180 2023-03-28 Jakub Jelinek <jakub@redhat.com>
26182 PR tree-optimization/109154
26183 * value-range.h (frange::flush_denormals_to_zero): Make it public
26184 rather than private.
26185 * value-range.cc (frange::set): Don't call flush_denormals_to_zero
26187 * range-op-float.cc (range_operator_float::fold_range): Call
26188 flush_denormals_to_zero.
26190 2023-03-28 Jakub Jelinek <jakub@redhat.com>
26192 PR middle-end/106190
26193 * sanopt.cc (pass_sanopt::execute): Return TODO_cleanup_cfg if any
26194 of the IFN_{UB,HWA,A}SAN_* internal fns are lowered.
26196 2023-03-28 Jakub Jelinek <jakub@redhat.com>
26198 * range-op-float.cc (float_widen_lhs_range): Use pass get_nan_state
26199 as 4th argument to set to avoid clear_nan and union_ calls.
26201 2023-03-28 Jakub Jelinek <jakub@redhat.com>
26204 * config/i386/i386.cc (assign_386_stack_local): For DImode
26205 with SLOT_FLOATxFDI_387 and -m32 -mpreferred-stack-boundary=2 pass
26206 align 32 rather than 0 to assign_stack_local.
26208 2023-03-28 Eric Botcazou <ebotcazou@adacore.com>
26211 * config/sparc/sparc.cc (sparc_expand_vcond): Call signed_condition
26212 on operand #3 to get the final condition code. Use std::swap.
26213 * config/sparc/sparc.md (vcondv8qiv8qi): New VIS 4 expander.
26214 (fucmp<gcond:code>8<P:mode>_vis): Move around.
26215 (fpcmpu<gcond:code><GCM:gcm_name><P:mode>_vis): Likewise.
26216 (vcondu<GCM:mode><GCM:mode>): New VIS 4 expander.
26218 2023-03-28 Eric Botcazou <ebotcazou@adacore.com>
26220 * doc/gm2.texi: Add missing Next, Previous and Top fields to most
26221 top-level sections.
26223 2023-03-28 Costas Argyris <costas.argyris@gmail.com>
26225 * config.host: Pull in i386/x-mingw32-utf8 Makefile
26226 fragment and reference utf8rc-mingw32.o explicitly
26228 * config/i386/sym-mingw32.cc: prevent name mangling of
26230 * config/i386/x-mingw32-utf8: Make utf8rc-mingw32.o
26231 depend on manifest file explicitly.
26233 2023-03-28 Richard Biener <rguenther@suse.de>
26236 2023-03-27 Richard Biener <rguenther@suse.de>
26238 PR rtl-optimization/109237
26239 * cfgcleanup.cc (bb_is_just_return): Walk insns backwards.
26241 2023-03-28 Richard Biener <rguenther@suse.de>
26243 * common.opt (gdwarf): Remove Negative(gdwarf-).
26245 2023-03-28 Richard Biener <rguenther@suse.de>
26247 * common.opt (gdwarf): Add RejectNegative.
26248 (gdwarf-): Likewise.
26252 2023-03-28 Hans-Peter Nilsson <hp@axis.com>
26254 * config/cris/constraints.md ("T"): Correct to
26255 define_memory_constraint.
26257 2023-03-28 Hans-Peter Nilsson <hp@axis.com>
26259 * config/cris/cris.md (BW2): New mode-iterator.
26260 (lra_szext_decomposed, lra_szext_decomposed_indirect_with_offset): New
26263 2023-03-28 Hans-Peter Nilsson <hp@axis.com>
26265 * config/cris/cris.md ("*add<mode>3_addi"): Improve to bail only
26266 for possible eliminable compares.
26268 2023-03-28 Hans-Peter Nilsson <hp@axis.com>
26270 * config/cris/constraints.md ("R"): Remove unused constraint.
26272 2023-03-27 Jonathan Wakely <jwakely@redhat.com>
26274 PR gcov-profile/109297
26275 * gcov-tool.cc (merge_usage): Fix "subcomand" typo.
26276 (merge_stream_usage): Likewise.
26277 (overlap_usage): Likewise.
26279 2023-03-27 Christoph Müllner <christoph.muellner@vrull.eu>
26282 * config/riscv/thead.md: Add missing mode specifiers.
26284 2023-03-27 Philipp Tomsich <philipp.tomsich@vrull.eu>
26285 Jiangning Liu <jiangning.liu@amperecomputing.com>
26286 Manolis Tsamis <manolis.tsamis@vrull.eu>
26288 * config/aarch64/aarch64.cc: Update vector costs for ampere1.
26290 2023-03-27 Richard Biener <rguenther@suse.de>
26292 PR rtl-optimization/109237
26293 * cfgcleanup.cc (bb_is_just_return): Walk insns backwards.
26295 2023-03-27 Richard Biener <rguenther@suse.de>
26298 * lto-wrapper.cc (run_gcc): Parse alternate debug options
26299 as well, they always enable debug.
26301 2023-03-27 Kewen Lin <linkw@linux.ibm.com>
26304 * config/rs6000/emmintrin.h (_mm_bslli_si128): Move the implementation
26306 (_mm_slli_si128): ... here. Change to call _mm_bslli_si128 directly.
26308 2023-03-27 Kewen Lin <linkw@linux.ibm.com>
26311 * config/rs6000/emmintrin.h (_mm_bslli_si128): Check __N is not less
26312 than zero when calling vec_sld.
26313 (_mm_bsrli_si128): Return __A if __N is zero, check __N is bigger than
26314 zero when calling vec_sld.
26315 (_mm_slli_si128): Return __A if _imm5 is zero, check _imm5 is bigger
26316 than zero when calling vec_sld.
26318 2023-03-27 Sandra Loosemore <sandra@codesourcery.com>
26320 * doc/generic.texi (OpenMP): Document OMP_SIMD, OMP_DISTRIBUTE,
26321 OMP_TASKLOOP, and OMP_LOOP with OMP_FOR. Document how collapsed
26322 loops are represented and which fields are vectors. Add
26323 documentation for OMP_FOR_PRE_BODY field. Document internal
26324 form of non-rectangular loops and OMP_FOR_NON_RECTANGULAR.
26325 * tree.def (OMP_FOR): Make documentation consistent with the
26326 Texinfo manual, to fill some gaps and correct errors.
26328 2023-03-26 Andreas Schwab <schwab@linux-m68k.org>
26331 * config/m68k/m68k.h (FINAL_PRESCAN_INSN): Define.
26332 * config/m68k/m68k.cc (m68k_final_prescan_insn): Define.
26333 (handle_move_double): Call it before handle_movsi.
26334 * config/m68k/m68k-protos.h: Declare it.
26336 2023-03-26 Jakub Jelinek <jakub@redhat.com>
26338 PR tree-optimization/109230
26339 * match.pd (fneg/fadd simplify): Verify also odd permutation indexes.
26341 2023-03-26 Jakub Jelinek <jakub@redhat.com>
26344 * predict.cc (compute_function_frequency): Don't call
26345 warn_function_cold if function already has cold attribute.
26347 2023-03-26 Gerald Pfeifer <gerald@pfeifer.com>
26349 * doc/install.texi: Remove anachronistic note
26350 related to languages built and separate source tarballs.
26352 2023-03-25 David Malcolm <dmalcolm@redhat.com>
26355 * diagnostic-format-sarif.cc (read_until_eof): Delete.
26356 (maybe_read_file): Delete.
26357 (sarif_builder::maybe_make_artifact_content_object): Use
26358 get_source_file_content rather than maybe_read_file.
26359 Reject it if it's not valid UTF-8.
26360 * input.cc (file_cache_slot::get_full_file_content): New.
26361 (get_source_file_content): New.
26362 (selftest::check_cpp_valid_utf8_p): New.
26363 (selftest::test_cpp_valid_utf8_p): New.
26364 (selftest::input_cc_tests): Call selftest::test_cpp_valid_utf8_p.
26365 * input.h (get_source_file_content): New prototype.
26367 2023-03-24 David Malcolm <dmalcolm@redhat.com>
26369 * doc/analyzer.texi (Debugging the Analyzer): Add notes on useful
26371 (Special Functions for Debugging the Analyzer): Convert to a
26372 table, and rewrite in places.
26373 (Other Debugging Techniques): Add notes on how to compare two
26374 different exploded graphs.
26376 2023-03-24 David Malcolm <dmalcolm@redhat.com>
26379 * json.cc: Update comments to indicate that we now preserve
26380 insertion order of keys within objects.
26381 (object::print): Traverse keys in insertion order.
26382 (object::set): Preserve insertion order of keys.
26383 (selftest::test_writing_objects): Add an additional key to verify
26384 that we preserve insertion order.
26385 * json.h (object::m_keys): New field.
26387 2023-03-24 Andrew MacLeod <amacleod@redhat.com>
26389 PR tree-optimization/109238
26390 * gimple-range-cache.cc (ranger_cache::resolve_dom): Ignore
26391 predecessors which this block dominates.
26393 2023-03-24 Richard Biener <rguenther@suse.de>
26395 PR tree-optimization/106912
26396 * tree-profile.cc (tree_profiling): Update stmts only when
26397 profiling or testing coverage. Make sure to update calls
26398 fntype, stripping 'const' there.
26400 2023-03-24 Jakub Jelinek <jakub@redhat.com>
26402 PR middle-end/109258
26403 * builtins.cc (inline_expand_builtin_bytecmp): Return NULL_RTX early
26404 if target == const0_rtx.
26406 2023-03-24 Alexandre Oliva <oliva@adacore.com>
26408 * doc/sourcebuild.texi (weak_undefined, posix_memalign):
26409 Document options and effective targets.
26411 2023-03-24 Costas Argyris <costas.argyris@gmail.com>
26413 * config/i386/x-mingw32-utf8: Make HOST_EXTRA_OBJS_SYMBOL
26416 2023-03-23 Pat Haugen <pthaugen@linux.ibm.com>
26418 * config/rs6000/rs6000.md (*mod<mode>3, umod<mode>3): Add
26419 non-earlyclobber alternative.
26421 2023-03-23 Andrew Pinski <apinski@marvell.com>
26424 * fold-const.cc (maybe_lvalue_p): Treat COMPOUND_LITERAL_EXPR
26427 2023-03-23 Richard Biener <rguenther@suse.de>
26429 PR tree-optimization/107569
26430 * tree-ssa-sccvn.cc (eliminate_dom_walker::eliminate_stmt):
26431 Do not push SSA names with zero uses as available leader.
26432 (process_bb): Likewise.
26434 2023-03-23 Richard Biener <rguenther@suse.de>
26436 PR tree-optimization/109262
26437 * tree-ssa-forwprop.cc (pass_forwprop::execute): When
26438 combining a piecewise complex load avoid touching loads
26439 that throw internally. Use fun, not cfun throughout.
26441 2023-03-23 Jakub Jelinek <jakub@redhat.com>
26443 * value-range.cc (irange::irange_union, irange::intersect): Fix
26444 comment spelling bugs.
26445 * gimple-range-trace.cc (range_tracer::do_header): Likewise.
26446 * gimple-range-trace.h: Likewise.
26447 * gimple-range-edge.cc: Likewise.
26448 (gimple_outgoing_range_stmt_p,
26449 gimple_outgoing_range::switch_edge_range,
26450 gimple_outgoing_range::edge_range_p): Likewise.
26451 * gimple-range.cc (gimple_ranger::prefill_stmt_dependencies,
26452 gimple_ranger::fold_stmt, gimple_ranger::register_transitive_infer,
26453 assume_query::assume_query, assume_query::calculate_phi): Likewise.
26454 * gimple-range-edge.h: Likewise.
26455 * value-range.h (Value_Range::set, Value_Range::lower_bound,
26456 Value_Range::upper_bound, frange::set_undefined): Likewise.
26457 * gimple-range-gori.h (range_def_chain::depend, gori_map::m_outgoing,
26458 gori_compute): Likewise.
26459 * gimple-range-fold.h (fold_using_range): Likewise.
26460 * gimple-range-path.cc (path_range_query::compute_ranges_in_phis):
26462 * gimple-range-gori.cc (range_def_chain::in_chain_p,
26463 range_def_chain::dump, gori_map::calculate_gori,
26464 gori_compute::compute_operand_range_switch,
26465 gori_compute::logical_combine, gori_compute::refine_using_relation,
26466 gori_compute::compute_operand1_range, gori_compute::may_recompute_p):
26468 * gimple-range.h: Likewise.
26469 (enable_ranger): Likewise.
26470 * range-op.h (empty_range_varying): Likewise.
26471 * value-query.h (value_query): Likewise.
26472 * gimple-range-cache.cc (block_range_cache::set_bb_range,
26473 block_range_cache::dump, ssa_global_cache::clear_global_range,
26474 temporal_cache::temporal_value, temporal_cache::current_p,
26475 ranger_cache::range_of_def, ranger_cache::propagate_updated_value,
26476 ranger_cache::range_from_dom, ranger_cache::register_inferred_value):
26478 * gimple-range-fold.cc (fur_edge::get_phi_operand,
26479 fur_stmt::get_operand, gimple_range_adjustment,
26480 fold_using_range::range_of_phi,
26481 fold_using_range::relation_fold_and_or): Likewise.
26482 * value-range-storage.h (irange_storage_slot::MAX_INTS): Likewise.
26483 * value-query.cc (range_query::value_of_expr,
26484 range_query::value_on_edge, range_query::query_relation): Likewise.
26485 * tree-vrp.cc (remove_unreachable::remove_and_update_globals,
26486 intersect_range_with_nonzero_bits): Likewise.
26487 * gimple-range-infer.cc (gimple_infer_range::check_assume_func,
26488 exit_range): Likewise.
26489 * value-relation.h: Likewise.
26490 (equiv_oracle, relation_trio::relation_trio, value_relation,
26491 value_relation::value_relation, pe_min): Likewise.
26492 * range-op-float.cc (range_operator_float::rv_fold,
26493 frange_arithmetic, foperator_unordered_equal::op1_range,
26494 foperator_div::rv_fold): Likewise.
26495 * gimple-range-op.cc (cfn_clz::fold_range): Likewise.
26496 * value-relation.cc (equiv_oracle::query_relation,
26497 equiv_oracle::register_equiv, equiv_oracle::add_equiv_to_block,
26498 value_relation::apply_transitive, relation_chain_head::find_relation,
26499 dom_oracle::query_relation, dom_oracle::find_relation_block,
26500 dom_oracle::find_relation_dom, path_oracle::register_equiv): Likewise.
26501 * range-op.cc (range_operator::wi_fold_in_parts_equiv,
26502 create_possibly_reversed_range, adjust_op1_for_overflow,
26503 operator_mult::wi_fold, operator_exact_divide::op1_range,
26504 operator_cast::lhs_op1_relation, operator_cast::fold_pair,
26505 operator_cast::fold_range, operator_abs::wi_fold, range_op_cast_tests,
26506 range_op_lshift_tests): Likewise.
26508 2023-03-23 Andrew Stubbs <ams@codesourcery.com>
26510 * config/gcn/gcn.cc (gcn_class_max_nregs): Handle vectors in SGPRs.
26511 (move_callee_saved_registers): Detect the bug condition early.
26513 2023-03-23 Andrew Stubbs <ams@codesourcery.com>
26515 * config/gcn/gcn-protos.h (gcn_stepped_zero_int_parallel_p): New.
26516 * config/gcn/gcn-valu.md (V_1REG_ALT): New.
26518 (vec_extract<V_1REG:mode><V_1REG_ALT:mode>_nop): New.
26519 (vec_extract<V_2REG:mode><V_2REG_ALT:mode>_nop): New.
26520 (vec_extract<V_ALL:mode><V_ALL_ALT:mode>): Use new patterns.
26521 * config/gcn/gcn.cc (gcn_stepped_zero_int_parallel_p): New.
26522 * config/gcn/predicates.md (ascending_zero_int_parallel): New.
26524 2023-03-23 Jakub Jelinek <jakub@redhat.com>
26526 PR tree-optimization/109176
26527 * tree-vect-generic.cc (expand_vector_condition): If a has
26528 vector boolean type and is a comparison, also check if both
26529 the comparison and VEC_COND_EXPR could be successfully expanded
26532 2023-03-23 Pan Li <pan2.li@intel.com>
26533 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26537 * config/riscv/riscv-modes.def (ADJUST_BYTESIZE): Adjust size
26538 for vector mask modes.
26539 * config/riscv/riscv.cc (riscv_v_adjust_bytesize): New.
26540 * config/riscv/riscv.h (riscv_v_adjust_bytesize): New.
26542 2023-03-23 Songhe Zhu <zhusonghe@eswincomputing.com>
26544 * config/riscv/multilib-generator: Adjusting the loop of 'alt' in 'alts'.
26546 2023-03-23 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26549 * config/riscv/riscv-protos.h (emit_vlmax_vsetvl): Define as global.
26550 (emit_vlmax_op): Ditto.
26551 * config/riscv/riscv-v.cc (get_sew): New function.
26552 (emit_vlmax_vsetvl): Adapt function.
26553 (emit_pred_op): Ditto.
26554 (emit_vlmax_op): Ditto.
26555 (emit_nonvlmax_op): Ditto.
26556 (legitimize_move): Fix LRA ICE.
26557 (gen_no_side_effects_vsetvl_rtx): Adapt function.
26558 * config/riscv/vector.md (@mov<V_FRACT:mode><P:mode>_lra): New pattern.
26559 (@mov<VB:mode><P:mode>_lra): Ditto.
26560 (*mov<V_FRACT:mode><P:mode>_lra): Ditto.
26561 (*mov<VB:mode><P:mode>_lra): Ditto.
26563 2023-03-23 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26566 * config/riscv/riscv-vector-builtins-bases.cc (class vlenb): Add
26567 __riscv_vlenb support.
26569 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
26570 * config/riscv/riscv-vector-builtins-functions.def (vlenb): Ditto.
26571 * config/riscv/riscv-vector-builtins-shapes.cc (struct vlenb_def): Ditto.
26573 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
26574 * config/riscv/riscv-vector-builtins.cc: Ditto.
26576 2023-03-23 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26577 kito-cheng <kito.cheng@sifive.com>
26579 * config/riscv/riscv-vsetvl.cc (reg_available_p): Fix bugs.
26580 (pass_vsetvl::compute_local_backward_infos): Fix bugs.
26581 (pass_vsetvl::need_vsetvl): Fix bugs.
26582 (pass_vsetvl::backward_demand_fusion): Fix bugs.
26583 (pass_vsetvl::demand_fusion): Fix bugs.
26584 (eliminate_insn): Fix bugs.
26585 (insert_vsetvl): Ditto.
26586 (pass_vsetvl::emit_local_forward_vsetvls): Ditto.
26587 * config/riscv/riscv-vsetvl.h (enum vsetvl_type): Ditto.
26588 * config/riscv/vector.md: Ditto.
26590 2023-03-23 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26591 kito-cheng <kito.cheng@sifive.com>
26593 * config/riscv/riscv-vector-builtins-bases.cc: Fix ternary bug.
26594 * config/riscv/vector-iterators.md (nmsac): Ditto.
26600 * config/riscv/vector.md (@pred_mul_<optab><mode>): Ditto.
26601 (@pred_mul_plus<mode>): Ditto.
26602 (*pred_madd<mode>): Ditto.
26603 (*pred_macc<mode>): Ditto.
26604 (*pred_mul_plus<mode>): Ditto.
26605 (@pred_mul_plus<mode>_scalar): Ditto.
26606 (*pred_madd<mode>_scalar): Ditto.
26607 (*pred_macc<mode>_scalar): Ditto.
26608 (*pred_mul_plus<mode>_scalar): Ditto.
26609 (*pred_madd<mode>_extended_scalar): Ditto.
26610 (*pred_macc<mode>_extended_scalar): Ditto.
26611 (*pred_mul_plus<mode>_extended_scalar): Ditto.
26612 (@pred_minus_mul<mode>): Ditto.
26613 (*pred_<madd_nmsub><mode>): Ditto.
26614 (*pred_nmsub<mode>): Ditto.
26615 (*pred_<macc_nmsac><mode>): Ditto.
26616 (*pred_nmsac<mode>): Ditto.
26617 (*pred_mul_<optab><mode>): Ditto.
26618 (*pred_minus_mul<mode>): Ditto.
26619 (@pred_mul_<optab><mode>_scalar): Ditto.
26620 (@pred_minus_mul<mode>_scalar): Ditto.
26621 (*pred_<madd_nmsub><mode>_scalar): Ditto.
26622 (*pred_nmsub<mode>_scalar): Ditto.
26623 (*pred_<macc_nmsac><mode>_scalar): Ditto.
26624 (*pred_nmsac<mode>_scalar): Ditto.
26625 (*pred_mul_<optab><mode>_scalar): Ditto.
26626 (*pred_minus_mul<mode>_scalar): Ditto.
26627 (*pred_<madd_nmsub><mode>_extended_scalar): Ditto.
26628 (*pred_nmsub<mode>_extended_scalar): Ditto.
26629 (*pred_<macc_nmsac><mode>_extended_scalar): Ditto.
26630 (*pred_nmsac<mode>_extended_scalar): Ditto.
26631 (*pred_mul_<optab><mode>_extended_scalar): Ditto.
26632 (*pred_minus_mul<mode>_extended_scalar): Ditto.
26633 (*pred_<madd_msub><mode>): Ditto.
26634 (*pred_<macc_msac><mode>): Ditto.
26635 (*pred_<madd_msub><mode>_scalar): Ditto.
26636 (*pred_<macc_msac><mode>_scalar): Ditto.
26637 (@pred_neg_mul_<optab><mode>): Ditto.
26638 (@pred_mul_neg_<optab><mode>): Ditto.
26639 (*pred_<nmadd_msub><mode>): Ditto.
26640 (*pred_<nmsub_nmadd><mode>): Ditto.
26641 (*pred_<nmacc_msac><mode>): Ditto.
26642 (*pred_<nmsac_nmacc><mode>): Ditto.
26643 (*pred_neg_mul_<optab><mode>): Ditto.
26644 (*pred_mul_neg_<optab><mode>): Ditto.
26645 (@pred_neg_mul_<optab><mode>_scalar): Ditto.
26646 (@pred_mul_neg_<optab><mode>_scalar): Ditto.
26647 (*pred_<nmadd_msub><mode>_scalar): Ditto.
26648 (*pred_<nmsub_nmadd><mode>_scalar): Ditto.
26649 (*pred_<nmacc_msac><mode>_scalar): Ditto.
26650 (*pred_<nmsac_nmacc><mode>_scalar): Ditto.
26651 (*pred_neg_mul_<optab><mode>_scalar): Ditto.
26652 (*pred_mul_neg_<optab><mode>_scalar): Ditto.
26653 (@pred_widen_neg_mul_<optab><mode>): Ditto.
26654 (@pred_widen_mul_neg_<optab><mode>): Ditto.
26655 (@pred_widen_neg_mul_<optab><mode>_scalar): Ditto.
26656 (@pred_widen_mul_neg_<optab><mode>_scalar): Ditto.
26658 2023-03-23 liuhongt <hongtao.liu@intel.com>
26660 * builtins.cc (builtin_memset_read_str): Replace
26661 targetm.gen_memset_scratch_rtx with gen_reg_rtx.
26662 (builtin_memset_gen_str): Ditto.
26663 * config/i386/i386-expand.cc
26664 (ix86_convert_const_wide_int_to_broadcast): Replace
26665 ix86_gen_scratch_sse_rtx with gen_reg_rtx.
26666 (ix86_expand_vector_move): Ditto.
26667 * config/i386/i386-protos.h (ix86_gen_scratch_sse_rtx):
26669 * config/i386/i386.cc (ix86_gen_scratch_sse_rtx): Removed.
26670 (TARGET_GEN_MEMSET_SCRATCH_RTX): Removed.
26671 * doc/tm.texi: Remove TARGET_GEN_MEMSET_SCRATCH_RTX.
26672 * doc/tm.texi.in: Ditto.
26673 * target.def: Ditto.
26675 2023-03-22 Vladimir N. Makarov <vmakarov@redhat.com>
26677 * lra.cc (lra): Do not repeat inheritance and live range splitting
26678 when asm error is found.
26680 2023-03-22 Andrew Jenner <andrew@codesourcery.com>
26682 * config/gcn/gcn-protos.h (gcn_expand_dpp_swap_pairs_insn)
26683 (gcn_expand_dpp_distribute_even_insn)
26684 (gcn_expand_dpp_distribute_odd_insn): Declare.
26685 * config/gcn/gcn-valu.md (@dpp_swap_pairs<mode>)
26686 (@dpp_distribute_even<mode>, @dpp_distribute_odd<mode>)
26687 (cmul<conj_op><mode>3, cml<addsub_as><mode>4, vec_addsub<mode>3)
26688 (cadd<rot><mode>3, vec_fmaddsub<mode>4, vec_fmsubadd<mode>4)
26689 (fms<mode>4<exec>, fms<mode>4_negop2<exec>, fms<mode>4)
26690 (fms<mode>4_negop2): New patterns.
26691 * config/gcn/gcn.cc (gcn_expand_dpp_swap_pairs_insn)
26692 (gcn_expand_dpp_distribute_even_insn)
26693 (gcn_expand_dpp_distribute_odd_insn): New functions.
26694 * config/gcn/gcn.md: Add entries to unspec enum.
26696 2023-03-22 Aldy Hernandez <aldyh@redhat.com>
26698 PR tree-optimization/109008
26699 * value-range.cc (frange::set): Add nan_state argument.
26700 * value-range.h (class nan_state): New.
26701 (frange::get_nan_state): New.
26703 2023-03-22 Martin Liska <mliska@suse.cz>
26705 * configure: Regenerate.
26707 2023-03-21 Joseph Myers <joseph@codesourcery.com>
26709 * stor-layout.cc (finalize_type_size): Copy TYPE_TYPELESS_STORAGE
26712 2023-03-21 Andrew MacLeod <amacleod@redhat.com>
26714 PR tree-optimization/109192
26715 * gimple-range-gori.cc (gori_compute::compute_operand_range):
26716 Terminate gori calculations if a relation is not relevant.
26717 * value-relation.h (value_relation::set_relation): Allow
26718 equality between op1 and op2 if they are the same.
26720 2023-03-21 Richard Biener <rguenther@suse.de>
26722 PR tree-optimization/109219
26723 * tree-vect-loop.cc (vectorizable_reduction): Check
26724 slp_node, not STMT_SLP_TYPE.
26725 * tree-vect-stmts.cc (vectorizable_condition): Likewise.
26726 * tree-vect-slp.cc (vect_slp_analyze_node_operations_1):
26727 Remove assertion on STMT_SLP_TYPE.
26729 2023-03-21 Jakub Jelinek <jakub@redhat.com>
26731 PR tree-optimization/109215
26732 * tree.h (enum special_array_member): Adjust comments for int_0
26734 * tree.cc (component_ref_sam_type): Clear zero_elts if memtype
26735 has zero sized element type and the array has variable number of
26736 elements or constant one or more elements.
26737 (component_ref_size): Adjust comments, formatting fix.
26739 2023-03-21 Arsen Arsenović <arsen@aarsen.me>
26741 * configure.ac: Add check for the Texinfo 6.8
26742 CONTENTS_OUTPUT_LOCATION customization variable and set it if
26744 * configure: Regenerate.
26745 * Makefile.in (MAKEINFO_TOC_INLINE_FLAG): New variable. Set by
26746 configure.ac to -c CONTENTS_OUTPUT_LOCATION=inline if
26747 CONTENTS_OUTPUT_LOCATION support is detected, empty otherwise.
26748 ($(build_htmldir)/%/index.html): Pass MAKEINFO_TOC_INLINE_FLAG.
26750 2023-03-21 Arsen Arsenović <arsen@aarsen.me>
26752 * doc/extend.texi: Associate use_hazard_barrier_return index
26753 entry with its attribute.
26754 * doc/invoke.texi: Associate -fcanon-prefix-map index entry with
26757 2023-03-21 Arsen Arsenović <arsen@aarsen.me>
26759 * doc/implement-c.texi: Remove usage of @gol.
26760 * doc/invoke.texi: Ditto.
26761 * doc/sourcebuild.texi: Ditto.
26762 * doc/include/gcc-common.texi: Remove @gol. In new Makeinfo and
26763 texinfo.tex versions, the bug it was working around appears to
26766 2023-03-21 Arsen Arsenović <arsen@aarsen.me>
26768 * doc/include/texinfo.tex: Update to 2023-01-17.19.
26770 2023-03-21 Arsen Arsenović <arsen@aarsen.me>
26772 * doc/include/gcc-common.texi: Add @defbuiltin{,x} and
26773 @enddefbuiltin for defining built-in functions.
26774 * doc/extend.texi: Apply @defbuiltin{,x} to many, but not all,
26775 places where it should be used.
26777 2023-03-21 Arsen Arsenović <arsen@aarsen.me>
26779 * doc/extend.texi (Formatted Output Function Checking): New
26780 subsection for grouping together printf et al.
26781 (Exception handling) Fix missing @ sign before copyright
26782 header, which lead to the copyright line leaking into
26783 '(gcc)Exception handling'.
26784 * doc/gcc.texi: Set document language to en_US.
26785 (@copying): Wrap front cover texts in quotations, move in manual
26788 2023-03-21 Arsen Arsenović <arsen@aarsen.me>
26790 * doc/gcc.texi: Add the Indices appendix, to make texinfo
26791 generate nice indices overview page.
26793 2023-03-21 Richard Biener <rguenther@suse.de>
26795 PR tree-optimization/109170
26796 * gimple-range-op.cc (cfn_pass_through_arg1): New.
26797 (gimple_range_op_handler::maybe_builtin_call): Handle
26798 __builtin_expect via cfn_pass_through_arg1.
26800 2023-03-20 Michael Meissner <meissner@linux.ibm.com>
26803 * config/rs6000/rs6000.cc (create_complex_muldiv): Delete.
26804 (init_float128_ieee): Delete code to switch complex multiply and divide
26806 (complex_multiply_builtin_code): New helper function.
26807 (complex_divide_builtin_code): Likewise.
26808 (rs6000_mangle_decl_assembler_name): Add support for mangling the name
26809 of complex 128-bit multiply and divide built-in functions.
26811 2023-03-20 Peter Bergner <bergner@linux.ibm.com>
26814 * config/rs6000/rs6000-builtin.cc (stv_expand_builtin): Use tmode.
26816 2023-03-19 Jonny Grant <jg@jguk.org>
26818 * doc/extend.texi (Common Function Attributes) <nonnull>:
26821 2023-03-18 Peter Bergner <bergner@linux.ibm.com>
26823 PR rtl-optimization/109179
26824 * lra-constraints.cc (combine_reload_insn): Enforce TO is not a debug
26825 insn or note. Move the tests earlier to guard lra_get_insn_recog_data.
26827 2023-03-17 Jakub Jelinek <jakub@redhat.com>
26830 * function.h (push_struct_function): Add ABSTRACT_P argument defaulted
26832 * function.cc (push_struct_function): Add ABSTRACT_P argument, pass it
26833 to allocate_struct_function instead of false.
26834 * tree-inline.cc (initialize_cfun): Don't copy DECL_ARGUMENTS
26835 nor DECL_RESULT here. Pass true as ABSTRACT_P to
26836 push_struct_function. Call targetm.target_option.relayout_function
26838 (tree_function_versioning): Formatting fix.
26840 2023-03-17 Vladimir N. Makarov <vmakarov@redhat.com>
26842 * lra-constraints.cc: Include hooks.h.
26843 (combine_reload_insn): New function.
26844 (lra_constraints): Call it.
26846 2023-03-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26847 kito-cheng <kito.cheng@sifive.com>
26849 * config/riscv/riscv-v.cc (legitimize_move): Allow undef value
26850 as legitimate value.
26851 * config/riscv/riscv-vector-builtins.cc
26852 (function_expander::use_ternop_insn): Fix bugs of ternary intrinsic.
26853 (function_expander::use_widen_ternop_insn): Ditto.
26854 * config/riscv/vector.md (@vundefined<mode>): New pattern.
26855 (pred_mul_<optab><mode>_undef_merge): Remove.
26856 (*pred_mul_<optab><mode>_undef_merge_scalar): Ditto.
26857 (*pred_mul_<optab><mode>_undef_merge_extended_scalar): Ditto.
26858 (pred_neg_mul_<optab><mode>_undef_merge): Ditto.
26859 (*pred_neg_mul_<optab><mode>_undef_merge_scalar): Ditto.
26861 2023-03-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26864 * config/riscv/riscv.md: Fix subreg bug.
26866 2023-03-17 Jakub Jelinek <jakub@redhat.com>
26868 PR middle-end/108685
26869 * omp-expand.cc (expand_omp_for_ordered_loops): Add L0_BB argument,
26870 use its loop_father rather than BODY_BB's loop_father.
26871 (expand_omp_for_generic): Adjust expand_omp_for_ordered_loops caller.
26872 If broken_loop with ordered > collapse and at least one of those
26873 extra loops aren't guaranteed to have at least one iteration, change
26874 l0_bb's loop_father to entry_bb's loop_father. Set cont_bb's
26875 loop_father to l0_bb's loop_father rather than l1_bb's.
26877 2023-03-17 Jakub Jelinek <jakub@redhat.com>
26880 * gdbhooks.py (TreePrinter.to_string): Wrap
26881 gdb.parse_and_eval('tree_code_type') in a try block, parse
26882 and eval 'tree_code_type_tmpl<0>::tree_code_type' instead if it
26883 raises exception. Update comments for the recent tree_code_type
26886 2023-03-17 Sandra Loosemore <sandra@codesourcery.com>
26888 * doc/extend.texi (BPF Built-in Functions): Fix numerous markup
26889 issues. Add more line breaks to example so it doesn't overflow
26892 2023-03-17 Sandra Loosemore <sandra@codesourcery.com>
26894 * doc/extend.texi (Common Function Attributes) <access>: Fix bad
26895 line breaks in examples.
26896 <malloc>: Fix bad line breaks in running text, also copy-edit
26898 (Extended Asm) <Generic Operand Modifiers>: Fix @multitable width.
26899 * doc/invoke.texi (Option Summary) <Developer Options>: Fix misplaced
26901 (C++ Dialect Options) <-fcontracts>: Add line break in example.
26902 <-Wctad-maybe-unsupported>: Likewise.
26903 <-Winvalid-constexpr>: Likewise.
26904 (Warning Options) <-Wdangling-pointer>: Likewise.
26905 <-Winterference-size>: Likewise.
26906 <-Wvla-parameter>: Likewise.
26907 (Static Analyzer Options): Fix bad line breaks in running text,
26908 plus add some missing markup.
26909 (Optimize Options) <openacc-privatization>: Fix more bad line
26910 breaks in running text.
26912 2023-03-16 Uros Bizjak <ubizjak@gmail.com>
26914 * config/i386/i386-expand.cc (expand_vec_perm_pblendv):
26915 Handle 8-byte modes only with TARGET_MMX_WITH_SSE.
26916 (expand_vec_perm_2perm_pblendv): Ditto.
26918 2023-03-16 Martin Liska <mliska@suse.cz>
26920 PR middle-end/106133
26921 * gcc.cc (driver_handle_option): Use x_main_input_basename
26922 if x_dump_base_name is null.
26923 * opts.cc (common_handle_option): Likewise.
26925 2023-03-16 Richard Biener <rguenther@suse.de>
26927 PR tree-optimization/109123
26928 * gimple-ssa-warn-access.cc (pass_waccess::warn_invalid_pointer):
26929 Do not emit -Wuse-after-free late.
26930 (pass_waccess::check_call): Always check call pointer uses.
26932 2023-03-16 Richard Biener <rguenther@suse.de>
26934 PR tree-optimization/109141
26935 * tree-dfa.h (renumber_gimple_stmt_uids_in_block): New.
26936 * tree-dfa.cc (renumber_gimple_stmt_uids_in_block): Split
26938 (renumber_gimple_stmt_uids): ... here and
26939 (renumber_gimple_stmt_uids_in_blocks): ... here.
26940 * gimple-ssa-warn-access.cc (pass_waccess::use_after_inval_p):
26941 Use renumber_gimple_stmt_uids_in_block to also assign UIDs
26943 (pass_waccess::check_pointer_uses): Process all PHIs.
26945 2023-03-15 David Malcolm <dmalcolm@redhat.com>
26948 * diagnostic-format-sarif.cc (class sarif_invocation): New.
26949 (class sarif_ice_notification): New.
26950 (sarif_builder::m_invocation_obj): New field.
26951 (sarif_invocation::add_notification_for_ice): New.
26952 (sarif_invocation::prepare_to_flush): New.
26953 (sarif_ice_notification::sarif_ice_notification): New.
26954 (sarif_builder::sarif_builder): Add m_invocation_obj.
26955 (sarif_builder::end_diagnostic): Special-case DK_ICE and
26957 (sarif_builder::flush_to_file): Call prepare_to_flush on
26958 m_invocation_obj. Pass the latter to make_top_level_object.
26959 (sarif_builder::make_result_object): Move creation of "locations"
26961 (sarif_builder::make_locations_arr): ...this new function.
26962 (sarif_builder::make_top_level_object): Add "invocation_obj" param
26963 and pass it to make_run_object.
26964 (sarif_builder::make_run_object): Add "invocation_obj" param and
26966 (sarif_ice_handler): New callback.
26967 (diagnostic_output_format_init_sarif): Wire up sarif_ice_handler.
26968 * diagnostic.cc (diagnostic_initialize): Initialize new field
26970 (diagnostic_action_after_output): If it is set, make one attempt
26971 to call ice_handler_cb.
26972 * diagnostic.h (diagnostic_context::ice_handler_cb): New field.
26974 2023-03-15 Uros Bizjak <ubizjak@gmail.com>
26976 * config/i386/i386-expand.cc (expand_vec_perm_blend):
26977 Handle 8-byte modes only with TARGET_MMX_WITH_SSE. Handle V2SFmode
26978 and fix V2HImode handling.
26979 (expand_vec_perm_1): Try to emit BLEND instruction
26980 before MOVSS/MOVSD.
26981 * config/i386/mmx.md (*mmx_blendps): New insn pattern.
26983 2023-03-15 Tobias Burnus <tobias@codesourcery.com>
26985 * omp-low.cc (omp_runtime_api_call): Add omp_in_explicit_task.
26987 2023-03-15 Richard Biener <rguenther@suse.de>
26989 * gimple-ssa-warn-access.cc (pass_waccess::check_pointer_uses):
26990 Do not diagnose clobbers.
26992 2023-03-15 Richard Biener <rguenther@suse.de>
26994 PR tree-optimization/109139
26995 * tree-ssa-live.cc (remove_unused_locals): Look at the
26996 base address for unused decls on the LHS of .DEFERRED_INIT.
26998 2023-03-15 Xi Ruoyao <xry111@xry111.site>
27001 * builtins.cc (inline_string_cmp): Force the character
27002 difference into "result" pseudo-register, instead of reassign
27003 the pseudo-register.
27005 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
27007 * config.gcc: Add thead.o to RISC-V extra_objs.
27008 * config/riscv/peephole.md: Add mempair peephole passes.
27009 * config/riscv/riscv-protos.h (riscv_split_64bit_move_p): New
27011 (th_mempair_operands_p): Likewise.
27012 (th_mempair_order_operands): Likewise.
27013 (th_mempair_prepare_save_restore_operands): Likewise.
27014 (th_mempair_save_restore_regs): Likewise.
27015 (th_mempair_output_move): Likewise.
27016 * config/riscv/riscv.cc (riscv_save_reg): Move code.
27017 (riscv_restore_reg): Move code.
27018 (riscv_for_each_saved_reg): Add code to emit mempair insns.
27019 * config/riscv/t-riscv: Add thead.cc.
27020 * config/riscv/thead.md (*th_mempair_load_<GPR:mode>2):
27022 (*th_mempair_store_<GPR:mode>2): Likewise.
27023 (*th_mempair_load_extendsidi2): Likewise.
27024 (*th_mempair_load_zero_extendsidi2): Likewise.
27025 * config/riscv/thead.cc: New file.
27027 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
27029 * config/riscv/constraints.md (TARGET_XTHEADFMV ? FP_REGS : NO_REGS)
27030 New constraint "th_f_fmv".
27031 (TARGET_XTHEADFMV ? GR_REGS : NO_REGS): New constraint
27033 * config/riscv/riscv.cc (riscv_split_doubleword_move):
27034 Add split code for XTheadFmv.
27035 (riscv_secondary_memory_needed): XTheadFmv does not need
27037 * config/riscv/riscv.md: Add new UNSPEC_XTHEADFMV and
27038 UNSPEC_XTHEADFMV_HW. Add support for XTheadFmv to
27039 movdf_hardfloat_rv32.
27040 * config/riscv/thead.md (th_fmv_hw_w_x): New INSN.
27041 (th_fmv_x_w): New INSN.
27042 (th_fmv_x_hw): New INSN.
27044 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
27046 * config/riscv/riscv.md (maddhisi4): New expand.
27047 (msubhisi4): New expand.
27048 * config/riscv/thead.md (*th_mula<mode>): New pattern.
27049 (*th_mulawsi): New pattern.
27050 (*th_mulawsi2): New pattern.
27051 (*th_maddhisi4): New pattern.
27052 (*th_sextw_maddhisi4): New pattern.
27053 (*th_muls<mode>): New pattern.
27054 (*th_mulswsi): New pattern.
27055 (*th_mulswsi2): New pattern.
27056 (*th_msubhisi4): New pattern.
27057 (*th_sextw_msubhisi4): New pattern.
27059 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
27061 * config/riscv/iterators.md (TARGET_64BIT): Add GPR2 iterator.
27062 * config/riscv/riscv-protos.h (riscv_expand_conditional_move):
27064 * config/riscv/riscv.cc (riscv_rtx_costs): Add costs for
27066 (riscv_expand_conditional_move): New function.
27067 (riscv_expand_conditional_move_onesided): New function.
27068 * config/riscv/riscv.md: Add support for XTheadCondMov.
27069 * config/riscv/thead.md (*th_cond_mov<GPR:mode><GPR2:mode>): Add
27070 support for XTheadCondMov.
27071 (*th_cond_gpr_mov<GPR:mode><GPR2:mode>): Likewise.
27073 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
27075 * config/riscv/bitmanip.md (clzdi2): New expand.
27076 (clzsi2): New expand.
27077 (ctz<mode>2): New expand.
27078 (popcount<mode>2): New expand.
27079 (<bitmanip_optab>si2): Rename INSN.
27080 (*<bitmanip_optab>si2): Hide INSN name.
27081 (<bitmanip_optab>di2): Rename INSN.
27082 (*<bitmanip_optab>di2): Hide INSN name.
27083 (rotrsi3): Remove INSN.
27084 (rotr<mode>3): Add expand.
27085 (*rotrsi3): New INSN.
27086 (rotrdi3): Rename INSN.
27087 (*rotrdi3): Hide INSN name.
27088 (rotrsi3_sext): Rename INSN.
27089 (*rotrsi3_sext): Hide INSN name.
27090 (bswap<mode>2): Remove INSN.
27091 (bswapdi2): Add expand.
27092 (bswapsi2): Add expand.
27093 (*bswap<mode>2): Hide INSN name.
27094 * config/riscv/riscv.cc (riscv_rtx_costs): Add costs for sign
27096 * config/riscv/riscv.md (extv<mode>): New expand.
27097 (extzv<mode>): New expand.
27098 * config/riscv/thead.md (*th_srri<mode>3): New INSN.
27099 (*th_ext<mode>): New INSN.
27100 (*th_extu<mode>): New INSN.
27101 (*th_clz<mode>2): New INSN.
27102 (*th_rev<mode>2): New INSN.
27104 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
27106 * config/riscv/riscv.cc (riscv_rtx_costs): Add xthead:tst cost.
27107 * config/riscv/thead.md (*th_tst<mode>3): New INSN.
27109 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
27111 * config/riscv/riscv.md: Include thead.md
27112 * config/riscv/thead.md: New file.
27114 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
27116 * config/riscv/riscv-cores.def (RISCV_CORE): Add "thead-c906".
27118 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
27120 * common/config/riscv/riscv-common.cc: Add xthead* extensions.
27121 * config/riscv/riscv-opts.h (MASK_XTHEADBA): New.
27122 (MASK_XTHEADBB): New.
27123 (MASK_XTHEADBS): New.
27124 (MASK_XTHEADCMO): New.
27125 (MASK_XTHEADCONDMOV): New.
27126 (MASK_XTHEADFMEMIDX): New.
27127 (MASK_XTHEADFMV): New.
27128 (MASK_XTHEADINT): New.
27129 (MASK_XTHEADMAC): New.
27130 (MASK_XTHEADMEMIDX): New.
27131 (MASK_XTHEADMEMPAIR): New.
27132 (MASK_XTHEADSYNC): New.
27133 (TARGET_XTHEADBA): New.
27134 (TARGET_XTHEADBB): New.
27135 (TARGET_XTHEADBS): New.
27136 (TARGET_XTHEADCMO): New.
27137 (TARGET_XTHEADCONDMOV): New.
27138 (TARGET_XTHEADFMEMIDX): New.
27139 (TARGET_XTHEADFMV): New.
27140 (TARGET_XTHEADINT): New.
27141 (TARGET_XTHEADMAC): New.
27142 (TARGET_XTHEADMEMIDX): New.
27143 (TARGET_XTHEADMEMPAIR): new.
27144 (TARGET_XTHEADSYNC): New.
27145 * config/riscv/riscv.opt: Add riscv_xthead_subext.
27147 2023-03-15 Hu, Lin1 <lin1.hu@intel.com>
27150 * config/i386/i386-builtin.def (__builtin_ia32_vaesdec_v16qi,
27151 __builtin_ia32_vaesdeclast_v16qi,__builtin_ia32_vaesenc_v16qi,
27152 __builtin_ia32_vaesenclast_v16qi): Require OPTION_MASK_ISA_AVX512VL.
27154 2023-03-14 Jakub Jelinek <jakub@redhat.com>
27157 * config/i386/i386-expand.cc (split_double_concat): Fix splitting
27158 when lo is equal to dhi and hi is a MEM which uses dlo register.
27160 2023-03-14 Martin Jambor <mjambor@suse.cz>
27163 * ipa-cp.cc (update_profiling_info): Drop counts of orig_node to
27164 global0 instead of zeroing when it does not have as many counts as
27167 2023-03-14 Martin Jambor <mjambor@suse.cz>
27170 * ipa-cp.cc (update_specialized_profile): Drop orig_node_count to
27171 ipa count, remove assert, lenient_count_portion_handling, dump
27172 also orig_node_count.
27174 2023-03-14 Uros Bizjak <ubizjak@gmail.com>
27176 * config/i386/i386-expand.cc (expand_vec_perm_movs):
27177 Handle V2SImode for TARGET_MMX_WITH_SSE.
27178 * config/i386/mmx.md (*mmx_movss_<mode>): Rename from *mmx_movss
27179 using V2FI mode iterator to handle both V2SI and V2SF modes.
27181 2023-03-14 Sam James <sam@gentoo.org>
27183 * config/riscv/genrvv-type-indexer.cc: Avoid calloc() poisoning on musl by
27184 including <sstream> earlier.
27185 * system.h: Add INCLUDE_SSTREAM.
27187 2023-03-14 Richard Biener <rguenther@suse.de>
27189 * tree-ssa-live.cc (remove_unused_locals): Do not treat
27190 the .DEFERRED_INIT of a variable as use, instead remove
27191 that if it is the only use.
27193 2023-03-14 Eric Botcazou <ebotcazou@adacore.com>
27195 PR rtl-optimization/107762
27196 * expr.cc (emit_group_store): Revert latest change.
27198 2023-03-14 Andre Vieira <andre.simoesdiasvieira@arm.com>
27200 PR tree-optimization/109005
27201 * tree-if-conv.cc (get_bitfield_rep): Replace BLKmode check with
27202 aggregate type check.
27204 2023-03-14 Jakub Jelinek <jakub@redhat.com>
27206 PR tree-optimization/109115
27207 * tree-vect-patterns.cc (vect_recog_divmod_pattern): Don't use
27208 r.upper_bound () on r.undefined_p () range.
27210 2023-03-14 Jan Hubicka <hubicka@ucw.cz>
27212 PR tree-optimization/106896
27213 * profile-count.cc (profile_count::to_sreal_scale): Synchronize
27214 implementatoin with probability_in; avoid some asserts.
27216 2023-03-13 Max Filippov <jcmvbkbc@gmail.com>
27218 * config/xtensa/linux.h (TARGET_ASM_FILE_END): New macro.
27220 2023-03-13 Sean Bright <sean@seanbright.com>
27222 * doc/invoke.texi (Warning Options): Remove errant 'See'
27225 2023-03-13 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
27227 * config/xtensa/xtensa.h (REG_OK_STRICT, REG_OK_FOR_INDEX_P,
27228 REG_OK_FOR_BASE_P): Remove.
27230 2023-03-13 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
27232 * config/riscv/vector-iterators.md (=vd,vr): Fine tune.
27233 (=vd,vd,vr,vr): Ditto.
27234 * config/riscv/vector.md: Ditto.
27236 2023-03-13 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
27238 * config/riscv/riscv-vector-builtins.cc
27239 (function_expander::use_compare_insn): Add operand predicate check.
27241 2023-03-13 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
27243 * config/riscv/vector.md: Fine tune RA constraints.
27245 2023-03-13 Tobias Burnus <tobias@codesourcery.com>
27247 * config/gcn/mkoffload.cc (main): Pass -save-temps on for the
27248 hsaco assemble/link.
27250 2023-03-13 Richard Biener <rguenther@suse.de>
27252 PR tree-optimization/109046
27253 * tree-ssa-forwprop.cc (pass_forwprop::execute): Combine
27254 piecewise complex loads.
27256 2023-03-12 Jakub Jelinek <jakub@redhat.com>
27258 * config/aarch64/aarch64.h (aarch64_bf16_type_node): Remove.
27259 (aarch64_bf16_ptr_type_node): Adjust comment.
27260 * config/aarch64/aarch64.cc (aarch64_gimplify_va_arg_expr): Use
27261 bfloat16_type_node rather than aarch64_bf16_type_node.
27262 (aarch64_libgcc_floating_mode_supported_p,
27263 aarch64_scalar_mode_supported_p): Also support BFmode.
27264 (aarch64_invalid_conversion, aarch64_invalid_unary_op): Remove.
27265 (aarch64_invalid_binary_op): Remove BFmode related rejections.
27266 (TARGET_INVALID_CONVERSION, TARGET_INVALID_UNARY_OP): Don't redefine.
27267 * config/aarch64/aarch64-builtins.cc (aarch64_bf16_type_node): Remove.
27268 (aarch64_int_or_fp_type): Use bfloat16_type_node rather than
27269 aarch64_bf16_type_node.
27270 (aarch64_init_simd_builtin_types): Likewise.
27271 (aarch64_init_bf16_types): Likewise. Don't create bfloat16_type_node,
27272 which is created in tree.cc already.
27273 * config/aarch64/aarch64-sve-builtins.def (svbfloat16_t): Likewise.
27275 2023-03-12 Roger Sayle <roger@nextmovesoftware.com>
27277 PR middle-end/109031
27278 * tree-chrec.cc (chrec_apply): When folding "{a, +, a} (x-1)",
27279 ensure that the type of x is as wide or wider than the type of a.
27281 2023-03-12 Tamar Christina <tamar.christina@arm.com>
27284 * config/aarch64/aarch64-simd.md (@aarch64_bitmask_udiv<mode>3): Remove.
27285 (*bitmask_shift_plus<mode>): New.
27286 * config/aarch64/aarch64-sve2.md (*bitmask_shift_plus<mode>): New.
27287 (@aarch64_bitmask_udiv<mode>3): Remove.
27288 * config/aarch64/aarch64.cc
27289 (aarch64_vectorize_can_special_div_by_constant,
27290 TARGET_VECTORIZE_CAN_SPECIAL_DIV_BY_CONST): Removed.
27291 (TARGET_VECTORIZE_PREFERRED_DIV_AS_SHIFTS_OVER_MULT,
27292 aarch64_vectorize_preferred_div_as_shifts_over_mult): New.
27294 2023-03-12 Tamar Christina <tamar.christina@arm.com>
27297 * target.def (preferred_div_as_shifts_over_mult): New.
27298 * doc/tm.texi.in: Document it.
27299 * doc/tm.texi: Regenerate.
27300 * targhooks.cc (default_preferred_div_as_shifts_over_mult): New.
27301 * targhooks.h (default_preferred_div_as_shifts_over_mult): New.
27302 * tree-vect-patterns.cc (vect_recog_divmod_pattern): Use it.
27304 2023-03-12 Tamar Christina <tamar.christina@arm.com>
27305 Richard Sandiford <richard.sandiford@arm.com>
27308 * tree-ssa-math-opts.cc (convert_mult_to_fma): Inhibit FMA in case not
27311 2023-03-12 Tamar Christina <tamar.christina@arm.com>
27312 Andrew MacLeod <amacleod@redhat.com>
27315 * gimple-range-op.h (gimple_range_op_handler): Add maybe_non_standard.
27316 * gimple-range-op.cc (gimple_range_op_handler::gimple_range_op_handler):
27318 (gimple_range_op_handler::maybe_non_standard): New.
27319 * range-op.cc (class operator_widen_plus_signed,
27320 operator_widen_plus_signed::wi_fold, class operator_widen_plus_unsigned,
27321 operator_widen_plus_unsigned::wi_fold, class operator_widen_mult_signed,
27322 operator_widen_mult_signed::wi_fold, class operator_widen_mult_unsigned,
27323 operator_widen_mult_unsigned::wi_fold,
27324 ptr_op_widen_mult_signed, ptr_op_widen_mult_unsigned,
27325 ptr_op_widen_plus_signed, ptr_op_widen_plus_unsigned): New.
27326 * range-op.h (ptr_op_widen_mult_signed, ptr_op_widen_mult_unsigned,
27327 ptr_op_widen_plus_signed, ptr_op_widen_plus_unsigned): New
27329 2023-03-12 Tamar Christina <tamar.christina@arm.com>
27332 * doc/tm.texi (TARGET_VECTORIZE_CAN_SPECIAL_DIV_BY_CONST): Remove.
27333 * doc/tm.texi.in: Likewise.
27334 * explow.cc (round_push, align_dynamic_address): Revert previous patch.
27335 * expmed.cc (expand_divmod): Likewise.
27336 * expmed.h (expand_divmod): Likewise.
27337 * expr.cc (force_operand, expand_expr_divmod): Likewise.
27338 * optabs.cc (expand_doubleword_mod, expand_doubleword_divmod): Likewise.
27339 * target.def (can_special_div_by_const): Remove.
27340 * target.h: Remove tree-core.h include
27341 * targhooks.cc (default_can_special_div_by_const): Remove.
27342 * targhooks.h (default_can_special_div_by_const): Remove.
27343 * tree-vect-generic.cc (expand_vector_operation): Remove hook.
27344 * tree-vect-patterns.cc (vect_recog_divmod_pattern): Remove hook.
27345 * tree-vect-stmts.cc (vectorizable_operation): Remove hook.
27347 2023-03-12 Sandra Loosemore <sandra@codesourcery.com>
27349 * doc/install.texi2html: Fix issue number typo in comment.
27351 2023-03-12 Gaius Mulley <gaiusmod2@gmail.com>
27353 * doc/gm2.texi (Elementary data types): Equivalence BOOLEAN with
27356 2023-03-12 Sandra Loosemore <sandra@codesourcery.com>
27358 * doc/invoke.texi (Optimize Options): Add markup to
27359 description of asan-kernel-mem-intrinsic-prefix, and clarify
27362 2023-03-11 Gerald Pfeifer <gerald@pfeifer.com>
27364 * doc/extend.texi (Named Address Spaces): Drop a redundant link
27367 2023-03-11 Jeff Law <jlaw@ventanamicro>
27370 * doc/extend.texi: Clarify Attribute Syntax a bit.
27372 2023-03-11 Sandra Loosemore <sandra@codesourcery.com>
27374 * doc/install.texi (Prerequisites): Suggest using newer versions
27376 (Final install): Clean up and modernize discussion of how to
27377 build or obtain the GCC manuals.
27378 * doc/install.texi2html: Update comment to point to the PR instead
27379 of "makeinfo 4.7 brokenness" (it's not specific to that version).
27381 2023-03-10 Jakub Jelinek <jakub@redhat.com>
27384 * optabs.cc (expand_fix): For conversions from BFmode to integral,
27385 use shifts to convert it to SFmode first and then convert SFmode
27388 2023-03-10 Andrew Pinski <apinski@marvell.com>
27390 * config/aarch64/aarch64.md: Add a new define_split
27393 2023-03-10 Richard Biener <rguenther@suse.de>
27395 * tree-ssa-structalias.cc (solve_graph): Immediately
27396 iterate self-cycles.
27398 2023-03-10 Jakub Jelinek <jakub@redhat.com>
27400 PR tree-optimization/109008
27401 * range-op-float.cc (float_widen_lhs_range): If not
27402 -frounding-math and not IBM double double format, extend lhs
27403 range just by 0.5ulp rather than 1ulp in each direction.
27405 2023-03-10 Jakub Jelinek <jakub@redhat.com>
27408 * config.gcc (x86_64-*-cygwin*): Don't add i386/t-cygwin-w64 into
27410 * config/i386/t-cygwin-w64: Remove.
27412 2023-03-10 Jakub Jelinek <jakub@redhat.com>
27415 * tree-core.h (tree_code_type, tree_code_length): For C++11 or
27416 C++14, don't declare as extern const arrays.
27417 (tree_code_type_tmpl, tree_code_length_tmpl): New types with
27418 static constexpr member arrays for C++11 or C++14.
27419 * tree.h (TREE_CODE_CLASS): For C++11 or C++14 use
27420 tree_code_type_tmpl <0>::tree_code_type instead of tree_code_type.
27421 (TREE_CODE_LENGTH): For C++11 or C++14 use
27422 tree_code_length_tmpl <0>::tree_code_length instead of
27424 * tree.cc (tree_code_type, tree_code_length): Remove.
27426 2023-03-10 Jakub Jelinek <jakub@redhat.com>
27429 * common.opt (fcanon-prefix-map): New option.
27430 * opts.cc: Include file-prefix-map.h.
27431 (flag_canon_prefix_map): New variable.
27432 (common_handle_option): Handle OPT_fcanon_prefix_map.
27433 (gen_command_line_string): Ignore OPT_fcanon_prefix_map.
27434 * file-prefix-map.h (flag_canon_prefix_map): Declare.
27435 * file-prefix-map.cc (struct file_prefix_map): Add canonicalize
27437 (add_prefix_map): Initialize canonicalize member from
27438 flag_canon_prefix_map, and if true canonicalize it using lrealpath.
27439 (remap_filename): Revert 2022-11-01 and 2022-11-07 changes,
27440 use lrealpath result only for map->canonicalize map entries.
27441 * lto-opts.cc (lto_write_options): Ignore OPT_fcanon_prefix_map.
27442 * opts-global.cc (handle_common_deferred_options): Clear
27443 flag_canon_prefix_map at the start and handle OPT_fcanon_prefix_map.
27444 * doc/invoke.texi (-fcanon-prefix-map): Document.
27445 (-ffile-prefix-map, -fdebug-prefix-map, -fprofile-prefix-map): Add
27446 see also for -fcanon-prefix-map.
27447 * doc/cppopts.texi (-fmacro-prefix-map): Likewise.
27449 2023-03-10 Jakub Jelinek <jakub@redhat.com>
27452 * cgraphunit.cc (check_global_declaration): Don't warn for unused
27453 variables which have OPT_Wunused_variable warning suppressed.
27455 2023-03-10 Jakub Jelinek <jakub@redhat.com>
27457 PR tree-optimization/109008
27458 * range-op-float.cc (float_widen_lhs_range): If lb is
27459 minimum representable finite number or ub is maximum
27460 representable finite number, instead of widening it to
27461 -inf or inf widen it to negative or positive 0x0.8p+(EMAX+1).
27462 Temporarily clear flag_finite_math_only when canonicalizing
27465 2023-03-10 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
27467 * config/riscv/riscv-builtins.cc (riscv_gimple_fold_builtin): New function.
27468 * config/riscv/riscv-protos.h (riscv_gimple_fold_builtin): Ditto.
27469 (gimple_fold_builtin): Ditto.
27470 * config/riscv/riscv-vector-builtins-bases.cc (class read_vl): New class.
27471 (class vleff): Ditto.
27473 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
27474 * config/riscv/riscv-vector-builtins-functions.def (read_vl): Ditto.
27476 * config/riscv/riscv-vector-builtins-shapes.cc (struct read_vl_def): Ditto.
27477 (struct fault_load_def): Ditto.
27479 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
27480 * config/riscv/riscv-vector-builtins.cc
27481 (rvv_arg_type_info::get_tree_type): Add size_ptr.
27482 (gimple_folder::gimple_folder): New class.
27483 (gimple_folder::fold): Ditto.
27484 (gimple_fold_builtin): New function.
27485 (get_read_vl_instance): Ditto.
27486 (get_read_vl_decl): Ditto.
27487 * config/riscv/riscv-vector-builtins.def (size_ptr): Add size_ptr.
27488 * config/riscv/riscv-vector-builtins.h (class gimple_folder): New class.
27489 (get_read_vl_instance): New function.
27490 (get_read_vl_decl): Ditto.
27491 * config/riscv/riscv-vsetvl.cc (fault_first_load_p): Ditto.
27492 (read_vl_insn_p): Ditto.
27493 (available_occurrence_p): Ditto.
27494 (backward_propagate_worthwhile_p): Ditto.
27495 (gen_vsetvl_pat): Adapt for vleff support.
27496 (get_forward_read_vl_insn): New function.
27497 (get_backward_fault_first_load_insn): Ditto.
27498 (source_equal_p): Adapt for vleff support.
27499 (first_ratio_invalid_for_second_sew_p): Remove.
27500 (first_ratio_invalid_for_second_lmul_p): Ditto.
27501 (first_lmul_less_than_second_lmul_p): Ditto.
27502 (first_ratio_less_than_second_ratio_p): Ditto.
27503 (support_relaxed_compatible_p): New function.
27504 (vector_insn_info::operator>): Remove.
27505 (vector_insn_info::operator>=): Refine.
27506 (vector_insn_info::parse_insn): Adapt for vleff support.
27507 (vector_insn_info::compatible_p): Ditto.
27508 (vector_insn_info::update_fault_first_load_avl): New function.
27509 (pass_vsetvl::transfer_after): Adapt for vleff support.
27510 (pass_vsetvl::demand_fusion): Ditto.
27511 (pass_vsetvl::cleanup_insns): Ditto.
27512 * config/riscv/riscv-vsetvl.def (DEF_INCOMPATIBLE_COND): Remove
27513 redundant condtions.
27514 * config/riscv/riscv-vsetvl.h (struct demands_cond): New function.
27515 * config/riscv/riscv.cc (TARGET_GIMPLE_FOLD_BUILTIN): New target hook.
27516 * config/riscv/riscv.md: Adapt for vleff support.
27517 * config/riscv/t-riscv: Ditto.
27518 * config/riscv/vector-iterators.md: New iterator.
27519 * config/riscv/vector.md (read_vlsi): New pattern.
27520 (read_vldi_zero_extend): Ditto.
27521 (@pred_fault_load<mode>): Ditto.
27523 2023-03-10 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
27525 * config/riscv/riscv-vector-builtins.cc
27526 (function_expander::use_ternop_insn): Use maybe_gen_insn instead.
27527 (function_expander::use_widen_ternop_insn): Ditto.
27528 * optabs.cc (maybe_gen_insn): Extend nops handling.
27530 2023-03-10 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
27532 * config/riscv/riscv-vector-builtins-bases.cc: Split indexed load
27533 patterns according to RVV ISA.
27534 * config/riscv/vector-iterators.md: New iterators.
27535 * config/riscv/vector.md
27536 (@pred_indexed_<order>load<VNX1_QHSD:mode><VNX1_QHSDI:mode>): Remove.
27537 (@pred_indexed_<order>load<mode>_same_eew): New pattern.
27538 (@pred_indexed_<order>load<mode>_x2_greater_eew): Ditto.
27539 (@pred_indexed_<order>load<mode>_x4_greater_eew): Ditto.
27540 (@pred_indexed_<order>load<mode>_x8_greater_eew): Ditto.
27541 (@pred_indexed_<order>load<mode>_x2_smaller_eew): Ditto.
27542 (@pred_indexed_<order>load<mode>_x4_smaller_eew): Ditto.
27543 (@pred_indexed_<order>load<mode>_x8_smaller_eew): Ditto.
27544 (@pred_indexed_<order>load<VNX2_QHSD:mode><VNX2_QHSDI:mode>): Remove.
27545 (@pred_indexed_<order>load<VNX4_QHSD:mode><VNX4_QHSDI:mode>): Ditto.
27546 (@pred_indexed_<order>load<VNX8_QHSD:mode><VNX8_QHSDI:mode>): Ditto.
27547 (@pred_indexed_<order>load<VNX16_QHS:mode><VNX16_QHSI:mode>): Ditto.
27548 (@pred_indexed_<order>load<VNX32_QH:mode><VNX32_QHI:mode>): Ditto.
27549 (@pred_indexed_<order>load<VNX64_Q:mode><VNX64_Q:mode>): Ditto.
27551 2023-03-10 Michael Collison <collison@rivosinc.com>
27553 * tree-vect-loop-manip.cc (vect_do_peeling): Use
27554 result of constant_lower_bound instead of vf for the lower
27555 bound of the epilog loop trip count.
27557 2023-03-09 Tamar Christina <tamar.christina@arm.com>
27559 * passes.cc (emergency_dump_function): Finish graph generation.
27561 2023-03-09 Tamar Christina <tamar.christina@arm.com>
27563 * config/aarch64/aarch64.md (tbranch_<code><mode>3): Restrict to SHORT
27564 and bottom bit only.
27566 2023-03-09 Andrew Pinski <apinski@marvell.com>
27568 PR tree-optimization/108980
27569 * gimple-array-bounds.cc (array_bounds_checker::check_array_ref):
27570 Reorgnize the call to warning for not strict flexible arrays
27571 to be before the check of warned.
27573 2023-03-09 Jason Merrill <jason@redhat.com>
27575 * doc/extend.texi: Comment out __is_deducible docs.
27577 2023-03-09 Jason Merrill <jason@redhat.com>
27580 * doc/extend.texi (Type Traits):: Document __is_deducible.
27582 2023-03-09 Costas Argyris <costas.argyris@gmail.com>
27585 * config.host: add object for x86_64-*-mingw*.
27586 * config/i386/sym-mingw32.cc: dummy file to attach
27588 * config/i386/utf8-mingw32.rc: windres resource file.
27589 * config/i386/winnt-utf8.manifest: XML manifest to
27591 * config/i386/x-mingw32: reference to x-mingw32-utf8.
27592 * config/i386/x-mingw32-utf8: Makefile fragment to
27593 embed UTF-8 manifest.
27595 2023-03-09 Vladimir N. Makarov <vmakarov@redhat.com>
27597 * lra-constraints.cc (process_alt_operands): Use operand modes for
27598 clobbered regs instead of the biggest access mode.
27600 2023-03-09 Richard Biener <rguenther@suse.de>
27602 PR middle-end/108995
27603 * fold-const.cc (extract_muldiv_1): Avoid folding
27604 (CST * b) / CST2 when sanitizing overflow and we rely on
27605 overflow being undefined.
27607 2023-03-09 Jakub Jelinek <jakub@redhat.com>
27608 Richard Biener <rguenther@suse.de>
27610 PR tree-optimization/109008
27611 * range-op-float.cc (float_widen_lhs_range): New function.
27612 (foperator_plus::op1_range, foperator_minus::op1_range,
27613 foperator_minus::op2_range, foperator_mult::op1_range,
27614 foperator_div::op1_range, foperator_div::op2_range): Use it.
27616 2023-03-07 Jonathan Grant <jg@jguk.org>
27619 * doc/invoke.texi (Instrumentation Options): Clarify
27620 LeakSanitizer behavior.
27622 2023-03-07 Benson Muite <benson_muite@emailplus.org>
27624 * doc/install.texi (Prerequisites): Add link to gmplib.org.
27626 2023-03-07 Pan Li <pan2.li@intel.com>
27627 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
27631 * config/riscv/riscv-modes.def (ADJUST_PRECISION): Adjust VNx*BI
27633 * config/riscv/riscv.cc (riscv_v_adjust_precision): New.
27634 * config/riscv/riscv.h (riscv_v_adjust_precision): New.
27635 * genmodes.cc (adj_precision): New.
27636 (ADJUST_PRECISION): New.
27637 (emit_mode_adjustments): Handle ADJUST_PRECISION.
27639 2023-03-07 Hans-Peter Nilsson <hp@axis.com>
27641 * doc/sourcebuild.texi: Document check_effective_target_tail_call.
27643 2023-03-06 Paul-Antoine Arras <pa@codesourcery.com>
27645 * config/gcn/gcn-valu.md (<expander><mode>3_exec): Add patterns for
27646 {s|u}{max|min} in QI, HI and DI modes.
27647 (<expander><mode>3): Add pattern for {s|u}{max|min} in DI mode.
27648 (cond_<fexpander><mode>): Add pattern for cond_f{max|min}.
27649 (cond_<expander><mode>): Add pattern for cond_{s|u}{max|min}.
27650 * config/gcn/gcn.cc (gcn_spill_class): Allow the exec register to be
27653 2023-03-06 Richard Biener <rguenther@suse.de>
27655 PR tree-optimization/109025
27656 * tree-vect-loop.cc (vect_is_simple_reduction): Verify
27657 the inner LC PHI use is the inner loop PHI latch definition
27658 before classifying an outer PHI as double reduction.
27660 2023-03-06 Jan Hubicka <hubicka@ucw.cz>
27663 * config/i386/x86-tune.def (X86_TUNE_USE_SCATTER_2PARTS): Enable for
27665 (X86_TUNE_USE_SCATTER_4PARTS): Likewise.
27666 (X86_TUNE_USE_SCATTER): Likewise.
27668 2023-03-06 Xi Ruoyao <xry111@xry111.site>
27671 * config/loongarch/loongarch.h (FP_RETURN): Use
27672 TARGET_*_FLOAT_ABI instead of TARGET_*_FLOAT.
27673 (UNITS_PER_FP_ARG): Likewise.
27675 2023-03-05 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
27677 * config/riscv/riscv-vsetvl.cc (reg_available_p): Fix bug.
27678 (pass_vsetvl::backward_demand_fusion): Ditto.
27680 2023-03-05 Liao Shihua <shihua@iscas.ac.cn>
27681 SiYu Wu <siyu@isrc.iscas.ac.cn>
27683 * config/riscv/crypto.md (riscv_sm3p0_<mode>): Add ZKSED's and ZKSH's
27685 (riscv_sm3p1_<mode>): New.
27686 (riscv_sm4ed_<mode>): New.
27687 (riscv_sm4ks_<mode>): New.
27688 * config/riscv/riscv-builtins.cc (AVAIL): Add ZKSED's and ZKSH's AVAIL.
27689 * config/riscv/riscv-scalar-crypto.def (RISCV_BUILTIN): Add ZKSED's and
27690 ZKSH's built-in functions.
27692 2023-03-05 Liao Shihua <shihua@iscas.ac.cn>
27693 SiYu Wu <siyu@isrc.iscas.ac.cn>
27695 * config/riscv/crypto.md (riscv_sha256sig0_<mode>): Add ZKNH's instructions.
27696 (riscv_sha256sig1_<mode>): New.
27697 (riscv_sha256sum0_<mode>): New.
27698 (riscv_sha256sum1_<mode>): New.
27699 (riscv_sha512sig0h): New.
27700 (riscv_sha512sig0l): New.
27701 (riscv_sha512sig1h): New.
27702 (riscv_sha512sig1l): New.
27703 (riscv_sha512sum0r): New.
27704 (riscv_sha512sum1r): New.
27705 (riscv_sha512sig0): New.
27706 (riscv_sha512sig1): New.
27707 (riscv_sha512sum0): New.
27708 (riscv_sha512sum1): New.
27709 * config/riscv/riscv-builtins.cc (AVAIL): And ZKNH's AVAIL.
27710 * config/riscv/riscv-scalar-crypto.def (RISCV_BUILTIN): And ZKNH's
27711 built-in functions.
27712 (DIRECT_BUILTIN): Add new.
27714 2023-03-05 Liao Shihua <shihua@iscas.ac.cn>
27715 SiYu Wu <siyu@isrc.iscas.ac.cn>
27717 * config/riscv/constraints.md (D03): Add constants of bs and rnum.
27719 * config/riscv/crypto.md (riscv_aes32dsi): Add ZKND's and ZKNE's instructions.
27720 (riscv_aes32dsmi): New.
27721 (riscv_aes64ds): New.
27722 (riscv_aes64dsm): New.
27723 (riscv_aes64im): New.
27724 (riscv_aes64ks1i): New.
27725 (riscv_aes64ks2): New.
27726 (riscv_aes32esi): New.
27727 (riscv_aes32esmi): New.
27728 (riscv_aes64es): New.
27729 (riscv_aes64esm): New.
27730 * config/riscv/riscv-builtins.cc (AVAIL): Add ZKND's and ZKNE's AVAIL.
27731 * config/riscv/riscv-scalar-crypto.def (DIRECT_BUILTIN): Add ZKND's and
27732 ZKNE's built-in functions.
27734 2023-03-05 Liao Shihua <shihua@iscas.ac.cn>
27735 SiYu Wu <siyu@isrc.iscas.ac.cn>
27737 * config/riscv/bitmanip.md: Add ZBKB's instructions.
27738 * config/riscv/riscv-builtins.cc (AVAIL): Add new.
27739 * config/riscv/riscv.md: Add new type for crypto instructions.
27740 * config/riscv/crypto.md: Add Scalar Cryptography extension's machine
27742 * config/riscv/riscv-scalar-crypto.def: Add Scalar Cryptography
27743 extension's built-in function file.
27745 2023-03-05 Liao Shihua <shihua@iscas.ac.cn>
27746 SiYu Wu <siyu@isrc.iscas.ac.cn>
27748 * config/riscv/riscv-builtins.cc (RISCV_FTYPE_NAME2): New.
27749 (RISCV_FTYPE_NAME3): New.
27750 (RISCV_ATYPE_QI): New.
27751 (RISCV_ATYPE_HI): New.
27752 (RISCV_FTYPE_ATYPES2): New.
27753 (RISCV_FTYPE_ATYPES3): New.
27754 * config/riscv/riscv-ftypes.def (2): New.
27757 2023-03-05 Vineet Gupta <vineetg@rivosinc.com>
27759 * config/riscv/riscv.cc (riscv_rtx_costs): Fixed IN_RANGE() to
27762 2023-03-05 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
27763 kito-cheng <kito.cheng@sifive.com>
27765 * config/riscv/predicates.md (vector_any_register_operand): New predicate.
27766 * config/riscv/riscv-c.cc (riscv_check_builtin_call): New function.
27767 (riscv_register_pragmas): Add builtin function check call.
27768 * config/riscv/riscv-protos.h (RVV_VUNDEF): Adapt macro.
27769 (check_builtin_call): New function.
27770 * config/riscv/riscv-vector-builtins-bases.cc (class vundefined): New class.
27771 (class vreinterpret): Ditto.
27772 (class vlmul_ext): Ditto.
27773 (class vlmul_trunc): Ditto.
27774 (class vset): Ditto.
27775 (class vget): Ditto.
27777 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
27778 * config/riscv/riscv-vector-builtins-functions.def (vluxei8): Change name.
27794 (vundefined): Add new intrinsic.
27795 (vreinterpret): Ditto.
27796 (vlmul_ext): Ditto.
27797 (vlmul_trunc): Ditto.
27800 * config/riscv/riscv-vector-builtins-shapes.cc (struct return_mask_def): New class.
27801 (struct narrow_alu_def): Ditto.
27802 (struct reduc_alu_def): Ditto.
27803 (struct vundefined_def): Ditto.
27804 (struct misc_def): Ditto.
27805 (struct vset_def): Ditto.
27806 (struct vget_def): Ditto.
27808 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
27809 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_EEW8_INTERPRET_OPS): New def.
27810 (DEF_RVV_EEW16_INTERPRET_OPS): Ditto.
27811 (DEF_RVV_EEW32_INTERPRET_OPS): Ditto.
27812 (DEF_RVV_EEW64_INTERPRET_OPS): Ditto.
27813 (DEF_RVV_X2_VLMUL_EXT_OPS): Ditto.
27814 (DEF_RVV_X4_VLMUL_EXT_OPS): Ditto.
27815 (DEF_RVV_X8_VLMUL_EXT_OPS): Ditto.
27816 (DEF_RVV_X16_VLMUL_EXT_OPS): Ditto.
27817 (DEF_RVV_X32_VLMUL_EXT_OPS): Ditto.
27818 (DEF_RVV_X64_VLMUL_EXT_OPS): Ditto.
27819 (DEF_RVV_LMUL1_OPS): Ditto.
27820 (DEF_RVV_LMUL2_OPS): Ditto.
27821 (DEF_RVV_LMUL4_OPS): Ditto.
27822 (vint16mf4_t): Ditto.
27823 (vint16mf2_t): Ditto.
27824 (vint16m1_t): Ditto.
27825 (vint16m2_t): Ditto.
27826 (vint16m4_t): Ditto.
27827 (vint16m8_t): Ditto.
27828 (vint32mf2_t): Ditto.
27829 (vint32m1_t): Ditto.
27830 (vint32m2_t): Ditto.
27831 (vint32m4_t): Ditto.
27832 (vint32m8_t): Ditto.
27833 (vint64m1_t): Ditto.
27834 (vint64m2_t): Ditto.
27835 (vint64m4_t): Ditto.
27836 (vint64m8_t): Ditto.
27837 (vuint16mf4_t): Ditto.
27838 (vuint16mf2_t): Ditto.
27839 (vuint16m1_t): Ditto.
27840 (vuint16m2_t): Ditto.
27841 (vuint16m4_t): Ditto.
27842 (vuint16m8_t): Ditto.
27843 (vuint32mf2_t): Ditto.
27844 (vuint32m1_t): Ditto.
27845 (vuint32m2_t): Ditto.
27846 (vuint32m4_t): Ditto.
27847 (vuint32m8_t): Ditto.
27848 (vuint64m1_t): Ditto.
27849 (vuint64m2_t): Ditto.
27850 (vuint64m4_t): Ditto.
27851 (vuint64m8_t): Ditto.
27852 (vint8mf4_t): Ditto.
27853 (vint8mf2_t): Ditto.
27854 (vint8m1_t): Ditto.
27855 (vint8m2_t): Ditto.
27856 (vint8m4_t): Ditto.
27857 (vint8m8_t): Ditto.
27858 (vuint8mf4_t): Ditto.
27859 (vuint8mf2_t): Ditto.
27860 (vuint8m1_t): Ditto.
27861 (vuint8m2_t): Ditto.
27862 (vuint8m4_t): Ditto.
27863 (vuint8m8_t): Ditto.
27864 (vint8mf8_t): Ditto.
27865 (vuint8mf8_t): Ditto.
27866 (vfloat32mf2_t): Ditto.
27867 (vfloat32m1_t): Ditto.
27868 (vfloat32m2_t): Ditto.
27869 (vfloat32m4_t): Ditto.
27870 (vfloat64m1_t): Ditto.
27871 (vfloat64m2_t): Ditto.
27872 (vfloat64m4_t): Ditto.
27873 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_TYPE): Ditto.
27874 (DEF_RVV_EEW8_INTERPRET_OPS): Ditto.
27875 (DEF_RVV_EEW16_INTERPRET_OPS): Ditto.
27876 (DEF_RVV_EEW32_INTERPRET_OPS): Ditto.
27877 (DEF_RVV_EEW64_INTERPRET_OPS): Ditto.
27878 (DEF_RVV_X2_VLMUL_EXT_OPS): Ditto.
27879 (DEF_RVV_X4_VLMUL_EXT_OPS): Ditto.
27880 (DEF_RVV_X8_VLMUL_EXT_OPS): Ditto.
27881 (DEF_RVV_X16_VLMUL_EXT_OPS): Ditto.
27882 (DEF_RVV_X32_VLMUL_EXT_OPS): Ditto.
27883 (DEF_RVV_X64_VLMUL_EXT_OPS): Ditto.
27884 (DEF_RVV_LMUL1_OPS): Ditto.
27885 (DEF_RVV_LMUL2_OPS): Ditto.
27886 (DEF_RVV_LMUL4_OPS): Ditto.
27887 (DEF_RVV_TYPE_INDEX): Ditto.
27888 (required_extensions_p): Adapt for new intrinsic support/
27889 (get_required_extensions): New function.
27890 (check_required_extensions): Ditto.
27891 (unsigned_base_type_p): Remove.
27892 (rvv_arg_type_info::get_scalar_ptr_type): New function.
27893 (get_mode_for_bitsize): Remove.
27894 (rvv_arg_type_info::get_scalar_const_ptr_type): New function.
27895 (rvv_arg_type_info::get_base_vector_type): Ditto.
27896 (rvv_arg_type_info::get_function_type_index): Ditto.
27897 (DEF_RVV_BASE_TYPE): New def.
27898 (function_builder::apply_predication): New class.
27899 (function_expander::mask_mode): Ditto.
27900 (function_checker::function_checker): Ditto.
27901 (function_checker::report_non_ice): Ditto.
27902 (function_checker::report_out_of_range): Ditto.
27903 (function_checker::require_immediate): Ditto.
27904 (function_checker::require_immediate_range): Ditto.
27905 (function_checker::check): Ditto.
27906 (check_builtin_call): Ditto.
27907 * config/riscv/riscv-vector-builtins.def (DEF_RVV_TYPE): New def.
27908 (DEF_RVV_BASE_TYPE): Ditto.
27909 (DEF_RVV_TYPE_INDEX): Ditto.
27910 (vbool64_t): Ditto.
27911 (vbool32_t): Ditto.
27912 (vbool16_t): Ditto.
27917 (vuint8mf8_t): Ditto.
27918 (vuint8mf4_t): Ditto.
27919 (vuint8mf2_t): Ditto.
27920 (vuint8m1_t): Ditto.
27921 (vuint8m2_t): Ditto.
27922 (vint8m4_t): Ditto.
27923 (vuint8m4_t): Ditto.
27924 (vint8m8_t): Ditto.
27925 (vuint8m8_t): Ditto.
27926 (vint16mf4_t): Ditto.
27927 (vuint16mf2_t): Ditto.
27928 (vuint16m1_t): Ditto.
27929 (vuint16m2_t): Ditto.
27930 (vuint16m4_t): Ditto.
27931 (vuint16m8_t): Ditto.
27932 (vint32mf2_t): Ditto.
27933 (vuint32m1_t): Ditto.
27934 (vuint32m2_t): Ditto.
27935 (vuint32m4_t): Ditto.
27936 (vuint32m8_t): Ditto.
27937 (vuint64m1_t): Ditto.
27938 (vuint64m2_t): Ditto.
27939 (vuint64m4_t): Ditto.
27940 (vuint64m8_t): Ditto.
27941 (vfloat32mf2_t): Ditto.
27942 (vfloat32m1_t): Ditto.
27943 (vfloat32m2_t): Ditto.
27944 (vfloat32m4_t): Ditto.
27945 (vfloat32m8_t): Ditto.
27946 (vfloat64m1_t): Ditto.
27947 (vfloat64m4_t): Ditto.
27948 (vector): Move it def.
27951 (signed_vector): Ditto.
27952 (unsigned_vector): Ditto.
27953 (unsigned_scalar): Ditto.
27954 (vector_ptr): Ditto.
27955 (scalar_ptr): Ditto.
27956 (scalar_const_ptr): Ditto.
27960 (unsigned_long): Ditto.
27962 (eew8_index): Ditto.
27963 (eew16_index): Ditto.
27964 (eew32_index): Ditto.
27965 (eew64_index): Ditto.
27966 (shift_vector): Ditto.
27967 (double_trunc_vector): Ditto.
27968 (quad_trunc_vector): Ditto.
27969 (oct_trunc_vector): Ditto.
27970 (double_trunc_scalar): Ditto.
27971 (double_trunc_signed_vector): Ditto.
27972 (double_trunc_unsigned_vector): Ditto.
27973 (double_trunc_unsigned_scalar): Ditto.
27974 (double_trunc_float_vector): Ditto.
27975 (float_vector): Ditto.
27976 (lmul1_vector): Ditto.
27977 (widen_lmul1_vector): Ditto.
27978 (eew8_interpret): Ditto.
27979 (eew16_interpret): Ditto.
27980 (eew32_interpret): Ditto.
27981 (eew64_interpret): Ditto.
27982 (vlmul_ext_x2): Ditto.
27983 (vlmul_ext_x4): Ditto.
27984 (vlmul_ext_x8): Ditto.
27985 (vlmul_ext_x16): Ditto.
27986 (vlmul_ext_x32): Ditto.
27987 (vlmul_ext_x64): Ditto.
27988 * config/riscv/riscv-vector-builtins.h (DEF_RVV_BASE_TYPE): New def.
27989 (struct function_type_info): New function.
27990 (struct rvv_arg_type_info): Ditto.
27991 (class function_checker): New class.
27992 (rvv_arg_type_info::get_scalar_type): New function.
27993 (rvv_arg_type_info::get_vector_type): Ditto.
27994 (function_expander::ret_mode): New function.
27995 (function_checker::arg_mode): Ditto.
27996 (function_checker::ret_mode): Ditto.
27997 * config/riscv/t-riscv: Add generator.
27998 * config/riscv/vector-iterators.md: New iterators.
27999 * config/riscv/vector.md (vundefined<mode>): New pattern.
28000 (@vundefined<mode>): Ditto.
28001 (@vreinterpret<mode>): Ditto.
28002 (@vlmul_extx2<mode>): Ditto.
28003 (@vlmul_extx4<mode>): Ditto.
28004 (@vlmul_extx8<mode>): Ditto.
28005 (@vlmul_extx16<mode>): Ditto.
28006 (@vlmul_extx32<mode>): Ditto.
28007 (@vlmul_extx64<mode>): Ditto.
28008 (*vlmul_extx2<mode>): Ditto.
28009 (*vlmul_extx4<mode>): Ditto.
28010 (*vlmul_extx8<mode>): Ditto.
28011 (*vlmul_extx16<mode>): Ditto.
28012 (*vlmul_extx32<mode>): Ditto.
28013 (*vlmul_extx64<mode>): Ditto.
28014 * config/riscv/genrvv-type-indexer.cc: New file.
28016 2023-03-05 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
28018 * config/riscv/riscv-protos.h (enum vlen_enum): New enum.
28019 (slide1_sew64_helper): New function.
28020 * config/riscv/riscv-v.cc (compute_vlmax): Ditto.
28021 (get_unknown_min_value): Ditto.
28022 (force_vector_length_operand): Ditto.
28023 (gen_no_side_effects_vsetvl_rtx): Ditto.
28024 (get_vl_x2_rtx): Ditto.
28025 (slide1_sew64_helper): Ditto.
28026 * config/riscv/riscv-vector-builtins-bases.cc (class slideop): New class.
28027 (class vrgather): Ditto.
28028 (class vrgatherei16): Ditto.
28029 (class vcompress): Ditto.
28031 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
28032 * config/riscv/riscv-vector-builtins-functions.def (vslideup): Ditto.
28033 (vslidedown): Ditto.
28034 (vslide1up): Ditto.
28035 (vslide1down): Ditto.
28036 (vfslide1up): Ditto.
28037 (vfslide1down): Ditto.
28039 (vrgatherei16): Ditto.
28040 (vcompress): Ditto.
28041 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_EI16_OPS): New macro.
28042 (vint8mf8_t): Ditto.
28043 (vint8mf4_t): Ditto.
28044 (vint8mf2_t): Ditto.
28045 (vint8m1_t): Ditto.
28046 (vint8m2_t): Ditto.
28047 (vint8m4_t): Ditto.
28048 (vint16mf4_t): Ditto.
28049 (vint16mf2_t): Ditto.
28050 (vint16m1_t): Ditto.
28051 (vint16m2_t): Ditto.
28052 (vint16m4_t): Ditto.
28053 (vint16m8_t): Ditto.
28054 (vint32mf2_t): Ditto.
28055 (vint32m1_t): Ditto.
28056 (vint32m2_t): Ditto.
28057 (vint32m4_t): Ditto.
28058 (vint32m8_t): Ditto.
28059 (vint64m1_t): Ditto.
28060 (vint64m2_t): Ditto.
28061 (vint64m4_t): Ditto.
28062 (vint64m8_t): Ditto.
28063 (vuint8mf8_t): Ditto.
28064 (vuint8mf4_t): Ditto.
28065 (vuint8mf2_t): Ditto.
28066 (vuint8m1_t): Ditto.
28067 (vuint8m2_t): Ditto.
28068 (vuint8m4_t): Ditto.
28069 (vuint16mf4_t): Ditto.
28070 (vuint16mf2_t): Ditto.
28071 (vuint16m1_t): Ditto.
28072 (vuint16m2_t): Ditto.
28073 (vuint16m4_t): Ditto.
28074 (vuint16m8_t): Ditto.
28075 (vuint32mf2_t): Ditto.
28076 (vuint32m1_t): Ditto.
28077 (vuint32m2_t): Ditto.
28078 (vuint32m4_t): Ditto.
28079 (vuint32m8_t): Ditto.
28080 (vuint64m1_t): Ditto.
28081 (vuint64m2_t): Ditto.
28082 (vuint64m4_t): Ditto.
28083 (vuint64m8_t): Ditto.
28084 (vfloat32mf2_t): Ditto.
28085 (vfloat32m1_t): Ditto.
28086 (vfloat32m2_t): Ditto.
28087 (vfloat32m4_t): Ditto.
28088 (vfloat32m8_t): Ditto.
28089 (vfloat64m1_t): Ditto.
28090 (vfloat64m2_t): Ditto.
28091 (vfloat64m4_t): Ditto.
28092 (vfloat64m8_t): Ditto.
28093 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_EI16_OPS): Ditto.
28094 * config/riscv/riscv.md: Adjust RVV instruction types.
28095 * config/riscv/vector-iterators.md (down): New iterator.
28096 (=vd,vr): New attribute.
28097 (UNSPEC_VSLIDE1UP): New unspec.
28098 * config/riscv/vector.md (@pred_slide<ud><mode>): New pattern.
28099 (*pred_slide<ud><mode>): Ditto.
28100 (*pred_slide<ud><mode>_extended): Ditto.
28101 (@pred_gather<mode>): Ditto.
28102 (@pred_gather<mode>_scalar): Ditto.
28103 (@pred_gatherei16<mode>): Ditto.
28104 (@pred_compress<mode>): Ditto.
28106 2023-03-05 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
28108 * config/riscv/riscv-vector-builtins.cc: Remove void_type_node.
28110 2023-03-05 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
28112 * config/riscv/constraints.md (Wb1): New constraint.
28113 * config/riscv/predicates.md
28114 (vector_least_significant_set_mask_operand): New predicate.
28115 (vector_broadcast_mask_operand): Ditto.
28116 * config/riscv/riscv-protos.h (enum vlmul_type): Adjust.
28117 (gen_scalar_move_mask): New function.
28118 * config/riscv/riscv-v.cc (gen_scalar_move_mask): Ditto.
28119 * config/riscv/riscv-vector-builtins-bases.cc (class vmv): New class.
28120 (class vmv_s): Ditto.
28122 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
28123 * config/riscv/riscv-vector-builtins-functions.def (vmv_x): Ditto.
28127 * config/riscv/riscv-vector-builtins-shapes.cc (struct scalar_move_def): Ditto.
28129 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
28130 * config/riscv/riscv-vector-builtins.cc (function_expander::mask_mode): Ditto.
28131 (function_expander::use_exact_insn): New function.
28132 (function_expander::use_contiguous_load_insn): New function.
28133 (function_expander::use_contiguous_store_insn): New function.
28134 (function_expander::use_ternop_insn): New function.
28135 (function_expander::use_widen_ternop_insn): New function.
28136 (function_expander::use_scalar_move_insn): New function.
28137 * config/riscv/riscv-vector-builtins.def (s): New operand suffix.
28138 * config/riscv/riscv-vector-builtins.h
28139 (function_expander::add_scalar_move_mask_operand): New class.
28140 * config/riscv/riscv-vsetvl.cc (ignore_vlmul_insn_p): New function.
28141 (scalar_move_insn_p): Ditto.
28142 (has_vsetvl_killed_avl_p): Ditto.
28143 (anticipatable_occurrence_p): Ditto.
28144 (insert_vsetvl): Ditto.
28145 (get_vl_vtype_info): Ditto.
28146 (calculate_sew): Ditto.
28147 (calculate_vlmul): Ditto.
28148 (incompatible_avl_p): Ditto.
28149 (different_sew_p): Ditto.
28150 (different_lmul_p): Ditto.
28151 (different_ratio_p): Ditto.
28152 (different_tail_policy_p): Ditto.
28153 (different_mask_policy_p): Ditto.
28154 (possible_zero_avl_p): Ditto.
28155 (first_ratio_invalid_for_second_sew_p): Ditto.
28156 (first_ratio_invalid_for_second_lmul_p): Ditto.
28157 (second_ratio_invalid_for_first_sew_p): Ditto.
28158 (second_ratio_invalid_for_first_lmul_p): Ditto.
28159 (second_sew_less_than_first_sew_p): Ditto.
28160 (first_sew_less_than_second_sew_p): Ditto.
28161 (compare_lmul): Ditto.
28162 (second_lmul_less_than_first_lmul_p): Ditto.
28163 (first_lmul_less_than_second_lmul_p): Ditto.
28164 (first_ratio_less_than_second_ratio_p): Ditto.
28165 (second_ratio_less_than_first_ratio_p): Ditto.
28166 (DEF_INCOMPATIBLE_COND): Ditto.
28167 (greatest_sew): Ditto.
28168 (first_sew): Ditto.
28169 (second_sew): Ditto.
28170 (first_vlmul): Ditto.
28171 (second_vlmul): Ditto.
28172 (first_ratio): Ditto.
28173 (second_ratio): Ditto.
28174 (vlmul_for_first_sew_second_ratio): Ditto.
28175 (ratio_for_second_sew_first_vlmul): Ditto.
28176 (DEF_SEW_LMUL_FUSE_RULE): Ditto.
28177 (always_unavailable): Ditto.
28178 (avl_unavailable_p): Ditto.
28179 (sew_unavailable_p): Ditto.
28180 (lmul_unavailable_p): Ditto.
28181 (ge_sew_unavailable_p): Ditto.
28182 (ge_sew_lmul_unavailable_p): Ditto.
28183 (ge_sew_ratio_unavailable_p): Ditto.
28184 (DEF_UNAVAILABLE_COND): Ditto.
28185 (same_sew_lmul_demand_p): Ditto.
28186 (propagate_avl_across_demands_p): Ditto.
28187 (reg_available_p): Ditto.
28188 (avl_info::has_non_zero_avl): Ditto.
28189 (vl_vtype_info::has_non_zero_avl): Ditto.
28190 (vector_insn_info::operator>=): Refactor.
28191 (vector_insn_info::parse_insn): Adjust for scalar move.
28192 (vector_insn_info::demand_vl_vtype): Remove.
28193 (vector_insn_info::compatible_p): New function.
28194 (vector_insn_info::compatible_avl_p): Ditto.
28195 (vector_insn_info::compatible_vtype_p): Ditto.
28196 (vector_insn_info::available_p): Ditto.
28197 (vector_insn_info::merge): Ditto.
28198 (vector_insn_info::fuse_avl): Ditto.
28199 (vector_insn_info::fuse_sew_lmul): Ditto.
28200 (vector_insn_info::fuse_tail_policy): Ditto.
28201 (vector_insn_info::fuse_mask_policy): Ditto.
28202 (vector_insn_info::dump): Ditto.
28203 (vector_infos_manager::release): Ditto.
28204 (pass_vsetvl::compute_local_backward_infos): Adjust for scalar move support.
28205 (pass_vsetvl::get_backward_fusion_type): Adjust for scalar move support.
28206 (pass_vsetvl::hard_empty_block_p): Ditto.
28207 (pass_vsetvl::backward_demand_fusion): Ditto.
28208 (pass_vsetvl::forward_demand_fusion): Ditto.
28209 (pass_vsetvl::refine_vsetvls): Ditto.
28210 (pass_vsetvl::cleanup_vsetvls): Ditto.
28211 (pass_vsetvl::commit_vsetvls): Ditto.
28212 (pass_vsetvl::propagate_avl): Ditto.
28213 * config/riscv/riscv-vsetvl.h (enum demand_status): New class.
28214 (struct demands_pair): Ditto.
28215 (struct demands_cond): Ditto.
28216 (struct demands_fuse_rule): Ditto.
28217 * config/riscv/vector-iterators.md: New iterator.
28218 * config/riscv/vector.md (@pred_broadcast<mode>): New pattern.
28219 (*pred_broadcast<mode>): Ditto.
28220 (*pred_broadcast<mode>_extended_scalar): Ditto.
28221 (@pred_extract_first<mode>): Ditto.
28222 (*pred_extract_first<mode>): Ditto.
28223 (@pred_extract_first_trunc<mode>): Ditto.
28224 * config/riscv/riscv-vsetvl.def: New file.
28226 2023-03-05 Lin Sinan <sinan.lin@linux.alibaba.com>
28228 * config/riscv/bitmanip.md: allow 0 constant in max/min
28231 2023-03-05 Lin Sinan <sinan.lin@linux.alibaba.com>
28233 * config/riscv/bitmanip.md: Fix wrong index in the check.
28235 2023-03-04 Jakub Jelinek <jakub@redhat.com>
28237 PR middle-end/109006
28238 * vec.cc (test_auto_alias): Adjust comment for removal of
28240 * read-rtl-function.cc (function_reader::parse_block): Likewise.
28241 * gdbhooks.py: Likewise.
28243 2023-03-04 Jakub Jelinek <jakub@redhat.com>
28245 PR testsuite/108973
28246 * selftest-diagnostic.cc
28247 (test_diagnostic_context::test_diagnostic_context): Set
28248 caret_max_width to 80.
28250 2023-03-03 Alexandre Oliva <oliva@adacore.com>
28252 * gimple-ssa-warn-access.cc
28253 (pass_waccess::check_dangling_stores): Skip non-stores.
28255 2023-03-03 Alexandre Oliva <oliva@adacore.com>
28257 * config/arm/vfp.md (*thumb2_movsi_vfp): Drop blank after tab
28258 after vmsr and vmrs, and lower the case of P0.
28260 2023-03-03 Jonathan Wakely <jwakely@redhat.com>
28262 PR middle-end/109006
28263 * gdbhooks.py (VecPrinter): Handle vec<T> as well as vec<T>*.
28265 2023-03-03 Jonathan Wakely <jwakely@redhat.com>
28267 PR middle-end/109006
28268 * gdbhooks.py (VecPrinter): Adjust for new vec layout.
28270 2023-03-03 Jakub Jelinek <jakub@redhat.com>
28273 * gimple-ssa-warn-access.cc (pass_waccess::maybe_check_access_sizes):
28274 Return immediately if OPT_Wnonnull or OPT_Wstringop_overflow_ is
28275 suppressed on stmt. For [static %E] warning, print access_nelts
28276 rather than access_size. Fix up comment wording.
28278 2023-03-03 Robin Dapp <rdapp@linux.ibm.com>
28280 * config/s390/driver-native.cc (s390_host_detect_local_cpu): Use
28281 arch14 instead of z16.
28283 2023-03-03 Anthony Green <green@moxielogic.com>
28285 * config/moxie/moxie.cc (TARGET_LRA_P): Remove.
28287 2023-03-03 Anthony Green <green@moxielogic.com>
28289 * config/moxie/constraints.md (A, B, W): Change
28290 define_constraint to define_memory_constraint.
28292 2023-03-03 Xi Ruoyao <xry111@xry111.site>
28294 * toplev.cc (process_options): Fix the spelling of
28295 "-fstack-clash-protection".
28297 2023-03-03 Richard Biener <rguenther@suse.de>
28299 PR tree-optimization/109002
28300 * tree-ssa-pre.cc (compute_partial_antic_aux): Properly
28301 PHI-translate ANTIC_IN.
28303 2023-03-03 Jakub Jelinek <jakub@redhat.com>
28305 PR tree-optimization/108988
28306 * gimple-fold.cc (gimple_fold_builtin_fputs): Fold len to
28307 size_type_node before passing it as argument to fwrite. Formatting
28310 2023-03-03 Richard Biener <rguenther@suse.de>
28313 * config/i386/i386.opt (--param x86-stv-max-visits): New param.
28314 * doc/invoke.texi (--param x86-stv-max-visits): Document it.
28315 * config/i386/i386-features.h (scalar_chain::max_visits): New.
28316 (scalar_chain::build): Add bitmap parameter, return boolean.
28317 (scalar_chain::add_insn): Likewise.
28318 (scalar_chain::analyze_register_chain): Likewise.
28319 * config/i386/i386-features.cc (scalar_chain::scalar_chain):
28320 Initialize max_visits.
28321 (scalar_chain::analyze_register_chain): When we exhaust
28322 max_visits, abort. Also abort when running into any
28324 (scalar_chain::add_insn): Propagate abort.
28325 (scalar_chain::build): Likewise. When aborting amend
28326 the set of disallowed insn with the insns set.
28327 (convert_scalars_to_vector): Adjust. Do not convert aborted
28330 2023-03-03 Richard Biener <rguenther@suse.de>
28333 * dwarf2out.cc (dwarf2out_late_global_decl): Do not
28334 generate a DIE for a function scope static.
28336 2023-03-03 Alexandre Oliva <oliva@adacore.com>
28338 * config/vx-common.h (WINT_TYPE): Alias to "wchar_t".
28340 2023-03-02 Jakub Jelinek <jakub@redhat.com>
28343 * target.h (emit_support_tinfos_callback): New typedef.
28344 * targhooks.h (default_emit_support_tinfos): Declare.
28345 * targhooks.cc (default_emit_support_tinfos): New function.
28346 * target.def (emit_support_tinfos): New target hook.
28347 * doc/tm.texi.in (emit_support_tinfos): Document it.
28348 * doc/tm.texi: Regenerated.
28349 * config/i386/i386.cc (ix86_emit_support_tinfos): New function.
28350 (TARGET_EMIT_SUPPORT_TINFOS): Redefine.
28352 2023-03-02 Vladimir N. Makarov <vmakarov@redhat.com>
28354 * ira-costs.cc: Include print-rtl.h.
28355 (record_reg_classes, scan_one_insn): Add code to print debug info.
28356 (record_operand_costs): Find and use smaller cost for hard reg
28359 2023-03-02 Kwok Cheung Yeung <kcy@codesourcery.com>
28360 Paul-Antoine Arras <pa@codesourcery.com>
28362 * builtins.cc (mathfn_built_in_explicit): New.
28363 * config/gcn/gcn.cc: Include case-cfn-macros.h.
28364 (mathfn_built_in_explicit): Add prototype.
28365 (gcn_vectorize_builtin_vectorized_function): New.
28366 (gcn_libc_has_function): New.
28367 (TARGET_LIBC_HAS_FUNCTION): Define.
28368 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Define.
28370 2023-03-02 Richard Sandiford <richard.sandiford@arm.com>
28372 PR tree-optimization/108979
28373 * tree-vect-stmts.cc (vectorizable_operation): Don't mask
28374 operations on invariants.
28376 2023-03-02 Robin Dapp <rdapp@linux.ibm.com>
28378 * config/s390/predicates.md (vll_bias_operand): Add -1 bias.
28379 * config/s390/s390.cc (s390_option_override_internal): Make
28380 partial vector usage the default from z13 on.
28381 * config/s390/vector.md (len_load_v16qi): Add.
28382 (len_store_v16qi): Add.
28384 2023-03-02 Andre Vieira <andre.simoesdiasvieira@arm.com>
28386 * simplify-rtx.cc (simplify_context::simplify_subreg): Use byte instead
28387 of constant 0 offset.
28389 2023-03-02 Robert Suchanek <robert.suchanek@imgtec.com>
28391 * config/mips/mips.cc (mips_set_text_contents_type): Use HOST_WIDE_INT
28393 * config/mips/mips-protos.h (mips_set_text_contents_type): Likewise.
28395 2023-03-02 Junxian Zhu <zhujunxian@oss.cipunited.com>
28397 * config.gcc: add -with-{no-}msa build option.
28398 * config/mips/mips.h: Likewise.
28399 * doc/install.texi: Likewise.
28401 2023-03-02 Richard Sandiford <richard.sandiford@arm.com>
28403 PR tree-optimization/108603
28404 * explow.cc (convert_memory_address_addr_space_1): Only wrap
28405 the result of a recursive call in a CONST if no instructions
28408 2023-03-02 Richard Sandiford <richard.sandiford@arm.com>
28410 PR tree-optimization/108430
28411 * tree-vect-stmts.cc (vectorizable_condition): Fix handling
28412 of inverted condition.
28414 2023-03-02 Jakub Jelinek <jakub@redhat.com>
28417 * fold-const.cc (native_interpret_expr) <case REAL_CST>: Before memcmp
28418 comparison copy the bytes from ptr to a temporary buffer and clearing
28419 padding bits in there.
28421 2023-03-01 Tobias Burnus <tobias@codesourcery.com>
28423 PR middle-end/108545
28424 * gimplify.cc (struct tree_operand_hash_no_se): New.
28425 (omp_index_mapping_groups_1, omp_index_mapping_groups,
28426 omp_reindex_mapping_groups, omp_mapped_by_containing_struct,
28427 omp_tsort_mapping_groups_1, omp_tsort_mapping_groups,
28428 oacc_resolve_clause_dependencies, omp_build_struct_sibling_lists,
28429 gimplify_scan_omp_clauses): Use tree_operand_hash_no_se instead
28430 of tree_operand_hash.
28432 2023-03-01 LIU Hao <lh_mouse@126.com>
28435 * config/i386/host-mingw32.cc (mingw32_gt_pch_get_address):
28436 Remove the size limit `pch_VA_max_size`
28438 2023-03-01 Tobias Burnus <tobias@codesourcery.com>
28440 PR middle-end/108546
28441 * omp-low.cc (lower_omp_target): Remove optional handling
28442 on the receiver side, i.e. inside target (data), for
28445 2023-03-01 Jakub Jelinek <jakub@redhat.com>
28448 * cfgexpand.cc (expand_debug_expr): Handle WIDEN_{PLUS,MINUS}_EXPR
28449 and VEC_WIDEN_{PLUS,MINUS}_{HI,LO}_EXPR.
28451 2023-03-01 Richard Biener <rguenther@suse.de>
28453 PR tree-optimization/108970
28454 * tree-vect-loop-manip.cc (slpeel_can_duplicate_loop_p):
28455 Check we can copy the BBs.
28456 (slpeel_tree_duplicate_loop_to_edge_cfg): Avoid redundant
28458 (vect_do_peeling): Streamline error handling.
28460 2023-03-01 Richard Biener <rguenther@suse.de>
28462 PR tree-optimization/108950
28463 * tree-vect-patterns.cc (vect_recog_widen_sum_pattern):
28464 Check oprnd0 is defined in the loop.
28465 * tree-vect-loop.cc (vectorizable_reduction): Record all
28466 operands vector types, compute that of invariants and
28467 properly update their SLP nodes.
28469 2023-03-01 Kewen Lin <linkw@linux.ibm.com>
28472 * config/rs6000/rs6000.cc (rs6000_option_override_internal): Allow
28473 implicit powerpc64 setting to be unset if 64 bit is enabled implicitly.
28475 2023-02-28 Qing Zhao <qing.zhao@oracle.com>
28477 PR middle-end/107411
28478 PR middle-end/107411
28479 * gimplify.cc (gimple_add_init_for_auto_var): Use sprintf to replace
28481 * tree-ssa-uninit.cc (warn_uninit): Handle the case when the
28482 LHS varaible of a .DEFERRED_INIT call doesn't have a DECL_NAME.
28484 2023-02-28 Jakub Jelinek <jakub@redhat.com>
28486 PR sanitizer/108894
28487 * ubsan.cc (ubsan_expand_bounds_ifn): Emit index >= bound
28488 comparison rather than index > bound.
28489 * gimple-fold.cc (gimple_fold_call): Use tree_int_cst_lt
28490 rather than tree_int_cst_le for IFN_UBSAN_BOUND comparison.
28491 * doc/invoke.texi (-fsanitize=bounds): Document that whether
28492 flexible array member-like arrays are instrumented or not depends
28493 on -fstrict-flex-arrays* options of strict_flex_array attributes.
28494 (-fsanitize=bounds-strict): Document that flexible array members
28495 are not instrumented.
28497 2023-02-27 Uroš Bizjak <ubizjak@gmail.com>
28501 * config/i386/i386.md (fmodxf3): Enable for flag_finite_math_only only.
28502 (fmod<mode>3): Ditto.
28503 (fpremxf4_i387): Ditto.
28504 (reminderxf3): Ditto.
28505 (reminder<mode>3): Ditto.
28506 (fprem1xf4_i387): Ditto.
28508 2023-02-27 Roger Sayle <roger@nextmovesoftware.com>
28510 * simplify-rtx.cc (simplify_unary_operation_1) <case FFS>: Avoid
28511 generating FFS with mismatched operand and result modes, by using
28512 an explicit SIGN_EXTEND/ZERO_EXTEND.
28513 <case POPCOUNT>: Likewise, for POPCOUNT of ZERO_EXTEND.
28514 <case PARITY>: Likewise, for PARITY of {ZERO,SIGN}_EXTEND.
28516 2023-02-27 Patrick Palka <ppalka@redhat.com>
28518 * hash-table.h (gt_pch_nx(hash_table<D>)): Remove static.
28519 * lra-int.h (lra_change_class): Likewise.
28520 * recog.h (which_op_alt): Likewise.
28521 * sel-sched-ir.h (sel_bb_empty_or_nop_p): Declare inline
28524 2023-02-27 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
28526 * config/xtensa/xtensa-protos.h (xtensa_match_CLAMPS_imms_p):
28528 * config/xtensa/xtensa.cc (xtensa_match_CLAMPS_imms_p):
28530 * config/xtensa/xtensa.h (TARGET_CLAMPS): New macro definition.
28531 * config/xtensa/xtensa.md (*xtensa_clamps): New insn pattern.
28533 2023-02-27 Max Filippov <jcmvbkbc@gmail.com>
28535 * config/xtensa/xtensa-dynconfig.cc (xtensa_get_config_v2)
28536 (xtensa_get_config_v3): New functions.
28538 2023-02-27 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
28540 * config/aarch64/aarch64-simd.md (aarch64_abs<mode>): Fix typo in comment.
28542 2023-02-27 Lulu Cheng <chenglulu@loongson.cn>
28544 * config/host-linux.cc (TRY_EMPTY_VM_SPACE): Modify the value of
28545 the macro to 0x1000000000.
28547 2023-02-25 Gaius Mulley <gaiusmod2@gmail.com>
28550 * doc/gm2.texi (-fm2-pathname): New option documented.
28551 (-fm2-pathnameI): New option documented.
28552 (-fm2-prefix=): New option documented.
28553 (-fruntime-modules=): Update default module list.
28555 2023-02-25 Max Filippov <jcmvbkbc@gmail.com>
28558 * config/xtensa/xtensa-protos.h
28559 (xtensa_prepare_expand_call): Rename to xtensa_expand_call.
28560 * config/xtensa/xtensa.cc (xtensa_prepare_expand_call): Rename
28561 to xtensa_expand_call.
28562 (xtensa_expand_call): Emit the call and add a clobber expression
28563 for the static chain to it in case of windowed ABI.
28564 * config/xtensa/xtensa.md (call, call_value, sibcall)
28565 (sibcall_value): Call xtensa_expand_call and complete expansion
28566 right after that call.
28568 2023-02-24 Richard Biener <rguenther@suse.de>
28570 * vec.h (vec<T, A, vl_embed>::m_vecdata): Remove.
28571 (vec<T, A, vl_embed>::m_vecpfx): Align as T to avoid
28572 changing alignment of vec<T, A, vl_embed> and simplifying
28574 (vec<T, A, vl_embed>::address): Compute as this + 1.
28575 (vec<T, A, vl_embed>::embedded_size): Use sizeof the
28576 vector instead of the offset of the m_vecdata member.
28577 (auto_vec<T, N>::m_data): Turn storage into
28578 uninitialized unsigned char.
28579 (auto_vec<T, N>::auto_vec): Allow allocation of one
28580 stack member. Initialize m_vec in a special way to
28581 avoid later stringop overflow diagnostics.
28582 * vec.cc (test_auto_alias): New.
28583 (vec_cc_tests): Call it.
28585 2023-02-24 Richard Biener <rguenther@suse.de>
28587 * vec.h (vec<T, A, vl_embed>::lower_bound): Adjust to
28588 take a const reference to the object, use address to
28590 (vec<T, A, vl_embed>::contains): Use address to access data.
28591 (vec<T, A, vl_embed>::operator[]): Use address instead of
28592 m_vecdata to access data.
28593 (vec<T, A, vl_embed>::iterate): Likewise.
28594 (vec<T, A, vl_embed>::copy): Likewise.
28595 (vec<T, A, vl_embed>::quick_push): Likewise.
28596 (vec<T, A, vl_embed>::pop): Likewise.
28597 (vec<T, A, vl_embed>::quick_insert): Likewise.
28598 (vec<T, A, vl_embed>::ordered_remove): Likewise.
28599 (vec<T, A, vl_embed>::unordered_remove): Likewise.
28600 (vec<T, A, vl_embed>::block_remove): Likewise.
28601 (vec<T, A, vl_heap>::address): Likewise.
28603 2023-02-24 Martin Liska <mliska@suse.cz>
28605 PR sanitizer/108834
28606 * asan.cc (asan_add_global): Use proper TU name for normal
28607 global variables (and aux_base_name for the artificial one).
28609 2023-02-24 Jakub Jelinek <jakub@redhat.com>
28611 * config/i386/i386-builtin.def: Update description of BDESC
28612 and BDESC_FIRST in file comment to include mask2.
28614 2023-02-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
28616 * config/aarch64/aarch64-cores.def (FLAGS): Update comment.
28618 2023-02-24 Jakub Jelinek <jakub@redhat.com>
28620 PR middle-end/108854
28621 * cgraphclones.cc (duplicate_thunk_for_node): If no parameter
28622 changes are needed, copy at least DECL_ARGUMENTS PARM_DECL
28623 nodes and adjust their DECL_CONTEXT.
28625 2023-02-24 Jakub Jelinek <jakub@redhat.com>
28628 * config/i386/i386-builtin.def (__builtin_ia32_cvtne2ps2bf16_v16bf,
28629 __builtin_ia32_cvtne2ps2bf16_v16bf_mask,
28630 __builtin_ia32_cvtne2ps2bf16_v16bf_maskz,
28631 __builtin_ia32_cvtne2ps2bf16_v8bf,
28632 __builtin_ia32_cvtne2ps2bf16_v8bf_mask,
28633 __builtin_ia32_cvtne2ps2bf16_v8bf_maskz,
28634 __builtin_ia32_cvtneps2bf16_v8sf_mask,
28635 __builtin_ia32_cvtneps2bf16_v8sf_maskz,
28636 __builtin_ia32_cvtneps2bf16_v4sf_mask,
28637 __builtin_ia32_cvtneps2bf16_v4sf_maskz,
28638 __builtin_ia32_dpbf16ps_v8sf, __builtin_ia32_dpbf16ps_v8sf_mask,
28639 __builtin_ia32_dpbf16ps_v8sf_maskz, __builtin_ia32_dpbf16ps_v4sf,
28640 __builtin_ia32_dpbf16ps_v4sf_mask,
28641 __builtin_ia32_dpbf16ps_v4sf_maskz): Require also
28642 OPTION_MASK_ISA_AVX512VL.
28644 2023-02-24 Sebastian Huber <sebastian.huber@embedded-brains.de>
28646 * config/riscv/t-rtems: Keep only -mcmodel=medany 64-bit multilibs.
28647 Add non-compact 32-bit multilibs.
28649 2023-02-24 Junxian Zhu <zhujunxian@oss.cipunited.com>
28651 * config/mips/mips.md (*clo<mode>2): New pattern.
28653 2023-02-24 Prachi Godbole <prachi.godbole@imgtec.com>
28655 * config/mips/mips.h (machine_function): New variable
28656 use_hazard_barrier_return_p.
28657 * config/mips/mips.md (UNSPEC_JRHB): New unspec.
28658 (mips_hb_return_internal): New insn pattern.
28659 * config/mips/mips.cc (mips_attribute_table): Add attribute
28660 use_hazard_barrier_return.
28661 (mips_use_hazard_barrier_return_p): New static function.
28662 (mips_function_attr_inlinable_p): Likewise.
28663 (mips_compute_frame_info): Set use_hazard_barrier_return_p.
28664 Emit error for unsupported architecture choice.
28665 (mips_function_ok_for_sibcall, mips_can_use_return_insn):
28666 Return false for use_hazard_barrier_return.
28667 (mips_expand_epilogue): Emit hazard barrier return.
28668 * doc/extend.texi: Document use_hazard_barrier_return.
28670 2023-02-23 Max Filippov <jcmvbkbc@gmail.com>
28672 * config/xtensa/xtensa-dynconfig.cc (config.h, system.h)
28673 (coretypes.h, diagnostic.h, intl.h): Use "..." instead of <...>
28674 for the gcc-internal headers.
28676 2023-02-23 Max Filippov <jcmvbkbc@gmail.com>
28678 * config/xtensa/t-xtensa (xtensa-dynconfig.o): Use $(COMPILE)
28679 and $(POSTCOMPILE) instead of manual dependency listing.
28680 * config/xtensa/xtensa-dynconfig.c: Rename to ...
28681 * config/xtensa/xtensa-dynconfig.cc: ... this.
28683 2023-02-23 Arsen Arsenović <arsen@aarsen.me>
28685 * doc/cfg.texi: Reorder index entries around @items.
28686 * doc/cpp.texi: Ditto.
28687 * doc/cppenv.texi: Ditto.
28688 * doc/cppopts.texi: Ditto.
28689 * doc/generic.texi: Ditto.
28690 * doc/install.texi: Ditto.
28691 * doc/extend.texi: Ditto.
28692 * doc/invoke.texi: Ditto.
28693 * doc/md.texi: Ditto.
28694 * doc/rtl.texi: Ditto.
28695 * doc/tm.texi.in: Ditto.
28696 * doc/trouble.texi: Ditto.
28697 * doc/tm.texi: Regenerate.
28699 2023-02-23 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
28701 * config/xtensa/xtensa.md: New peephole2 pattern that eliminates
28702 the occurrence of general-purpose register used only once and for
28703 transferring intermediate value.
28705 2023-02-23 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
28707 * config/xtensa/xtensa.cc (machine_function): Add new member
28708 'eliminated_callee_saved_bmp'.
28709 (xtensa_can_eliminate_callee_saved_reg_p): New function to
28710 determine whether the register can be eliminated or not.
28711 (xtensa_expand_prologue): Add invoking the above function and
28712 elimination the use of callee-saved register by using its stack
28713 slot through the stack pointer (or the frame pointer if needed)
28715 (xtensa_expand_prologue): Modify to not emit register restoration
28716 insn from its stack slot if the register is already eliminated.
28718 2023-02-23 Jakub Jelinek <jakub@redhat.com>
28720 PR translation/108890
28721 * config/xtensa/xtensa-dynconfig.c (xtensa_load_config): Drop _()s
28722 around fatal_error format strings.
28724 2023-02-23 Richard Biener <rguenther@suse.de>
28726 * tree-ssa-structalias.cc (handle_lhs_call): Do not
28727 re-create rhsc, only truncate it.
28729 2023-02-23 Jakub Jelinek <jakub@redhat.com>
28731 PR middle-end/106258
28732 * ipa-prop.cc (try_make_edge_direct_virtual_call): Handle
28733 BUILT_IN_UNREACHABLE_TRAP like BUILT_IN_UNREACHABLE.
28735 2023-02-23 Richard Biener <rguenther@suse.de>
28737 * tree-if-conv.cc (tree_if_conversion): Properly manage
28738 memory of refs and the contained data references.
28740 2023-02-23 Richard Biener <rguenther@suse.de>
28742 PR tree-optimization/108888
28743 * tree-if-conv.cc (if_convertible_stmt_p): Set PLF_2 on
28744 calls to predicate.
28745 (predicate_statements): Only predicate calls with PLF_2.
28747 2023-02-23 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
28749 * config/xtensa/xtensa.md
28750 (zero_cost_loop_start, zero_cost_loop_end, loop_end):
28751 Add missing "SI:" to PLUS RTXes.
28753 2023-02-23 Max Filippov <jcmvbkbc@gmail.com>
28756 * config/xtensa/xtensa.cc (xtensa_expand_epilogue):
28757 Emit (use (reg:SI A0_REG)) at the end in the sibling call
28758 (i.e. the same place as (return) in the normal call).
28760 2023-02-23 Max Filippov <jcmvbkbc@gmail.com>
28763 2023-02-21 Max Filippov <jcmvbkbc@gmail.com>
28766 * config/xtensa/xtensa.cc (xtensa_expand_epilogue): Drop emit_use
28768 * config/xtensa/xtensa.md (sibcall, sibcall_internal)
28769 (sibcall_value, sibcall_value_internal): Add 'use' expression
28772 2023-02-23 Arsen Arsenović <arsen@aarsen.me>
28774 * doc/cppdiropts.texi: Reorder @opindex commands to precede
28775 @items they relate to.
28776 * doc/cppopts.texi: Ditto.
28777 * doc/cppwarnopts.texi: Ditto.
28778 * doc/invoke.texi: Ditto.
28779 * doc/lto.texi: Ditto.
28781 2023-02-22 Andrew Stubbs <ams@codesourcery.com>
28783 * internal-fn.cc (expand_MASK_CALL): New.
28784 * internal-fn.def (MASK_CALL): New.
28785 * internal-fn.h (expand_MASK_CALL): New prototype.
28786 * omp-simd-clone.cc (simd_clone_adjust_argument_types): Set vector_type
28787 for mask arguments also.
28788 * tree-if-conv.cc: Include cgraph.h.
28789 (if_convertible_stmt_p): Do if conversions for calls to SIMD calls.
28790 (predicate_statements): Convert functions to IFN_MASK_CALL.
28791 * tree-vect-loop.cc (vect_get_datarefs_in_loop): Recognise
28792 IFN_MASK_CALL as a SIMD function call.
28793 * tree-vect-stmts.cc (vectorizable_simd_clone_call): Handle
28794 IFN_MASK_CALL as an inbranch SIMD function call.
28795 Generate the mask vector arguments.
28797 2023-02-22 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
28799 * config/riscv/riscv-vector-builtins-bases.cc (class reducop): New class.
28800 (class widen_reducop): Ditto.
28801 (class freducop): Ditto.
28802 (class widen_freducop): Ditto.
28804 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
28805 * config/riscv/riscv-vector-builtins-functions.def (vredsum): Add reduction support.
28814 (vwredsumu): Ditto.
28815 (vfredusum): Ditto.
28816 (vfredosum): Ditto.
28819 (vfwredosum): Ditto.
28820 (vfwredusum): Ditto.
28821 * config/riscv/riscv-vector-builtins-shapes.cc (struct reduc_alu_def): Ditto.
28823 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
28824 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_WI_OPS): New macro.
28825 (DEF_RVV_WU_OPS): Ditto.
28826 (DEF_RVV_WF_OPS): Ditto.
28827 (vint8mf8_t): Ditto.
28828 (vint8mf4_t): Ditto.
28829 (vint8mf2_t): Ditto.
28830 (vint8m1_t): Ditto.
28831 (vint8m2_t): Ditto.
28832 (vint8m4_t): Ditto.
28833 (vint8m8_t): Ditto.
28834 (vint16mf4_t): Ditto.
28835 (vint16mf2_t): Ditto.
28836 (vint16m1_t): Ditto.
28837 (vint16m2_t): Ditto.
28838 (vint16m4_t): Ditto.
28839 (vint16m8_t): Ditto.
28840 (vint32mf2_t): Ditto.
28841 (vint32m1_t): Ditto.
28842 (vint32m2_t): Ditto.
28843 (vint32m4_t): Ditto.
28844 (vint32m8_t): Ditto.
28845 (vuint8mf8_t): Ditto.
28846 (vuint8mf4_t): Ditto.
28847 (vuint8mf2_t): Ditto.
28848 (vuint8m1_t): Ditto.
28849 (vuint8m2_t): Ditto.
28850 (vuint8m4_t): Ditto.
28851 (vuint8m8_t): Ditto.
28852 (vuint16mf4_t): Ditto.
28853 (vuint16mf2_t): Ditto.
28854 (vuint16m1_t): Ditto.
28855 (vuint16m2_t): Ditto.
28856 (vuint16m4_t): Ditto.
28857 (vuint16m8_t): Ditto.
28858 (vuint32mf2_t): Ditto.
28859 (vuint32m1_t): Ditto.
28860 (vuint32m2_t): Ditto.
28861 (vuint32m4_t): Ditto.
28862 (vuint32m8_t): Ditto.
28863 (vfloat32mf2_t): Ditto.
28864 (vfloat32m1_t): Ditto.
28865 (vfloat32m2_t): Ditto.
28866 (vfloat32m4_t): Ditto.
28867 (vfloat32m8_t): Ditto.
28868 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_WI_OPS): Ditto.
28869 (DEF_RVV_WU_OPS): Ditto.
28870 (DEF_RVV_WF_OPS): Ditto.
28871 (required_extensions_p): Add reduction support.
28872 (rvv_arg_type_info::get_base_vector_type): Ditto.
28873 (rvv_arg_type_info::get_tree_type): Ditto.
28874 * config/riscv/riscv-vector-builtins.h (enum rvv_base_type): Ditto.
28875 * config/riscv/riscv.md: Ditto.
28876 * config/riscv/vector-iterators.md (minu): Ditto.
28877 * config/riscv/vector.md (@pred_reduc_<reduc><mode><vlmul1>): New patern.
28878 (@pred_reduc_<reduc><mode><vlmul1_zve32>): Ditto.
28879 (@pred_widen_reduc_plus<v_su><mode><vwlmul1>): Ditto.
28880 (@pred_widen_reduc_plus<v_su><mode><vwlmul1_zve32>):Ditto.
28881 (@pred_reduc_plus<order><mode><vlmul1>): Ditto.
28882 (@pred_reduc_plus<order><mode><vlmul1_zve32>): Ditto.
28883 (@pred_widen_reduc_plus<order><mode><vwlmul1>): Ditto.
28885 2023-02-22 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
28887 * config/riscv/iterators.md: New iterator.
28888 * config/riscv/riscv-vector-builtins-bases.cc (class widen_binop): New class.
28889 (enum ternop_type): New enum.
28890 (class vmacc): New class.
28891 (class imac): Ditto.
28892 (class vnmsac): Ditto.
28893 (enum widen_ternop_type): New enum.
28894 (class vmadd): Ditto.
28895 (class vnmsub): Ditto.
28896 (class iwmac): Ditto.
28897 (class vwmacc): Ditto.
28898 (class vwmaccu): Ditto.
28899 (class vwmaccsu): Ditto.
28900 (class vwmaccus): Ditto.
28901 (class reverse_binop): Ditto.
28902 (class vfmacc): Ditto.
28903 (class vfnmsac): Ditto.
28904 (class vfmadd): Ditto.
28905 (class vfnmsub): Ditto.
28906 (class vfnmacc): Ditto.
28907 (class vfmsac): Ditto.
28908 (class vfnmadd): Ditto.
28909 (class vfmsub): Ditto.
28910 (class vfwmacc): Ditto.
28911 (class vfwnmacc): Ditto.
28912 (class vfwmsac): Ditto.
28913 (class vfwnmsac): Ditto.
28914 (class float_misc): Ditto.
28915 (class fcmp): Ditto.
28916 (class vfclass): Ditto.
28917 (class vfcvt_x): Ditto.
28918 (class vfcvt_rtz_x): Ditto.
28919 (class vfcvt_f): Ditto.
28920 (class vfwcvt_x): Ditto.
28921 (class vfwcvt_rtz_x): Ditto.
28922 (class vfwcvt_f): Ditto.
28923 (class vfncvt_x): Ditto.
28924 (class vfncvt_rtz_x): Ditto.
28925 (class vfncvt_f): Ditto.
28926 (class vfncvt_rod_f): Ditto.
28928 * config/riscv/riscv-vector-builtins-bases.h:
28929 * config/riscv/riscv-vector-builtins-functions.def (vzext): Ditto.
28973 (vfcvt_rtz_x): Ditto.
28974 (vfcvt_rtz_xu): Ditto.
28977 (vfwcvt_xu): Ditto.
28978 (vfwcvt_rtz_x): Ditto.
28979 (vfwcvt_rtz_xu): Ditto.
28982 (vfncvt_xu): Ditto.
28983 (vfncvt_rtz_x): Ditto.
28984 (vfncvt_rtz_xu): Ditto.
28986 (vfncvt_rod_f): Ditto.
28987 * config/riscv/riscv-vector-builtins-shapes.cc (struct alu_def): Ditto.
28988 (struct move_def): Ditto.
28989 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_WEXTF_OPS): New macro.
28990 (DEF_RVV_CONVERT_I_OPS): Ditto.
28991 (DEF_RVV_CONVERT_U_OPS): Ditto.
28992 (DEF_RVV_WCONVERT_I_OPS): Ditto.
28993 (DEF_RVV_WCONVERT_U_OPS): Ditto.
28994 (DEF_RVV_WCONVERT_F_OPS): Ditto.
28995 (vfloat64m1_t): Ditto.
28996 (vfloat64m2_t): Ditto.
28997 (vfloat64m4_t): Ditto.
28998 (vfloat64m8_t): Ditto.
28999 (vint32mf2_t): Ditto.
29000 (vint32m1_t): Ditto.
29001 (vint32m2_t): Ditto.
29002 (vint32m4_t): Ditto.
29003 (vint32m8_t): Ditto.
29004 (vint64m1_t): Ditto.
29005 (vint64m2_t): Ditto.
29006 (vint64m4_t): Ditto.
29007 (vint64m8_t): Ditto.
29008 (vuint32mf2_t): Ditto.
29009 (vuint32m1_t): Ditto.
29010 (vuint32m2_t): Ditto.
29011 (vuint32m4_t): Ditto.
29012 (vuint32m8_t): Ditto.
29013 (vuint64m1_t): Ditto.
29014 (vuint64m2_t): Ditto.
29015 (vuint64m4_t): Ditto.
29016 (vuint64m8_t): Ditto.
29017 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_CONVERT_I_OPS): Ditto.
29018 (DEF_RVV_CONVERT_U_OPS): Ditto.
29019 (DEF_RVV_WCONVERT_I_OPS): Ditto.
29020 (DEF_RVV_WCONVERT_U_OPS): Ditto.
29021 (DEF_RVV_WCONVERT_F_OPS): Ditto.
29022 (DEF_RVV_F_OPS): Ditto.
29023 (DEF_RVV_WEXTF_OPS): Ditto.
29024 (required_extensions_p): Adjust for floating-point support.
29025 (check_required_extensions): Ditto.
29026 (unsigned_base_type_p): Ditto.
29027 (get_mode_for_bitsize): Ditto.
29028 (rvv_arg_type_info::get_base_vector_type): Ditto.
29029 (rvv_arg_type_info::get_tree_type): Ditto.
29030 * config/riscv/riscv-vector-builtins.def (v_f): New define.
29033 (xu_v): New define.
29035 (xu_w): New define.
29036 * config/riscv/riscv-vector-builtins.h (enum rvv_base_type): New enum.
29037 (function_expander::arg_mode): New function.
29038 * config/riscv/vector-iterators.md (sof): New iterator.
29044 (fixuns_trunc): Ditto.
29046 * config/riscv/vector.md (@pred_broadcast<mode>): New pattern.
29047 (@pred_<optab><mode>): Ditto.
29048 (@pred_<optab><mode>_scalar): Ditto.
29049 (@pred_<optab><mode>_reverse_scalar): Ditto.
29050 (@pred_<copysign><mode>): Ditto.
29051 (@pred_<copysign><mode>_scalar): Ditto.
29052 (@pred_mul_<optab><mode>): Ditto.
29053 (pred_mul_<optab><mode>_undef_merge): Ditto.
29054 (*pred_<madd_nmsub><mode>): Ditto.
29055 (*pred_<macc_nmsac><mode>): Ditto.
29056 (*pred_mul_<optab><mode>): Ditto.
29057 (@pred_mul_<optab><mode>_scalar): Ditto.
29058 (*pred_mul_<optab><mode>_undef_merge_scalar): Ditto.
29059 (*pred_<madd_nmsub><mode>_scalar): Ditto.
29060 (*pred_<macc_nmsac><mode>_scalar): Ditto.
29061 (*pred_mul_<optab><mode>_scalar): Ditto.
29062 (@pred_neg_mul_<optab><mode>): Ditto.
29063 (pred_neg_mul_<optab><mode>_undef_merge): Ditto.
29064 (*pred_<nmadd_msub><mode>): Ditto.
29065 (*pred_<nmacc_msac><mode>): Ditto.
29066 (*pred_neg_mul_<optab><mode>): Ditto.
29067 (@pred_neg_mul_<optab><mode>_scalar): Ditto.
29068 (*pred_neg_mul_<optab><mode>_undef_merge_scalar): Ditto.
29069 (*pred_<nmadd_msub><mode>_scalar): Ditto.
29070 (*pred_<nmacc_msac><mode>_scalar): Ditto.
29071 (*pred_neg_mul_<optab><mode>_scalar): Ditto.
29072 (@pred_<misc_op><mode>): Ditto.
29073 (@pred_class<mode>): Ditto.
29074 (@pred_dual_widen_<optab><mode>): Ditto.
29075 (@pred_dual_widen_<optab><mode>_scalar): Ditto.
29076 (@pred_single_widen_<plus_minus:optab><mode>): Ditto.
29077 (@pred_single_widen_<plus_minus:optab><mode>_scalar): Ditto.
29078 (@pred_widen_mul_<optab><mode>): Ditto.
29079 (@pred_widen_mul_<optab><mode>_scalar): Ditto.
29080 (@pred_widen_neg_mul_<optab><mode>): Ditto.
29081 (@pred_widen_neg_mul_<optab><mode>_scalar): Ditto.
29082 (@pred_cmp<mode>): Ditto.
29083 (*pred_cmp<mode>): Ditto.
29084 (*pred_cmp<mode>_narrow): Ditto.
29085 (@pred_cmp<mode>_scalar): Ditto.
29086 (*pred_cmp<mode>_scalar): Ditto.
29087 (*pred_cmp<mode>_scalar_narrow): Ditto.
29088 (@pred_eqne<mode>_scalar): Ditto.
29089 (*pred_eqne<mode>_scalar): Ditto.
29090 (*pred_eqne<mode>_scalar_narrow): Ditto.
29091 (@pred_merge<mode>_scalar): Ditto.
29092 (@pred_fcvt_x<v_su>_f<mode>): Ditto.
29093 (@pred_<fix_cvt><mode>): Ditto.
29094 (@pred_<float_cvt><mode>): Ditto.
29095 (@pred_widen_fcvt_x<v_su>_f<mode>): Ditto.
29096 (@pred_widen_<fix_cvt><mode>): Ditto.
29097 (@pred_widen_<float_cvt><mode>): Ditto.
29098 (@pred_extend<mode>): Ditto.
29099 (@pred_narrow_fcvt_x<v_su>_f<mode>): Ditto.
29100 (@pred_narrow_<fix_cvt><mode>): Ditto.
29101 (@pred_narrow_<float_cvt><mode>): Ditto.
29102 (@pred_trunc<mode>): Ditto.
29103 (@pred_rod_trunc<mode>): Ditto.
29105 2023-02-22 Jakub Jelinek <jakub@redhat.com>
29107 PR middle-end/106258
29108 * cgraph.cc (cgraph_edge::redirect_call_stmt_to_callee,
29109 cgraph_update_edges_for_call_stmt_node, cgraph_node::verify_node):
29110 Handle BUILT_IN_UNREACHABLE_TRAP like BUILT_IN_UNREACHABLE.
29111 * cgraphclones.cc (cgraph_node::create_clone): Likewise.
29113 2023-02-22 Thomas Schwinge <thomas@codesourcery.com>
29115 * common.opt (-Wcomplain-wrong-lang): New.
29116 * doc/invoke.texi (-Wno-complain-wrong-lang): Document it.
29117 * opts-common.cc (prune_options): Handle it.
29118 * opts-global.cc (complain_wrong_lang): Use it.
29120 2023-02-21 David Malcolm <dmalcolm@redhat.com>
29123 * doc/invoke.texi: Document -fno-analyzer-suppress-followups.
29125 2023-02-21 Max Filippov <jcmvbkbc@gmail.com>
29128 * config/xtensa/xtensa.cc (xtensa_expand_epilogue): Drop emit_use
29130 * config/xtensa/xtensa.md (sibcall, sibcall_internal)
29131 (sibcall_value, sibcall_value_internal): Add 'use' expression
29134 2023-02-21 Richard Biener <rguenther@suse.de>
29136 PR tree-optimization/108691
29137 * tree-ssa-dce.cc (eliminate_unnecessary_stmts): Remove
29138 assert about calls_setjmp not becoming true when it was false.
29140 2023-02-21 Richard Biener <rguenther@suse.de>
29142 PR tree-optimization/108793
29143 * tree-ssa-loop-niter.cc (number_of_iterations_until_wrap):
29144 Use convert operands to niter_type when computing num.
29146 2023-02-21 Richard Biener <rguenther@suse.de>
29149 2023-02-13 Richard Biener <rguenther@suse.de>
29151 PR tree-optimization/108691
29152 * tree-cfg.cc (notice_special_calls): When the CFG is built
29153 honor gimple_call_ctrl_altering_p.
29154 * cfgexpand.cc (expand_call_stmt): Clear cfun->calls_setjmp
29155 temporarily if the call is not control-altering.
29156 * calls.cc (emit_call_1): Do not add REG_SETJMP if
29157 cfun->calls_setjmp is not set. Do not alter cfun->calls_setjmp.
29159 2023-02-21 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
29161 * config/xtensa/xtensa.cc (xtensa_call_save_reg): Change to return
29162 true if register A0 (return address register) when -Og is specified.
29164 2023-02-20 Uroš Bizjak <ubizjak@gmail.com>
29166 * config/i386/predicates.md
29167 (general_x64constmem_operand): New predicate.
29168 * config/i386/i386.md (*cmpqi_ext<mode>_1):
29169 Use nonimm_x64constmem_operand.
29170 (*cmpqi_ext<mode>_3): Use general_x64constmem_operand.
29171 (*addqi_ext<mode>_1): Ditto.
29172 (*testqi_ext<mode>_1): Ditto.
29173 (*andqi_ext<mode>_1): Ditto.
29174 (*andqi_ext<mode>_1_cc): Ditto.
29175 (*<any_or:code>qi_ext<mode>_1): Ditto.
29176 (*xorqi_ext<mode>_1_cc): Ditto.
29178 2023-02-20 Jakub Jelinek <jakub2redhat.com>
29181 * config/rs6000/rs6000.md (umaddditi4): Swap gen_maddlddi4 with
29182 gen_umadddi4_highpart{,_le}.
29184 2023-02-20 Kito Cheng <kito.cheng@sifive.com>
29186 * config/riscv/riscv.md (prefetch): Use r instead of p for the
29188 (riscv_prefetchi_<mode>): Ditto.
29190 2023-02-20 Richard Biener <rguenther@suse.de>
29192 PR tree-optimization/108816
29193 * tree-vect-loop-manip.cc (vect_loop_versioning): Adjust
29194 versioning condition split prerequesite, assert required
29197 2023-02-20 Richard Biener <rguenther@suse.de>
29199 PR tree-optimization/108825
29200 * tree-ssa-loop-manip.cc (verify_loop_closed_ssa): For
29201 loop-local verfication only verify there's no pending SSA
29204 2023-02-20 Richard Biener <rguenther@suse.de>
29206 PR tree-optimization/108819
29207 * tree-ssa-loop-niter.cc (number_of_iterations_cltz): Check
29208 we have an SSA name as iv_2 as expected.
29210 2023-02-18 Jakub Jelinek <jakub@redhat.com>
29212 PR tree-optimization/108819
29213 * tree-ssa-reassoc.cc (update_ops): Fold new stmt in place.
29215 2023-02-18 Jakub Jelinek <jakub@redhat.com>
29218 * config/i386/i386-protos.h (ix86_replace_reg_with_reg): Declare.
29219 * config/i386/i386-expand.cc (ix86_replace_reg_with_reg): New
29221 * config/i386/i386.md: Replace replace_rtx calls in all peephole2s
29222 with ix86_replace_reg_with_reg.
29224 2023-02-18 Gerald Pfeifer <gerald@pfeifer.com>
29226 * doc/invoke.texi (AVR Options): Update link to AVR-LibC.
29228 2023-02-18 Xi Ruoyao <xry111@xry111.site>
29230 * config.gcc (triplet_abi): Set its value based on $with_abi,
29231 instead of $target.
29232 (la_canonical_triplet): Set it after $triplet_abi is set
29234 * config/loongarch/t-linux (MULTILIB_OSDIRNAMES): Make the
29235 multiarch tuple for lp64d "loongarch64-linux-gnu" (without
29238 2023-02-18 Andrew Pinski <apinski@marvell.com>
29240 * match.pd: Remove #if GIMPLE around the
29243 2023-02-18 Andrew Pinski <apinski@marvell.com>
29245 * value-query.h (get_range_query): Return the global ranges
29246 for a nullptr func.
29248 2023-02-17 Siddhesh Poyarekar <siddhesh@gotplt.org>
29250 * doc/invoke.texi (@item -Wall): Fix typo in
29253 2023-02-17 Uroš Bizjak <ubizjak@gmail.com>
29256 * config/i386/predicates.md
29257 (nonimm_x64constmem_operand): New predicate.
29258 * config/i386/i386.md (*addqi_ext<mode>_0): New insn pattern.
29259 (*subqi_ext<mode>_0): Ditto.
29260 (*andqi_ext<mode>_0): Ditto.
29261 (*<any_or:code>qi_ext<mode>_0): Ditto.
29263 2023-02-17 Uroš Bizjak <ubizjak@gmail.com>
29266 * simplify-rtx.cc (simplify_context::simplify_subreg): Use
29267 int_outermode instead of GET_MODE (tem) to prevent
29268 VOIDmode from entering simplify_gen_subreg.
29270 2023-02-17 Richard Biener <rguenther@suse.de>
29272 PR tree-optimization/108821
29273 * tree-ssa-loop-im.cc (sm_seq_valid_bb): We can also not
29274 move volatile accesses.
29276 2023-02-17 Richard Biener <rguenther@suse.de>
29278 * tree-ssa.cc (ssa_undefined_value_p): Assert we are not
29279 called on virtual operands.
29280 * tree-ssa-sccvn.cc (vn_phi_lookup): Guard
29281 ssa_undefined_value_p calls.
29282 (vn_phi_insert): Likewise.
29283 (set_ssa_val_to): Likewise.
29284 (visit_phi): Avoid extra work with equivalences for
29285 virtual operand PHIs.
29287 2023-02-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29289 * config/riscv/riscv-vector-builtins-bases.cc (class mask_logic): New
29291 (class mask_nlogic): Ditto.
29292 (class mask_notlogic): Ditto.
29293 (class vmmv): Ditto.
29294 (class vmclr): Ditto.
29295 (class vmset): Ditto.
29296 (class vmnot): Ditto.
29297 (class vcpop): Ditto.
29298 (class vfirst): Ditto.
29299 (class mask_misc): Ditto.
29300 (class viota): Ditto.
29301 (class vid): Ditto.
29303 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
29304 * config/riscv/riscv-vector-builtins-functions.def (vmand): Ditto.
29323 * config/riscv/riscv-vector-builtins-shapes.cc (struct alu_def): Ditto.
29324 (struct mask_alu_def): Ditto.
29326 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
29327 * config/riscv/riscv-vector-builtins.cc: Ditto.
29328 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::cleanup_insns): Fix bug
29329 for dest it scalar RVV intrinsics.
29330 * config/riscv/vector-iterators.md (sof): New iterator.
29331 * config/riscv/vector.md (@pred_<optab>n<mode>): New pattern.
29332 (@pred_<optab>not<mode>): New pattern.
29333 (@pred_popcount<VB:mode><P:mode>): New pattern.
29334 (@pred_ffs<VB:mode><P:mode>): New pattern.
29335 (@pred_<misc_op><mode>): New pattern.
29336 (@pred_iota<mode>): New pattern.
29337 (@pred_series<mode>): New pattern.
29339 2023-02-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29341 * config/riscv/riscv-vector-builtins-functions.def (vadc): Rename.
29345 * config/riscv/riscv-vector-builtins.cc: Ditto.
29347 2023-02-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29348 kito-cheng <kito.cheng@sifive.com>
29350 * config/riscv/riscv-protos.h (sew64_scalar_helper): New function.
29351 * config/riscv/riscv-v.cc (has_vi_variant_p): Adjust.
29352 (sew64_scalar_helper): New function.
29353 * config/riscv/vector.md: Normalization.
29355 2023-02-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29357 * config/riscv/riscv-vector-builtins-functions.def (vsetvlmax): Rearrange.
29419 2023-02-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29421 * config/riscv/vector.md (@pred_<optab><mode>): Rearrange.
29422 (@pred_<optab><mode>_scalar): Ditto.
29423 (*pred_<optab><mode>_scalar): Ditto.
29424 (*pred_<optab><mode>_extended_scalar): Ditto.
29426 2023-02-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29428 * config/riscv/riscv-protos.h (riscv_run_selftests): Remove 'extern'.
29429 (init_builtins): Ditto.
29430 (mangle_builtin_type): Ditto.
29431 (verify_type_context): Ditto.
29432 (handle_pragma_vector): Ditto.
29433 (builtin_decl): Ditto.
29434 (expand_builtin): Ditto.
29435 (const_vec_all_same_in_range_p): Ditto.
29436 (legitimize_move): Ditto.
29437 (emit_vlmax_op): Ditto.
29438 (emit_nonvlmax_op): Ditto.
29439 (get_vlmul): Ditto.
29440 (get_ratio): Ditto.
29443 (get_avl_type): Ditto.
29444 (calculate_ratio): Ditto.
29445 (enum vlmul_type): Ditto.
29447 (neg_simm5_p): Ditto.
29448 (has_vi_variant_p): Ditto.
29450 2023-02-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29452 * config/riscv/riscv-protos.h (simm32_p): Remove.
29453 * config/riscv/riscv-v.cc (simm32_p): Ditto.
29454 * config/riscv/vector.md: Use immediate_operand
29455 instead of riscv_vector::simm32_p.
29457 2023-02-16 Gerald Pfeifer <gerald@pfeifer.com>
29459 * doc/invoke.texi (Optimize Options): Reword the explanation
29460 getting minimal, maximal and default values of a parameter.
29462 2023-02-16 Patrick Palka <ppalka@redhat.com>
29464 * addresses.h: Mechanically drop 'static' from 'static inline'
29465 functions via s/^static inline/inline/g.
29466 * asan.h: Likewise.
29467 * attribs.h: Likewise.
29468 * basic-block.h: Likewise.
29469 * bitmap.h: Likewise.
29470 * cfghooks.h: Likewise.
29471 * cfgloop.h: Likewise.
29472 * cgraph.h: Likewise.
29473 * cselib.h: Likewise.
29474 * data-streamer.h: Likewise.
29475 * debug.h: Likewise.
29477 * diagnostic.h: Likewise.
29478 * dominance.h: Likewise.
29479 * dumpfile.h: Likewise.
29480 * emit-rtl.h: Likewise.
29481 * except.h: Likewise.
29482 * expmed.h: Likewise.
29483 * expr.h: Likewise.
29484 * fixed-value.h: Likewise.
29485 * gengtype.h: Likewise.
29486 * gimple-expr.h: Likewise.
29487 * gimple-iterator.h: Likewise.
29488 * gimple-predict.h: Likewise.
29489 * gimple-range-fold.h: Likewise.
29490 * gimple-ssa.h: Likewise.
29491 * gimple.h: Likewise.
29492 * graphite.h: Likewise.
29493 * hard-reg-set.h: Likewise.
29494 * hash-map.h: Likewise.
29495 * hash-set.h: Likewise.
29496 * hash-table.h: Likewise.
29497 * hwint.h: Likewise.
29498 * input.h: Likewise.
29499 * insn-addr.h: Likewise.
29500 * internal-fn.h: Likewise.
29501 * ipa-fnsummary.h: Likewise.
29502 * ipa-icf-gimple.h: Likewise.
29503 * ipa-inline.h: Likewise.
29504 * ipa-modref.h: Likewise.
29505 * ipa-prop.h: Likewise.
29506 * ira-int.h: Likewise.
29508 * lra-int.h: Likewise.
29510 * lto-streamer.h: Likewise.
29511 * memmodel.h: Likewise.
29512 * omp-general.h: Likewise.
29513 * optabs-query.h: Likewise.
29514 * optabs.h: Likewise.
29515 * plugin.h: Likewise.
29516 * pretty-print.h: Likewise.
29517 * range.h: Likewise.
29518 * read-md.h: Likewise.
29519 * recog.h: Likewise.
29520 * regs.h: Likewise.
29521 * rtl-iter.h: Likewise.
29523 * sbitmap.h: Likewise.
29524 * sched-int.h: Likewise.
29525 * sel-sched-ir.h: Likewise.
29526 * sese.h: Likewise.
29527 * sparseset.h: Likewise.
29528 * ssa-iterators.h: Likewise.
29529 * system.h: Likewise.
29530 * target-globals.h: Likewise.
29531 * target.h: Likewise.
29532 * timevar.h: Likewise.
29533 * tree-chrec.h: Likewise.
29534 * tree-data-ref.h: Likewise.
29535 * tree-iterator.h: Likewise.
29536 * tree-outof-ssa.h: Likewise.
29537 * tree-phinodes.h: Likewise.
29538 * tree-scalar-evolution.h: Likewise.
29539 * tree-sra.h: Likewise.
29540 * tree-ssa-alias.h: Likewise.
29541 * tree-ssa-live.h: Likewise.
29542 * tree-ssa-loop-manip.h: Likewise.
29543 * tree-ssa-loop.h: Likewise.
29544 * tree-ssa-operands.h: Likewise.
29545 * tree-ssa-propagate.h: Likewise.
29546 * tree-ssa-sccvn.h: Likewise.
29547 * tree-ssa.h: Likewise.
29548 * tree-ssanames.h: Likewise.
29549 * tree-streamer.h: Likewise.
29550 * tree-switch-conversion.h: Likewise.
29551 * tree-vectorizer.h: Likewise.
29552 * tree.h: Likewise.
29553 * wide-int.h: Likewise.
29555 2023-02-16 Jakub Jelinek <jakub@redhat.com>
29557 PR tree-optimization/108657
29558 * tree-ssa-dse.cc (initialize_ao_ref_for_dse): If lhs of stmt
29559 exists and is not a SSA_NAME, call ao_ref_init even if the stmt
29560 is a call to internal or builtin function.
29562 2023-02-16 Jonathan Wakely <jwakely@redhat.com>
29564 * doc/invoke.texi (C++ Dialect Options): Suggest adding a
29565 using-declaration to unhide functions.
29567 2023-02-16 Jakub Jelinek <jakub@redhat.com>
29569 PR tree-optimization/108783
29570 * tree-ssa-reassoc.cc (eliminate_redundant_comparison): If lcode
29571 is equal to TREE_CODE (t), op1 to newop1 and op2 to newop2, set
29572 t to curr->op. Otherwise, punt if either newop1 or newop2 are
29573 SSA_NAME_OCCURS_IN_ABNORMAL_PHI SSA_NAMEs.
29575 2023-02-16 Richard Biener <rguenther@suse.de>
29577 PR tree-optimization/108791
29578 * tree-ssa-forwprop.cc (optimize_vector_load): Build
29579 the ADDR_EXPR of a TARGET_MEM_REF using a more meaningful
29582 2023-02-15 Eric Botcazou <ebotcazou@adacore.com>
29585 * config/i386/i386.cc (ix86_compute_frame_layout): Disable the
29586 effects of -fstack-clash-protection for TARGET_STACK_PROBE.
29587 (ix86_expand_prologue): Likewise.
29589 2023-02-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
29591 * config/bpf/bpf.cc (bpf_option_override): Fix doubled space.
29593 2023-02-15 Uroš Bizjak <ubizjak@gmail.com>
29595 * config/i386/i386.md (*cmpqi_ext<mode>_1): Use
29596 int248_register_operand predicate in zero_extract sub-RTX.
29597 (*cmpqi_ext<mode>_2): Ditto.
29598 (*cmpqi_ext<mode>_3): Ditto.
29599 (*cmpqi_ext<mode>_4): Ditto.
29600 (*extzvqi_mem_rex64): Ditto.
29602 (*insvqi_1_mem_rex64): Ditto.
29603 (@insv<mode>_1): Ditto.
29604 (*insvqi_1): Ditto.
29605 (*insvqi_2): Ditto.
29606 (*insvqi_3): Ditto.
29607 (*extendqi<SWI24:mode>_ext_1): Ditto.
29608 (*addqi_ext<mode>_1): Ditto.
29609 (*addqi_ext<mode>_2): Ditto.
29610 (*subqi_ext<mode>_2): Ditto.
29611 (*testqi_ext<mode>_1): Ditto.
29612 (*testqi_ext<mode>_2): Ditto.
29613 (*andqi_ext<mode>_1): Ditto.
29614 (*andqi_ext<mode>_1_cc): Ditto.
29615 (*andqi_ext<mode>_2): Ditto.
29616 (*<any_or:code>qi_ext<mode>_1): Ditto.
29617 (*<any_or:code>qi_ext<mode>_2): Ditto.
29618 (*xorqi_ext<mode>_1_cc): Ditto.
29619 (*negqi_ext<mode>_2): Ditto.
29620 (*ashlqi_ext<mode>_2): Ditto.
29621 (*<any_shiftrt:insn>qi_ext<mode>_2): Ditto.
29623 2023-02-15 Uroš Bizjak <ubizjak@gmail.com>
29625 * config/i386/predicates.md (int248_register_operand):
29626 Rename from extr_register_operand.
29627 * config/i386/i386.md (*extv<mode>): Update for renamed predicate.
29628 (*extzx<mode>): Ditto.
29629 (*ashl<dwi>3_doubleword_mask): Use int248_register_operand predicate.
29630 (*ashl<mode>3_mask): Ditto.
29631 (*<any_shiftrt:insn><mode>3_mask): Ditto.
29632 (*<any_shiftrt:insn><dwi>3_doubleword_mask): Ditto.
29633 (*<any_rotate:insn><mode>3_mask): Ditto.
29634 (*<btsc><mode>_mask): Ditto.
29635 (*btr<mode>_mask): Ditto.
29636 (*jcc_bt<mode>_mask_1): Ditto.
29638 2023-02-15 Richard Biener <rguenther@suse.de>
29640 PR middle-end/26854
29641 * df-core.cc (df_worklist_propagate_forward): Put later
29642 blocks on worklist and only earlier blocks on pending.
29643 (df_worklist_propagate_backward): Likewise.
29644 (df_worklist_dataflow_doublequeue): Change the iteration
29645 to process new blocks in the same iteration if that
29646 maintains the iteration order.
29648 2023-02-15 Marek Polacek <polacek@redhat.com>
29650 PR middle-end/106080
29651 * gimple-ssa-warn-access.cc (is_auto_decl): Remove. Use auto_var_p
29654 2023-02-15 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29656 * config/riscv/predicates.md: Refine codes.
29657 * config/riscv/riscv-protos.h (RVV_VUNDEF): New macro.
29658 * config/riscv/riscv-v.cc: Refine codes.
29659 * config/riscv/riscv-vector-builtins-bases.cc (enum ternop_type): New
29661 (class imac): New class.
29662 (enum widen_ternop_type): New enum.
29663 (class iwmac): New class.
29665 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
29666 * config/riscv/riscv-vector-builtins-functions.def (vmacc): Ditto.
29674 * config/riscv/riscv-vector-builtins.cc
29675 (function_builder::apply_predication): Adjust for multiply-add support.
29676 (function_expander::add_vundef_operand): Refine codes.
29677 (function_expander::use_ternop_insn): New function.
29678 (function_expander::use_widen_ternop_insn): Ditto.
29679 * config/riscv/riscv-vector-builtins.h: New function.
29680 * config/riscv/vector.md (@pred_mul_<optab><mode>): New pattern.
29681 (pred_mul_<optab><mode>_undef_merge): Ditto.
29682 (*pred_<madd_nmsub><mode>): Ditto.
29683 (*pred_<macc_nmsac><mode>): Ditto.
29684 (*pred_mul_<optab><mode>): Ditto.
29685 (@pred_mul_<optab><mode>_scalar): Ditto.
29686 (*pred_mul_<optab><mode>_undef_merge_scalar): Ditto.
29687 (*pred_<madd_nmsub><mode>_scalar): Ditto.
29688 (*pred_<macc_nmsac><mode>_scalar): Ditto.
29689 (*pred_mul_<optab><mode>_scalar): Ditto.
29690 (*pred_mul_<optab><mode>_undef_merge_extended_scalar): Ditto.
29691 (*pred_<madd_nmsub><mode>_extended_scalar): Ditto.
29692 (*pred_<macc_nmsac><mode>_extended_scalar): Ditto.
29693 (*pred_mul_<optab><mode>_extended_scalar): Ditto.
29694 (@pred_widen_mul_plus<su><mode>): Ditto.
29695 (@pred_widen_mul_plus<su><mode>_scalar): Ditto.
29696 (@pred_widen_mul_plussu<mode>): Ditto.
29697 (@pred_widen_mul_plussu<mode>_scalar): Ditto.
29698 (@pred_widen_mul_plusus<mode>_scalar): Ditto.
29700 2023-02-15 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29702 * config/riscv/predicates.md (vector_mask_operand): Refine the codes.
29703 (vector_all_trues_mask_operand): New predicate.
29704 (vector_undef_operand): New predicate.
29705 (ltge_operator): New predicate.
29706 (comparison_except_ltge_operator): New predicate.
29707 (comparison_except_eqge_operator): New predicate.
29708 (ge_operator): New predicate.
29709 * config/riscv/riscv-v.cc (has_vi_variant_p): Add compare support.
29710 * config/riscv/riscv-vector-builtins-bases.cc (class icmp): New class.
29712 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
29713 * config/riscv/riscv-vector-builtins-functions.def (vmseq): Ditto.
29723 * config/riscv/riscv-vector-builtins-shapes.cc
29724 (struct return_mask_def): Adjust for compare support.
29725 * config/riscv/riscv-vector-builtins.cc
29726 (function_expander::use_compare_insn): New function.
29727 * config/riscv/riscv-vector-builtins.h
29728 (function_expander::add_integer_operand): Ditto.
29729 * config/riscv/riscv.cc (riscv_print_operand): Add compare support.
29730 * config/riscv/riscv.md: Add vector min/max attributes.
29731 * config/riscv/vector-iterators.md (xnor): New iterator.
29732 * config/riscv/vector.md (@pred_cmp<mode>): New pattern.
29733 (*pred_cmp<mode>): Ditto.
29734 (*pred_cmp<mode>_narrow): Ditto.
29735 (@pred_ltge<mode>): Ditto.
29736 (*pred_ltge<mode>): Ditto.
29737 (*pred_ltge<mode>_narrow): Ditto.
29738 (@pred_cmp<mode>_scalar): Ditto.
29739 (*pred_cmp<mode>_scalar): Ditto.
29740 (*pred_cmp<mode>_scalar_narrow): Ditto.
29741 (@pred_eqne<mode>_scalar): Ditto.
29742 (*pred_eqne<mode>_scalar): Ditto.
29743 (*pred_eqne<mode>_scalar_narrow): Ditto.
29744 (*pred_cmp<mode>_extended_scalar): Ditto.
29745 (*pred_cmp<mode>_extended_scalar_narrow): Ditto.
29746 (*pred_eqne<mode>_extended_scalar): Ditto.
29747 (*pred_eqne<mode>_extended_scalar_narrow): Ditto.
29748 (@pred_ge<mode>_scalar): Ditto.
29749 (@pred_<optab><mode>): Ditto.
29750 (@pred_n<optab><mode>): Ditto.
29751 (@pred_<optab>n<mode>): Ditto.
29752 (@pred_not<mode>): Ditto.
29754 2023-02-15 Martin Jambor <mjambor@suse.cz>
29757 * ipa-sra.cc (push_param_adjustments_for_index): Do not omit
29758 creation of non-scalar replacements even if IPA-CP knows their
29761 2023-02-15 Jakub Jelinek <jakub@redhat.com>
29765 * config/rs6000/rs6000.md (<u>maddditi4): Change into umaddditi4 only
29766 expander, change operand 3 to be TImode, emit maddlddi4 and
29767 umadddi4_highpart{,_le} with its low half and finally add the high
29768 half to the result.
29770 2023-02-15 Martin Liska <mliska@suse.cz>
29772 * doc/invoke.texi: Document --param=asan-kernel-mem-intrinsic-prefix.
29774 2023-02-15 Richard Biener <rguenther@suse.de>
29776 * sanopt.cc (sanitize_asan_mark_unpoison): Use bitmap
29777 for with_poison and alias worklist to it.
29778 (sanitize_asan_mark_poison): Likewise.
29780 2023-02-15 Richard Biener <rguenther@suse.de>
29783 * config/i386/i386-features.cc (scalar_chain::add_to_queue):
29784 Combine bitmap test and set.
29785 (scalar_chain::add_insn): Likewise.
29786 (scalar_chain::analyze_register_chain): Remove redundant
29787 attempt to add to queue and instead strengthen assert.
29788 Sink common attempts to mark the def dual-mode.
29789 (scalar_chain::add_to_queue): Remove redundant insn bitmap
29792 2023-02-15 Richard Biener <rguenther@suse.de>
29795 * config/i386/i386-features.cc (convert_scalars_to_vector):
29796 Switch candidates bitmaps to tree view before building the chains.
29798 2023-02-15 Hans-Peter Nilsson <hp@axis.com>
29800 * reload1.cc (gen_reload): Correct rtx parameter for fatal_insn
29801 "failure trying to reload" call.
29803 2023-02-15 Hans-Peter Nilsson <hp@axis.com>
29805 * gdbinit.in (phrs): New command.
29806 * sel-sched-dump.cc (debug_hard_reg_set): Remove debug-function.
29807 * ira-color.cc (debug_hard_reg_set): New, calling print_hard_reg_set.
29809 2023-02-14 David Faust <david.faust@oracle.com>
29812 * config/bpf/constraints.md (q): New memory constraint.
29813 * config/bpf/bpf.md (zero_extendhidi2): Use it here.
29814 (zero_extendqidi2): Likewise.
29815 (zero_extendsidi2): Likewise.
29816 (*mov<MM:mode>): Likewise.
29818 2023-02-14 Andrew Pinski <apinski@marvell.com>
29820 PR tree-optimization/108355
29821 PR tree-optimization/96921
29822 * match.pd: Add pattern for "1 - bool_val".
29824 2023-02-14 Richard Biener <rguenther@suse.de>
29826 * tree-ssa-sccvn.cc (vn_phi_compute_hash): Key skipping
29827 basic block index hashing on the availability of ->cclhs.
29828 (vn_phi_eq): Avoid re-doing sanity checks for CSE but
29829 rely on ->cclhs availability.
29830 (vn_phi_lookup): Set ->cclhs only when we are eventually
29831 going to CSE the PHI.
29832 (vn_phi_insert): Likewise.
29834 2023-02-14 Eric Botcazou <ebotcazou@adacore.com>
29836 * gimplify.cc (gimplify_save_expr): Add missing guard.
29838 2023-02-14 Richard Biener <rguenther@suse.de>
29840 PR tree-optimization/108782
29841 * tree-vect-loop.cc (vect_phi_first_order_recurrence_p):
29842 Make sure we're not vectorizing an inner loop.
29844 2023-02-14 Jakub Jelinek <jakub@redhat.com>
29846 PR sanitizer/108777
29847 * params.opt (-param=asan-kernel-mem-intrinsic-prefix=): New param.
29848 * asan.h (asan_memfn_rtl): Declare.
29849 * asan.cc (asan_memfn_rtls): New variable.
29850 (asan_memfn_rtl): New function.
29851 * builtins.cc (expand_builtin): If
29852 param_asan_kernel_mem_intrinsic_prefix and function is
29853 kernel-{,hw}address sanitized, emit calls to
29854 __{,hw}asan_{memcpy,memmove,memset} rather than
29855 {memcpy,memmove,memset}. Use sanitize_flags_p (SANITIZE_ADDRESS)
29856 instead of flag_sanitize & SANITIZE_ADDRESS to check if
29857 asan_intercepted_p functions shouldn't be expanded inline.
29859 2023-02-14 Richard Sandiford <richard.sandiford@arm.com>
29861 PR tree-optimization/96373
29862 * tree-vect-stmts.cc (vectorizable_operation): Predicate trapping
29863 operations on the loop mask. Reject partial vectors if this isn't
29866 2023-02-13 Richard Sandiford <richard.sandiford@arm.com>
29868 PR rtl-optimization/108681
29869 * lra-spills.cc (lra_final_code_change): Extend subreg replacement
29870 code to handle bare uses and clobbers.
29872 2023-02-13 Vladimir N. Makarov <vmakarov@redhat.com>
29874 * ira.cc (ira_update_equiv_info_by_shuffle_insn): Clear equiv
29875 caller_save_p flag when clearing defined_p flag.
29876 (setup_reg_equiv): Ditto.
29877 * lra-constraints.cc (lra_constraints): Ditto.
29879 2023-02-13 Uroš Bizjak <ubizjak@gmail.com>
29882 * config/i386/predicates.md (extr_register_operand):
29883 New special predicate.
29884 * config/i386/i386.md (*extv<mode>): Use extr_register_operand
29885 as operand 1 predicate.
29886 (*exzv<mode>): Ditto.
29887 (*extendqi<SWI24:mode>_ext_1): New insn pattern.
29889 2023-02-13 Richard Biener <rguenther@suse.de>
29891 PR tree-optimization/28614
29892 * tree-ssa-sccvn.cc (can_track_predicate_on_edge): Avoid
29893 walking all edges in most cases.
29894 (vn_nary_op_insert_pieces_predicated): Avoid repeated
29895 calls to can_track_predicate_on_edge unless checking is
29897 (process_bb): Instead call it once here for each edge
29898 we register possibly multiple predicates on.
29900 2023-02-13 Richard Biener <rguenther@suse.de>
29902 PR tree-optimization/108691
29903 * tree-cfg.cc (notice_special_calls): When the CFG is built
29904 honor gimple_call_ctrl_altering_p.
29905 * cfgexpand.cc (expand_call_stmt): Clear cfun->calls_setjmp
29906 temporarily if the call is not control-altering.
29907 * calls.cc (emit_call_1): Do not add REG_SETJMP if
29908 cfun->calls_setjmp is not set. Do not alter cfun->calls_setjmp.
29910 2023-02-13 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
29913 * config/s390/s390.cc (s390_bb_fallthru_entry_likely): Remove.
29914 (struct s390_sched_state): Initialise to zero.
29915 (s390_sched_variable_issue): For better debuggability also emit
29917 (s390_sched_init): Unconditionally reset scheduler state.
29919 2023-02-13 Richard Sandiford <richard.sandiford@arm.com>
29921 * ifcvt.h (noce_if_info::cond_inverted): New field.
29922 * ifcvt.cc (cond_move_convert_if_block): Swap the then and else
29923 values when cond_inverted is true.
29924 (noce_find_if_block): Allow the condition to be inverted when
29925 handling conditional moves.
29927 2023-02-13 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
29929 * config/s390/predicates.md (execute_operation): Use
29930 constrain_operands instead of extract_constrain_insn in order to
29931 determine wheter there exists a valid alternative.
29933 2023-02-13 Claudiu Zissulescu <claziss@gmail.com>
29935 * common/config/arc/arc-common.cc (arc_option_optimization_table):
29936 Remove millicode from list.
29938 2023-02-13 Martin Liska <mliska@suse.cz>
29940 * doc/invoke.texi: Document ira-simple-lra-insn-threshold.
29942 2023-02-13 Richard Biener <rguenther@suse.de>
29944 PR tree-optimization/106722
29945 * tree-ssa-dce.cc (mark_last_stmt_necessary): Return
29946 whether we marked a stmt.
29947 (mark_control_dependent_edges_necessary): When
29948 mark_last_stmt_necessary didn't mark any stmt make sure
29949 to mark its control dependent edges.
29950 (propagate_necessity): Likewise.
29952 2023-02-13 Kito Cheng <kito.cheng@sifive.com>
29954 * config/riscv/riscv.h (RISCV_DWARF_VLENB): New.
29955 (DWARF_FRAME_REGISTERS): New.
29956 (DWARF_REG_TO_UNWIND_COLUMN): New.
29958 2023-02-12 Gerald Pfeifer <gerald@pfeifer.com>
29960 * doc/sourcebuild.texi: Remove (broken) direct reference to
29961 "The GNU configure and build system".
29963 2023-02-12 Jin Ma <jinma@linux.alibaba.com>
29965 * config/riscv/riscv.cc (riscv_adjust_libcall_cfi_prologue): Change
29966 gen_add3_insn to gen_rtx_SET.
29967 (riscv_adjust_libcall_cfi_epilogue): Likewise.
29969 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29971 * config/riscv/riscv-vector-builtins-bases.cc (class sat_op): New class.
29972 (class vnclip): Ditto.
29974 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
29975 * config/riscv/riscv-vector-builtins-functions.def (vaadd): Ditto.
29984 * config/riscv/vector-iterators.md (su): Add instruction.
29987 * config/riscv/vector.md (@pred_<sat_op><mode>): New pattern.
29988 (@pred_<sat_op><mode>_scalar): Ditto.
29989 (*pred_<sat_op><mode>_scalar): Ditto.
29990 (*pred_<sat_op><mode>_extended_scalar): Ditto.
29991 (@pred_narrow_clip<v_su><mode>): Ditto.
29992 (@pred_narrow_clip<v_su><mode>_scalar): Ditto.
29994 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29996 * config/riscv/constraints.md (Wbr): Remove unused constraint.
29997 * config/riscv/predicates.md: Fix move operand predicate.
29998 * config/riscv/riscv-vector-builtins-bases.cc (class vnshift): New class.
29999 (class vncvt_x): Ditto.
30000 (class vmerge): Ditto.
30001 (class vmv_v): Ditto.
30003 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
30004 * config/riscv/riscv-vector-builtins-functions.def (vsra): Ditto.
30011 * config/riscv/riscv-vector-builtins-shapes.cc (struct narrow_alu_def): Ditto.
30012 (struct move_def): Ditto.
30014 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
30015 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_WEXTI_OPS): New variable.
30016 (DEF_RVV_WEXTU_OPS): Ditto
30017 * config/riscv/riscv-vector-builtins.def (x_x_w): Fix type for suffix.
30022 * config/riscv/riscv.cc (riscv_print_operand): Refine ASM printting rule.
30023 * config/riscv/vector-iterators.md (nmsac):New iterator.
30024 (nmsub): New iterator.
30025 * config/riscv/vector.md (@pred_merge<mode>): New pattern.
30026 (@pred_merge<mode>_scalar): New pattern.
30027 (*pred_merge<mode>_scalar): New pattern.
30028 (*pred_merge<mode>_extended_scalar): New pattern.
30029 (@pred_narrow_<optab><mode>): New pattern.
30030 (@pred_narrow_<optab><mode>_scalar): New pattern.
30031 (@pred_trunc<mode>): New pattern.
30033 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30035 * config/riscv/riscv-vector-builtins-bases.cc (class vmadc): New class.
30036 (class vmsbc): Ditto.
30037 (BASE): Define new class.
30038 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
30039 * config/riscv/riscv-vector-builtins-functions.def (vmadc): New define.
30041 * config/riscv/riscv-vector-builtins-shapes.cc (struct return_mask_def):
30044 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
30045 * config/riscv/riscv-vector-builtins.cc
30046 (function_expander::use_exact_insn): Adjust for new support
30047 * config/riscv/riscv-vector-builtins.h
30048 (function_base::has_merge_operand_p): New function.
30049 * config/riscv/vector-iterators.md: New iterator.
30050 * config/riscv/vector.md (@pred_madc<mode>): New pattern.
30051 (@pred_msbc<mode>): Ditto.
30052 (@pred_madc<mode>_scalar): Ditto.
30053 (@pred_msbc<mode>_scalar): Ditto.
30054 (*pred_madc<mode>_scalar): Ditto.
30055 (*pred_madc<mode>_extended_scalar): Ditto.
30056 (*pred_msbc<mode>_scalar): Ditto.
30057 (*pred_msbc<mode>_extended_scalar): Ditto.
30058 (@pred_madc<mode>_overflow): Ditto.
30059 (@pred_msbc<mode>_overflow): Ditto.
30060 (@pred_madc<mode>_overflow_scalar): Ditto.
30061 (@pred_msbc<mode>_overflow_scalar): Ditto.
30062 (*pred_madc<mode>_overflow_scalar): Ditto.
30063 (*pred_madc<mode>_overflow_extended_scalar): Ditto.
30064 (*pred_msbc<mode>_overflow_scalar): Ditto.
30065 (*pred_msbc<mode>_overflow_extended_scalar): Ditto.
30067 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30069 * config/riscv/riscv-protos.h (simm5_p): Add vadc/vsbc support.
30070 * config/riscv/riscv-v.cc (simm32_p): Ditto.
30071 * config/riscv/riscv-vector-builtins-bases.cc (class vadc): New class.
30072 (class vsbc): Ditto.
30074 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
30075 * config/riscv/riscv-vector-builtins-functions.def (vadc): Ditto.
30077 * config/riscv/riscv-vector-builtins-shapes.cc
30078 (struct no_mask_policy_def): Ditto.
30080 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
30081 * config/riscv/riscv-vector-builtins.cc
30082 (rvv_arg_type_info::get_base_vector_type): Add vadc/vsbc support.
30083 (rvv_arg_type_info::get_tree_type): Ditto.
30084 (function_expander::use_exact_insn): Ditto.
30085 * config/riscv/riscv-vector-builtins.h (enum rvv_base_type): Ditto.
30086 (function_base::use_mask_predication_p): New function.
30087 * config/riscv/vector-iterators.md: New iterator.
30088 * config/riscv/vector.md (@pred_adc<mode>): New pattern.
30089 (@pred_sbc<mode>): Ditto.
30090 (@pred_adc<mode>_scalar): Ditto.
30091 (@pred_sbc<mode>_scalar): Ditto.
30092 (*pred_adc<mode>_scalar): Ditto.
30093 (*pred_adc<mode>_extended_scalar): Ditto.
30094 (*pred_sbc<mode>_scalar): Ditto.
30095 (*pred_sbc<mode>_extended_scalar): Ditto.
30097 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30099 * config/riscv/vector.md: use "zero" reg.
30101 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30103 * config/riscv/riscv-vector-builtins-bases.cc (class widen_binop): New
30105 (class vwmulsu): Ditto.
30106 (class vwcvt): Ditto.
30107 (BASE): Add integer widening support.
30108 * config/riscv/riscv-vector-builtins-bases.h: Ditto
30109 * config/riscv/riscv-vector-builtins-functions.def (vwadd): New class.
30110 (vwsub): New class.
30111 (vwmul): New class.
30112 (vwmulu): New class.
30113 (vwmulsu): New class.
30114 (vwaddu): New class.
30115 (vwsubu): New class.
30116 (vwcvt_x): New class.
30117 (vwcvtu_x): New class.
30118 * config/riscv/riscv-vector-builtins-shapes.cc (struct alu_def): New
30120 (struct widen_alu_def): New class.
30121 (SHAPE): New class.
30122 * config/riscv/riscv-vector-builtins-shapes.h: New class.
30123 * config/riscv/riscv-vector-builtins.cc
30124 (rvv_arg_type_info::get_base_vector_type): Add integer widening support.
30125 (rvv_arg_type_info::get_tree_type): Ditto.
30126 * config/riscv/riscv-vector-builtins.def (x_x_v): Change into "x_v"
30128 * config/riscv/riscv-vector-builtins.h (enum rvv_base_type): Add integer
30130 * config/riscv/riscv-vsetvl.cc (change_insn): Fix reg_equal use bug.
30131 * config/riscv/riscv.h (X0_REGNUM): New constant.
30132 * config/riscv/vector-iterators.md: New iterators.
30133 * config/riscv/vector.md
30134 (@pred_dual_widen_<any_widen_binop:optab><any_extend:su><mode>): New
30136 (@pred_dual_widen_<any_widen_binop:optab><any_extend:su><mode>_scalar):
30138 (@pred_single_widen_<plus_minus:optab><any_extend:su><mode>): Ditto.
30139 (@pred_single_widen_<plus_minus:optab><any_extend:su><mode>_scalar):
30141 (@pred_widen_mulsu<mode>): Ditto.
30142 (@pred_widen_mulsu<mode>_scalar): Ditto.
30143 (@pred_<optab><mode>): Ditto.
30145 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30146 kito-cheng <kito.cheng@sifive.com>
30148 * common/config/riscv/riscv-common.cc: Add flag for 'V' extension.
30149 * config/riscv/riscv-vector-builtins-bases.cc (class vmulh): New class.
30151 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
30152 * config/riscv/riscv-vector-builtins-functions.def (vmulh): Add vmulh
30156 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_FULL_V_I_OPS):
30158 (DEF_RVV_FULL_V_U_OPS): Ditto.
30159 (vint8mf8_t): Ditto.
30160 (vint8mf4_t): Ditto.
30161 (vint8mf2_t): Ditto.
30162 (vint8m1_t): Ditto.
30163 (vint8m2_t): Ditto.
30164 (vint8m4_t): Ditto.
30165 (vint8m8_t): Ditto.
30166 (vint16mf4_t): Ditto.
30167 (vint16mf2_t): Ditto.
30168 (vint16m1_t): Ditto.
30169 (vint16m2_t): Ditto.
30170 (vint16m4_t): Ditto.
30171 (vint16m8_t): Ditto.
30172 (vint32mf2_t): Ditto.
30173 (vint32m1_t): Ditto.
30174 (vint32m2_t): Ditto.
30175 (vint32m4_t): Ditto.
30176 (vint32m8_t): Ditto.
30177 (vint64m1_t): Ditto.
30178 (vint64m2_t): Ditto.
30179 (vint64m4_t): Ditto.
30180 (vint64m8_t): Ditto.
30181 (vuint8mf8_t): Ditto.
30182 (vuint8mf4_t): Ditto.
30183 (vuint8mf2_t): Ditto.
30184 (vuint8m1_t): Ditto.
30185 (vuint8m2_t): Ditto.
30186 (vuint8m4_t): Ditto.
30187 (vuint8m8_t): Ditto.
30188 (vuint16mf4_t): Ditto.
30189 (vuint16mf2_t): Ditto.
30190 (vuint16m1_t): Ditto.
30191 (vuint16m2_t): Ditto.
30192 (vuint16m4_t): Ditto.
30193 (vuint16m8_t): Ditto.
30194 (vuint32mf2_t): Ditto.
30195 (vuint32m1_t): Ditto.
30196 (vuint32m2_t): Ditto.
30197 (vuint32m4_t): Ditto.
30198 (vuint32m8_t): Ditto.
30199 (vuint64m1_t): Ditto.
30200 (vuint64m2_t): Ditto.
30201 (vuint64m4_t): Ditto.
30202 (vuint64m8_t): Ditto.
30203 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_FULL_V_I_OPS): Ditto.
30204 (DEF_RVV_FULL_V_U_OPS): Ditto.
30205 (check_required_extensions): Add vmulh support.
30206 (rvv_arg_type_info::get_tree_type): Ditto.
30207 * config/riscv/riscv-vector-builtins.h (RVV_REQUIRE_FULL_V): Ditto.
30208 (enum rvv_base_type): Ditto.
30209 * config/riscv/riscv.opt: Add 'V' extension flag.
30210 * config/riscv/vector-iterators.md (su): New iterator.
30211 * config/riscv/vector.md (@pred_mulh<v_su><mode>): New pattern.
30212 (@pred_mulh<v_su><mode>_scalar): Ditto.
30213 (*pred_mulh<v_su><mode>_scalar): Ditto.
30214 (*pred_mulh<v_su><mode>_extended_scalar): Ditto.
30216 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30218 * config/riscv/iterators.md: Add sign_extend/zero_extend.
30219 * config/riscv/riscv-vector-builtins-bases.cc (class ext): New class.
30221 * config/riscv/riscv-vector-builtins-bases.h: Add vsext/vzext support.
30222 * config/riscv/riscv-vector-builtins-functions.def (vsext): New macro
30225 * config/riscv/riscv-vector-builtins-shapes.cc (struct alu_def): Adjust
30226 for vsext/vzext support.
30227 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_WEXTI_OPS): New
30229 (DEF_RVV_QEXTI_OPS): Ditto.
30230 (DEF_RVV_OEXTI_OPS): Ditto.
30231 (DEF_RVV_WEXTU_OPS): Ditto.
30232 (DEF_RVV_QEXTU_OPS): Ditto.
30233 (DEF_RVV_OEXTU_OPS): Ditto.
30234 (vint16mf4_t): Ditto.
30235 (vint16mf2_t): Ditto.
30236 (vint16m1_t): Ditto.
30237 (vint16m2_t): Ditto.
30238 (vint16m4_t): Ditto.
30239 (vint16m8_t): Ditto.
30240 (vint32mf2_t): Ditto.
30241 (vint32m1_t): Ditto.
30242 (vint32m2_t): Ditto.
30243 (vint32m4_t): Ditto.
30244 (vint32m8_t): Ditto.
30245 (vint64m1_t): Ditto.
30246 (vint64m2_t): Ditto.
30247 (vint64m4_t): Ditto.
30248 (vint64m8_t): Ditto.
30249 (vuint16mf4_t): Ditto.
30250 (vuint16mf2_t): Ditto.
30251 (vuint16m1_t): Ditto.
30252 (vuint16m2_t): Ditto.
30253 (vuint16m4_t): Ditto.
30254 (vuint16m8_t): Ditto.
30255 (vuint32mf2_t): Ditto.
30256 (vuint32m1_t): Ditto.
30257 (vuint32m2_t): Ditto.
30258 (vuint32m4_t): Ditto.
30259 (vuint32m8_t): Ditto.
30260 (vuint64m1_t): Ditto.
30261 (vuint64m2_t): Ditto.
30262 (vuint64m4_t): Ditto.
30263 (vuint64m8_t): Ditto.
30264 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_WEXTI_OPS): Ditto.
30265 (DEF_RVV_QEXTI_OPS): Ditto.
30266 (DEF_RVV_OEXTI_OPS): Ditto.
30267 (DEF_RVV_WEXTU_OPS): Ditto.
30268 (DEF_RVV_QEXTU_OPS): Ditto.
30269 (DEF_RVV_OEXTU_OPS): Ditto.
30270 (rvv_arg_type_info::get_base_vector_type): Add sign_exted/zero_extend
30272 (rvv_arg_type_info::get_tree_type): Ditto.
30273 * config/riscv/riscv-vector-builtins.h (enum rvv_base_type): Ditto.
30274 * config/riscv/vector-iterators.md (z): New attribute.
30275 * config/riscv/vector.md (@pred_<optab><mode>_vf2): New pattern.
30276 (@pred_<optab><mode>_vf4): Ditto.
30277 (@pred_<optab><mode>_vf8): Ditto.
30279 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30281 * config/riscv/iterators.md: Add saturating Addition && Subtraction.
30282 * config/riscv/riscv-v.cc (has_vi_variant_p): Ditto.
30283 * config/riscv/riscv-vector-builtins-bases.cc (BASE): Ditto.
30284 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
30285 * config/riscv/riscv-vector-builtins-functions.def (vsadd): New def.
30289 * config/riscv/vector-iterators.md (sll.vi): Adjust for Saturating
30294 * config/riscv/vector.md (@pred_<optab><mode>): New pattern.
30295 (@pred_<optab><mode>_scalar): New pattern.
30296 (*pred_<optab><mode>_scalar): New pattern.
30297 (*pred_<optab><mode>_extended_scalar): New pattern.
30299 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30301 * config/riscv/iterators.md: Add neg and not.
30302 * config/riscv/riscv-vector-builtins-bases.cc (class unop): New class.
30304 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
30305 * config/riscv/riscv-vector-builtins-functions.def (vadd): Rename binop
30326 * config/riscv/riscv-vector-builtins-shapes.cc (struct binop_def): Ditto.
30327 (struct alu_def): Ditto.
30329 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
30330 * config/riscv/riscv-vector-builtins.cc: Support unary C/C/++.
30331 * config/riscv/vector-iterators.md: New iterator.
30332 * config/riscv/vector.md (@pred_<optab><mode>): New pattern
30334 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30336 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::compute_probabilities): Skip exit block.
30338 2023-02-11 Jakub Jelinek <jakub@redhat.com>
30341 * ipa-cp.cc (ipa_agg_value_from_jfunc): Return NULL_TREE also if
30342 item->offset bit position is too large to be representable as
30343 unsigned int byte position.
30345 2023-02-11 Gerald Pfeifer <gerald@pfeifer.com>
30347 * doc/extend.texi (Other Builtins): Adjust link to WG14 N965.
30349 2023-02-10 Vladimir N. Makarov <vmakarov@redhat.com>
30351 * ira.cc (update_equiv_regs): Set up ira_reg_equiv for
30352 valid_combine only when ira_use_lra_p is true.
30354 2023-02-10 Vladimir N. Makarov <vmakarov@redhat.com>
30356 * params.opt (ira-simple-lra-insn-threshold): Add new param.
30357 * ira.cc (ira): Use the param to switch on simple LRA.
30359 2023-02-10 Andrew MacLeod <amacleod@redhat.com>
30361 PR tree-optimization/108687
30362 * gimple-range-cache.cc (ranger_cache::range_on_edge): Revert
30363 back to RFD_NONE mode for calculations.
30364 (ranger_cache::propagate_cache): Call the internal edge range API
30365 with RFD_READ_ONLY instead of changing the external routine.
30367 2023-02-10 Andrew MacLeod <amacleod@redhat.com>
30369 PR tree-optimization/108520
30370 * gimple-range-infer.cc (check_assume_func): Invoke
30371 gimple_range_global directly instead using global_range_query.
30372 * value-query.cc (get_range_global): Add function context and
30373 avoid calling nonnull_arg_p if not cfun.
30374 (gimple_range_global): Add function context pointer.
30375 * value-query.h (imple_range_global): Add function context.
30377 2023-02-10 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30379 * config/riscv/constraints.md (Wdm): Adjust constraint.
30380 (Wbr): New constraint.
30381 * config/riscv/predicates.md (reg_or_int_operand): New predicate.
30382 * config/riscv/riscv-protos.h (emit_pred_op): Remove function.
30383 (emit_vlmax_op): New function.
30384 (emit_nonvlmax_op): Ditto.
30386 (neg_simm5_p): Ditto.
30387 (has_vi_variant_p): Ditto.
30388 * config/riscv/riscv-v.cc (emit_pred_op): Adjust function.
30389 (emit_vlmax_op): New function.
30390 (emit_nonvlmax_op): Ditto.
30391 (expand_const_vector): Adjust function.
30392 (legitimize_move): Ditto.
30393 (simm32_p): New function.
30395 (neg_simm5_p): Ditto.
30396 (has_vi_variant_p): Ditto.
30397 * config/riscv/riscv-vector-builtins-bases.cc (class vrsub): New class.
30399 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
30400 * config/riscv/riscv-vector-builtins-functions.def (vmin): Remove
30403 (vminu): Remove signed cases.
30405 (vdiv): Remove unsigned cases.
30407 (vdivu): Remove signed cases.
30411 (vrsub): New class.
30416 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_U_OPS): New macro.
30417 * config/riscv/riscv.h: change VL/VTYPE as fixed reg.
30418 * config/riscv/vector-iterators.md: New iterators.
30419 * config/riscv/vector.md (@pred_broadcast<mode>): Adjust pattern for vx
30421 (@pred_<optab><mode>_scalar): New pattern.
30422 (@pred_sub<mode>_reverse_scalar): Ditto.
30423 (*pred_<optab><mode>_scalar): Ditto.
30424 (*pred_<optab><mode>_extended_scalar): Ditto.
30425 (*pred_sub<mode>_reverse_scalar): Ditto.
30426 (*pred_sub<mode>_extended_reverse_scalar): Ditto.
30428 2023-02-10 Richard Biener <rguenther@suse.de>
30430 PR tree-optimization/108724
30431 * tree-vect-stmts.cc (vectorizable_operation): Avoid
30432 using word_mode vectors when vector lowering will
30433 decompose them to elementwise operations.
30435 2023-02-10 Jakub Jelinek <jakub@redhat.com>
30438 2023-02-09 Martin Liska <mliska@suse.cz>
30441 * doc/extend.texi: Document that the function
30442 does not work correctly for old VIA processors.
30444 2023-02-10 Andrew Pinski <apinski@marvell.com>
30445 Andrew Macleod <amacleod@redhat.com>
30447 PR tree-optimization/108684
30448 * tree-ssa-dce.cc (simple_dce_from_worklist):
30449 Check all ssa names and not just non-vdef ones
30450 before accepting the inline-asm.
30451 Call unlink_stmt_vdef on the statement before
30454 2023-02-09 Vladimir N. Makarov <vmakarov@redhat.com>
30456 * ira.h (struct ira_reg_equiv_s): Add new field caller_save_p.
30457 * ira.cc (validate_equiv_mem): Check memref address variance.
30458 (no_equiv): Clear caller_save_p flag.
30459 (update_equiv_regs): Define caller save equivalence for
30461 (setup_reg_equiv): Clear defined_p flag for caller save equivalence.
30462 * lra-constraints.cc (lra_copy_reg_equiv): Add new arg
30463 call_save_p. Use caller save equivalence depending on the arg.
30464 (split_reg): Adjust the call.
30466 2023-02-09 Jakub Jelinek <jakub@redhat.com>
30469 * common/config/i386/cpuinfo.h (get_zhaoxin_cpu): Formatting fixes.
30470 (cpu_indicator_init): Call get_available_features for all CPUs with
30471 max_level >= 1, rather than just Intel, AMD or Zhaoxin. Formatting
30474 2023-02-09 Jakub Jelinek <jakub@redhat.com>
30476 PR tree-optimization/108688
30477 * match.pd (bit_field_ref [bit_insert]): Simplify BIT_FIELD_REF
30478 of BIT_INSERT_EXPR extracting exactly all inserted bits even
30479 when without mode precision. Formatting fixes.
30481 2023-02-09 Andrew Pinski <apinski@marvell.com>
30483 PR tree-optimization/108688
30484 * match.pd (bit_field_ref [bit_insert]): Avoid generating
30485 BIT_FIELD_REFs of non-mode-precision integral operands.
30487 2023-02-09 Martin Liska <mliska@suse.cz>
30490 * doc/extend.texi: Document that the function
30491 does not work correctly for old VIA processors.
30493 2023-02-09 Andreas Schwab <schwab@suse.de>
30495 * lto-wrapper.cc (merge_and_complain): Handle
30496 -funwind-tables and -fasynchronous-unwind-tables.
30497 (append_compiler_options): Likewise.
30499 2023-02-09 Richard Biener <rguenther@suse.de>
30501 PR tree-optimization/26854
30502 * tree-into-ssa.cc (update_ssa): Turn blocks_to_update to tree
30503 view around insert_updated_phi_nodes_for.
30504 * tree-ssa-alias.cc (maybe_skip_until): Allocate visited bitmap
30506 (walk_aliased_vdefs_1): Likewise.
30508 2023-02-08 Gerald Pfeifer <gerald@pfeifer.com>
30510 * doc/include/gpl_v3.texi: Change fsf.org to www.fsf.org.
30512 2023-02-08 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
30515 * config.gcc (tm_mlib_file): Define new variable.
30517 2023-02-08 Jakub Jelinek <jakub@redhat.com>
30519 PR tree-optimization/108692
30520 * tree-vect-patterns.cc (vect_widened_op_tree): If rhs_code is
30521 widened_code which is different from code, don't call
30522 vect_look_through_possible_promotion but instead just check op is
30523 SSA_NAME with integral type for which vect_is_simple_use is true
30524 and call set_op on this_unprom.
30526 2023-02-08 Andrea Corallo <andrea.corallo@arm.com>
30528 * config/aarch64/aarch64-protos.h (aarch_ra_sign_key): Remove
30530 * config/aarch64/aarch64.cc (aarch_ra_sign_key): Remove
30532 * config/aarch64/aarch64.opt (aarch64_ra_sign_key): Rename
30533 to 'aarch_ra_sign_key'.
30534 * config/arm/aarch-common.cc (aarch_ra_sign_key): Remove
30536 * config/arm/arm-protos.h (aarch_ra_sign_key): Likewise.
30537 * config/arm/arm.cc (enum aarch_key_type): Remove definition.
30538 * config/arm/arm.opt: Define.
30540 2023-02-08 Richard Sandiford <richard.sandiford@arm.com>
30542 PR tree-optimization/108316
30543 * tree-vect-stmts.cc (get_load_store_type): When using
30544 internal functions for gather/scatter, make sure that the type
30545 of the offset argument is consistent with the offset vector type.
30547 2023-02-08 Vladimir N. Makarov <vmakarov@redhat.com>
30550 2023-02-07 Vladimir N. Makarov <vmakarov@redhat.com>
30552 * ira.h (struct ira_reg_equiv_s): Add new field caller_save_p.
30553 * ira.cc (validate_equiv_mem): Check memref address variance.
30554 (update_equiv_regs): Define caller save equivalence for
30556 (setup_reg_equiv): Clear defined_p flag for caller save equivalence.
30557 * lra-constraints.cc (lra_copy_reg_equiv): Add new arg
30558 call_save_p. Use caller save equivalence depending on the arg.
30559 (split_reg): Adjust the call.
30561 2023-02-08 Jakub Jelinek <jakub@redhat.com>
30563 * tree.def (SAD_EXPR): Remove outdated comment about missing
30566 2023-02-07 Marek Polacek <polacek@redhat.com>
30568 * doc/invoke.texi: Update -fchar8_t documentation.
30570 2023-02-07 Vladimir N. Makarov <vmakarov@redhat.com>
30572 * ira.h (struct ira_reg_equiv_s): Add new field caller_save_p.
30573 * ira.cc (validate_equiv_mem): Check memref address variance.
30574 (update_equiv_regs): Define caller save equivalence for
30576 (setup_reg_equiv): Clear defined_p flag for caller save equivalence.
30577 * lra-constraints.cc (lra_copy_reg_equiv): Add new arg
30578 call_save_p. Use caller save equivalence depending on the arg.
30579 (split_reg): Adjust the call.
30581 2023-02-07 Richard Biener <rguenther@suse.de>
30583 PR tree-optimization/26854
30584 * gimple-fold.cc (has_use_on_stmt): Look at stmt operands
30585 instead of immediate uses.
30587 2023-02-07 Jakub Jelinek <jakub@redhat.com>
30589 PR tree-optimization/106923
30590 * ipa-split.cc (execute_split_functions): Don't split returns_twice
30593 2023-02-07 Jakub Jelinek <jakub@redhat.com>
30595 PR tree-optimization/106433
30596 * cgraph.cc (set_const_flag_1): Recurse on simd clones too.
30597 (cgraph_node::set_pure_flag): Call set_pure_flag_1 on simd clones too.
30599 2023-02-07 Jan Hubicka <jh@suse.cz>
30601 * config/i386/x86-tune.def (X86_TUNE_AVX256_OPTIMAL): Turn off
30604 2023-02-06 Andrew Stubbs <ams@codesourcery.com>
30606 * config/gcn/mkoffload.cc (gcn_stack_size): New global variable.
30607 (process_asm): Create a constructor for GCN_STACK_SIZE.
30608 (main): Parse the -mstack-size option.
30610 2023-02-06 Alex Coplan <alex.coplan@arm.com>
30613 * config/aarch64/aarch64-simd.md (aarch64_bfmlal<bt>_lane<q>v4sf):
30614 Use correct constraint for operand 3.
30616 2023-02-06 Martin Jambor <mjambor@suse.cz>
30618 * ipa-sra.cc (adjust_parameter_descriptions): Fix a typo in a dump.
30620 2023-02-06 Xi Ruoyao <xry111@xry111.site>
30622 * config/loongarch/loongarch.md (bytepick_w_ashift_amount):
30623 New define_int_iterator.
30624 (bytepick_d_ashift_amount): Likewise.
30625 (bytepick_imm): New define_int_attr.
30626 (bytepick_w_lshiftrt_amount): Likewise.
30627 (bytepick_d_lshiftrt_amount): Likewise.
30628 (bytepick_w_<bytepick_imm>): New define_insn template.
30629 (bytepick_w_<bytepick_imm>_extend): Likewise.
30630 (bytepick_d_<bytepick_imm>): Likewise.
30631 (bytepick_w): Remove unused define_insn.
30632 (bytepick_d): Likewise.
30633 (UNSPEC_BYTEPICK_W): Remove unused unspec.
30634 (UNSPEC_BYTEPICK_D): Likewise.
30635 * config/loongarch/predicates.md (const_0_to_3_operand):
30636 Remove unused define_predicate.
30637 (const_0_to_7_operand): Likewise.
30639 2023-02-06 Jakub Jelinek <jakub@redhat.com>
30641 PR tree-optimization/108655
30642 * ubsan.cc (sanitize_unreachable_fn): For -funreachable-traps
30643 or -fsanitize=unreachable -fsanitize-trap=unreachable return
30644 BUILT_IN_UNREACHABLE_TRAP decl rather than BUILT_IN_TRAP.
30646 2023-02-05 Gerald Pfeifer <gerald@pfeifer.com>
30648 * doc/install.texi (Specific): Remove PW32.
30650 2023-02-03 Jakub Jelinek <jakub@redhat.com>
30652 PR tree-optimization/108647
30653 * range-op.cc (operator_equal::op1_range,
30654 operator_not_equal::op1_range): Don't test op2 bound
30655 equality if op2.undefined_p (), instead set_varying.
30656 (operator_lt::op1_range, operator_le::op1_range,
30657 operator_gt::op1_range, operator_ge::op1_range): Return false if
30658 op2.undefined_p ().
30659 (operator_lt::op2_range, operator_le::op2_range,
30660 operator_gt::op2_range, operator_ge::op2_range): Return false if
30661 op1.undefined_p ().
30663 2023-02-03 Aldy Hernandez <aldyh@redhat.com>
30665 PR tree-optimization/108639
30666 * value-range.cc (irange::legacy_equal_p): Compare nonzero bits as
30668 (irange::operator==): Same.
30670 2023-02-03 Aldy Hernandez <aldyh@redhat.com>
30672 PR tree-optimization/108647
30673 * range-op-float.cc (foperator_lt::op1_range): Handle undefined ranges.
30674 (foperator_lt::op2_range): Same.
30675 (foperator_le::op1_range): Same.
30676 (foperator_le::op2_range): Same.
30677 (foperator_gt::op1_range): Same.
30678 (foperator_gt::op2_range): Same.
30679 (foperator_ge::op1_range): Same.
30680 (foperator_ge::op2_range): Same.
30681 (foperator_unordered_lt::op1_range): Same.
30682 (foperator_unordered_lt::op2_range): Same.
30683 (foperator_unordered_le::op1_range): Same.
30684 (foperator_unordered_le::op2_range): Same.
30685 (foperator_unordered_gt::op1_range): Same.
30686 (foperator_unordered_gt::op2_range): Same.
30687 (foperator_unordered_ge::op1_range): Same.
30688 (foperator_unordered_ge::op2_range): Same.
30690 2023-02-03 Andrew MacLeod <amacleod@redhat.com>
30692 PR tree-optimization/107570
30693 * tree-vrp.cc (remove_and_update_globals): Reset SCEV.
30695 2023-02-03 Gaius Mulley <gaiusmod2@gmail.com>
30697 * doc/gm2.texi (Internals): Remove from menu.
30698 (Using): Comment out ifnohtml conditional.
30699 (Documentation): Use gcc url.
30700 (License): Node simplified.
30701 (Copying): New node. Include gpl_v3_without_node.
30702 (Contributing): Node simplified.
30703 (Internals): Commented out.
30704 (Libraries): Node simplified.
30707 (Functions): Ditto.
30709 2023-02-03 Christophe Lyon <christophe.lyon@arm.com>
30711 * config/arm/mve.md (mve_vabavq_p_<supf><mode>): Add length
30713 (mve_vqshluq_m_n_s<mode>): Likewise.
30714 (mve_vshlq_m_<supf><mode>): Likewise.
30715 (mve_vsriq_m_n_<supf><mode>): Likewise.
30716 (mve_vsubq_m_<supf><mode>): Likewise.
30718 2023-02-03 Martin Jambor <mjambor@suse.cz>
30721 * ipa-sra.cc (push_param_adjustments_for_index): Remove a size check
30722 when comparing to an IPA-CP value.
30723 (dump_list_of_param_indices): New function.
30724 (adjust_parameter_descriptions): Check for mismatching IPA-CP values.
30725 Dump removed candidates using dump_list_of_param_indices.
30726 * ipa-param-manipulation.cc
30727 (ipa_param_body_adjustments::modify_expression): Add assert checking
30728 sizes of a VIEW_CONVERT_EXPR will match.
30729 (ipa_param_body_adjustments::modify_assignment): Likewise.
30731 2023-02-03 Monk Chiang <monk.chiang@sifive.com>
30733 * config/riscv/riscv.h: Remove VL_REGS, VTYPE_REGS class.
30734 * config/riscv/riscv.cc: Ditto.
30736 2023-02-03 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30738 * config/riscv/vector-iterators.md (sll.vi): Fix constraint bug.
30742 * config/riscv/vector.md: Ditto.
30744 2023-02-03 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30746 * config/riscv/predicates.md (pmode_reg_or_uimm5_operand): New predicate.
30747 * config/riscv/riscv-vector-builtins-bases.cc: New class.
30748 * config/riscv/riscv-vector-builtins-functions.def (vsll): Ditto.
30751 * config/riscv/riscv-vector-builtins.cc: Ditto.
30752 * config/riscv/vector.md (@pred_<optab><mode>_scalar): New pattern.
30754 2023-02-02 Iain Sandoe <iain@sandoe.co.uk>
30756 * toplev.cc (toplev::main): Only print the version information header
30757 from toplevel main().
30759 2023-02-02 Paul-Antoine Arras <pa@codesourcery.com>
30761 * config/gcn/gcn-valu.md (cond_<expander><mode>): Add
30762 cond_{ashl|ashr|lshr}
30764 2023-02-02 Richard Sandiford <richard.sandiford@arm.com>
30766 PR rtl-optimization/108086
30767 * rtl-ssa/insns.h (insn_info): Make m_num_defs a full unsigned int.
30768 Adjust size-related commentary accordingly.
30770 2023-02-02 Richard Sandiford <richard.sandiford@arm.com>
30772 PR rtl-optimization/108508
30773 * rtl-ssa/accesses.cc (function_info::split_clobber_group): When
30774 the splay tree search gives the first clobber in the second group,
30775 make sure that the root of the first clobber group is updated
30776 correctly. Enter the new clobber group into the definition splay
30779 2023-02-02 Jin Ma <jinma@linux.alibaba.com>
30781 * common/config/riscv/riscv-common.cc (riscv_compute_multilib):
30782 Fix finding best match score.
30784 2023-02-02 Jakub Jelinek <jakub@redhat.com>
30787 PR rtl-optimization/108463
30789 * cselib.cc (cselib_current_insn): Move declaration earlier.
30790 (cselib_hasher::equal): For debug only locs, temporarily override
30791 cselib_current_insn to their l->setting_insn for the
30792 rtx_equal_for_cselib_1 call, so that unsuccessful comparisons don't
30793 promote some debug locs.
30794 * sched-deps.cc (sched_analyze_2) <case MEM>: For MEMs in DEBUG_INSNs
30795 when using cselib call cselib_lookup_from_insn on the address but
30796 don't substitute it.
30798 2023-02-02 Richard Biener <rguenther@suse.de>
30800 PR middle-end/108625
30801 * genmatch.cc (expr::gen_transform): Also disallow resimplification
30802 from pushing to lseq with force_leaf.
30803 (dt_simplify::gen_1): Likewise.
30805 2023-02-02 Andrew Stubbs <ams@codesourcery.com>
30807 * config/gcn/gcn-run.cc: Include libgomp-gcn.h.
30808 (struct kernargs): Replace the common content with kernargs_abi.
30809 (struct heap): Delete.
30810 (main): Read GCN_STACK_SIZE envvar.
30811 Allocate space for the device stacks.
30812 Write the new kernargs fields.
30813 * config/gcn/gcn.cc (gcn_option_override): Remove stack_size_opt.
30814 (default_requested_args): Remove PRIVATE_SEGMENT_BUFFER_ARG and
30815 PRIVATE_SEGMENT_WAVE_OFFSET_ARG.
30816 (gcn_addr_space_convert): Mask the QUEUE_PTR_ARG content.
30817 (gcn_expand_prologue): Move the TARGET_PACKED_WORK_ITEMS to the top.
30818 Set up the stacks from the values in the kernargs, not private.
30819 (gcn_expand_builtin_1): Match the stack configuration in the prologue.
30820 (gcn_hsa_declare_function_name): Turn off the private segment.
30821 (gcn_conditional_register_usage): Ensure QUEUE_PTR is fixed.
30822 * config/gcn/gcn.h (FIXED_REGISTERS): Fix the QUEUE_PTR register.
30823 * config/gcn/gcn.opt (mstack-size): Change the description.
30825 2023-02-02 Andre Vieira <andre.simoesdiasvieira@arm.com>
30828 * config/arm/arm.h (VALID_MVE_PRED_MODE): Add V2QI.
30829 * config/arm/arm.cc (thumb2_legitimate_address_p): Use HImode for
30830 addressing MVE predicate modes.
30831 (mve_bool_vec_to_const): Change to represent correct MVE predicate
30833 (arm_hard_regno_mode_ok): Use VALID_MVE_PRED_MODE instead of checking
30835 (arm_vector_mode_supported_p): Likewise.
30836 (arm_mode_to_pred_mode): Add V2QI.
30837 * config/arm/arm-builtins.cc (UNOP_PRED_UNONE_QUALIFIERS): New
30839 (UNOP_PRED_PRED_QUALIFIERS): New qualifier
30840 (BINOP_PRED_UNONE_PRED_QUALIFIERS): New qualifier.
30841 (v2qi_UP): New macro.
30842 (v4bi_UP): New macro.
30843 (v8bi_UP): New macro.
30844 (v16bi_UP): New macro.
30845 (arm_expand_builtin_args): Make it able to expand the new predicate
30847 * config/arm/arm-modes.def (V2QI): New mode.
30848 * config/arm/arm-simd-builtin-types.def (Pred1x16_t, Pred2x8_t
30849 Pred4x4_t): Remove unused predicate builtin types.
30850 * config/arm/arm_mve.h (__arm_vctp16q, __arm_vctp32q, __arm_vctp64q,
30851 __arm_vctp8q, __arm_vpnot, __arm_vctp8q_m, __arm_vctp64q_m,
30852 __arm_vctp32q_m, __arm_vctp16q_m): Use predicate modes.
30853 * config/arm/arm_mve_builtins.def (vctp16q, vctp32q, vctp64q, vctp8q,
30854 vpnot, vctp8q_m, vctp16q_m, vctp32q_m, vctp64q_m): Likewise.
30855 * config/arm/constraints.md (DB): Check for VALID_MVE_PRED_MODE instead
30856 of MODE_VECTOR_BOOL.
30857 * config/arm/iterators.md (MVE_7, MVE_7_HI): Add V2QI
30858 (MVE_VPRED): Likewise.
30859 (MVE_vpred): Add V2QI and map upper case predicate modes to lower case.
30860 (MVE_vctp): New mode attribute.
30864 * config/arm/mve.md (mve_vctp<mode1>qhi): Rename this...
30865 (mve_vctp<MVE_vctp>q<MVE_vpred>): ... to this. And use new mode
30867 (mve_vpnothi): Rename this...
30868 (mve_vpnotv16bi): ... to this.
30869 (mve_vctp<mode1>q_mhi): Rename this...
30870 (mve_vctp<MVE_vctp>q_m<MVE_vpred>):... to this.
30871 (mve_vldrdq_gather_base_z_<supf>v2di,
30872 mve_vldrdq_gather_offset_z_<supf>v2di,
30873 mve_vldrdq_gather_shifted_offset_z_<supf>v2di,
30874 mve_vstrdq_scatter_base_p_<supf>v2di,
30875 mve_vstrdq_scatter_offset_p_<supf>v2di,
30876 mve_vstrdq_scatter_offset_p_<supf>v2di_insn,
30877 mve_vstrdq_scatter_shifted_offset_p_<supf>v2di,
30878 mve_vstrdq_scatter_shifted_offset_p_<supf>v2di_insn,
30879 mve_vstrdq_scatter_base_wb_p_<supf>v2di,
30880 mve_vldrdq_gather_base_wb_z_<supf>v2di,
30881 mve_vldrdq_gather_base_nowb_z_<supf>v2di,
30882 mve_vldrdq_gather_base_wb_z_<supf>v2di_insn): Use V2QI insead of HI for
30884 * config/arm/unspecs.md (VCTP8Q, VCTP16Q, VCTP32Q, VCTP64Q): Replace
30886 (VCTP): ... with this.
30887 (VCTP8Q_M, VCTP16Q_M, VCTP32Q_M, VCTP64Q_M): Replace these...
30888 (VCTP_M): ... with this.
30889 * config/arm/vfp.md (*thumb2_movhi_vfp, *thumb2_movhi_fp16): Use
30890 VALID_MVE_PRED_MODE instead of checking for MODE_VECTOR_BOOL class.
30892 2023-02-02 Andre Vieira <andre.simoesdiasvieira@arm.com>
30895 * config/arm/arm.cc (arm_hard_regno_mode_ok): Use new MACRO.
30896 (arm_modes_tieable_p): Make MVE predicate modes tieable.
30897 * config/arm/arm.h (VALID_MVE_PRED_MODE): New define.
30898 * simplify-rtx.cc (simplify_context::simplify_subreg): Teach
30899 simplify_subreg to simplify subregs where the outermode is not scalar.
30901 2023-02-02 Andre Vieira <andre.simoesdiasvieira@arm.com>
30904 * config/arm/arm-builtins.cc (arm_simd_builtin_type): Rewrite to use
30905 new qualifiers parameter and use unsigned short type for MVE predicate.
30906 (arm_init_builtin): Call arm_simd_builtin_type with qualifiers
30908 (arm_init_crypto_builtins): Likewise.
30910 2023-02-02 Jakub Jelinek <jakub@redhat.com>
30913 * builtins.def (BUILT_IN_UNREACHABLE_TRAP): New builtin.
30914 * internal-fn.def (TRAP): Remove.
30915 * internal-fn.cc (expand_TRAP): Remove.
30916 * tree.cc (build_common_builtin_nodes): Define
30917 BUILT_IN_UNREACHABLE_TRAP if not yet defined.
30918 (builtin_decl_unreachable): Use BUILT_IN_UNREACHABLE_TRAP
30919 instead of BUILT_IN_TRAP.
30920 * gimple.cc (gimple_build_builtin_unreachable): Remove
30921 emitting internal function for BUILT_IN_TRAP.
30922 * asan.cc (maybe_instrument_call): Handle BUILT_IN_UNREACHABLE_TRAP.
30923 * cgraph.cc (cgraph_edge::verify_corresponds_to_fndecl): Handle
30924 BUILT_IN_UNREACHABLE_TRAP instead of BUILT_IN_TRAP.
30925 * ipa-devirt.cc (possible_polymorphic_call_target_p): Handle
30926 BUILT_IN_UNREACHABLE_TRAP.
30927 * builtins.cc (expand_builtin, is_inexpensive_builtin): Likewise.
30928 * tree-cfg.cc (verify_gimple_call,
30929 pass_warn_function_return::execute): Likewise.
30930 * attribs.cc (decl_attributes): Don't report exclusions on
30931 BUILT_IN_UNREACHABLE_TRAP either.
30933 2023-02-02 liuhongt <hongtao.liu@intel.com>
30935 PR tree-optimization/108601
30936 * tree-vectorizer.h (vect_can_peel_nonlinear_iv_p): Removed.
30937 * tree-vect-loop.cc
30938 (vectorizable_nonlinear_induction): Remove
30939 vect_can_peel_nonlinear_iv_p.
30940 (vect_can_peel_nonlinear_iv_p): Don't peel
30941 nonlinear iv(mult or shift) for epilog when vf is not
30942 constant and moved the defination to ..
30943 * tree-vect-loop-manip.cc (vect_can_peel_nonlinear_iv_p):
30946 2023-02-02 Jakub Jelinek <jakub@redhat.com>
30948 PR middle-end/108435
30949 * tree-nested.cc (convert_nonlocal_omp_clauses)
30950 <case OMP_CLAUSE_LASTPRIVATE>: If info->new_local_var_chain and *seq
30951 is not a GIMPLE_BIND, wrap the sequence into a new GIMPLE_BIND
30952 before calling declare_vars.
30953 (convert_nonlocal_omp_clauses) <case OMP_CLAUSE_LINEAR>: Merge
30954 with the OMP_CLAUSE_LASTPRIVATE handling except for whether
30955 seq is initialized to &OMP_CLAUSE_LASTPRIVATE_GIMPLE_SEQ (clause)
30956 or &OMP_CLAUSE_LINEAR_GIMPLE_SEQ (clause).
30958 2023-02-01 Tamar Christina <tamar.christina@arm.com>
30960 * common/config/aarch64/aarch64-common.cc
30961 (struct aarch64_option_extension): Add native_detect and document struct
30963 (all_extensions): Set new field native_detect.
30964 * config/aarch64/aarch64.cc (struct aarch64_option_extension): Delete
30967 2023-02-01 Martin Liska <mliska@suse.cz>
30969 * ipa-devirt.cc (odr_types_equivalent_p): Respect *warned
30972 2023-02-01 Andrew MacLeod <amacleod@redhat.com>
30974 PR tree-optimization/108356
30975 * gimple-range-cache.cc (ranger_cache::range_on_edge): Always
30976 do a search of the DOM tree for a range.
30978 2023-02-01 Martin Liska <mliska@suse.cz>
30981 * cgraphunit.cc (walk_polymorphic_call_targets): Insert
30982 ony non-null values.
30983 * ipa.cc (walk_polymorphic_call_targets): Likewise.
30985 2023-02-01 Martin Liska <mliska@suse.cz>
30988 * gcc.cc (LINK_COMPRESS_DEBUG_SPEC): Report error only for
30991 2023-02-01 Jakub Jelinek <jakub@redhat.com>
30994 * ree.cc (combine_reaching_defs): Don't return false for paradoxical
30995 subregs in DEBUG_INSNs.
30997 2023-02-01 Richard Sandiford <richard.sandiford@arm.com>
30999 * compare-elim.cc (find_flags_uses_in_insn): Guard use of SET_SRC.
31001 2023-02-01 Andreas Krebbel <krebbel@linux.ibm.com>
31003 * config/s390/s390.cc (s390_restore_gpr_p): New function.
31004 (s390_preserve_gpr_arg_in_range_p): New function.
31005 (s390_preserve_gpr_arg_p): New function.
31006 (s390_preserve_fpr_arg_p): New function.
31007 (s390_register_info_stdarg_fpr): Rename to ...
31008 (s390_register_info_arg_fpr): ... this. Add -mpreserve-args handling.
31009 (s390_register_info_stdarg_gpr): Rename to ...
31010 (s390_register_info_arg_gpr): ... this. Add -mpreserve-args handling.
31011 (s390_register_info): Use the renamed functions above.
31012 (s390_optimize_register_info): Likewise.
31013 (save_fpr): Generate CFI for -mpreserve-args.
31014 (save_gprs): Generate CFI for -mpreserve-args. Drop return value.
31015 (s390_emit_prologue): Adjust to changed calling convention of save_gprs.
31016 (s390_optimize_prologue): Likewise.
31017 * config/s390/s390.opt: New option -mpreserve-args
31019 2023-02-01 Andreas Krebbel <krebbel@linux.ibm.com>
31021 * config/s390/s390.cc (save_gprs): Use gen_frame_mem.
31022 (restore_gprs): Likewise.
31023 (s390_emit_stack_tie): Make the stack_tie to be dependent on the
31024 frame pointer if a frame-pointer is used.
31025 (s390_emit_prologue): Emit stack_tie when frame-pointer is needed.
31026 * config/s390/s390.md (stack_tie): Add a register operand and
31028 (@stack_tie<mode>): ... this.
31030 2023-02-01 Andreas Krebbel <krebbel@linux.ibm.com>
31032 * dwarf2cfi.cc (dwarf2out_frame_debug_cfa_restore): Add
31033 EMIT_CFI parameter.
31034 (dwarf2out_frame_debug): Add case for REG_CFA_NORESTORE.
31035 * reg-notes.def (REG_CFA_NOTE): New reg note definition.
31037 2023-02-01 Richard Biener <rguenther@suse.de>
31039 PR middle-end/108500
31040 * dominance.cc (assign_dfs_numbers): Replace recursive DFS
31041 with tree traversal algorithm.
31043 2023-02-01 Jason Merrill <jason@redhat.com>
31045 * doc/invoke.texi: Document -Wno-changes-meaning.
31047 2023-02-01 David Malcolm <dmalcolm@redhat.com>
31049 * doc/invoke.texi (Static Analyzer Options): Add notes about
31050 limitations of -fanalyzer.
31052 2023-01-31 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31054 * config/riscv/constraints.md (vj): New.
31056 * config/riscv/iterators.md: Add more opcode.
31057 * config/riscv/predicates.md (vector_arith_operand): New.
31058 (vector_neg_arith_operand): New.
31059 (vector_shift_operand): New.
31060 * config/riscv/riscv-vector-builtins-bases.cc (class binop): New.
31061 * config/riscv/riscv-vector-builtins-bases.h: (vadd): New.
31078 * config/riscv/riscv-vector-builtins-functions.def (vadd): New.
31095 * config/riscv/riscv-vector-builtins-shapes.cc (struct binop_def): New.
31096 * config/riscv/riscv-vector-builtins-shapes.h (binop): New.
31097 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_I_OPS): New.
31098 (DEF_RVV_U_OPS): New.
31099 (rvv_arg_type_info::get_base_vector_type): Handle
31100 RVV_BASE_shift_vector.
31101 (rvv_arg_type_info::get_tree_type): Ditto.
31102 * config/riscv/riscv-vector-builtins.h (enum rvv_base_type): Add
31103 RVV_BASE_shift_vector.
31104 * config/riscv/riscv.cc (riscv_print_operand): Handle 'V'.
31105 * config/riscv/vector-iterators.md: Handle more opcode.
31106 * config/riscv/vector.md (@pred_<optab><mode>): New.
31108 2023-01-31 Philipp Tomsich <philipp.tomsich@vrull.eu>
31111 * config/aarch64/aarch64.cc (aarch_macro_fusion_pair_p): Check
31114 2023-01-31 Richard Sandiford <richard.sandiford@arm.com>
31116 PR tree-optimization/108608
31117 * tree-vect-loop.cc (vect_transform_reduction): Handle single
31118 def-use cycles that involve function calls rather than tree codes.
31120 2023-01-31 Andrew MacLeod <amacleod@redhat.com>
31122 PR tree-optimization/108385
31123 * gimple-range-gori.cc (gori_compute::compute_operand_range):
31124 Allow VARYING computations to continue if there is a relation.
31125 * range-op.cc (pointer_plus_operator::op2_range): New.
31127 2023-01-31 Andrew MacLeod <amacleod@redhat.com>
31129 PR tree-optimization/108359
31130 * range-op.cc (range_operator::wi_fold_in_parts_equiv): New.
31131 (range_operator::fold_range): If op1 is equivalent to op2 then
31132 invoke new fold_in_parts_equiv to operate on sub-components.
31133 * range-op.h (wi_fold_in_parts_equiv): New prototype.
31135 2023-01-31 Andrew MacLeod <amacleod@redhat.com>
31137 * gimple-range-gori.cc (gori_compute::compute_operand_range): Do
31138 not abort calculations if there is a valid relation available.
31139 (gori_compute::refine_using_relation): Pass correct relation trio.
31140 (gori_compute::compute_operand1_range): Create trio and use it.
31141 (gori_compute::compute_operand2_range): Ditto.
31142 * range-op.cc (operator_plus::op1_range): Use correct trio member.
31143 (operator_minus::op1_range): Use correct trio member.
31144 * value-relation.cc (value_relation::create_trio): New.
31145 * value-relation.h (value_relation::create_trio): New prototype.
31147 2023-01-31 Jakub Jelinek <jakub@redhat.com>
31150 * config/i386/i386-expand.cc
31151 (ix86_convert_const_wide_int_to_broadcast): Return nullptr if
31152 CONST_WIDE_INT_NUNITS (op) times HOST_BITS_PER_WIDE_INT isn't
31153 equal to bitsize of mode.
31155 2023-01-31 Jakub Jelinek <jakub@redhat.com>
31157 PR rtl-optimization/108596
31158 * bb-reorder.cc (fix_up_fall_thru_edges): Handle the case where cur_bb
31159 ends with asm goto and has a crossing fallthrough edge to the same bb
31160 that contains at least one of its labels by restoring EDGE_CROSSING
31161 flag even on possible edge from cur_bb to new_bb successor.
31163 2023-01-31 Jakub Jelinek <jakub@redhat.com>
31166 * config/i386/avx512erintrin.h (_mm512_exp2a23_round_pd,
31167 _mm512_exp2a23_round_ps, _mm512_rcp28_round_pd, _mm512_rcp28_round_ps,
31168 _mm512_rsqrt28_round_pd, _mm512_rsqrt28_round_ps): Use
31169 _mm512_undefined_pd () or _mm512_undefined_ps () instead of using
31170 uninitialized automatic variable __W.
31172 2023-01-31 Gerald Pfeifer <gerald@pfeifer.com>
31174 * doc/include/fdl.texi: Change fsf.org to www.fsf.org.
31176 2023-01-30 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31178 * config/riscv/riscv-protos.h (get_vector_mode): New function.
31179 * config/riscv/riscv-v.cc (get_vector_mode): Ditto.
31180 * config/riscv/riscv-vector-builtins-bases.cc (enum lst_type): New enum.
31181 (class loadstore): Adjust for indexed loads/stores support.
31183 * config/riscv/riscv-vector-builtins-bases.h: New function declare.
31184 * config/riscv/riscv-vector-builtins-functions.def (vluxei8): Ditto.
31200 * config/riscv/riscv-vector-builtins-shapes.cc
31201 (struct indexed_loadstore_def): New class.
31203 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
31204 * config/riscv/riscv-vector-builtins.cc (required_extensions_p): Adjust
31205 for indexed loads/stores support.
31206 (check_required_extensions): Ditto.
31207 (rvv_arg_type_info::get_base_vector_type): New function.
31208 (rvv_arg_type_info::get_tree_type): Ditto.
31209 (function_builder::add_unique_function): Adjust for indexed loads/stores
31211 (function_expander::use_exact_insn): New function.
31212 * config/riscv/riscv-vector-builtins.h (enum rvv_base_type): Adjust for
31213 indexed loads/stores support.
31214 (struct rvv_arg_type_info): Ditto.
31215 (function_expander::index_mode): New function.
31216 (function_base::apply_tail_policy_p): Ditto.
31217 (function_base::apply_mask_policy_p): Ditto.
31218 * config/riscv/vector-iterators.md (unspec): New unspec.
31219 * config/riscv/vector.md (unspec): Ditto.
31220 (@pred_indexed_<order>load<VNX1_QHSD:mode><VNX1_QHSDI:mode>): New
31222 (@pred_indexed_<order>store<VNX1_QHSD:mode><VNX1_QHSDI:mode>): Ditto.
31223 (@pred_indexed_<order>load<VNX2_QHSD:mode><VNX2_QHSDI:mode>): Ditto.
31224 (@pred_indexed_<order>store<VNX2_QHSD:mode><VNX2_QHSDI:mode>): Ditto.
31225 (@pred_indexed_<order>load<VNX4_QHSD:mode><VNX4_QHSDI:mode>): Ditto.
31226 (@pred_indexed_<order>store<VNX4_QHSD:mode><VNX4_QHSDI:mode>): Ditto.
31227 (@pred_indexed_<order>load<VNX8_QHSD:mode><VNX8_QHSDI:mode>): Ditto.
31228 (@pred_indexed_<order>store<VNX8_QHSD:mode><VNX8_QHSDI:mode>): Ditto.
31229 (@pred_indexed_<order>load<VNX16_QHS:mode><VNX16_QHSI:mode>): Ditto.
31230 (@pred_indexed_<order>store<VNX16_QHS:mode><VNX16_QHSI:mode>): Ditto.
31231 (@pred_indexed_<order>load<VNX32_QH:mode><VNX32_QHI:mode>): Ditto.
31232 (@pred_indexed_<order>store<VNX32_QH:mode><VNX32_QHI:mode>): Ditto.
31233 (@pred_indexed_<order>load<VNX64_Q:mode><VNX64_Q:mode>): Ditto.
31234 (@pred_indexed_<order>store<VNX64_Q:mode><VNX64_Q:mode>): Ditto.
31236 2023-01-30 Flavio Cruz <flaviocruz@gmail.com>
31238 * config.gcc: Recognize x86_64-*-gnu* targets and include
31240 * config/i386/gnu64.h: Define configuration for new target
31241 including ld.so location.
31243 2023-01-30 Philipp Tomsich <philipp.tomsich@vrull.eu>
31245 * config/aarch64/aarch64-cores.def (AARCH64_CORE): Update
31246 ampere1a to include SM4.
31248 2023-01-30 Andrew Pinski <apinski@marvell.com>
31250 PR tree-optimization/108582
31251 * tree-ssa-phiopt.cc (match_simplify_replacement): Add check
31252 for middlebb to have no phi nodes.
31254 2023-01-30 Richard Biener <rguenther@suse.de>
31256 PR tree-optimization/108574
31257 * tree-ssa-sccvn.cc (visit_phi): Instead of swapping
31258 sameval and def, ignore the equivalence if there's the
31259 danger of oscillating between two values.
31261 2023-01-30 Andreas Schwab <schwab@suse.de>
31263 * common/config/riscv/riscv-common.cc
31264 (riscv_option_optimization_table)
31265 [TARGET_DEFAULT_ASYNC_UNWIND_TABLES]: Enable
31266 -fasynchronous-unwind-tables and -funwind-tables.
31267 * config.gcc (riscv*-*-linux*): Define
31268 TARGET_DEFAULT_ASYNC_UNWIND_TABLES.
31270 2023-01-30 YunQiang Su <yunqiang.su@cipunited.com>
31272 * Makefile.in (CROSS_SYSTEM_HEADER_DIR): set according the
31273 value of includedir.
31275 2023-01-30 Richard Biener <rguenther@suse.de>
31278 * cgraph.cc (possibly_call_in_translation_unit_p): Relax
31281 2023-01-30 liuhongt <hongtao.liu@intel.com>
31283 * config/i386/i386.opt: Change AVX512FP16 to AVX512-FP16.
31284 * doc/invoke.texi: Ditto.
31286 2023-01-29 Jan Hubicka <hubicka@ucw.cz>
31288 * ipa-utils.cc: Include calls.h, cfgloop.h and cfganal.h
31289 (stmt_may_terminate_function_p): If assuming return or EH
31290 volatile asm is safe.
31291 (find_always_executed_bbs): Fix handling of terminating BBS and
31292 infinite loops; add debug output.
31293 * tree-ssa-alias.cc (stmt_kills_ref_p): Fix debug output
31295 2023-01-28 Philipp Tomsich <philipp.tomsich@vrull.eu>
31297 * config/aarch64/aarch64.cc (aarch64_uxt_size): fix an
31298 off-by-one in checking the permissible shift-amount.
31300 2023-01-28 Gerald Pfeifer <gerald@pfeifer.com>
31302 * doc/extend.texi (Named Address Spaces): Update link to the
31305 2023-01-28 Gerald Pfeifer <gerald@pfeifer.com>
31307 * doc/standards.texi (Standards): Fix markup.
31309 2023-01-28 Gerald Pfeifer <gerald@pfeifer.com>
31311 * doc/standards.texi (Standards): Update link to Objective-C book.
31313 2023-01-28 Gerald Pfeifer <gerald@pfeifer.com>
31315 * doc/invoke.texi (Instrumentation Options): Update reference to
31318 2023-01-28 Gerald Pfeifer <gerald@pfeifer.com>
31320 * doc/standards.texi: Update Go1 link.
31322 2023-01-28 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31324 * config/riscv/predicates.md (pmode_reg_or_0_operand): New predicate.
31325 * config/riscv/riscv-vector-builtins-bases.cc (class loadstore):
31328 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
31329 * config/riscv/riscv-vector-builtins-functions.def (vlse): New class.
31331 * config/riscv/riscv-vector-builtins.cc
31332 (function_expander::use_contiguous_load_insn): Support vlse/vsse.
31333 * config/riscv/vector.md (@pred_strided_load<mode>): New md pattern.
31334 (@pred_strided_store<mode>): Ditto.
31336 2023-01-28 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31338 * config/riscv/vector.md (tail_policy_op_idx): Remove.
31339 (mask_policy_op_idx): Remove.
31340 (avl_type_op_idx): Remove.
31342 2023-01-27 Richard Sandiford <richard.sandiford@arm.com>
31344 PR tree-optimization/96373
31345 * tree.h (sign_mask_for): Declare.
31346 * tree.cc (sign_mask_for): New function.
31347 (signed_or_unsigned_type_for): For vector types, try to use the
31348 related_int_vector_mode.
31349 * genmatch.cc (commutative_op): Handle conditional internal functions.
31350 * match.pd: Fold an IFN_COND_MUL+copysign into an IFN_COND_XOR+and.
31352 2023-01-27 Richard Sandiford <richard.sandiford@arm.com>
31354 * tree-vectorizer.cc (vector_costs::compare_inside_loop_cost):
31355 Use the likely minimum VF when bounding the denominators to
31356 the estimated number of iterations.
31358 2023-01-27 Richard Biener <rguenther@suse.de>
31361 * doc/invoke.texi (-shared): Clarify effect on -ffast-math
31362 and -Ofast FP environment side-effects.
31364 2023-01-27 Richard Biener <rguenther@suse.de>
31367 * config/mips/gnu-user.h (GNU_USER_TARGET_MATHFILE_SPEC):
31368 Don't add crtfastmath.o for -shared.
31370 2023-01-27 Richard Biener <rguenther@suse.de>
31373 * config/ia64/linux.h (ENDFILE_SPEC): Don't add crtfastmath.o
31376 2023-01-27 Richard Biener <rguenther@suse.de>
31379 * config/alpha/linux.h (ENDFILE_SPEC): Don't add
31380 crtfastmath.o for -shared.
31382 2023-01-27 Andrew MacLeod <amacleod@redhat.com>
31384 PR tree-optimization/108306
31385 * range-op.cc (operator_lshift::fold_range): Return [0, 0] not
31386 varying for shifts that are always out of void range.
31387 (operator_rshift::fold_range): Return [0, 0] not
31388 varying for shifts that are always out of void range.
31390 2023-01-27 Andrew MacLeod <amacleod@redhat.com>
31392 PR tree-optimization/108447
31393 * gimple-range-fold.cc (old_using_range::relation_fold_and_or):
31394 Do not attempt to fold HONOR_NAN types.
31396 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31398 * config/riscv/riscv-vector-builtins-shapes.cc (struct loadstore_def):
31399 Remove _m suffix for "vop_m" C++ overloaded API name.
31401 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31403 * config/riscv/riscv-vector-builtins-bases.cc (BASE): Add vlm/vsm support.
31404 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
31405 * config/riscv/riscv-vector-builtins-functions.def (vlm): New define.
31407 * config/riscv/riscv-vector-builtins-shapes.cc (struct loadstore_def): Add vlm/vsm support.
31408 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_B_OPS): Ditto.
31409 (vbool64_t): Ditto.
31410 (vbool32_t): Ditto.
31411 (vbool16_t): Ditto.
31416 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_B_OPS): Ditto.
31417 (rvv_arg_type_info::get_tree_type): Ditto.
31418 (function_expander::use_contiguous_load_insn): Ditto.
31419 * config/riscv/vector.md (@pred_store<mode>): Ditto.
31421 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31423 * config/riscv/riscv-vsetvl.cc (vsetvl_insn_p): Add condition to avoid ICE.
31424 (vsetvl_discard_result_insn_p): New function.
31425 (reg_killed_by_bb_p): rename to find_reg_killed_by.
31426 (find_reg_killed_by): New name.
31427 (get_vl): allow it to be called by more functions.
31428 (has_vsetvl_killed_avl_p): Add condition.
31429 (get_avl): allow it to be called by more functions.
31430 (insn_should_be_added_p): New function.
31431 (get_all_nonphi_defs): Refine function.
31432 (get_all_sets): Ditto.
31433 (get_same_bb_set): New function.
31434 (any_insn_in_bb_p): Ditto.
31435 (any_set_in_bb_p): Ditto.
31436 (get_vl_vtype_info): Add VLMAX forward optimization.
31437 (source_equal_p): Fix issues.
31438 (extract_single_source): Refine.
31439 (avl_info::multiple_source_equal_p): New function.
31440 (avl_info::operator==): Adjust for final version.
31441 (vl_vtype_info::operator==): Ditto.
31442 (vl_vtype_info::same_avl_p): Ditto.
31443 (vector_insn_info::parse_insn): Ditto.
31444 (vector_insn_info::available_p): New function.
31445 (vector_insn_info::merge): Adjust for final version.
31446 (vector_insn_info::dump): Add hard_empty.
31447 (pass_vsetvl::hard_empty_block_p): New function.
31448 (pass_vsetvl::backward_demand_fusion): Adjust for final version.
31449 (pass_vsetvl::forward_demand_fusion): Ditto.
31450 (pass_vsetvl::demand_fusion): Ditto.
31451 (pass_vsetvl::cleanup_illegal_dirty_blocks): New function.
31452 (pass_vsetvl::compute_local_properties): Adjust for final version.
31453 (pass_vsetvl::can_refine_vsetvl_p): Ditto.
31454 (pass_vsetvl::refine_vsetvls): Ditto.
31455 (pass_vsetvl::commit_vsetvls): Ditto.
31456 (pass_vsetvl::propagate_avl): New function.
31457 (pass_vsetvl::lazy_vsetvl): Adjust for new version.
31458 * config/riscv/riscv-vsetvl.h (enum def_type): New enum.
31460 2023-01-27 Jakub Jelinek <jakub@redhat.com>
31463 * doc/extend.texi: Fix up return type of __builtin_va_arg_pack_len
31464 from size_t to int.
31466 2023-01-27 Jakub Jelinek <jakub@redhat.com>
31469 * cgraph.cc (cgraph_edge::verify_corresponds_to_fndecl): Allow
31470 redirection of calls to __builtin_trap in addition to redirection
31471 to __builtin_unreachable.
31473 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31475 * config/riscv/riscv-vsetvl.cc (before_p): Fix bug.
31477 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31479 * config/riscv/riscv-vsetvl.cc (gen_vsetvl_pat): Refine function args.
31480 (emit_vsetvl_insn): Ditto.
31482 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31484 * config/riscv/vector.md: Fix constraints.
31486 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31488 * config/riscv/vector-iterators.md: Add TARGET_MIN_VLEN > 32 predicates.
31490 2023-01-27 Patrick Palka <ppalka@redhat.com>
31491 Jakub Jelinek <jakub@redhat.com>
31493 * tree-core.h (tree_code_type, tree_code_length): For
31494 C++17 and later, add inline keyword, otherwise don't define
31495 the arrays, but declare extern arrays.
31496 * tree.cc (tree_code_type, tree_code_length): Define these
31497 arrays for C++14 and older.
31499 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31501 * config/riscv/riscv-vsetvl.h: Change it into public.
31503 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31505 * config/riscv/riscv-passes.def (INSERT_PASS_BEFORE): Reorder VSETVL
31508 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31510 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::execute): Always call split_all_insns.
31512 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31514 * config/riscv/vector.md: Fix incorrect attributes.
31516 2023-01-27 Richard Biener <rguenther@suse.de>
31519 * config/loongarch/gnu-user.h (GNU_USER_TARGET_MATHFILE_SPEC):
31520 Don't add crtfastmath.o for -shared.
31522 2023-01-27 Alexandre Oliva <oliva@gnu.org>
31524 * doc/options.texi (option, RejectNegative): Mention that
31525 -g-started options are also implicitly negatable.
31527 2023-01-26 Kito Cheng <kito.cheng@sifive.com>
31529 * config/riscv/riscv-vector-builtins.cc (register_builtin_types):
31530 Use get_typenode_from_name to get fixed-width integer type
31532 * config/riscv/riscv-vector-builtins.def: Update define with
31533 fixed-width integer type nodes.
31535 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31537 * config/riscv/riscv-vsetvl.cc (same_bb_and_before_p): Remove it.
31538 (real_insn_and_same_bb_p): New function.
31539 (same_bb_and_after_or_equal_p): Remove it.
31540 (before_p): New function.
31541 (reg_killed_by_bb_p): Ditto.
31542 (has_vsetvl_killed_avl_p): Ditto.
31543 (get_vl): Move location so that we can call it.
31544 (anticipatable_occurrence_p): Fix issue of AVL=REG support.
31545 (available_occurrence_p): Ditto.
31546 (dominate_probability_p): Remove it.
31547 (can_backward_propagate_p): Remove it.
31548 (get_all_nonphi_defs): New function.
31549 (get_all_predecessors): Ditto.
31550 (any_insn_in_bb_p): Ditto.
31551 (insert_vsetvl): Adjust AVL REG.
31552 (source_equal_p): New function.
31553 (extract_single_source): Ditto.
31554 (avl_info::single_source_equal_p): Ditto.
31555 (avl_info::operator==): Adjust for AVL=REG.
31556 (vl_vtype_info::same_avl_p): Ditto.
31557 (vector_insn_info::set_demand_info): Remove it.
31558 (vector_insn_info::compatible_p): Adjust for AVL=REG.
31559 (vector_insn_info::compatible_avl_p): New function.
31560 (vector_insn_info::merge): Adjust AVL=REG.
31561 (vector_insn_info::dump): Ditto.
31562 (pass_vsetvl::merge_successors): Remove it.
31563 (enum fusion_type): New enum.
31564 (pass_vsetvl::get_backward_fusion_type): New function.
31565 (pass_vsetvl::backward_demand_fusion): Adjust for AVL=REG.
31566 (pass_vsetvl::forward_demand_fusion): Ditto.
31567 (pass_vsetvl::demand_fusion): Ditto.
31568 (pass_vsetvl::prune_expressions): Ditto.
31569 (pass_vsetvl::compute_local_properties): Ditto.
31570 (pass_vsetvl::cleanup_vsetvls): Ditto.
31571 (pass_vsetvl::commit_vsetvls): Ditto.
31572 (pass_vsetvl::init): Ditto.
31573 * config/riscv/riscv-vsetvl.h (enum fusion_type): New enum.
31574 (enum merge_type): New enum.
31576 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31578 * config/riscv/riscv-vsetvl.cc
31579 (vector_infos_manager::vector_infos_manager): Add probability.
31580 (vector_infos_manager::dump): Ditto.
31581 (pass_vsetvl::compute_probabilities): Ditto.
31582 * config/riscv/riscv-vsetvl.h (struct vector_block_info): Ditto.
31584 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31586 * config/riscv/riscv-vsetvl.cc (vector_insn_info::operator==): Remove dirty_pat.
31587 (vector_insn_info::merge): Ditto.
31588 (vector_insn_info::dump): Ditto.
31589 (pass_vsetvl::merge_successors): Ditto.
31590 (pass_vsetvl::backward_demand_fusion): Ditto.
31591 (pass_vsetvl::forward_demand_fusion): Ditto.
31592 (pass_vsetvl::commit_vsetvls): Ditto.
31593 * config/riscv/riscv-vsetvl.h: Ditto.
31595 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31597 * config/riscv/riscv-vsetvl.cc (add_label_notes): Rename insn to
31600 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31602 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::backward_demand_fusion): Refine codes.
31604 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31606 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::forward_demand_fusion):
31607 Add pre-check for redundant flow.
31609 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31611 * config/riscv/riscv-vsetvl.cc (vector_infos_manager::create_bitmap_vectors): New function.
31612 (vector_infos_manager::free_bitmap_vectors): Ditto.
31613 (pass_vsetvl::pre_vsetvl): Adjust codes.
31614 * config/riscv/riscv-vsetvl.h: New function declaration.
31616 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31618 * config/riscv/riscv-vsetvl.cc (can_backward_propagate_p): Fix for null iter_bb.
31619 (vector_insn_info::set_demand_info): New function.
31620 (pass_vsetvl::emit_local_forward_vsetvls): Adjust for refinement of Phase 3.
31621 (pass_vsetvl::merge_successors): Ditto.
31622 (pass_vsetvl::compute_global_backward_infos): Ditto.
31623 (pass_vsetvl::backward_demand_fusion): Ditto.
31624 (pass_vsetvl::forward_demand_fusion): Ditto.
31625 (pass_vsetvl::demand_fusion): New function.
31626 (pass_vsetvl::lazy_vsetvl): Adjust for refinement of phase 3.
31627 * config/riscv/riscv-vsetvl.h: New function declaration.
31629 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31631 * config/riscv/riscv-vsetvl.cc (vector_insn_info::operator>=): Fix available condition.
31633 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31635 * config/riscv/riscv-vsetvl.cc (change_vsetvl_insn): New function.
31636 (pass_vsetvl::compute_global_backward_infos): Simplify codes.
31638 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31640 * config/riscv/riscv-vsetvl.cc (loop_basic_block_p): Adjust function.
31641 (backward_propagate_worthwhile_p): Fix non-worthwhile.
31643 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31645 * config/riscv/riscv-vsetvl.cc (change_insn): Adjust in_group in validate_change.
31647 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31649 * config/riscv/riscv-vsetvl.cc (vector_infos_manager::all_same_avl_p): New function.
31650 (pass_vsetvl::can_refine_vsetvl_p): Add AVL check.
31651 (pass_vsetvl::commit_vsetvls): Ditto.
31652 * config/riscv/riscv-vsetvl.h: New function declaration.
31654 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31656 * config/riscv/vector.md:
31658 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31660 * config/riscv/riscv-vector-builtins-bases.cc (class loadstore): use
31661 pred_store for vse.
31662 * config/riscv/riscv-vector-builtins.cc
31663 (function_expander::add_mem_operand): Refine function.
31664 (function_expander::use_contiguous_load_insn): Adjust new
31666 (function_expander::use_contiguous_store_insn): Ditto.
31667 * config/riscv/riscv-vector-builtins.h: Refine function.
31668 * config/riscv/vector.md (@pred_store<mode>): New pattern.
31670 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
31672 * config/riscv/riscv-vector-builtins.cc: Change to scalar pointer.
31674 2023-01-26 Marek Polacek <polacek@redhat.com>
31676 PR middle-end/108543
31677 * opts.cc (parse_sanitizer_options): Don't always clear SANITIZE_ADDRESS
31678 if it was previously set.
31680 2023-01-26 Jakub Jelinek <jakub@redhat.com>
31682 PR tree-optimization/108540
31683 * range-op-float.cc (foperator_equal::fold_range): If both op1 and op2
31684 are singletons, use range_true even if op1 != op2
31685 when one range is [-0.0, -0.0] and another [0.0, 0.0]. Similarly,
31686 even if intersection of the ranges is empty and one has
31687 zero low bound and another zero high bound, use range_true_and_false
31688 rather than range_false.
31689 (foperator_not_equal::fold_range): If both op1 and op2
31690 are singletons, use range_false even if op1 != op2
31691 when one range is [-0.0, -0.0] and another [0.0, 0.0]. Similarly,
31692 even if intersection of the ranges is empty and one has
31693 zero low bound and another zero high bound, use range_true_and_false
31694 rather than range_true.
31696 2023-01-26 Jakub Jelinek <jakub@redhat.com>
31698 * value-relation.cc (kind_string): Add const.
31699 (rr_negate_table, rr_swap_table, rr_intersect_table,
31700 rr_union_table, rr_transitive_table): Add static const, change
31701 element type from relation_kind to unsigned char.
31702 (relation_negate, relation_swap, relation_intersect, relation_union,
31703 relation_transitive): Cast rr_*_table element to relation_kind.
31704 (relation_to_code): Add static const.
31705 (relation_tests): Assert VREL_LAST is smaller than UCHAR_MAX.
31707 2023-01-26 Richard Biener <rguenther@suse.de>
31709 PR tree-optimization/108547
31710 * gimple-predicate-analysis.cc (value_sat_pred_p):
31713 2023-01-26 Siddhesh Poyarekar <siddhesh@gotplt.org>
31715 PR tree-optimization/108522
31716 * tree-object-size.cc (compute_object_offset): Make EXPR
31717 argument non-const. Call component_ref_field_offset.
31719 2023-01-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
31721 * config/aarch64/aarch64-option-extensions.def (cssc): Specify
31722 FEATURE_STRING field.
31724 2023-01-26 Gerald Pfeifer <gerald@pfeifer.com>
31726 * doc/sourcebuild.texi: Refer to projects as GCC and GDB.
31728 2023-01-25 Iain Sandoe <iain@sandoe.co.uk>
31732 * gcc.cc: Provide default specs for Modula-2 so that when the
31733 language is not built-in better diagnostics are emitted for
31734 attempts to use .mod or .m2i file extensions.
31736 2023-01-25 Andrea Corallo <andrea.corallo@arm.com>
31738 * config/arm/mve.md (mve_vqnegq_s<mode>): Fix spacing.
31740 2023-01-25 Andrea Corallo <andrea.corallo@arm.com>
31742 * config/arm/mve.md (mve_vqabsq_s<mode>): Fix spacing.
31744 2023-01-25 Andrea Corallo <andrea.corallo@arm.com>
31746 * config/arm/mve.md (mve_vnegq_f<mode>, mve_vnegq_s<mode>):
31749 2023-01-25 Andrea Corallo <andrea.corallo@arm.com>
31751 * config/arm/mve.md (@mve_vclzq_s<mode>): Fix spacing.
31753 2023-01-25 Andrea Corallo <andrea.corallo@arm.com>
31755 * config/arm/mve.md (mve_vclsq_s<mode>): Fix spacing.
31757 2023-01-25 Richard Biener <rguenther@suse.de>
31759 PR tree-optimization/108523
31760 * tree-ssa-sccvn.cc (visit_phi): Avoid using the exclusive
31761 backedge value for the result when using predication to
31764 2023-01-25 Richard Biener <rguenther@suse.de>
31766 * doc/lto.texi (Command line options): Reword and update reference
31767 to removed lto_read_all_file_options.
31769 2023-01-25 Richard Sandiford <richard.sandiford@arm.com>
31771 * config/aarch64/aarch64.md (umax<mode>3): Separate the CNT and CSSC
31774 2023-01-25 Gerald Pfeifer <gerald@pfeifer.com>
31776 * doc/contrib.texi: Add Jose E. Marchesi.
31778 2023-01-25 Jakub Jelinek <jakub@redhat.com>
31780 PR tree-optimization/108498
31781 * gimple-ssa-store-merging.cc (class store_operand_info):
31782 End coment with full stop rather than comma.
31783 (split_group): Likewise.
31784 (merged_store_group::apply_stores): Clear string_concatenation if
31785 start or end aren't on a byte boundary.
31787 2023-01-25 Siddhesh Poyarekar <siddhesh@gotplt.org>
31788 Jakub Jelinek <jakub@redhat.com>
31790 PR tree-optimization/108522
31791 * tree-object-size.cc (compute_object_offset): Use
31792 TREE_OPERAND(ref, 2) for COMPONENT_REF when available.
31794 2023-01-24 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
31796 * config/xtensa/xtensa.md:
31797 Fix exit from loops detecting references before overwriting in the
31800 2023-01-24 Vladimir N. Makarov <vmakarov@redhat.com>
31802 * lra-constraints.cc (get_hard_regno): Remove final_p arg. Always
31803 do elimination but only for hard register.
31804 (operands_match_p, uses_hard_regs_p, process_alt_operands): Adjust
31805 calls of get_hard_regno.
31807 2023-01-24 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
31809 * config/s390/s390-d.cc (s390_d_target_versions): Fix detection
31812 2023-01-24 Andre Vieira <andre.simoesdiasvieira@arm.com>
31815 * config/arm/mve.md (mve_vstrbq_p_<supf><mode>, mve_vstrhq_p_fv8hf,
31816 mve_vstrhq_p_<supf><mode>, mve_vstrwq_p_<supf>v4si): Add memory operand
31819 2023-01-24 Xianmiao Qu <cooper.qu@linux.alibaba.com>
31821 * config.gcc(csky-*-linux*): Define CSKY_ENABLE_MULTILIB
31822 and only include 'csky/t-csky-linux' when enable multilib.
31823 * config/csky/csky-linux-elf.h(SYSROOT_SUFFIX_SPEC): Don't
31824 define it when disable multilib.
31826 2023-01-24 Richard Biener <rguenther@suse.de>
31828 PR tree-optimization/108500
31829 * dominance.h (calculate_dominance_info): Add parameter
31830 to indicate fast-query compute, defaulted to true.
31831 * dominance.cc (calculate_dominance_info): Honor
31832 fast-query compute parameter.
31833 * tree-cfgcleanup.cc (cleanup_tree_cfg_noloop): Do
31834 not compute the dominator fast-query DFS numbers.
31836 2023-01-24 Eric Biggers <ebiggers@google.com>
31839 * optc-save-gen.awk: Fix copy-and-paste error.
31841 2023-01-24 Jakub Jelinek <jakub@redhat.com>
31844 * cgraphbuild.cc: Include gimplify.h.
31845 (record_reference): Replace VAR_DECLs with DECL_HAS_VALUE_EXPR_P with
31846 their corresponding DECL_VALUE_EXPR expressions after unsharing.
31848 2023-01-24 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
31851 * config.gcc (tm_file): Move the variable out of loop.
31853 2023-01-24 Lulu Cheng <chenglulu@loongson.cn>
31854 Yang Yujie <yangyujie@loongson.cn>
31857 * config/loongarch/loongarch.cc (loongarch_classify_address):
31858 Add precessint for CONST_INT.
31859 (loongarch_print_operand_reloc): Operand modifier 'c' is supported.
31860 (loongarch_print_operand): Increase the processing of '%c'.
31861 * doc/extend.texi: Adds documents for LoongArch operand modifiers.
31862 And port the public operand modifiers information to this document.
31864 2023-01-23 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
31866 * doc/invoke.texi (-mbranch-protection): Update documentation.
31868 2023-01-23 Richard Biener <rguenther@suse.de>
31871 * config/sparc/freebsd.h (ENDFILE_SPEC): Don't add crtfastmath.o
31873 * config/sparc/linux.h (ENDFILE_SPEC): Likewise.
31874 * config/sparc/linux64.h (ENDFILE_SPEC): Likewise.
31875 * config/sparc/sp-elf.h (ENDFILE_SPEC): Likewise.
31876 * config/sparc/sp64-elf.h (ENDFILE_SPEC): Likewise.
31878 2023-01-23 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
31880 * config/arm/aout.h (ra_auth_code): Add entry in enum.
31881 * config/arm/arm.cc (emit_multi_reg_push): Add RA_AUTH_CODE register
31882 to dwarf frame expression.
31883 (arm_emit_multi_reg_pop): Restore RA_AUTH_CODE register.
31884 (arm_expand_prologue): Update frame related information and reg notes
31885 for pac/pacbit insn.
31886 (arm_regno_class): Check for pac pseudo reigster.
31887 (arm_dbx_register_number): Assign ra_auth_code register number in dwarf.
31888 (arm_init_machine_status): Set pacspval_needed to zero.
31889 (arm_debugger_regno): Check for PAC register.
31890 (arm_unwind_emit_sequence): Print .save directive with ra_auth_code
31892 (arm_unwind_emit_set): Add entry for IP_REGNUM in switch case.
31893 (arm_unwind_emit): Update REG_CFA_REGISTER case._
31894 * config/arm/arm.h (FIRST_PSEUDO_REGISTER): Modify.
31895 (DWARF_PAC_REGNUM): Define.
31896 (IS_PAC_REGNUM): Likewise.
31897 (enum reg_class): Add PAC_REG entry.
31898 (machine_function): Add pacbti_needed state to structure.
31899 * config/arm/arm.md (RA_AUTH_CODE): Define.
31901 2023-01-23 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
31903 * config.gcc ($tm_file): Update variable.
31904 * config/arm/arm-mlib.h: Create new header file.
31905 * config/arm/t-rmprofile (MULTI_ARCH_DIRS_RM): Rename mbranch-protection
31906 multilib arch directory.
31907 (MULTILIB_REUSE): Add multilib reuse rules.
31908 (MULTILIB_MATCHES): Add multilib match rules.
31910 2023-01-23 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
31912 * config/arm/arm-cpus.in (cortex-m85): Define new CPU.
31913 * config/arm/arm-tables.opt: Regenerate.
31914 * config/arm/arm-tune.md: Likewise.
31915 * doc/invoke.texi (Arm Options): Document -mcpu=cortex-m85.
31916 * (-mfix-cmse-cve-2021-35465): Likewise.
31918 2023-01-23 Richard Biener <rguenther@suse.de>
31920 PR tree-optimization/108482
31921 * tree-vect-generic.cc (expand_vector_operations): Fold remaining
31922 .LOOP_DIST_ALIAS calls.
31924 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31926 * config.gcc (arm*-*-*): Add 'aarch-bti-insert.o' object.
31927 * config/arm/arm-protos.h: Update.
31928 * config/arm/aarch-common-protos.h: Declare
31929 'aarch_bti_arch_check'.
31930 * config/arm/arm.cc (aarch_bti_enabled) Update.
31931 (aarch_bti_j_insn_p, aarch_pac_insn_p, aarch_gen_bti_c)
31932 (aarch_gen_bti_j, aarch_bti_arch_check): New functions.
31933 * config/arm/arm.md (bti_nop): New insn.
31934 * config/arm/t-arm (PASSES_EXTRA): Add 'arm-passes.def'.
31935 (aarch-bti-insert.o): New target.
31936 * config/arm/unspecs.md (VUNSPEC_BTI_NOP): New unspec.
31937 * config/arm/aarch-bti-insert.cc (rest_of_insert_bti): Verify arch
31939 (gate): Make use of 'aarch_bti_arch_check'.
31940 * config/arm/arm-passes.def: New file.
31941 * config/aarch64/aarch64.cc (aarch_bti_arch_check): New function.
31943 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31945 * config.gcc (aarch64*-*-*): Rename 'aarch64-bti-insert.o' into
31946 'aarch-bti-insert.o'.
31947 * config/aarch64/aarch64-protos.h: Remove 'aarch64_bti_enabled'
31949 * config/aarch64/aarch64.cc (aarch_bti_enabled): Rename.
31950 (aarch_bti_j_insn_p, aarch_pac_insn_p): New functions.
31951 (aarch64_output_mi_thunk)
31952 (aarch64_print_patchable_function_entry)
31953 (aarch64_file_end_indicate_exec_stack): Update renamed function
31954 calls to renamed functions.
31955 * config/aarch64/aarch64-c.cc (aarch64_update_cpp_builtins): Likewise.
31956 * config/aarch64/t-aarch64 (aarch-bti-insert.o): Update
31958 * config/aarch64/aarch64-bti-insert.cc: Delete.
31959 * config/arm/aarch-bti-insert.cc: New file including and
31960 generalizing code from aarch64-bti-insert.cc.
31961 * config/arm/aarch-common-protos.h: Update.
31963 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31965 * config/arm/arm.h (arm_arch8m_main): Declare it.
31966 * config/arm/arm-protos.h (arm_current_function_pac_enabled_p):
31968 * config/arm/arm.cc (arm_arch8m_main): Define it.
31969 (arm_option_reconfigure_globals): Set arm_arch8m_main.
31970 (arm_compute_frame_layout, arm_expand_prologue)
31971 (thumb2_expand_return, arm_expand_epilogue)
31972 (arm_conditional_register_usage): Update for pac codegen.
31973 (arm_current_function_pac_enabled_p): New function.
31974 (aarch_bti_enabled) New function.
31975 (use_return_insn): Return zero when pac is enabled.
31976 * config/arm/arm.md (pac_ip_lr_sp, pacbti_ip_lr_sp, aut_ip_lr_sp):
31978 * config/arm/unspecs.md (UNSPEC_PAC_NOP)
31979 (VUNSPEC_PACBTI_NOP, VUNSPEC_AUT_NOP): Add unspecs.
31981 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31983 * config/arm/t-rmprofile: Add multilib rules for march +pacbti and
31984 mbranch-protection.
31986 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31987 Tejas Belagod <tbelagod@arm.com>
31989 * config/arm/arm.cc (arm_file_start): Emit EABI attributes for
31990 Tag_PAC_extension, Tag_BTI_extension, TAG_BTI_use, TAG_PACRET_use.
31992 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31993 Tejas Belagod <tbelagod@arm.com>
31994 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
31996 * ginclude/unwind-arm-common.h (_Unwind_VRS_RegClass): Introduce
31997 new pseudo register class _UVRSC_PAC.
31999 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
32000 Tejas Belagod <tbelagod@arm.com>
32002 * config/arm/arm-c.cc (arm_cpu_builtins): Define
32003 __ARM_FEATURE_BTI_DEFAULT, __ARM_FEATURE_PAC_DEFAULT,
32004 __ARM_FEATURE_PAUTH and __ARM_FEATURE_BTI.
32006 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
32007 Tejas Belagod <tbelagod@arm.com>
32009 * doc/sourcebuild.texi: Document arm_pacbti_hw.
32011 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
32012 Tejas Belagod <tbelagod@arm.com>
32013 Richard Earnshaw <Richard.Earnshaw@arm.com>
32015 * config/arm/arm.cc (arm_configure_build_target): Parse and validate
32016 -mbranch-protection option and initialize appropriate data structures.
32017 * config/arm/arm.opt (-mbranch-protection): New option.
32018 * doc/invoke.texi (Arm Options): Document it.
32020 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
32021 Tejas Belagod <tbelagod@arm.com>
32023 * config/arm/arm.h (TARGET_HAVE_PACBTI): New macro.
32024 * config/arm/arm-cpus.in (pacbti): New feature.
32025 * doc/invoke.texi (Arm Options): Document it.
32027 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
32028 Tejas Belagod <tbelagod@arm.com>
32030 * common/config/aarch64/aarch64-common.cc: Include aarch-common.h.
32031 (all_architectures): Fix comment.
32032 (aarch64_parse_extension): Rename return type, enum value names.
32033 * config/aarch64/aarch64-c.cc (aarch64_update_cpp_builtins): Rename
32034 factored out aarch_ra_sign_scope and aarch_ra_sign_key variables.
32035 Also rename corresponding enum values.
32036 * config/aarch64/aarch64-opts.h (aarch64_function_type): Factor
32037 out aarch64_function_type and move it to common code as
32038 aarch_function_type in aarch-common.h.
32039 * config/aarch64/aarch64-protos.h: Include common types header,
32040 move out types aarch64_parse_opt_result and aarch64_key_type to
32042 * config/aarch64/aarch64.cc: Move mbranch-protection parsing types
32043 and functions out into aarch-common.h and aarch-common.cc. Fix up
32044 all the name changes resulting from the move.
32045 * config/aarch64/aarch64.md: Fix up aarch64_ra_sign_key type name change
32047 * config/aarch64/aarch64.opt: Include aarch-common.h to import
32048 type move. Fix up name changes from factoring out common code and
32050 * config/arm/aarch-common-protos.h: Export factored out routines to both
32052 * config/arm/aarch-common.cc: Include newly factored out types.
32053 Move all mbranch-protection code and data structures from
32055 * config/arm/aarch-common.h: New header that declares types shared
32056 between aarch32 and aarch64 backends.
32057 * config/arm/arm-protos.h: Declare types and variables that are
32058 made common to aarch64 and aarch32 backends - aarch_ra_sign_key,
32059 aarch_ra_sign_scope and aarch_enable_bti.
32060 * config/arm/arm.opt (config/arm/aarch-common.h): Include header.
32061 (aarch_ra_sign_scope, aarch_enable_bti): Declare variable.
32062 * config/arm/arm.cc: Add missing includes.
32064 2023-01-23 Tobias Burnus <tobias@codesourcery.com>
32066 * doc/install.texi (amdgcn, nvptx): Require newlib 4.3.0.
32068 2023-01-23 Richard Biener <rguenther@suse.de>
32070 PR tree-optimization/108449
32071 * cgraphunit.cc (check_global_declaration): Do not turn
32072 undefined statics into externs.
32074 2023-01-22 Dimitar Dimitrov <dimitar@dinux.eu>
32076 * config/pru/pru.h (CLZ_DEFINED_VALUE_AT_ZERO): Fix value for QI
32077 and HI input modes.
32078 * config/pru/pru.md (clz): Fix generated code for QI and HI
32081 2023-01-22 Cupertino Miranda <cupertino.miranda@oracle.com>
32083 * config/v850/v850.cc (v850_select_section): Put const volatile
32084 objects into read-only sections.
32086 2023-01-20 Tejas Belagod <tejas.belagod@arm.com>
32088 * config/aarch64/arm_neon.h (vmull_p64, vmull_high_p64, vaeseq_u8,
32089 vaesdq_u8, vaesmcq_u8, vaesimcq_u8): Gate under "nothing+aes".
32090 (vsha1*_u32, vsha256*_u32): Gate under "nothing+sha2".
32092 2023-01-20 Jakub Jelinek <jakub@redhat.com>
32094 PR tree-optimization/108457
32095 * tree-ssa-loop-niter.cc (build_cltz_expr): Use
32096 SCALAR_INT_TYPE_MODE (utype) directly as C[LT]Z_DEFINED_VALUE_AT_ZERO
32097 argument instead of a temporary. Formatting fixes.
32099 2023-01-19 Jakub Jelinek <jakub@redhat.com>
32101 PR tree-optimization/108447
32102 * value-relation.cc (rr_union_table): Fix VREL_UNDEFINED row order.
32103 (relation_tests): Add self-tests for relation_{intersect,union}
32105 * selftest.h (relation_tests): Declare.
32106 * function-tests.cc (test_ranges): Call it.
32108 2023-01-19 H.J. Lu <hjl.tools@gmail.com>
32111 * config/i386/i386-expand.cc (ix86_expand_builtin): Check
32112 invalid third argument to __builtin_ia32_prefetch.
32114 2023-01-19 Jakub Jelinek <jakub@redhat.com>
32116 PR middle-end/108459
32117 * omp-expand.cc (expand_omp_for_init_counts): Use fold_build1 rather
32118 than fold_unary for NEGATE_EXPR.
32120 2023-01-19 Christophe Lyon <christophe.lyon@arm.com>
32123 * config/aarch64/aarch64.cc (aarch64_layout_arg): Improve
32124 comment. Move assert about alignment a bit later.
32126 2023-01-19 Jakub Jelinek <jakub@redhat.com>
32128 PR tree-optimization/108440
32129 * tree-ssa-forwprop.cc: Include gimple-range.h.
32130 (simplify_rotate): For the forms with T2 wider than T and shift counts of
32131 Y and B - Y add & (B - 1) masking for the rotate count if Y could be equal
32132 to B. For the forms with T2 wider than T and shift counts of
32133 Y and (-Y) & (B - 1), don't punt if range could be [B, B2], but only if
32134 range doesn't guarantee Y < B or Y = N * B. If range doesn't guarantee
32135 Y < B, also add & (B - 1) masking for the rotate count. Use lazily created
32136 pass specific ranger instead of get_global_range_query.
32137 (pass_forwprop::execute): Disable that ranger at the end of pass if it has
32140 2023-01-19 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
32142 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_set<mode>): Use
32143 exact_log2 (INTVAL (operands[2])) >= 0 as condition for gating
32145 (aarch64_simd_vec_copy_lane<mode>): Likewise.
32146 (aarch64_simd_vec_copy_lane_<vswap_width_name><mode>): Likewise.
32148 2023-01-19 Alexandre Oliva <oliva@adacore.com>
32151 * sched-deps.cc (sched_analyze_2): Skip cselib address lookup
32152 within debug insns.
32154 2023-01-18 Martin Jambor <mjambor@suse.cz>
32157 * cgraph.cc (cgraph_node::remove): Check whether nodes up the
32158 lcone_of chain also do not need the body.
32160 2023-01-18 Richard Biener <rguenther@suse.de>
32163 2022-12-16 Richard Biener <rguenther@suse.de>
32165 PR middle-end/108086
32166 * tree-inline.cc (remap_ssa_name): Do not unshare the
32167 result from the decl_map.
32169 2023-01-18 Murray Steele <murray.steele@arm.com>
32172 * config/arm/arm_mve.h (__arm_vst1q_p_u8): Use prefixed intrinsic
32174 (__arm_vst1q_p_s8): Likewise.
32175 (__arm_vld1q_z_u8): Likewise.
32176 (__arm_vld1q_z_s8): Likewise.
32177 (__arm_vst1q_p_u16): Likewise.
32178 (__arm_vst1q_p_s16): Likewise.
32179 (__arm_vld1q_z_u16): Likewise.
32180 (__arm_vld1q_z_s16): Likewise.
32181 (__arm_vst1q_p_u32): Likewise.
32182 (__arm_vst1q_p_s32): Likewise.
32183 (__arm_vld1q_z_u32): Likewise.
32184 (__arm_vld1q_z_s32): Likewise.
32185 (__arm_vld1q_z_f16): Likewise.
32186 (__arm_vst1q_p_f16): Likewise.
32187 (__arm_vld1q_z_f32): Likewise.
32188 (__arm_vst1q_p_f32): Likewise.
32190 2023-01-18 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
32192 * config/xtensa/xtensa.md (xorsi3_internal):
32193 Rename from the original of "xorsi3".
32194 (xorsi3): New expansion pattern that emits addition rather than
32195 bitwise-XOR when the second source is a constant of -2147483648
32198 2023-01-18 Kewen Lin <linkw@linux.ibm.com>
32199 Andrew Pinski <apinski@marvell.com>
32202 * config/rs6000/rs6000-overload.def (VEC_VSUBCUQ): Fix typo
32203 vec_vsubcuqP with vec_vsubcuq.
32205 2023-01-18 Kewen Lin <linkw@linux.ibm.com>
32208 * config/rs6000/rs6000.cc (rs6000_opaque_type_invalid_use_p): Add the
32209 support for invalid uses of MMA opaque type in function arguments.
32211 2023-01-18 liuhongt <hongtao.liu@intel.com>
32214 * config/i386/cygwin.h (ENDFILE_SPEC): Link crtfastmath.o
32215 whenever -mdaz-ftz is specified. Don't link crtfastmath.o when
32216 -share or -mno-daz-ftz is specified.
32217 * config/i386/darwin.h (ENDFILE_SPEC): Ditto.
32218 * config/i386/mingw32.h (ENDFILE_SPEC): Ditto.
32220 2023-01-17 Jose E. Marchesi <jose.marchesi@oracle.com>
32222 * config/bpf/bpf.cc (bpf_option_override): Disable
32225 2023-01-17 Jakub Jelinek <jakub@redhat.com>
32227 PR tree-optimization/106523
32228 * tree-ssa-forwprop.cc (simplify_rotate): For the
32229 patterns with (-Y) & (B - 1) in one operand's shift
32230 count and Y in another, if T2 has wider precision than T,
32231 punt if Y could have a value in [B, B2 - 1] range.
32233 2023-01-16 H.J. Lu <hjl.tools@gmail.com>
32236 * config/i386/i386.cc (x86_output_mi_thunk): Disable
32237 -mforce-indirect-call for PIC in 32-bit mode.
32239 2023-01-16 Jan Hubicka <hubicka@ucw.cz>
32242 * ipa-modref.cc (modref_access_analysis::analyze): Use
32243 find_always_executed_bbs.
32244 * ipa-sra.cc (process_scan_results): Likewise.
32245 * ipa-utils.cc (stmt_may_terminate_function_p): New function.
32246 (find_always_executed_bbs): New function.
32247 * ipa-utils.h (stmt_may_terminate_function_p): Declare.
32248 (find_always_executed_bbs): Declare.
32250 2023-01-16 Jan Hubicka <jh@suse.cz>
32252 * config/i386/i386.cc (ix86_vectorize_builtin_scatter): Guard scatter
32253 by TARGET_USE_SCATTER.
32254 * config/i386/i386.h (TARGET_USE_SCATTER_2PARTS,
32255 TARGET_USE_SCATTER_4PARTS, TARGET_USE_SCATTER): New macros.
32256 * config/i386/x86-tune.def (TARGET_USE_SCATTER_2PARTS,
32257 TARGET_USE_SCATTER_4PARTS, TARGET_USE_SCATTER): New tunes.
32258 (X86_TUNE_AVOID_256FMA_CHAINS, X86_TUNE_AVOID_512FMA_CHAINS): Disable
32259 for znver4. (X86_TUNE_USE_GATHER): Disable for zen4.
32261 2023-01-16 Richard Biener <rguenther@suse.de>
32264 * config/sol2.h (ENDFILE_SPEC): Don't add crtfastmath.o for -shared.
32266 2023-01-16 Stam Markianos-Wright <stam.markianos-wright@arm.com>
32270 * config/arm/arm_mve.h (__ARM_mve_coerce2): Split types.
32271 (__ARM_mve_coerce3): Likewise.
32273 2023-01-16 Andrew Carlotti <andrew.carlotti@arm.com>
32275 * tree-ssa-loop-niter.cc (build_popcount_expr): Add IFN support.
32277 2023-01-16 Andrew Carlotti <andrew.carlotti@arm.com>
32279 * tree-ssa-loop-niter.cc (number_of_iterations_cltz): New.
32280 (number_of_iterations_bitcount): Add call to the above.
32281 (number_of_iterations_exit_assumptions): Add EQ_EXPR case for
32282 c[lt]z idiom recognition.
32284 2023-01-16 Andrew Carlotti <andrew.carlotti@arm.com>
32286 * doc/sourcebuild.texi: Add missing target attributes.
32288 2023-01-16 Andrew Carlotti <andrew.carlotti@arm.com>
32290 PR tree-optimization/94793
32291 * tree-scalar-evolution.cc (expression_expensive_p): Add checks
32293 * tree-ssa-loop-niter.cc (build_cltz_expr): New.
32294 (number_of_iterations_cltz_complement): New.
32295 (number_of_iterations_bitcount): Add call to the above.
32297 2023-01-16 Jonathan Wakely <jwakely@redhat.com>
32299 * doc/extend.texi (Common Function Attributes): Fix grammar.
32301 2023-01-16 Jakub Jelinek <jakub@redhat.com>
32304 * config/riscv/riscv-vsetvl.h: Add space in between Copyright and (C).
32305 * config/riscv/riscv-vsetvl.cc: Likewise.
32307 2023-01-16 Jakub Jelinek <jakub@redhat.com>
32310 * config/i386/xmmintrin.h (_mm_undefined_ps): Temporarily
32311 disable -Winit-self using pragma GCC diagnostic ignored.
32312 * config/i386/emmintrin.h (_mm_undefined_pd, _mm_undefined_si128):
32314 * config/i386/avxintrin.h (_mm256_undefined_pd, _mm256_undefined_ps,
32315 _mm256_undefined_si256): Likewise.
32316 * config/i386/avx512fintrin.h (_mm512_undefined_pd,
32317 _mm512_undefined_ps, _mm512_undefined_epi32): Likewise.
32318 * config/i386/avx512fp16intrin.h (_mm_undefined_ph,
32319 _mm256_undefined_ph, _mm512_undefined_ph): Likewise.
32321 2023-01-16 Kewen Lin <linkw@linux.ibm.com>
32324 * config/rs6000/rs6000.cc (rs6000_opaque_type_invalid_use_p): Add the
32325 support for invalid uses in inline asm, factor out the checking and
32326 erroring to lambda function check_and_error_invalid_use.
32328 2023-01-15 Aldy Hernandez <aldyh@redhat.com>
32330 PR tree-optimization/107608
32331 * range-op-float.cc (range_operator_float::fold_range): Avoid
32332 folding into INF when flag_trapping_math.
32333 * value-range.h (frange::known_isinf): Return false for possible NANs.
32335 2023-01-15 Xianmiao Qu <cooper.qu@linux.alibaba.com>
32337 * config.gcc (csky-*-*): Support --with-float=softfp.
32339 2023-01-14 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
32341 * config/xtensa/xtensa-protos.h (order_regs_for_local_alloc):
32342 Rename to xtensa_adjust_reg_alloc_order.
32343 * config/xtensa/xtensa.cc (xtensa_adjust_reg_alloc_order):
32344 Ditto. And also remove code to reorder register numbers for
32345 leaf functions, rename the tables, and adjust the allocation
32346 order for the call0 ABI to use register A0 more.
32347 (xtensa_leaf_regs): Remove.
32348 * config/xtensa/xtensa.h (REG_ALLOC_ORDER): Cosmetics.
32349 (order_regs_for_local_alloc): Rename as the above.
32350 (LEAF_REGISTERS, LEAF_REG_REMAP, leaf_function): Remove.
32352 2023-01-14 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
32354 * config/aarch64/aarch64-sve.md (aarch64_vec_duplicate_vq<mode>_le):
32355 Change to define_insn_and_split to fold ldr+dup to ld1rq.
32356 * config/aarch64/predicates.md (aarch64_sve_dup_ld1rq_operand): New.
32358 2023-01-14 Alexandre Oliva <oliva@adacore.com>
32360 * hash-table.h (is_deleted): Precheck !is_empty.
32361 (mark_deleted): Postcheck !is_empty.
32362 (copy constructor): Test is_empty before is_deleted.
32364 2023-01-14 Alexandre Oliva <oliva@adacore.com>
32367 * config/arm/arm.md (movmisaligndi): Prefer aligned SImode
32370 2023-01-13 Eric Botcazou <ebotcazou@adacore.com>
32372 PR rtl-optimization/108274
32373 * function.cc (thread_prologue_and_epilogue_insns): Also update the
32374 DF information for calls in a few more cases.
32376 2023-01-13 John David Anglin <danglin@gcc.gnu.org>
32378 * config/pa/pa-linux.h (TARGET_SYNC_LIBCALL): Delete define.
32379 * config/pa/pa.cc (pa_init_libfuncs): Use MAX_SYNC_LIBFUNC_SIZE
32381 * config/pa/pa.h (TARGET_SYNC_LIBCALLS): Use flag_sync_libcalls.
32382 (MAX_SYNC_LIBFUNC_SIZE): Define.
32383 (TARGET_CPU_CPP_BUILTINS): Define __SOFTFP__ when soft float is
32385 * config/pa/pa.md (atomic_storeqi): Emit __atomic_exchange_1
32386 libcall when sync libcalls are disabled.
32387 (atomic_storehi, atomic_storesi, atomic_storedi): Likewise.
32388 (atomic_loaddi): Emit __atomic_load_8 libcall when sync libcalls
32389 are disabled on 32-bit target.
32390 * config/pa/pa.opt (matomic-libcalls): New option.
32391 * doc/invoke.texi (HPPA Options): Update.
32393 2023-01-13 Alexander Monakov <amonakov@ispras.ru>
32395 PR rtl-optimization/108117
32396 PR rtl-optimization/108132
32397 * sched-deps.cc (deps_analyze_insn): Do not schedule across
32398 calls before reload.
32400 2023-01-13 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
32402 * common/config/arm/arm-common.cc (arm_canon_arch_option_1): Ignore cde
32403 options for -mlibarch.
32404 * config/arm/arm-cpus.in (begin cpu cortex-m55): Add cde options.
32405 * doc/invoke.texi (CDE): Document options for Cortex-M55 CPU.
32407 2023-01-13 Qing Zhao <qing.zhao@oracle.com>
32409 * attribs.cc (strict_flex_array_level_of): Move this function to ...
32410 * attribs.h (strict_flex_array_level_of): Remove the declaration.
32411 * gimple-array-bounds.cc (array_bounds_checker::check_array_ref):
32412 replace the referece to strict_flex_array_level_of with
32413 DECL_NOT_FLEXARRAY.
32414 * tree.cc (component_ref_size): Likewise.
32416 2023-01-13 Richard Biener <rguenther@suse.de>
32419 * config/arm/linux-eabi.h (ENDFILE_SPEC): Don't add
32420 crtfastmath.o for -shared.
32421 * config/arm/unknown-elf.h (STARTFILE_SPEC): Likewise.
32423 2023-01-13 Richard Biener <rguenther@suse.de>
32426 * config/aarch64/aarch64-elf-raw.h (ENDFILE_SPEC): Don't add
32427 crtfastmath.o for -shared.
32428 * config/aarch64/aarch64-freebsd.h (GNU_USER_TARGET_MATHFILE_SPEC):
32430 * config/aarch64/aarch64-linux.h (GNU_USER_TARGET_MATHFILE_SPEC):
32433 2023-01-13 Richard Sandiford <richard.sandiford@arm.com>
32435 * config/aarch64/aarch64.cc (aarch64_dwarf_frame_reg_mode): New
32437 (TARGET_DWARF_FRAME_REG_MODE): Define.
32439 2023-01-13 Richard Biener <rguenther@suse.de>
32442 * config/aarch64/aarch64.cc (aarch64_gimple_fold_builtin): Don't
32443 update EH info on the fly.
32445 2023-01-13 Richard Biener <rguenther@suse.de>
32447 PR tree-optimization/108387
32448 * tree-ssa-sccvn.cc (visit_nary_op): Check for SSA_NAME
32449 value before inserting expression into the tables.
32451 2023-01-12 Andrew Pinski <apinski@marvell.com>
32452 Roger Sayle <roger@nextmovesoftware.com>
32454 PR tree-optimization/92342
32455 * match.pd ((m1 CMP m2) * d -> (m1 CMP m2) ? d : 0):
32456 Use tcc_comparison and :c for the multiply.
32457 (b & -(a CMP c) -> (a CMP c)?b:0): New pattern.
32459 2023-01-12 Christophe Lyon <christophe.lyon@arm.com>
32460 Richard Sandiford <richard.sandiford@arm.com>
32463 * config/aarch64/aarch64.cc (aarch64_function_arg_alignment):
32464 Check DECL_PACKED for bitfield.
32465 (aarch64_layout_arg): Warn when parameter passing ABI changes.
32466 (aarch64_function_arg_boundary): Do not warn here.
32467 (aarch64_gimplify_va_arg_expr): Warn when parameter passing ABI
32470 2023-01-12 Christophe Lyon <christophe.lyon@arm.com>
32471 Richard Sandiford <richard.sandiford@arm.com>
32473 * config/aarch64/aarch64.cc (aarch64_function_arg_alignment): Fix
32475 (aarch64_layout_arg): Factorize warning conditions.
32476 (aarch64_function_arg_boundary): Fix typo.
32477 * function.cc (currently_expanding_function_start): New variable.
32478 (expand_function_start): Handle
32479 currently_expanding_function_start.
32480 * function.h (currently_expanding_function_start): Declare.
32482 2023-01-12 Richard Biener <rguenther@suse.de>
32484 PR tree-optimization/99412
32485 * tree-ssa-reassoc.cc (is_phi_for_stmt): Remove.
32486 (swap_ops_for_binary_stmt): Remove reduction handling.
32487 (rewrite_expr_tree_parallel): Adjust.
32488 (reassociate_bb): Likewise.
32489 * tree-parloops.cc (build_new_reduction): Handle MINUS_EXPR.
32491 2023-01-12 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
32493 * config/xtensa/xtensa.md (ctzsi2, ffssi2):
32494 Rearrange the emitting codes.
32496 2023-01-12 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
32498 * config/xtensa/xtensa.md (*btrue):
32499 Correct value of the attribute "length" that depends on
32500 TARGET_DENSITY and operands, and add '?' character to the register
32501 constraint of the compared operand.
32503 2023-01-12 Alexandre Oliva <oliva@adacore.com>
32505 * hash-table.h (expand): Check elements and deleted counts.
32506 (verify): Likewise.
32508 2023-01-11 Roger Sayle <roger@nextmovesoftware.com>
32510 PR tree-optimization/71343
32511 * tree-ssa-sccvn.cc (visit_nary_op) <case LSHIFT_EXPR>: Make
32512 the value number of the expression X << C the same as the value
32513 number for the multiplication X * (1<<C).
32515 2023-01-11 David Faust <david.faust@oracle.com>
32518 * config/bpf/bpf.cc (bpf_print_operand): Correct handling for
32519 floating point modes.
32521 2023-01-11 Eric Botcazou <ebotcazou@adacore.com>
32523 PR tree-optimization/108199
32524 * tree-sra.cc (sra_modify_expr): Deal with reverse storage order
32525 for bit-field references.
32527 2023-01-11 Kewen Lin <linkw@linux.ibm.com>
32529 * config/rs6000/rs6000.cc (rs6000_option_override_internal): Make
32530 OPTION_MASK_P10_FUSION implicit setting honour Power10 tuning setting.
32531 * config/rs6000/rs6000-cpus.def (ISA_3_1_MASKS_SERVER): Remove
32532 OPTION_MASK_P10_FUSION.
32534 2023-01-11 Richard Biener <rguenther@suse.de>
32536 PR tree-optimization/107767
32537 * tree-cfgcleanup.cc (phi_alternatives_equal): Export.
32538 * tree-cfgcleanup.h (phi_alternatives_equal): Declare.
32539 * tree-switch-conversion.cc (switch_conversion::collect):
32540 Count unique non-default targets accounting for later
32541 merging opportunities.
32543 2023-01-11 Martin Liska <mliska@suse.cz>
32545 PR middle-end/107976
32546 * params.opt: Limit JT params.
32547 * stmt.cc (emit_case_dispatch_table): Use auto_vec.
32549 2023-01-11 Richard Biener <rguenther@suse.de>
32551 PR tree-optimization/108352
32552 * tree-ssa-threadbackward.cc
32553 (back_threader_profitability::profitable_path_p): Adjust
32554 heuristic that allows non-multi-way branch threads creating
32556 * doc/invoke.texi (--param fsm-scale-path-blocks): Remove.
32557 (--param fsm-scale-path-stmts): Adjust.
32558 * params.opt (--param=fsm-scale-path-blocks=): Remove.
32559 (-param=fsm-scale-path-stmts=): Adjust description.
32561 2023-01-11 Richard Biener <rguenther@suse.de>
32563 PR tree-optimization/108353
32564 * tree-ssa-propagate.cc (cfg_blocks_back, ssa_edge_worklist_back):
32566 (add_ssa_edge): Simplify.
32567 (add_control_edge): Likewise.
32568 (ssa_prop_init): Likewise.
32569 (ssa_prop_fini): Likewise.
32570 (ssa_propagation_engine::ssa_propagate): Likewise.
32572 2023-01-11 Andreas Krebbel <krebbel@linux.ibm.com>
32574 * config/s390/s390.md (*not<mode>): New pattern.
32576 2023-01-11 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
32578 * config/xtensa/xtensa.cc (xtensa_insn_cost):
32579 Let insn cost for size be obtained by applying COSTS_N_INSNS()
32580 to instruction length and then dividing by 3.
32582 2023-01-10 Richard Biener <rguenther@suse.de>
32584 PR tree-optimization/106293
32585 * tree-ssa-dse.cc (dse_classify_store): Use a worklist to
32586 process degenerate PHI defs.
32588 2023-01-10 Roger Sayle <roger@nextmovesoftware.com>
32590 PR rtl-optimization/106421
32591 * cprop.cc (bypass_block): Check that DEST is local to this
32592 function (non-NULL) before calling find_edge.
32594 2023-01-10 Martin Jambor <mjambor@suse.cz>
32597 * ipa-param-manipulation.h (ipa_param_body_adjustments): New members
32598 sort_replacements, lookup_first_base_replacement and
32599 m_sorted_replacements_p.
32600 * ipa-param-manipulation.cc: Define INCLUDE_ALGORITHM.
32601 (ipa_param_body_adjustments::register_replacement): Set
32602 m_sorted_replacements_p to false.
32603 (compare_param_body_replacement): New function.
32604 (ipa_param_body_adjustments::sort_replacements): Likewise.
32605 (ipa_param_body_adjustments::common_initialization): Call
32607 (ipa_param_body_adjustments::ipa_param_body_adjustments): Initialize
32608 m_sorted_replacements_p.
32609 (ipa_param_body_adjustments::lookup_replacement_1): Rework to use
32611 (ipa_param_body_adjustments::lookup_first_base_replacement): New
32613 (ipa_param_body_adjustments::modify_call_stmt): Use
32614 lookup_first_base_replacement.
32615 * omp-simd-clone.cc (ipa_simd_modify_function_body): Call
32616 adjustments->sort_replacements.
32618 2023-01-10 Richard Biener <rguenther@suse.de>
32620 PR tree-optimization/108314
32621 * tree-vect-stmts.cc (vectorizable_condition): Do not
32622 perform BIT_NOT_EXPR optimization for EXTRACT_LAST_REDUCTION.
32624 2023-01-10 Xianmiao Qu <cooper.qu@linux.alibaba.com>
32626 * config/csky/csky-linux-elf.h (SYSROOT_SUFFIX_SPEC): New.
32628 2023-01-10 Xianmiao Qu <cooper.qu@linux.alibaba.com>
32630 * config/csky/csky.h (MULTILIB_DEFAULTS): Fix float abi option.
32632 2023-01-10 Xianmiao Qu <cooper.qu@linux.alibaba.com>
32634 * config/csky/csky.cc (csky_cpu_cpp_builtins): Add builtin
32635 defines for soft float abi.
32637 2023-01-10 Xianmiao Qu <cooper.qu@linux.alibaba.com>
32639 * config/csky/csky.md (smart_bseti): Change condition to CSKY_ISA_FEATURE (E1).
32640 (smart_bclri): Likewise.
32641 (fast_bseti): Change condition to CSKY_ISA_FEATURE (E2).
32642 (fast_bclri): Likewise.
32643 (fast_cmpnesi_i): Likewise.
32644 (*fast_cmpltsi_i): Likewise.
32645 (*fast_cmpgeusi_i): Likewise.
32647 2023-01-10 Xianmiao Qu <cooper.qu@linux.alibaba.com>
32649 * config/csky/csky_insn_fpuv3.md (l<frm_pattern><fixsuop><mode>si2): Test
32650 flag_fp_int_builtin_inexact || !flag_trapping_math.
32651 (<frm_pattern><mode>2): Likewise.
32653 2023-01-10 Andreas Krebbel <krebbel@linux.ibm.com>
32655 * config/s390/s390.cc (s390_register_info): Check call_used_regs
32656 instead of hard-coding the register numbers for call saved
32658 (s390_optimize_register_info): Likewise.
32660 2023-01-09 Eric Botcazou <ebotcazou@adacore.com>
32662 * doc/gm2.texi (Overview): Fix @node markers.
32663 (Using): Likewise. Remove subsections that were moved to Overview
32664 from the menu and move others around.
32666 2023-01-09 Richard Biener <rguenther@suse.de>
32668 PR middle-end/108209
32669 * genmatch.cc (commutative_op): Fix return value for
32670 user-id with non-commutative first replacement.
32672 2023-01-09 Jakub Jelinek <jakub@redhat.com>
32675 * calls.cc (expand_call): For calls with
32676 TYPE_NO_NAMED_ARGS_STDARG_P (funtype) use zero for n_named_args.
32679 2023-01-09 Richard Biener <rguenther@suse.de>
32681 PR middle-end/69482
32682 * cfgexpand.cc (discover_nonconstant_array_refs_r): Volatile
32683 qualified accesses also force objects to memory.
32685 2023-01-09 Martin Liska <mliska@suse.cz>
32688 * lto-cgraph.cc (compute_ltrans_boundary): Do not insert
32689 NULL (deleleted value) to a hash_set.
32691 2023-01-08 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
32693 * config/xtensa/xtensa.md (*splice_bits):
32694 New insn_and_split pattern.
32696 2023-01-07 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
32698 * config/xtensa/xtensa.cc
32699 (xtensa_split_imm_two_addends, xtensa_emit_add_imm):
32700 New helper functions.
32701 (xtensa_set_return_address, xtensa_output_mi_thunk):
32702 Change to use the helper function.
32703 (xtensa_emit_adjust_stack_ptr): Ditto.
32704 And also change to try reusing the content of scratch register
32705 A9 if the register is not modified in the function body.
32707 2023-01-07 LIU Hao <lh_mouse@126.com>
32709 PR middle-end/108300
32710 * config/xtensa/xtensa-dynconfig.c: Define `WIN32_LEAN_AND_MEAN`
32711 before <windows.h>.
32712 * diagnostic-color.cc: Likewise.
32713 * plugin.cc: Likewise.
32714 * prefix.cc: Likewise.
32716 2023-01-06 Joseph Myers <joseph@codesourcery.com>
32718 * doc/extend.texi (__builtin_tgmath): Do not restate standard rule
32719 for handling real integer types.
32721 2023-01-06 Tamar Christina <tamar.christina@arm.com>
32724 2022-12-12 Tamar Christina <tamar.christina@arm.com>
32726 * config/aarch64/aarch64-simd.md (*aarch64_simd_movv2hf): New.
32727 (mov<mode>, movmisalign<mode>, aarch64_dup_lane<mode>,
32728 aarch64_store_lane0<mode>, aarch64_simd_vec_set<mode>,
32729 @aarch64_simd_vec_copy_lane<mode>, vec_set<mode>,
32730 reduc_<optab>_scal_<mode>, reduc_<fmaxmin>_scal_<mode>,
32731 aarch64_reduc_<optab>_internal<mode>, aarch64_get_lane<mode>,
32732 vec_init<mode><Vel>, vec_extract<mode><Vel>): Support V2HF.
32733 (aarch64_simd_dupv2hf): New.
32734 * config/aarch64/aarch64.cc (aarch64_classify_vector_mode):
32736 * config/aarch64/iterators.md (VHSDF_P): New.
32737 (V2F, VMOVE, nunits, Vtype, Vmtype, Vetype, stype, VEL,
32738 Vel, q, vp): Add V2HF.
32739 * config/arm/types.md (neon_fp_reduc_add_h): New.
32741 2023-01-06 Martin Liska <mliska@suse.cz>
32743 PR middle-end/107966
32744 * doc/options.texi: Fix Var documentation in internal manual.
32746 2023-01-05 Roger Sayle <roger@nextmovesoftware.com>
32749 2023-01-03 Roger Sayle <roger@nextmovesoftware.com>
32751 * config/i386/i386-expand.cc (ix86_expand_int_movcc): Rewrite
32752 RTL expansion to allow condition (mask) to be shared/reused,
32753 by avoiding overwriting pseudos and adding REG_EQUAL notes.
32755 2023-01-05 Iain Sandoe <iain@sandoe.co.uk>
32757 * common.opt: Add -static-libgm2.
32758 * config/darwin.h (LINK_SPEC): Handle static-libgm2.
32759 * doc/gm2.texi: Document static-libgm2.
32760 * gcc.cc (driver_handle_option): Allow static-libgm2.
32762 2023-01-05 Tejas Joshi <TejasSanjay.Joshi@amd.com>
32764 * common/config/i386/i386-common.cc (processor_alias_table):
32765 Use CPU_ZNVER4 for znver4.
32766 * config/i386/i386.md: Add znver4.md.
32767 * config/i386/znver4.md: New.
32769 2023-01-04 Jakub Jelinek <jakub@redhat.com>
32771 PR tree-optimization/108253
32772 * tree-vrp.cc (maybe_set_nonzero_bits): Handle var with pointer
32775 2023-01-04 Jakub Jelinek <jakub@redhat.com>
32777 PR middle-end/108237
32778 * generic-match-head.cc: Include tree-pass.h.
32779 (canonicalize_math_p, optimize_vectors_before_lowering_p): Define
32780 to false if cfun and cfun->curr_properties has PROP_gimple_opt_math
32781 resp. PROP_gimple_lvec property set.
32783 2023-01-04 Jakub Jelinek <jakub@redhat.com>
32785 PR sanitizer/108256
32786 * convert.cc (do_narrow): Punt for MULT_EXPR if original
32787 type doesn't wrap around and -fsanitize=signed-integer-overflow
32789 * fold-const.cc (fold_unary_loc) <CASE_CONVERT>: Likewise.
32791 2023-01-04 Hu, Lin1 <lin1.hu@intel.com>
32793 * common/config/i386/cpuinfo.h (get_intel_cpu): Handle Emeraldrapids.
32794 * common/config/i386/i386-common.cc: Add Emeraldrapids.
32796 2023-01-04 Hu, Lin1 <lin1.hu@intel.com>
32798 * common/config/i386/cpuinfo.h (get_intel_cpu): Remove case 0xb5
32801 2023-01-03 Sandra Loosemore <sandra@codesourcery.com>
32803 * cgraph.h (struct cgraph_node): Add gc_candidate bit, modify
32804 default constructor to initialize it.
32805 * cgraphunit.cc (expand_all_functions): Save gc_candidate functions
32806 for last and iterate to handle recursive calls. Delete leftover
32807 candidates at the end.
32808 * omp-simd-clone.cc (simd_clone_create): Set gc_candidate bit
32810 * tree-vect-stmts.cc (vectorizable_simd_clone_call): Clear
32811 gc_candidate bit when a clone is used.
32813 2023-01-03 Florian Weimer <fweimer@redhat.com>
32816 2023-01-02 Florian Weimer <fweimer@redhat.com>
32818 * dwarf2cfi.cc (init_return_column_size): Remove.
32819 (init_one_dwarf_reg_size): Adjust.
32820 (generate_dwarf_reg_sizes): New function. Extracted
32821 from expand_builtin_init_dwarf_reg_sizes.
32822 (expand_builtin_init_dwarf_reg_sizes): Call
32823 generate_dwarf_reg_sizes.
32824 * target.def (init_dwarf_reg_sizes_extra): Adjust
32826 * config/msp430/msp430.cc
32827 (msp430_init_dwarf_reg_sizes_extra): Adjust.
32828 * config/rs6000/rs6000.cc
32829 (rs6000_init_dwarf_reg_sizes_extra): Likewise.
32830 * doc/tm.texi: Update.
32832 2023-01-03 Florian Weimer <fweimer@redhat.com>
32835 2023-01-02 Florian Weimer <fweimer@redhat.com>
32837 * debug.h (dwarf_reg_sizes_constant): Declare.
32838 * dwarf2cfi.cc (dwarf_reg_sizes_constant): New function.
32840 2023-01-03 Siddhesh Poyarekar <siddhesh@gotplt.org>
32842 PR tree-optimization/105043
32843 * doc/extend.texi (Object Size Checking): Split out into two
32844 subsections and mention _FORTIFY_SOURCE.
32846 2023-01-03 Roger Sayle <roger@nextmovesoftware.com>
32848 * config/i386/i386-expand.cc (ix86_expand_int_movcc): Rewrite
32849 RTL expansion to allow condition (mask) to be shared/reused,
32850 by avoiding overwriting pseudos and adding REG_EQUAL notes.
32852 2023-01-03 Roger Sayle <roger@nextmovesoftware.com>
32855 * config/i386/i386-features.cc
32856 (general_scalar_chain::compute_convert_gain) <case PLUS>: Consider
32857 the gain/cost of converting a MEM operand.
32859 2023-01-03 Jakub Jelinek <jakub@redhat.com>
32861 PR middle-end/108264
32862 * expr.cc (store_expr): For stores into SUBREG_PROMOTED_* targets
32863 from source which doesn't have scalar integral mode first convert
32866 2023-01-03 Jakub Jelinek <jakub@redhat.com>
32868 PR rtl-optimization/108263
32869 * cfgrtl.cc (fixup_reorder_chain): Avoid trying to redirect
32872 2023-01-02 Alexander Monakov <amonakov@ispras.ru>
32875 * config/i386/lujiazui.md (lujiazui_div): New automaton.
32876 (lua_div): New unit.
32877 (lua_idiv_qi): Correct unit in the reservation.
32878 (lua_idiv_qi_load): Ditto.
32879 (lua_idiv_hi): Ditto.
32880 (lua_idiv_hi_load): Ditto.
32881 (lua_idiv_si): Ditto.
32882 (lua_idiv_si_load): Ditto.
32883 (lua_idiv_di): Ditto.
32884 (lua_idiv_di_load): Ditto.
32885 (lua_fdiv_SF): Ditto.
32886 (lua_fdiv_SF_load): Ditto.
32887 (lua_fdiv_DF): Ditto.
32888 (lua_fdiv_DF_load): Ditto.
32889 (lua_fdiv_XF): Ditto.
32890 (lua_fdiv_XF_load): Ditto.
32891 (lua_ssediv_SF): Ditto.
32892 (lua_ssediv_load_SF): Ditto.
32893 (lua_ssediv_V4SF): Ditto.
32894 (lua_ssediv_load_V4SF): Ditto.
32895 (lua_ssediv_V8SF): Ditto.
32896 (lua_ssediv_load_V8SF): Ditto.
32897 (lua_ssediv_SD): Ditto.
32898 (lua_ssediv_load_SD): Ditto.
32899 (lua_ssediv_V2DF): Ditto.
32900 (lua_ssediv_load_V2DF): Ditto.
32901 (lua_ssediv_V4DF): Ditto.
32902 (lua_ssediv_load_V4DF): Ditto.
32904 2023-01-02 Florian Weimer <fweimer@redhat.com>
32906 * debug.h (dwarf_reg_sizes_constant): Declare.
32907 * dwarf2cfi.cc (dwarf_reg_sizes_constant): New function.
32909 2023-01-02 Florian Weimer <fweimer@redhat.com>
32911 * dwarf2cfi.cc (init_return_column_size): Remove.
32912 (init_one_dwarf_reg_size): Adjust.
32913 (generate_dwarf_reg_sizes): New function. Extracted
32914 from expand_builtin_init_dwarf_reg_sizes.
32915 (expand_builtin_init_dwarf_reg_sizes): Call
32916 generate_dwarf_reg_sizes.
32917 * target.def (init_dwarf_reg_sizes_extra): Adjust
32919 * config/msp430/msp430.cc
32920 (msp430_init_dwarf_reg_sizes_extra): Adjust.
32921 * config/rs6000/rs6000.cc
32922 (rs6000_init_dwarf_reg_sizes_extra): Likewise.
32923 * doc/tm.texi: Update.
32925 2023-01-02 Jakub Jelinek <jakub@redhat.com>
32927 * gcc.cc (process_command): Update copyright notice dates.
32928 * gcov-dump.cc (print_version): Ditto.
32929 * gcov.cc (print_version): Ditto.
32930 * gcov-tool.cc (print_version): Ditto.
32931 * gengtype.cc (create_file): Ditto.
32932 * doc/cpp.texi: Bump @copying's copyright year.
32933 * doc/cppinternals.texi: Ditto.
32934 * doc/gcc.texi: Ditto.
32935 * doc/gccint.texi: Ditto.
32936 * doc/gcov.texi: Ditto.
32937 * doc/install.texi: Ditto.
32938 * doc/invoke.texi: Ditto.
32940 2023-01-01 Roger Sayle <roger@nextmovesoftware.com>
32941 Uroš Bizjak <ubizjak@gmail.com>
32943 * config/i386/i386.md (extendditi2): New define_insn.
32944 (define_split): Use DWIH mode iterator to treat new extendditi2
32945 identically to existing extendsidi2_1.
32946 (define_peephole2): Likewise.
32947 (define_peephole2): Likewise.
32948 (define_Split): Likewise.
32951 Copyright (C) 2023 Free Software Foundation, Inc.
32953 Copying and distribution of this file, with or without modification,
32954 are permitted in any medium without royalty provided the copyright
32955 notice and this notice are preserved.