1 2023-08-18 Uros Bizjak <ubizjak@gmail.com>
4 * config/i386/i386-expand.cc (ix86_split_mmx_punpck):
6 (ix86_expand_sse_extend): New function.
7 * config/i386/i386-protos.h (ix86_expand_sse_extend): New prototype.
8 * config/i386/mmx.md (<any_extend:insn>v4qiv4hi2): Enable for
9 TARGET_SSE2. Expand through ix86_expand_sse_extend for !TARGET_SSE4_1.
10 (<any_extend:insn>v2hiv2si2): Ditto.
11 (<any_extend:insn>v2qiv2hi2): Ditto.
12 * config/i386/sse.md (<any_extend:insn>v8qiv8hi2): Ditto.
13 (<any_extend:insn>v4hiv4si2): Ditto.
14 (<any_extend:insn>v2siv2di2): Ditto.
16 2023-08-18 Aldy Hernandez <aldyh@redhat.com>
19 * value-range.cc (irange::union_bitmask): Return FALSE if updated
20 bitmask is semantically equivalent to the original mask.
21 (irange::intersect_bitmask): Same.
22 (irange::get_bitmask): Add comment.
24 2023-08-18 Richard Biener <rguenther@suse.de>
26 PR tree-optimization/111019
27 * tree-ssa-loop-im.cc (gather_mem_refs_stmt): When canonicalizing
28 also scrap base and offset in case the ref is indirect.
30 2023-08-18 Jose E. Marchesi <jose.marchesi@oracle.com>
32 * config/bpf/bpf.opt (mframe-limit): Set default to 32767.
34 2023-08-18 Kewen Lin <linkw@linux.ibm.com>
37 * Makefile.in (TM_P_H): Add $(TREE_H) as dependence.
39 2023-08-18 Kewen Lin <linkw@linux.ibm.com>
41 * tree-vect-stmts.cc (vect_build_scatter_store_calls): New, factor
43 (vectorizable_store): ... here.
45 2023-08-18 Richard Biener <rguenther@suse.de>
47 PR tree-optimization/111048
48 * fold-const.cc (fold_vec_perm_cst): Check for non-VLA
51 2023-08-18 Haochen Jiang <haochen.jiang@intel.com>
54 * config/i386/avx512vldqintrin.h: Push AVX2 when AVX2 is
57 2023-08-18 Kewen Lin <linkw@linux.ibm.com>
59 * tree-vect-stmts.cc (vectorizable_load): Move the handlings on
60 VMAT_GATHER_SCATTER in the final loop nest to its own loop,
61 and update the final nest accordingly.
63 2023-08-18 Andrew Pinski <apinski@marvell.com>
65 * doc/md.texi (Standard patterns): Document cond_neg, cond_one_cmpl,
66 cond_len_neg and cond_len_one_cmpl.
68 2023-08-18 Lehua Ding <lehua.ding@rivai.ai>
70 * config/riscv/iterators.md (TARGET_HARD_FLOAT || TARGET_ZFINX): New.
71 * config/riscv/pic.md (*local_pic_load<ANYF:mode>): Change ANYF.
72 (*local_pic_load<ANYLSF:mode>): To ANYLSF.
73 (*local_pic_load_32d<ANYF:mode>): Ditto.
74 (*local_pic_load_32d<ANYLSF:mode>): Ditto.
75 (*local_pic_store<ANYF:mode>): Ditto.
76 (*local_pic_store<ANYLSF:mode>): Ditto.
77 (*local_pic_store_32d<ANYF:mode>): Ditto.
78 (*local_pic_store_32d<ANYLSF:mode>): Ditto.
80 2023-08-18 Lehua Ding <lehua.ding@rivai.ai>
81 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
83 * config/riscv/predicates.md (vector_const_0_operand): New.
84 * config/riscv/vector.md (*pred_broadcast<mode>_zero): Ditto.
86 2023-08-18 Lehua Ding <lehua.ding@rivai.ai>
88 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::backward_demand_fusion):
91 2023-08-17 Andrew MacLeod <amacleod@redhat.com>
93 PR tree-optimization/111009
94 * range-op.cc (operator_addr_expr::op1_range): Be more restrictive.
96 2023-08-17 Vladimir N. Makarov <vmakarov@redhat.com>
98 * lra-spills.cc (assign_stack_slot_num_and_sort_pseudos): Moving
99 slots_num initialization from here ...
100 (lra_spill): ... to here before the 1st call of
101 assign_stack_slot_num_and_sort_pseudos. Add the 2nd call after
104 2023-08-17 Jose E. Marchesi <jose.marchesi@oracle.com>
107 * doc/invoke.texi (Option Summary): Mention
108 -Wcompare-distinct-pointer-types under `Warning Options'.
109 (Warning Options): Document -Wcompare-distinct-pointer-types.
111 2023-08-17 Jan-Benedict Glaw <jbglaw@lug-owl.de>
113 * recog.cc (memory_address_addr_space_p): Mark possibly unused
116 2023-08-17 Richard Biener <rguenther@suse.de>
118 PR tree-optimization/111039
119 * tree-ssa-ifcombine.cc (ifcombine_ifandif): Check for
120 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
122 2023-08-17 Alex Coplan <alex.coplan@arm.com>
124 * doc/rtl.texi: Fix up sample code for RTL-SSA insn changes.
126 2023-08-17 Jose E. Marchesi <jose.marchesi@oracle.com>
129 * config/bpf/bpf.cc (bpf_attribute_table): Add entry for the
130 `naked' function attribute.
131 (bpf_warn_func_return): New function.
132 (TARGET_WARN_FUNC_RETURN): Define.
133 (bpf_expand_prologue): Add preventive comment.
134 (bpf_expand_epilogue): Likewise.
135 * doc/extend.texi (BPF Function Attributes): Document the `naked'
138 2023-08-17 Richard Biener <rguenther@suse.de>
140 * tree-vect-slp.cc (vect_slp_check_for_roots): Use
141 !needs_fold_left_reduction_p to decide whether we can
142 handle the reduction with association.
143 (vectorize_slp_instance_root_stmt): For TYPE_OVERFLOW_UNDEFINED
144 reductions perform all arithmetic in an unsigned type.
146 2023-08-17 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
148 * configure.ac (gcc_cv_ld64_version): Allow for dyld in ld -v
150 * configure: Regenerate.
152 2023-08-17 Pan Li <pan2.li@intel.com>
154 * config/riscv/riscv-vector-builtins-bases.cc
155 (widen_freducop): Add frm_opt_type template arg.
156 (vfwredosum_frm_obj): New declaration.
158 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
159 * config/riscv/riscv-vector-builtins-functions.def
160 (vfwredosum_frm): New intrinsic function def.
162 2023-08-17 Pan Li <pan2.li@intel.com>
164 * config/riscv/riscv-vector-builtins-bases.cc
165 (vfredosum_frm_obj): New declaration.
167 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
168 * config/riscv/riscv-vector-builtins-functions.def
169 (vfredosum_frm): New intrinsic function def.
171 2023-08-17 Pan Li <pan2.li@intel.com>
173 * config/riscv/riscv-vector-builtins-bases.cc
174 (class freducop): Add frm_op_type template arg.
175 (vfredusum_frm_obj): New declaration.
177 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
178 * config/riscv/riscv-vector-builtins-functions.def
179 (vfredusum_frm): New intrinsic function def.
180 * config/riscv/riscv-vector-builtins-shapes.cc
181 (struct reduc_alu_frm_def): New class for frm shape.
182 (SHAPE): New declaration.
183 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
185 2023-08-17 Pan Li <pan2.li@intel.com>
187 * config/riscv/riscv-vector-builtins-bases.cc
188 (class vfncvt_f): Add frm_op_type template arg.
189 (vfncvt_f_frm_obj): New declaration.
191 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
192 * config/riscv/riscv-vector-builtins-functions.def
193 (vfncvt_f_frm): New intrinsic function def.
195 2023-08-17 Pan Li <pan2.li@intel.com>
197 * config/riscv/riscv-vector-builtins-bases.cc
198 (vfncvt_xu_frm_obj): New declaration.
200 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
201 * config/riscv/riscv-vector-builtins-functions.def
202 (vfncvt_xu_frm): New intrinsic function def.
204 2023-08-17 Pan Li <pan2.li@intel.com>
206 * config/riscv/riscv-vector-builtins-bases.cc
207 (class vfncvt_x): Add frm_op_type template arg.
208 (BASE): New declaration.
209 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
210 * config/riscv/riscv-vector-builtins-functions.def
211 (vfncvt_x_frm): New intrinsic function def.
212 * config/riscv/riscv-vector-builtins-shapes.cc
213 (struct narrow_alu_frm_def): New shape function for frm.
214 (SHAPE): New declaration.
215 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
217 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
219 * config/i386/avx512vldqintrin.h: Remove target attribute.
220 * config/i386/i386-builtin.def (BDESC):
221 Add OPTION_MASK_ISA2_AVX10_1.
222 * config/i386/sse.md (VF_AVX512VLDQ_AVX10_1): New.
223 (VFH_AVX512VLDQ_AVX10_1): Ditto.
224 (VF1_AVX512VLDQ_AVX10_1): Ditto.
225 (<mask_codefor>reducep<mode><mask_name><round_saeonly_name>):
226 Change iterator to VFH_AVX512VLDQ_AVX10_1. Remove target check.
227 (vec_pack<floatprefix>_float_<mode>): Change iterator to
228 VI8_AVX512VLDQ_AVX10_1. Remove target check.
229 (vec_unpack_<fixprefix>fix_trunc_lo_<mode>): Change iterator to
230 VF1_AVX512VLDQ_AVX10_1. Remove target check.
231 (vec_unpack_<fixprefix>fix_trunc_hi_<mode>): Ditto.
232 (VI48F_256_DQVL_AVX10_1): Rename from VI48F_256_DQ.
233 (avx512vl_vextractf128<mode>): Change iterator to
234 VI48F_256_DQVL_AVX10_1. Remove target check.
235 (vec_extract_hi_<mode>_mask): Add TARGET_AVX10_1.
236 (vec_extract_hi_<mode>): Ditto.
237 (avx512vl_vinsert<mode>): Ditto.
238 (vec_set_lo_<mode><mask_name>): Ditto.
239 (vec_set_hi_<mode><mask_name>): Ditto.
240 (avx512dq_rangep<mode><mask_name><round_saeonly_name>): Change
241 iterator to VF_AVX512VLDQ_AVX10_1. Remove target check.
242 (avx512dq_fpclass<mode><mask_scalar_merge_name>): Change
243 iterator to VFH_AVX512VLDQ_AVX10_1. Remove target check.
244 * config/i386/subst.md (mask_avx512dq_condition): Add
246 (mask_scalar_merge): Ditto.
248 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
250 * config/i386/avx512vldqintrin.h: Remove target attribute.
251 * config/i386/i386-builtin.def (BDESC):
252 Add OPTION_MASK_ISA2_AVX10_1.
253 * config/i386/i386.cc (standard_sse_constant_opcode): Add TARGET_AVX10_1.
254 * config/i386/sse.md: (VI48_AVX512VL_AVX10_1): New.
255 (VI48_AVX512VLDQ_AVX10_1): Ditto.
256 (VF2_AVX512VL): Remove.
257 (VI8_256_512VLDQ_AVX10_1): Rename from VI8_256_512.
259 (*<code><mode>3<mask_name>): Change isa attribute to
260 avx10_1_or_avx512dq. Add TARGET_AVX10_1.
261 (<code><mode>3): Add TARGET_AVX10_1. Change isa attr
262 to avx10_1_or_avx512vl.
263 (<mask_codefor>avx512dq_cvtps2qq<mode><mask_name><round_name>):
264 Change iterator to VI8_256_512VLDQ_AVX10_1. Remove target check.
265 (<mask_codefor>avx512dq_cvtps2qqv2di<mask_name>):
267 (<mask_codefor>avx512dq_cvtps2uqq<mode><mask_name><round_name>):
268 Change iterator to VI8_256_512VLDQ_AVX10_1. Remove target check.
269 (<mask_codefor>avx512dq_cvtps2uqqv2di<mask_name>):
271 (float<floatunssuffix><sseintvecmodelower><mode>2<mask_name><round_name>):
272 Change iterator to VF2_AVX512VLDQ_AVX10_1. Remove target check.
273 (float<floatunssuffix><sselongvecmodelower><mode>2<mask_name><round_name>):
274 Change iterator to VF1_128_256VLDQ_AVX10_1. Remove target check.
275 (float<floatunssuffix>v4div4sf2<mask_name>):
277 (avx512dq_float<floatunssuffix>v2div2sf2): Ditto.
278 (*avx512dq_float<floatunssuffix>v2div2sf2): Ditto.
279 (float<floatunssuffix>v2div2sf2): Ditto.
280 (float<floatunssuffix>v2div2sf2_mask): Ditto.
281 (*float<floatunssuffix>v2div2sf2_mask): Ditto.
282 (*float<floatunssuffix>v2div2sf2_mask_1): Ditto.
283 (<avx512>_cvt<ssemodesuffix>2mask<mode>):
284 Change iterator to VI48_AVX512VLDQ_AVX10_1. Remove target check.
285 (<avx512>_cvtmask2<ssemodesuffix><mode>): Ditto.
286 (*<avx512>_cvtmask2<ssemodesuffix><mode>):
287 Change iterator to VI48_AVX512VL_AVX10_1. Remove target check.
288 Change when constraint is enabled.
290 2023-08-17 Juzhe-Zhong <juzhe.zhong@rivai.ai>
293 * config/riscv/riscv-vsetvl.cc (float_insn_valid_sew_p): New function.
294 (second_sew_less_than_first_sew_p): Fix bug.
295 (first_sew_less_than_second_sew_p): Ditto.
297 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
299 * config/i386/avx512vldqintrin.h: Remove target attribute.
300 * config/i386/i386-builtin.def (BDESC):
301 Add OPTION_MASK_ISA2_AVX10_1.
302 * config/i386/i386-builtins.cc (def_builtin): Handle AVX10_1.
303 * config/i386/i386-expand.cc
304 (ix86_check_builtin_isa_match): Ditto.
305 (ix86_expand_sse2_mulvxdi3): Add TARGET_AVX10_1.
306 * config/i386/i386.md: Add new isa attribute avx10_1_or_avx512dq
307 and avx10_1_or_avx512vl.
308 * config/i386/sse.md: (VF2_AVX512VLDQ_AVX10_1): New.
309 (VF1_128_256VLDQ_AVX10_1): Ditto.
310 (VI8_AVX512VLDQ_AVX10_1): Ditto.
311 (<sse>_andnot<mode>3<mask_name>):
312 Add TARGET_AVX10_1 and change isa attr from avx512dq to
314 (*andnot<mode>3): Add TARGET_AVX10_1 and change isa attr from
315 avx512vl to avx10_1_or_avx512vl.
316 (fix<fixunssuffix>_trunc<mode><sseintvecmodelower>2<mask_name><round_saeonly_name>):
317 Change iterator to VF2_AVX512VLDQ_AVX10_1. Remove target check.
318 (fix_notrunc<mode><sseintvecmodelower>2<mask_name><round_name>):
320 (ufix_notrunc<mode><sseintvecmodelower>2<mask_name><round_name>):
322 (fix<fixunssuffix>_trunc<mode><sselongvecmodelower>2<mask_name><round_saeonly_name>):
323 Change iterator to VF1_128_256VLDQ_AVX10_1. Remove target check.
324 (avx512dq_fix<fixunssuffix>_truncv2sfv2di2<mask_name>):
326 (fix<fixunssuffix>_truncv2sfv2di2): Ditto.
327 (cond_mul<mode>): Change iterator to VI8_AVX10_1_AVX512DQVL.
329 (avx512dq_mul<mode>3<mask_name>): Ditto.
330 (*avx512dq_mul<mode>3<mask_name>): Ditto.
331 (VI4F_BRCST32x2): Add TARGET_AVX512DQ and TARGET_AVX10_1.
332 (<mask_codefor>avx512dq_broadcast<mode><mask_name>):
334 (VI8F_BRCST64x2): Add TARGET_AVX512DQ and TARGET_AVX10_1.
335 (<mask_codefor>avx512dq_broadcast<mode><mask_name>_1):
337 * config/i386/subst.md (mask_mode512bit_condition): Add TARGET_AVX10_1.
338 (mask_avx512vl_condition): Ditto.
341 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
343 * common/config/i386/i386-common.cc
344 (ix86_check_avx10_vector_width): New function to check isa_flags
345 to emit a warning when there is a conflict in AVX10 options for
347 (ix86_handle_option): Add check for avx10.1-256 and avx10.1-512.
348 * config/i386/driver-i386.cc (host_detect_local_cpu):
349 Do not append -mno-avx10-max-512bit for -march=native.
351 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
353 * common/config/i386/i386-common.cc
354 (ix86_check_avx10): New function to check isa_flags and
355 isa_flags_explicit to emit warning when AVX10 is enabled
357 (ix86_check_avx512): New function to check isa_flags and
358 isa_flags_explicit to emit warning when AVX512 is enabled
360 (ix86_handle_option): Do not change the flags when warning
362 * config/i386/driver-i386.cc (host_detect_local_cpu):
363 Do not append -mno-avx10.1 for -march=native.
365 2023-08-17 Haochen Jiang <haochen.jiang@intel.com>
367 * common/config/i386/cpuinfo.h (get_available_features):
368 Add avx10_set and version and detect avx10.1.
369 (cpu_indicator_init): Handle avx10.1-512.
370 * common/config/i386/i386-common.cc
371 (OPTION_MASK_ISA2_AVX10_512BIT_SET): New.
372 (OPTION_MASK_ISA2_AVX10_1_SET): Ditto.
373 (OPTION_MASK_ISA2_AVX10_512BIT_UNSET): Ditto.
374 (OPTION_MASK_ISA2_AVX10_1_UNSET): Ditto.
375 (OPTION_MASK_ISA2_AVX2_UNSET): Modify for AVX10_1.
376 (ix86_handle_option): Handle -mavx10.1, -mavx10.1-256 and
378 * common/config/i386/i386-cpuinfo.h (enum processor_features):
379 Add FEATURE_AVX10_512BIT, FEATURE_AVX10_1 and
380 FEATURE_AVX10_512BIT.
381 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
382 AVX10_512BIT, AVX10_1 and AVX10_1_512.
383 * config/i386/constraints.md (Yk): Add AVX10_1.
386 * config/i386/cpuid.h (bit_AVX10): New.
387 (bit_AVX10_256): Ditto.
388 (bit_AVX10_512): Ditto.
389 * config/i386/i386-c.cc (ix86_target_macros_internal):
390 Define AVX10_512BIT and AVX10_1.
391 * config/i386/i386-isa.def
392 (AVX10_512BIT): Add DEF_PTA(AVX10_512BIT).
393 (AVX10_1): Add DEF_PTA(AVX10_1).
394 * config/i386/i386-options.cc (isa2_opts): Add -mavx10.1.
395 (ix86_valid_target_attribute_inner_p): Handle avx10-512bit, avx10.1
397 (ix86_option_override_internal): Enable AVX512{F,VL,BW,DQ,CD,BF16,
398 FP16,VBMI,VBMI2,VNNI,IFMA,BITALG,VPOPCNTDQ} features for avx10.1-512.
399 (ix86_valid_target_attribute_inner_p): Handle AVX10_1.
400 * config/i386/i386.cc (ix86_get_ssemov): Add AVX10_1.
401 (ix86_conditional_register_usage): Ditto.
402 (ix86_hard_regno_mode_ok): Ditto.
403 (ix86_rtx_costs): Ditto.
404 * config/i386/i386.h (VALID_MASK_AVX10_MODE): New macro.
405 * config/i386/i386.opt: Add option -mavx10.1, -mavx10.1-256 and
407 * doc/extend.texi: Document avx10.1, avx10.1-256 and avx10.1-512.
408 * doc/invoke.texi: Document -mavx10.1, -mavx10.1-256 and -mavx10.1-512.
409 * doc/sourcebuild.texi: Document target avx10.1, avx10.1-256
412 2023-08-17 Sergei Trofimovich <siarheit@google.com>
414 * flag-types.h (vrp_mode): Remove unused.
416 2023-08-17 Yanzhang Wang <yanzhang.wang@intel.com>
418 * simplify-rtx.cc (simplify_context::simplify_binary_operation_1): Use
421 2023-08-17 Andrew Pinski <apinski@marvell.com>
423 * internal-fn.def (COND_NOT): New internal function.
424 * match.pd (UNCOND_UNARY, COND_UNARY): Add bit_not/not
426 (`vec (a ? -1 : 0) ^ b`): New pattern to convert
427 into conditional not.
428 * optabs.def (cond_one_cmpl): New optab.
429 (cond_len_one_cmpl): Likewise.
431 2023-08-16 Surya Kumari Jangala <jskumari@linux.ibm.com>
433 PR rtl-optimization/110254
434 * ira-color.cc (improve_allocation): Update array
435 allocated_hard_reg_p.
437 2023-08-16 Vladimir N. Makarov <vmakarov@redhat.com>
439 * lra-int.h (lra_update_fp2sp_elimination): Change the prototype.
440 * lra-eliminations.cc (spill_pseudos): Record spilled pseudos.
441 (lra_update_fp2sp_elimination): Ditto.
442 (update_reg_eliminate): Adjust spill_pseudos call.
443 * lra-spills.cc (lra_spill): Assign stack slots to pseudos spilled
444 in lra_update_fp2sp_elimination.
446 2023-08-16 Richard Ball <richard.ball@arm.com>
448 * config/aarch64/aarch64-cores.def (AARCH64_CORE): Add Cortex-A720 CPU.
449 * config/aarch64/aarch64-tune.md: Regenerate.
450 * doc/invoke.texi: Document Cortex-A720 CPU.
452 2023-08-16 Robin Dapp <rdapp@ventanamicro.com>
454 * config/riscv/autovec.md (<u>avg<v_double_trunc>3_floor):
456 (<u>avg<v_double_trunc>3_ceil): Ditto.
457 * config/riscv/vector-iterators.md (ashiftrt): New iterator.
460 2023-08-16 Robin Dapp <rdapp@ventanamicro.com>
462 * internal-fn.cc (vec_extract_direct): Change type argument
464 (expand_vec_extract_optab_fn): Call convert_optab_fn.
465 (direct_vec_extract_optab_supported_p): Use
466 convert_optab_supported_p.
468 2023-08-16 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
469 Richard Sandiford <richard.sandiford@arm.com>
471 * fold-const.cc (INCLUDE_ALGORITHM): Add Include.
472 (valid_mask_for_fold_vec_perm_cst_p): New function.
473 (fold_vec_perm_cst): Likewise.
474 (fold_vec_perm): Adjust assert and call fold_vec_perm_cst.
475 (test_fold_vec_perm_cst): New namespace.
476 (test_fold_vec_perm_cst::build_vec_cst_rand): New function.
477 (test_fold_vec_perm_cst::validate_res): Likewise.
478 (test_fold_vec_perm_cst::validate_res_vls): Likewise.
479 (test_fold_vec_perm_cst::builder_push_elems): Likewise.
480 (test_fold_vec_perm_cst::test_vnx4si_v4si): Likewise.
481 (test_fold_vec_perm_cst::test_v4si_vnx4si): Likewise.
482 (test_fold_vec_perm_cst::test_all_nunits): Likewise.
483 (test_fold_vec_perm_cst::test_nunits_min_2): Likewise.
484 (test_fold_vec_perm_cst::test_nunits_min_4): Likewise.
485 (test_fold_vec_perm_cst::test_nunits_min_8): Likewise.
486 (test_fold_vec_perm_cst::test_nunits_max_4): Likewise.
487 (test_fold_vec_perm_cst::is_simple_vla_size): Likewise.
488 (test_fold_vec_perm_cst::test): Likewise.
489 (fold_const_cc_tests): Call test_fold_vec_perm_cst::test.
491 2023-08-16 Pan Li <pan2.li@intel.com>
493 * config/riscv/riscv-vector-builtins-bases.cc
494 (BASE): New declaration.
495 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
496 * config/riscv/riscv-vector-builtins-functions.def
497 (vfwcvt_xu_frm): New intrinsic function def.
499 2023-08-16 Pan Li <pan2.li@intel.com>
501 * config/riscv/riscv-vector-builtins-bases.cc: Use explicit argument.
503 2023-08-16 Pan Li <pan2.li@intel.com>
505 * config/riscv/riscv-vector-builtins-bases.cc
506 (BASE): New declaration.
507 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
508 * config/riscv/riscv-vector-builtins-functions.def
509 (vfwcvt_x_frm): New intrinsic function def.
511 2023-08-16 Pan Li <pan2.li@intel.com>
513 * config/riscv/riscv-vector-builtins-bases.cc (BASE): New declaration.
514 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
515 * config/riscv/riscv-vector-builtins-functions.def
516 (vfcvt_f_frm): New intrinsic function def.
518 2023-08-16 Pan Li <pan2.li@intel.com>
520 * config/riscv/riscv-vector-builtins-bases.cc
521 (BASE): New declaration.
522 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
523 * config/riscv/riscv-vector-builtins-functions.def
524 (vfcvt_xu_frm): New intrinsic function def..
526 2023-08-16 Haochen Gui <guihaoc@gcc.gnu.org>
529 * config/rs6000/vsx.md (*vsx_extract_<mode>_store_p9): Skip vector
530 extract when the element is 7 on BE while 8 on LE for byte or 3 on
531 BE while 4 on LE for halfword.
533 2023-08-16 Haochen Gui <guihaoc@gcc.gnu.org>
536 * config/rs6000/vsx.md (expand vsx_extract_<mode>): Set it only
538 (vsx_extract_v4si): New expand for V4SI extraction.
539 (vsx_extract_v4si_w1): New insn pattern for V4SI extraction on
540 word 1 from BE order.
541 (*mfvsrwz): New insn pattern for mfvsrwz.
542 (*vsx_extract_<mode>_di_p9): Assert that it won't be generated on
543 word 1 from BE order.
544 (*vsx_extract_si): Remove.
545 (*vsx_extract_v4si_w023): New insn and split pattern on word 0, 2,
548 2023-08-16 Juzhe-Zhong <juzhe.zhong@rivai.ai>
550 * config/riscv/autovec.md (vec_mask_len_load_lanes<mode><vsingle>):
552 (vec_mask_len_store_lanes<mode><vsingle>): Ditto.
553 * config/riscv/riscv-protos.h (expand_lanes_load_store): New function.
554 * config/riscv/riscv-v.cc (get_mask_mode): Add tuple mask mode.
555 (expand_lanes_load_store): New function.
556 * config/riscv/vector-iterators.md: New iterator.
558 2023-08-16 Juzhe-Zhong <juzhe.zhong@rivai.ai>
560 * internal-fn.cc (internal_load_fn_p): Apply
561 MASK_LEN_{LOAD_LANES,STORE_LANES} into vectorizer.
562 (internal_store_fn_p): Ditto.
563 (internal_fn_len_index): Ditto.
564 (internal_fn_mask_index): Ditto.
565 (internal_fn_stored_value_index): Ditto.
566 * tree-vect-data-refs.cc (vect_store_lanes_supported): Ditto.
567 (vect_load_lanes_supported): Ditto.
568 * tree-vect-loop.cc: Ditto.
569 * tree-vect-slp.cc (vect_slp_prefer_store_lanes_p): Ditto.
570 * tree-vect-stmts.cc (check_load_store_for_partial_vectors): Ditto.
571 (get_group_load_store_type): Ditto.
572 (vectorizable_store): Ditto.
573 (vectorizable_load): Ditto.
574 * tree-vectorizer.h (vect_store_lanes_supported): Ditto.
575 (vect_load_lanes_supported): Ditto.
577 2023-08-16 Pan Li <pan2.li@intel.com>
579 * config/riscv/riscv-vector-builtins-bases.cc
580 (enum frm_op_type): New type for frm.
581 (BASE): New declaration.
582 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
583 * config/riscv/riscv-vector-builtins-functions.def
584 (vfcvt_x_frm): New intrinsic function def.
586 2023-08-16 liuhongt <hongtao.liu@intel.com>
588 * config/i386/i386-builtins.cc
589 (ix86_vectorize_builtin_gather): Adjust for use_gather_8parts.
590 * config/i386/i386-options.cc (parse_mtune_ctrl_str):
591 Set/Clear tune features use_{gather,scatter}_{2parts, 4parts,
592 8parts} for -mtune-crtl={,^}{use_gather,use_scatter}.
593 * config/i386/i386.cc (ix86_vectorize_builtin_scatter): Adjust
594 for use_scatter_8parts
595 * config/i386/i386.h (TARGET_USE_GATHER): Rename to ..
596 (TARGET_USE_GATHER_8PARTS): .. this.
597 (TARGET_USE_SCATTER): Rename to ..
598 (TARGET_USE_SCATTER_8PARTS): .. this.
599 * config/i386/x86-tune.def (X86_TUNE_USE_GATHER): Rename to
600 (X86_TUNE_USE_GATHER_8PARTS): .. this.
601 (X86_TUNE_USE_SCATTER): Rename to
602 (X86_TUNE_USE_SCATTER_8PARTS): .. this.
603 * config/i386/i386.opt: Add new options mgather, mscatter.
605 2023-08-16 liuhongt <hongtao.liu@intel.com>
607 * config/i386/i386-options.cc (m_GDS): New macro.
608 * config/i386/x86-tune.def (X86_TUNE_USE_GATHER_2PARTS): Don't
610 (X86_TUNE_USE_GATHER_4PARTS): Ditto.
611 (X86_TUNE_USE_GATHER): Ditto.
613 2023-08-16 liuhongt <hongtao.liu@intel.com>
615 * config/i386/i386.md (movdf_internal): Generate vmovapd instead of
616 vmovsd when moving DFmode between SSE_REGS.
617 (movhi_internal): Generate vmovdqa instead of vmovsh when
618 moving HImode between SSE_REGS.
619 (mov<mode>_internal): Use vmovaps instead of vmovsh when
620 moving HF/BFmode between SSE_REGS.
622 2023-08-15 David Faust <david.faust@oracle.com>
624 * config/bpf/bpf.md (extendsisi2): Delete useless define_insn.
626 2023-08-15 David Faust <david.faust@oracle.com>
629 * config/bpf/bpf.cc (bpf_print_register): Print 'w' registers
630 for any mode 32-bits or smaller, not just SImode.
632 2023-08-15 Martin Jambor <mjambor@suse.cz>
636 * ipa-prop.h (ipcp_get_aggregate_const): Declare.
637 * ipa-prop.cc (ipcp_get_aggregate_const): New function.
638 (ipcp_transform_function): Do not deallocate transformation info.
639 * tree-ssa-sccvn.cc: Include alloc-pool.h, symbol-summary.h and
641 (vn_reference_lookup_2): When hitting default-def vuse, query
642 IPA-CP transformation info for any known constants.
644 2023-08-15 Chung-Lin Tang <cltang@codesourcery.com>
645 Thomas Schwinge <thomas@codesourcery.com>
647 * gimplify.cc (oacc_region_type_name): New function.
648 (oacc_default_clause): If no 'default' clause appears on this
649 compute construct, see if one appears on a lexically containing
651 (gimplify_scan_omp_clauses): Upon OMP_CLAUSE_DEFAULT case, set
652 ctx->oacc_default_clause_ctx to current context.
654 2023-08-15 Juzhe-Zhong <juzhe.zhong@rivai.ai>
657 * config/riscv/predicates.md: Fix predicate.
659 2023-08-15 Richard Biener <rguenther@suse.de>
661 * tree-vect-slp.cc (vect_analyze_slp_instance): Remove
662 slp_inst_kind_ctor handling.
663 (vect_analyze_slp): Simplify.
664 (vect_build_slp_instance): Dump when we analyze a CTOR.
665 (vect_slp_check_for_constructors): Rename to ...
666 (vect_slp_check_for_roots): ... this. Register a
667 slp_root for CONSTRUCTORs instead of shoving them to
668 the set of grouped stores.
669 (vect_slp_analyze_bb_1): Adjust.
671 2023-08-15 Richard Biener <rguenther@suse.de>
673 * tree-vectorizer.h (_slp_instance::remain_stmts): Change
675 (_slp_instance::remain_defs): ... this.
676 (SLP_INSTANCE_REMAIN_STMTS): Rename to ...
677 (SLP_INSTANCE_REMAIN_DEFS): ... this.
678 (slp_root::remain): New.
679 (slp_root::slp_root): Adjust.
680 * tree-vect-slp.cc (vect_free_slp_instance): Adjust.
681 (vect_build_slp_instance): Get extra remain parameter,
682 adjust former handling of a cut off stmt.
683 (vect_analyze_slp_instance): Adjust.
684 (vect_analyze_slp): Likewise.
685 (_bb_vec_info::~_bb_vec_info): Likewise.
686 (vectorizable_bb_reduc_epilogue): Dump something if we fail.
687 (vect_slp_check_for_constructors): Handle non-internal
688 defs as remain defs of a reduction.
689 (vectorize_slp_instance_root_stmt): Adjust.
691 2023-08-15 Richard Biener <rguenther@suse.de>
693 * tree-ssa-loop-ivcanon.cc: Include tree-vectorizer.h
694 (canonicalize_loop_induction_variables): Use find_loop_location.
696 2023-08-15 Hans-Peter Nilsson <hp@axis.com>
699 * config/cris/cris-protos.h: Revert recent change.
700 * config/cris/cris.cc (cris_legitimate_address_p): Remove
701 code_helper unused parameter.
702 (cris_legitimate_address_p_hook): New wrapper function.
703 (TARGET_LEGITIMATE_ADDRESS_P): Change to
704 cris_legitimate_address_p_hook.
706 2023-08-15 Richard Biener <rguenther@suse.de>
708 PR tree-optimization/110963
709 * tree-ssa-pre.cc (do_pre_regular_insertion): Also insert
710 a PHI node when the expression is available on all edges
711 and we insert at most one copy from a constant.
713 2023-08-15 Richard Biener <rguenther@suse.de>
715 PR tree-optimization/110991
716 * tree-ssa-loop-ivcanon.cc (constant_after_peeling): Handle
717 VIEW_CONVERT_EXPR <op>, handle more simple IV-like SSA cycles
718 that will end up constant.
720 2023-08-15 Kewen Lin <linkw@linux.ibm.com>
723 * Makefile.in (RECOG_H): Add $(TREE_H) as dependence.
725 2023-08-15 Kewen Lin <linkw@linux.ibm.com>
727 * tree-vect-stmts.cc (vectorizable_load): Move the handlings on
728 VMAT_LOAD_STORE_LANES in the final loop nest to its own loop,
729 and update the final nest accordingly.
731 2023-08-15 Kewen Lin <linkw@linux.ibm.com>
733 * tree-vect-stmts.cc (vectorizable_load): Remove some useless checks
736 2023-08-15 Pan Li <pan2.li@intel.com>
738 * mode-switching.cc (create_pre_exit): Add SET insn check.
740 2023-08-15 Pan Li <pan2.li@intel.com>
742 * config/riscv/riscv-vector-builtins-bases.cc
743 (class vfrec7_frm): New class for frm.
744 (vfrec7_frm_obj): New declaration.
746 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
747 * config/riscv/riscv-vector-builtins-functions.def
748 (vfrec7_frm): New intrinsic function definition.
749 * config/riscv/vector-iterators.md
750 (VFMISC): Remove VFREC7.
752 (float_insn_type): Ditto.
753 (VFMISC_FRM): New int iterator.
754 (misc_frm_op): New op for frm.
755 (float_frm_insn_type): New type for frm.
756 * config/riscv/vector.md (@pred_<misc_frm_op><mode>):
757 New pattern for misc frm.
759 2023-08-14 Vladimir N. Makarov <vmakarov@redhat.com>
761 * lra-constraints.cc (curr_insn_transform): Process output stack
762 pointer reloads before emitting reload insns.
764 2023-08-14 benjamin priour <vultkayn@gcc.gnu.org>
767 * doc/invoke.texi: Add documentation of
768 fanalyzer-show-events-in-system-headers
770 2023-08-14 Jan Hubicka <jh@suse.cz>
772 PR gcov-profile/110988
773 * tree-cfg.cc (fold_loop_internal_call): Avoid division by zero.
775 2023-08-14 Jiawei <jiawei@iscas.ac.cn>
777 * config/riscv/riscv-c.cc (riscv_cpu_cpp_builtins):
778 Enable compressed builtins when ZC* extensions enabled.
779 * config/riscv/riscv-shorten-memrefs.cc:
780 Enable shorten_memrefs pass when ZC* extensions enabled.
781 * config/riscv/riscv.cc (riscv_compressed_reg_p):
782 Enable compressible registers when ZC* extensions enabled.
783 (riscv_rtx_costs): Allow adjusting rtx costs when ZC* extensions enabled.
784 (riscv_address_cost): Allow adjusting address cost when ZC* extensions enabled.
785 (riscv_first_stack_step): Allow compression of the register saves
786 without adding extra instructions.
787 * config/riscv/riscv.h (FUNCTION_BOUNDARY): Adjusts function boundary
788 to 16 bits when ZC* extensions enabled.
790 2023-08-14 Jiawei <jiawei@iscas.ac.cn>
792 * common/config/riscv/riscv-common.cc (riscv_subset_list::parse): New extensions.
793 * config/riscv/riscv-opts.h (MASK_ZCA): New mask.
800 (TARGET_ZCA): New target.
805 (TARGET_ZCMP): Ditto.
806 (TARGET_ZCMT): Ditto.
807 * config/riscv/riscv.opt: New target variable.
809 2023-08-14 Juzhe-Zhong <juzhe.zhong@rivai.ai>
812 2023-05-17 Jin Ma <jinma@linux.alibaba.com>
814 * genrecog.cc (print_nonbool_test): Fix type error of
815 switch (SUBREG_BYTE (op))'.
817 2023-08-14 Richard Biener <rguenther@suse.de>
819 * tree-cfg.cc (print_loop_info): Dump to 'file', not 'dump_file'.
821 2023-08-14 Pan Li <pan2.li@intel.com>
823 * config/riscv/riscv-vector-builtins-bases.cc
824 (class unop_frm): New class for frm.
825 (vfsqrt_frm_obj): New declaration.
827 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
828 * config/riscv/riscv-vector-builtins-functions.def
829 (vfsqrt_frm): New intrinsic function definition.
831 2023-08-14 Pan Li <pan2.li@intel.com>
833 * config/riscv/riscv-vector-builtins-bases.cc
834 (class vfwnmsac_frm): New class for frm.
835 (vfwnmsac_frm_obj): New declaration.
837 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
838 * config/riscv/riscv-vector-builtins-functions.def
839 (vfwnmsac_frm): New intrinsic function definition.
841 2023-08-14 Pan Li <pan2.li@intel.com>
843 * config/riscv/riscv-vector-builtins-bases.cc
844 (class vfwmsac_frm): New class for frm.
845 (vfwmsac_frm_obj): New declaration.
847 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
848 * config/riscv/riscv-vector-builtins-functions.def
849 (vfwmsac_frm): New intrinsic function definition.
851 2023-08-14 Pan Li <pan2.li@intel.com>
853 * config/riscv/riscv-vector-builtins-bases.cc
854 (class vfwnmacc_frm): New class for frm.
855 (vfwnmacc_frm_obj): New declaration.
857 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
858 * config/riscv/riscv-vector-builtins-functions.def
859 (vfwnmacc_frm): New intrinsic function definition.
861 2023-08-14 Cui, Lili <lili.cui@intel.com>
863 * common/config/i386/cpuinfo.h (get_intel_cpu): Add model value 0xba
866 2023-08-14 Hans-Peter Nilsson <hp@axis.com>
868 * config/mmix/predicates.md (mmix_address_operand): Use
869 lra_in_progress, not reload_in_progress.
871 2023-08-14 Hans-Peter Nilsson <hp@axis.com>
873 * config/mmix/mmix.cc: Re-enable LRA.
875 2023-08-14 Hans-Peter Nilsson <hp@axis.com>
877 * config/mmix/predicates.md (frame_pointer_operand): Handle FP+offset
878 when lra_in_progress.
880 2023-08-14 Hans-Peter Nilsson <hp@axis.com>
882 * config/mmix/mmix.cc: Disable LRA for MMIX.
884 2023-08-14 Pan Li <pan2.li@intel.com>
886 * config/riscv/riscv-vector-builtins-bases.cc
887 (class vfwmacc_frm): New class for vfwmacc frm.
888 (vfwmacc_frm_obj): New declaration.
890 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
891 * config/riscv/riscv-vector-builtins-functions.def
892 (vfwmacc_frm): Function definition for vfwmacc.
893 * config/riscv/riscv-vector-builtins.cc
894 (function_expander::use_widen_ternop_insn): Add frm support.
896 2023-08-14 Pan Li <pan2.li@intel.com>
898 * config/riscv/riscv-vector-builtins-bases.cc
899 (class vfnmsub_frm): New class for vfnmsub frm.
900 (vfnmsub_frm): New declaration.
902 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
903 * config/riscv/riscv-vector-builtins-functions.def
904 (vfnmsub_frm): New function declaration.
906 2023-08-14 Vladimir N. Makarov <vmakarov@redhat.com>
908 * lra-constraints.cc (curr_insn_transform): Set done_p up and
909 check it on true after processing output stack pointer reload.
911 2023-08-12 Jakub Jelinek <jakub@redhat.com>
913 * Makefile.in (USER_H): Add stdckdint.h.
914 * ginclude/stdckdint.h: New file.
916 2023-08-12 Juzhe-Zhong <juzhe.zhong@rivai.ai>
919 * config/riscv/riscv-opts.h (TARGET_VECTOR_VLS): Add TARGET_VETOR.
921 2023-08-12 Patrick Palka <ppalka@redhat.com>
923 * tree-pretty-print.cc (dump_generic_node) <case TREE_VEC>:
924 Delimit output with braces.
926 2023-08-12 Juzhe-Zhong <juzhe.zhong@rivai.ai>
929 * config/riscv/riscv-v.cc (expand_vec_series): Refactor the expander.
931 2023-08-12 Juzhe-Zhong <juzhe.zhong@rivai.ai>
933 * config/riscv/autovec.md: Add VLS CONST_VECTOR.
934 * config/riscv/riscv.cc (riscv_const_insns): Ditto.
935 * config/riscv/vector.md: Ditto.
937 2023-08-11 David Malcolm <dmalcolm@redhat.com>
940 * doc/analyzer.texi (__analyzer_get_strlen): New.
941 * doc/invoke.texi: Add -Wanalyzer-unterminated-string.
943 2023-08-11 Jeff Law <jlaw@ventanamicro.com>
945 * config/rx/rx.md (subdi3): Fix test for borrow.
947 2023-08-11 Juzhe-Zhong <juzhe.zhong@rivai.ai>
950 * tree-vect-stmts.cc (vectorizable_store): Replace iv_type with sizetype.
951 (vectorizable_load): Ditto.
953 2023-08-11 Jose E. Marchesi <jose.marchesi@oracle.com>
955 * config/bpf/bpf.md (allocate_stack): Define.
956 * config/bpf/bpf.h (FIRST_PSEUDO_REGISTER): Make room for fake
957 stack pointer register.
958 (FIXED_REGISTERS): Adjust accordingly.
959 (CALL_USED_REGISTERS): Likewise.
960 (REG_CLASS_CONTENTS): Likewise.
961 (REGISTER_NAMES): Likewise.
962 * config/bpf/bpf.cc (bpf_compute_frame_layout): Do not reserve
963 space for callee-saved registers.
964 (bpf_expand_prologue): Do not save callee-saved registers in xbpf.
965 (bpf_expand_epilogue): Do not restore callee-saved registers in
968 2023-08-11 Jose E. Marchesi <jose.marchesi@oracle.com>
970 * config/bpf/bpf.cc (bpf_function_arg_advance): Do not complain
971 about too many arguments if function is always inlined.
973 2023-08-11 Patrick Palka <ppalka@redhat.com>
975 * tree-pretty-print.cc (dump_generic_node) <case COMPONENT_REF>:
976 Don't call component_ref_field_offset if the RHS isn't a decl.
978 2023-08-11 John David Anglin <danglin@gcc.gnu.org>
981 * gensupport.cc(class conlist): Use strtol instead of std::stoi.
983 2023-08-11 Vladimir N. Makarov <vmakarov@redhat.com>
985 * lra-constraints.cc (goal_alt_out_sp_reload_p): New flag.
986 (process_alt_operands): Set the flag.
987 (curr_insn_transform): Modify stack pointer offsets if output
988 stack pointer reload is generated.
990 2023-08-11 Joseph Myers <joseph@codesourcery.com>
992 * configure: Regenerate.
994 2023-08-11 Richard Biener <rguenther@suse.de>
996 PR tree-optimization/110979
997 * tree-vect-loop.cc (vectorizable_reduction): For
998 FOLD_LEFT_REDUCTION without target support make sure
999 we don't need to honor signed zeros and sign dependent rounding.
1001 2023-08-11 Richard Biener <rguenther@suse.de>
1003 * tree-vect-slp.cc (vect_slp_region): Provide opt-info for all SLP
1004 subgraph entries. Dump the used vector size based on the
1005 SLP subgraph entry root vector type.
1007 2023-08-11 Pan Li <pan2.li@intel.com>
1009 * config/riscv/riscv-vector-builtins-bases.cc
1010 (class vfmsub_frm): New class for vfmsub frm.
1011 (vfmsub_frm): New declaration.
1013 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1014 * config/riscv/riscv-vector-builtins-functions.def
1015 (vfmsub_frm): New function declaration.
1017 2023-08-11 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1019 * doc/md.texi: Add vec_mask_len_{load_lanes,store_lanes} patterns.
1020 * internal-fn.cc (expand_partial_load_optab_fn): Ditto.
1021 (expand_partial_store_optab_fn): Ditto.
1022 * internal-fn.def (MASK_LEN_LOAD_LANES): Ditto.
1023 (MASK_LEN_STORE_LANES): Ditto.
1024 * optabs.def (OPTAB_CD): Ditto.
1026 2023-08-11 Pan Li <pan2.li@intel.com>
1028 * config/riscv/riscv-vector-builtins-bases.cc
1029 (class vfnmadd_frm): New class for vfnmadd frm.
1030 (vfnmadd_frm): New declaration.
1032 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1033 * config/riscv/riscv-vector-builtins-functions.def
1034 (vfnmadd_frm): New function declaration.
1036 2023-08-11 Drew Ross <drross@redhat.com>
1037 Jakub Jelinek <jakub@redhat.com>
1039 PR tree-optimization/109938
1040 * match.pd (((x ^ y) & z) | x -> (z & y) | x): New simplification.
1042 2023-08-11 Pan Li <pan2.li@intel.com>
1044 * config/riscv/riscv-vector-builtins-bases.cc
1045 (class vfmadd_frm): New class for vfmadd frm.
1046 (vfmadd_frm_obj): New declaration.
1048 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1049 * config/riscv/riscv-vector-builtins-functions.def
1050 (vfmadd_frm): New function definition.
1052 2023-08-11 Pan Li <pan2.li@intel.com>
1054 * config/riscv/riscv-vector-builtins-bases.cc
1055 (class vfnmsac_frm): New class for vfnmsac frm.
1056 (vfnmsac_frm_obj): New declaration.
1058 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1059 * config/riscv/riscv-vector-builtins-functions.def
1060 (vfnmsac_frm): New function definition.
1062 2023-08-11 Jakub Jelinek <jakub@redhat.com>
1064 * doc/extend.texi (Typeof): Document typeof_unqual
1065 and __typeof_unqual__.
1067 2023-08-11 Andrew Pinski <apinski@marvell.com>
1069 PR tree-optimization/110954
1070 * generic-match-head.cc (bitwise_inverted_equal_p): Add
1071 wascmp argument and set it accordingly.
1072 * gimple-match-head.cc (bitwise_inverted_equal_p): Add
1073 wascmp argument to the macro.
1074 (gimple_bitwise_inverted_equal_p): Add
1075 wascmp argument and set it accordingly.
1076 * match.pd (`a & ~a`, `a ^| ~a`): Update call
1077 to bitwise_inverted_equal_p and handle wascmp case.
1078 (`(~x | y) & x`, `(~x | y) & x`, `a?~t:t`): Update
1079 call to bitwise_inverted_equal_p and check to see
1080 if was !wascmp or if precision was 1.
1082 2023-08-11 Martin Uecker <uecker@tugraz.at>
1085 * doc/invoke.texi: Update.
1087 2023-08-11 Pan Li <pan2.li@intel.com>
1089 * config/riscv/riscv-vector-builtins-bases.cc
1090 (class vfmsac_frm): New class for vfmsac frm.
1091 (vfmsac_frm_obj): New declaration.
1093 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1094 * config/riscv/riscv-vector-builtins-functions.def
1095 (vfmsac_frm): New function definition
1097 2023-08-10 Jan Hubicka <jh@suse.cz>
1099 PR middle-end/110923
1100 * tree-ssa-loop-split.cc (split_loop): Watch for division by zero.
1102 2023-08-10 Patrick O'Neill <patrick@rivosinc.com>
1104 * common/config/riscv/riscv-common.cc: Add Ztso and mark Ztso as
1105 dependent on 'a' extension.
1106 * config/riscv/riscv-opts.h (MASK_ZTSO): New mask.
1107 (TARGET_ZTSO): New target.
1108 * config/riscv/riscv.cc (riscv_memmodel_needs_amo_acquire): Add
1110 (riscv_memmodel_needs_amo_release): Add Ztso case.
1111 (riscv_print_operand): Add Ztso case for LR/SC annotations.
1112 * config/riscv/riscv.md: Import sync-rvwmo.md and sync-ztso.md.
1113 * config/riscv/riscv.opt: Add Ztso target variable.
1114 * config/riscv/sync.md (mem_thread_fence_1): Expand to RVWMO or
1116 (atomic_load<mode>): Expand to RVWMO or Ztso specific insn.
1117 (atomic_store<mode>): Expand to RVWMO or Ztso specific insn.
1118 * config/riscv/sync-rvwmo.md: New file. Seperate out RVWMO
1119 specific load/store/fence mappings.
1120 * config/riscv/sync-ztso.md: New file. Seperate out Ztso
1121 specific load/store/fence mappings.
1123 2023-08-10 Jan Hubicka <jh@suse.cz>
1125 * cfgloopmanip.cc (duplicate_loop_body_to_header_edge): Special case loops with
1128 2023-08-10 Jan Hubicka <jh@suse.cz>
1130 * tree-ssa-threadupdate.cc (ssa_fix_duplicate_block_edges): Fix profile update.
1132 2023-08-10 Jan Hubicka <jh@suse.cz>
1134 * profile-count.cc (profile_count::differs_from_p): Fix overflow and
1135 handling of undefined values.
1137 2023-08-10 Jakub Jelinek <jakub@redhat.com>
1140 * tree-ssa-phiopt.cc (single_non_singleton_phi_for_edges): Never
1141 return virtual phis and return NULL if there is a virtual phi
1142 where the arguments from E0 and E1 edges aren't equal.
1144 2023-08-10 Richard Biener <rguenther@suse.de>
1146 * internal-fn.def (VCOND, VCONDU, VCONDEQ, VCOND_MASK,
1147 VEC_SET, VEC_EXTRACT): Make ECF_CONST | ECF_NOTHROW.
1149 2023-08-10 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1152 * config/riscv/autovec.md (vec_duplicate<mode>): New pattern.
1154 2023-08-10 Pan Li <pan2.li@intel.com>
1156 * config/riscv/riscv-vector-builtins-bases.cc
1157 (class vfnmacc_frm): New class for vfnmacc.
1158 (vfnmacc_frm_obj): New declaration.
1160 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1161 * config/riscv/riscv-vector-builtins-functions.def
1162 (vfnmacc_frm): New function definition.
1164 2023-08-10 Pan Li <pan2.li@intel.com>
1166 * config/riscv/riscv-vector-builtins-bases.cc
1167 (class vfmacc_frm): New class for vfmacc frm.
1168 (vfmacc_frm_obj): New declaration.
1170 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
1171 * config/riscv/riscv-vector-builtins-functions.def
1172 (vfmacc_frm): New function definition.
1174 2023-08-10 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1177 * config/riscv/riscv-v.cc (expand_cond_len_ternop): Add integer ternary.
1179 2023-08-10 Richard Biener <rguenther@suse.de>
1181 * tree-vectorizer.h (vectorizable_live_operation): Remove
1182 gimple_stmt_iterator * argument.
1183 * tree-vect-loop.cc (vectorizable_live_operation): Likewise.
1184 Adjust plumbing around vect_get_loop_mask.
1185 (vect_analyze_loop_operations): Adjust.
1186 * tree-vect-slp.cc (vect_slp_analyze_node_operations_1): Likewise.
1187 (vect_bb_slp_mark_live_stmts): Likewise.
1188 (vect_schedule_slp_node): Likewise.
1189 * tree-vect-stmts.cc (can_vectorize_live_stmts): Likewise.
1190 Remove gimple_stmt_iterator * argument.
1191 (vect_transform_stmt): Adjust.
1193 2023-08-10 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1195 * config/riscv/vector-iterators.md: Add missing modes.
1197 2023-08-10 Jakub Jelinek <jakub@redhat.com>
1200 * lto-streamer-in.cc (lto_input_tree_1): Assert TYPE_PRECISION
1201 is up to WIDE_INT_MAX_PRECISION rather than MAX_BITSIZE_MODE_ANY_INT.
1203 2023-08-10 Jakub Jelinek <jakub@redhat.com>
1206 * expr.cc (expand_expr_real_1) <case MEM_REF>: Add an early return for
1207 EXPAND_WRITE or EXPAND_MEMORY modifiers to avoid testing it multiple
1210 2023-08-10 liuhongt <hongtao.liu@intel.com>
1213 * config/i386/mmx.md: (movq_<mode>_to_sse): Also do not
1214 sanitize upper part of V4HFmode register with
1216 (<insn>v4hf3): Enable for ix86_partial_vec_fp_math.
1218 (<insn>v2hf3): Ditto.
1220 (movd_v2hf_to_sse): Do not sanitize upper part of V2HFmode
1221 register with -fno-trapping-math.
1223 2023-08-10 Pan Li <pan2.li@intel.com>
1224 Kito Cheng <kito.cheng@sifive.com>
1226 * config/riscv/riscv-protos.h
1227 (enum floating_point_rounding_mode): Add NONE, DYN_EXIT and DYN_CALL.
1228 (get_frm_mode): New declaration.
1229 * config/riscv/riscv-v.cc (get_frm_mode): New function to get frm mode.
1230 * config/riscv/riscv-vector-builtins.cc
1231 (function_expander::use_ternop_insn): Take care of frm reg.
1232 * config/riscv/riscv.cc (riscv_static_frm_mode_p): Migrate to FRM_XXX.
1233 (riscv_emit_frm_mode_set): Ditto.
1234 (riscv_emit_mode_set): Ditto.
1235 (riscv_frm_adjust_mode_after_call): Ditto.
1236 (riscv_frm_mode_needed): Ditto.
1237 (riscv_frm_mode_after): Ditto.
1238 (riscv_mode_entry): Ditto.
1239 (riscv_mode_exit): Ditto.
1240 * config/riscv/riscv.h (NUM_MODES_FOR_MODE_SWITCHING): Ditto.
1241 * config/riscv/vector.md
1242 (rne,rtz,rdn,rup,rmm,dyn,dyn_exit,dyn_call,none): Removed
1243 (symbol_ref): * config/riscv/vector.md: Set frm_mode attr explicitly.
1245 2023-08-09 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1247 * config/riscv/riscv-vsetvl.cc (anticipatable_occurrence_p): Fix
1248 incorrect anticipate info.
1250 2023-08-09 Tsukasa OI <research_trasio@irq.a4lg.com>
1252 * common/config/riscv/riscv-common.cc (riscv_ext_version_table):
1253 Remove 'Zve32d' from the version list.
1255 2023-08-09 Jin Ma <jinma@linux.alibaba.com>
1257 * config/riscv/riscv.cc (riscv_sched_variable_issue): New function.
1258 (TARGET_SCHED_VARIABLE_ISSUE): New macro.
1259 Co-authored-by: Philipp Tomsich <philipp.tomsich@vrull.eu>
1260 Co-authored-by: Jeff Law <jlaw@ventanamicro.com>
1262 2023-08-09 Jivan Hakobyan <jivanhakobyan9@gmail.com>
1264 * config/riscv/riscv.cc (riscv_legitimize_address): Handle folding.
1265 (mem_shadd_or_shadd_rtx_p): New function.
1267 2023-08-09 Andrew Pinski <apinski@marvell.com>
1269 PR tree-optimization/110937
1270 PR tree-optimization/100798
1271 * match.pd (`a ? ~b : b`): Handle this
1274 2023-08-09 Uros Bizjak <ubizjak@gmail.com>
1276 * config/i386/i386.opt (mpartial-vector-fp-math): Add dot.
1278 2023-08-09 Richard Ball <richard.ball@arm.com>
1280 * config/aarch64/aarch64-cores.def (AARCH64_CORE): Add Cortex-A520 CPU.
1281 * config/aarch64/aarch64-tune.md: Regenerate.
1282 * doc/invoke.texi: Document Cortex-A520 CPU.
1284 2023-08-09 Carl Love <cel@us.ibm.com>
1286 * config/rs6000/rs6000-builtins.def (vcmpneb, vcmpneh, vcmpnew):
1287 Move definitions to Altivec stanza.
1288 * config/rs6000/altivec.md (vcmpneb, vcmpneh, vcmpnew): New
1291 2023-08-09 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1294 * config/riscv/riscv-v.cc (expand_const_vector): Add NPATTERNS = 1
1295 stepped vector support.
1297 2023-08-09 liuhongt <hongtao.liu@intel.com>
1299 * common/config/i386/cpuinfo.h (get_available_features):
1300 Rename local variable subleaf_level to max_subleaf_level.
1302 2023-08-09 Richard Biener <rguenther@suse.de>
1304 PR rtl-optimization/110587
1305 * lra-assigns.cc (find_hard_regno_for_1): Re-order checks.
1307 2023-08-09 Kewen Lin <linkw@linux.ibm.com>
1309 PR tree-optimization/110248
1310 * config/rs6000/rs6000.cc (rs6000_legitimate_address_p): Check if
1311 the given code is for ifn LEN_{LOAD,STORE}, if yes then make it not
1312 legitimate when outer code is PLUS.
1314 2023-08-09 Kewen Lin <linkw@linux.ibm.com>
1316 PR tree-optimization/110248
1317 * recog.cc (memory_address_addr_space_p): Add one more argument ch of
1318 type code_helper and pass it to targetm.addr_space.legitimate_address_p
1319 instead of ERROR_MARK.
1320 (offsettable_address_addr_space_p): Update one function pointer with
1321 one more argument of type code_helper as its assignees
1322 memory_address_addr_space_p and strict_memory_address_addr_space_p
1323 have been adjusted, and adjust some call sites with ERROR_MARK.
1324 * recog.h (tree.h): New include header file for tree_code ERROR_MARK.
1325 (memory_address_addr_space_p): Adjust with one more unnamed argument
1326 of type code_helper with default ERROR_MARK.
1327 (strict_memory_address_addr_space_p): Likewise.
1328 * reload.cc (strict_memory_address_addr_space_p): Add one unnamed
1329 argument of type code_helper.
1330 * tree-ssa-address.cc (valid_mem_ref_p): Add one more argument ch of
1331 type code_helper and pass it to memory_address_addr_space_p.
1332 * tree-ssa-address.h (valid_mem_ref_p): Adjust the declaration with
1333 one more unnamed argument of type code_helper with default value
1335 * tree-ssa-loop-ivopts.cc (get_address_cost): Use ERROR_MARK as code
1336 by default, change it with ifn code for USE_PTR_ADDRESS type use, and
1337 pass it to all valid_mem_ref_p calls.
1339 2023-08-09 Kewen Lin <linkw@linux.ibm.com>
1341 PR tree-optimization/110248
1342 * coretypes.h (class code_helper): Add forward declaration.
1343 * doc/tm.texi: Regenerate.
1344 * lra-constraints.cc (valid_address_p): Call target hook
1345 targetm.addr_space.legitimate_address_p with an extra parameter
1346 ERROR_MARK as its prototype changes.
1347 * recog.cc (memory_address_addr_space_p): Likewise.
1348 * reload.cc (strict_memory_address_addr_space_p): Likewise.
1349 * target.def (legitimate_address_p, addr_space.legitimate_address_p):
1350 Extend with one more argument of type code_helper, update the
1351 documentation accordingly.
1352 * targhooks.cc (default_legitimate_address_p): Adjust for the
1353 new code_helper argument.
1354 (default_addr_space_legitimate_address_p): Likewise.
1355 * targhooks.h (default_legitimate_address_p): Likewise.
1356 (default_addr_space_legitimate_address_p): Likewise.
1357 * config/aarch64/aarch64.cc (aarch64_legitimate_address_hook_p): Adjust
1358 with extra unnamed code_helper argument with default ERROR_MARK.
1359 * config/alpha/alpha.cc (alpha_legitimate_address_p): Likewise.
1360 * config/arc/arc.cc (arc_legitimate_address_p): Likewise.
1361 * config/arm/arm-protos.h (arm_legitimate_address_p): Likewise.
1362 (tree.h): New include for tree_code ERROR_MARK.
1363 * config/arm/arm.cc (arm_legitimate_address_p): Adjust with extra
1364 unnamed code_helper argument with default ERROR_MARK.
1365 * config/avr/avr.cc (avr_addr_space_legitimate_address_p): Likewise.
1366 * config/bfin/bfin.cc (bfin_legitimate_address_p): Likewise.
1367 * config/bpf/bpf.cc (bpf_legitimate_address_p): Likewise.
1368 * config/c6x/c6x.cc (c6x_legitimate_address_p): Likewise.
1369 * config/cris/cris-protos.h (cris_legitimate_address_p): Likewise.
1370 (tree.h): New include for tree_code ERROR_MARK.
1371 * config/cris/cris.cc (cris_legitimate_address_p): Adjust with extra
1372 unnamed code_helper argument with default ERROR_MARK.
1373 * config/csky/csky.cc (csky_legitimate_address_p): Likewise.
1374 * config/epiphany/epiphany.cc (epiphany_legitimate_address_p):
1376 * config/frv/frv.cc (frv_legitimate_address_p): Likewise.
1377 * config/ft32/ft32.cc (ft32_addr_space_legitimate_address_p): Likewise.
1378 * config/gcn/gcn.cc (gcn_addr_space_legitimate_address_p): Likewise.
1379 * config/h8300/h8300.cc (h8300_legitimate_address_p): Likewise.
1380 * config/i386/i386.cc (ix86_legitimate_address_p): Likewise.
1381 * config/ia64/ia64.cc (ia64_legitimate_address_p): Likewise.
1382 * config/iq2000/iq2000.cc (iq2000_legitimate_address_p): Likewise.
1383 * config/lm32/lm32.cc (lm32_legitimate_address_p): Likewise.
1384 * config/loongarch/loongarch.cc (loongarch_legitimate_address_p):
1386 * config/m32c/m32c.cc (m32c_legitimate_address_p): Likewise.
1387 (m32c_addr_space_legitimate_address_p): Likewise.
1388 * config/m32r/m32r.cc (m32r_legitimate_address_p): Likewise.
1389 * config/m68k/m68k.cc (m68k_legitimate_address_p): Likewise.
1390 * config/mcore/mcore.cc (mcore_legitimate_address_p): Likewise.
1391 * config/microblaze/microblaze-protos.h (tree.h): New include for
1392 tree_code ERROR_MARK.
1393 (microblaze_legitimate_address_p): Adjust with extra unnamed
1394 code_helper argument with default ERROR_MARK.
1395 * config/microblaze/microblaze.cc (microblaze_legitimate_address_p):
1397 * config/mips/mips.cc (mips_legitimate_address_p): Likewise.
1398 * config/mmix/mmix.cc (mmix_legitimate_address_p): Likewise.
1399 * config/mn10300/mn10300.cc (mn10300_legitimate_address_p): Likewise.
1400 * config/moxie/moxie.cc (moxie_legitimate_address_p): Likewise.
1401 * config/msp430/msp430.cc (msp430_legitimate_address_p): Likewise.
1402 (msp430_addr_space_legitimate_address_p): Adjust with extra code_helper
1403 argument with default ERROR_MARK and adjust the call to function
1404 msp430_legitimate_address_p.
1405 * config/nds32/nds32.cc (nds32_legitimate_address_p): Adjust with extra
1406 unnamed code_helper argument with default ERROR_MARK.
1407 * config/nios2/nios2.cc (nios2_legitimate_address_p): Likewise.
1408 * config/nvptx/nvptx.cc (nvptx_legitimate_address_p): Likewise.
1409 * config/or1k/or1k.cc (or1k_legitimate_address_p): Likewise.
1410 * config/pa/pa.cc (pa_legitimate_address_p): Likewise.
1411 * config/pdp11/pdp11.cc (pdp11_legitimate_address_p): Likewise.
1412 * config/pru/pru.cc (pru_addr_space_legitimate_address_p): Likewise.
1413 * config/riscv/riscv.cc (riscv_legitimate_address_p): Likewise.
1414 * config/rl78/rl78-protos.h (rl78_as_legitimate_address): Likewise.
1415 (tree.h): New include for tree_code ERROR_MARK.
1416 * config/rl78/rl78.cc (rl78_as_legitimate_address): Adjust with
1417 extra unnamed code_helper argument with default ERROR_MARK.
1418 * config/rs6000/rs6000.cc (rs6000_legitimate_address_p): Likewise.
1419 (rs6000_debug_legitimate_address_p): Adjust with extra code_helper
1420 argument and adjust the call to function rs6000_legitimate_address_p.
1421 * config/rx/rx.cc (rx_is_legitimate_address): Adjust with extra
1422 unnamed code_helper argument with default ERROR_MARK.
1423 * config/s390/s390.cc (s390_legitimate_address_p): Likewise.
1424 * config/sh/sh.cc (sh_legitimate_address_p): Likewise.
1425 * config/sparc/sparc.cc (sparc_legitimate_address_p): Likewise.
1426 * config/v850/v850.cc (v850_legitimate_address_p): Likewise.
1427 * config/vax/vax.cc (vax_legitimate_address_p): Likewise.
1428 * config/visium/visium.cc (visium_legitimate_address_p): Likewise.
1429 * config/xtensa/xtensa.cc (xtensa_legitimate_address_p): Likewise.
1430 * config/stormy16/stormy16-protos.h (xstormy16_legitimate_address_p):
1432 (tree.h): New include for tree_code ERROR_MARK.
1433 * config/stormy16/stormy16.cc (xstormy16_legitimate_address_p):
1434 Adjust with extra unnamed code_helper argument with default
1437 2023-08-09 liuhongt <hongtao.liu@intel.com>
1439 * common/config/i386/cpuinfo.h (get_available_features): Check
1440 EAX for valid subleaf before use CPUID.
1442 2023-08-08 Jeff Law <jlaw@ventanamicro.com>
1444 * config/riscv/riscv.cc (riscv_expand_conditional_move): Use word_mode
1445 for the temporary when canonicalizing the condition.
1447 2023-08-08 Cupertino Miranda <cupertino.miranda@oracle.com>
1449 * config/bpf/core-builtins.cc: Cleaned include headers.
1450 (struct cr_builtins): Added GTY.
1451 (cr_builtins_ref): Created.
1452 (builtins_data) Changed to GC root.
1453 (allocate_builtin_data): Changed.
1454 Included gt-core-builtins.h.
1455 * config/bpf/coreout.cc: (bpf_core_extra) Added GTY.
1456 (bpf_core_extra_ref): Created.
1457 (bpf_comment_info): Changed to GC root.
1458 (bpf_core_reloc_add, output_btfext_header, btf_ext_init): Changed.
1460 2023-08-08 Uros Bizjak <ubizjak@gmail.com>
1463 * config/i386/i386.opt (mpartial-vector-fp-math): New option.
1464 * config/i386/mmx.md (movq_<mode>_to_sse): Do not sanitize
1465 upper part of V2SFmode register with -fno-trapping-math.
1466 (<plusminusmult:insn>v2sf3): Enable for ix86_partial_vec_fp_math.
1468 (<smaxmin:code>v2sf3): Ditto.
1470 (*mmx_haddv2sf3_low): Ditto.
1471 (*mmx_hsubv2sf3_low): Ditto.
1472 (vec_addsubv2sf3): Ditto.
1473 (vec_cmpv2sfv2si): Ditto.
1474 (vcond<V2FI:mode>v2sf): Ditto.
1479 (fix_truncv2sfv2si2): Ditto.
1480 (fixuns_truncv2sfv2si2): Ditto.
1481 (floatv2siv2sf2): Ditto.
1482 (floatunsv2siv2sf2): Ditto.
1483 (nearbyintv2sf2): Ditto.
1485 (lrintv2sfv2si2): Ditto.
1487 (lceilv2sfv2si2): Ditto.
1488 (floorv2sf2): Ditto.
1489 (lfloorv2sfv2si2): Ditto.
1490 (btruncv2sf2): Ditto.
1491 (roundv2sf2): Ditto.
1492 (lroundv2sfv2si2): Ditto.
1493 * doc/invoke.texi (x86 Options): Document
1494 -mpartial-vector-fp-math option.
1496 2023-08-08 Andrew Pinski <apinski@marvell.com>
1498 PR tree-optimization/103281
1499 PR tree-optimization/28794
1500 * vr-values.cc (simplify_using_ranges::simplify_cond_using_ranges_1): Split out
1502 (simplify_using_ranges::simplify_compare_using_ranges_1): Here.
1503 (simplify_using_ranges::simplify_casted_cond): Rename to ...
1504 (simplify_using_ranges::simplify_casted_compare): This
1505 and change arguments to take op0 and op1.
1506 (simplify_using_ranges::simplify_compare_assign_using_ranges_1): New method.
1507 (simplify_using_ranges::simplify): For tcc_comparison assignments call
1508 simplify_compare_assign_using_ranges_1.
1509 * vr-values.h (simplify_using_ranges): Add
1510 new methods, simplify_compare_using_ranges_1 and simplify_compare_assign_using_ranges_1.
1511 Rename simplify_casted_cond and simplify_casted_compare and
1512 update argument types.
1514 2023-08-08 Andrzej Turko <andrzej.turko@gmail.com>
1516 * genmatch.cc: Log line numbers indirectly.
1518 2023-08-08 Andrzej Turko <andrzej.turko@gmail.com>
1520 * genmatch.cc: Make sinfo map ordered.
1521 * Makefile.in: Require the ordered map header for genmatch.o.
1523 2023-08-08 Andrzej Turko <andrzej.turko@gmail.com>
1525 * ordered-hash-map.h: Add get_or_insert.
1526 * ordered-hash-map-tests.cc: Use get_or_insert in tests.
1528 2023-08-08 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1530 * config/riscv/autovec.md (cond_<optab><mode>): New pattern.
1531 (cond_len_<optab><mode>): Ditto.
1532 (cond_fma<mode>): Ditto.
1533 (cond_len_fma<mode>): Ditto.
1534 (cond_fnma<mode>): Ditto.
1535 (cond_len_fnma<mode>): Ditto.
1536 (cond_fms<mode>): Ditto.
1537 (cond_len_fms<mode>): Ditto.
1538 (cond_fnms<mode>): Ditto.
1539 (cond_len_fnms<mode>): Ditto.
1540 * config/riscv/riscv-protos.h (riscv_get_v_regno_alignment): Export
1542 (enum insn_type): Add new enum type.
1543 (prepare_ternary_operands): New function.
1544 * config/riscv/riscv-v.cc (emit_vlmax_masked_fp_mu_insn): Ditto.
1545 (emit_nonvlmax_tumu_insn): Ditto.
1546 (emit_nonvlmax_fp_tumu_insn): Ditto.
1547 (expand_cond_len_binop): Add condtional operations.
1548 (expand_cond_len_ternop): Ditto.
1549 (prepare_ternary_operands): New function.
1550 * config/riscv/riscv.cc (riscv_memmodel_needs_amo_release): Export
1551 riscv_get_v_regno_alignment as global scope.
1552 * config/riscv/vector.md: Fix ternary bugs.
1554 2023-08-08 Richard Biener <rguenther@suse.de>
1556 PR tree-optimization/49955
1557 * tree-vectorizer.h (_slp_instance::remain_stmts): New.
1558 (SLP_INSTANCE_REMAIN_STMTS): Likewise.
1559 * tree-vect-slp.cc (vect_free_slp_instance): Release
1560 SLP_INSTANCE_REMAIN_STMTS.
1561 (vect_build_slp_instance): Make the number of lanes of
1562 a BB reduction even.
1563 (vectorize_slp_instance_root_stmt): Handle unvectorized
1564 defs of a BB reduction.
1566 2023-08-08 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
1568 * internal-fn.cc (get_len_internal_fn): New function.
1569 (DEF_INTERNAL_COND_FN): Ditto.
1570 (DEF_INTERNAL_SIGNED_COND_FN): Ditto.
1571 * internal-fn.h (get_len_internal_fn): Ditto.
1572 * tree-vect-stmts.cc (vectorizable_call): Add CALL auto-vectorization.
1574 2023-08-08 Richard Biener <rguenther@suse.de>
1576 PR tree-optimization/110924
1577 * tree-ssa-live.h (virtual_operand_live): Update comment.
1578 * tree-ssa-live.cc (virtual_operand_live::get_live_in): Remove
1579 optimization, look at each predecessor.
1580 * tree-ssa-sink.cc (pass_sink_code::execute): Mark backedges.
1582 2023-08-08 yulong <shiyulong@iscas.ac.cn>
1584 * config/riscv/riscv-v.cc (slide1_sew64_helper): Modify.
1586 2023-08-08 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1588 * config/riscv/autovec-vls.md (<optab><mode>2): Add VLS neg.
1589 * config/riscv/vector.md: Ditto.
1591 2023-08-08 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1593 * config/riscv/autovec.md: Add VLS shift.
1595 2023-08-07 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1597 * config/riscv/autovec-vls.md (<optab><mode>3): Add VLS modes.
1598 * config/riscv/vector-iterators.md: Ditto.
1599 * config/riscv/vector.md: Ditto.
1601 2023-08-07 Jonathan Wakely <jwakely@redhat.com>
1603 * config/i386/i386.cc (ix86_invalid_conversion): Fix grammar.
1605 2023-08-07 Nick Alcock <nick.alcock@oracle.com>
1607 * configure: Regenerate.
1609 2023-08-07 John Ericson <git@JohnEricson.me>
1611 * configure: Regenerate.
1613 2023-08-07 Alan Modra <amodra@gmail.com>
1615 * configure: Regenerate.
1617 2023-08-07 Alexander von Gluck IV <kallisti5@unixzen.com>
1619 * configure: Regenerate.
1621 2023-08-07 Nick Alcock <nick.alcock@oracle.com>
1623 * configure: Regenerate.
1625 2023-08-07 Nick Alcock <nick.alcock@oracle.com>
1627 * configure: Regenerate.
1629 2023-08-07 H.J. Lu <hjl.tools@gmail.com>
1631 * configure: Regenerate.
1633 2023-08-07 H.J. Lu <hjl.tools@gmail.com>
1635 * configure: Regenerate.
1637 2023-08-07 Jeff Law <jlaw@ventanamicro.com>
1639 * config/riscv/riscv.cc (riscv_expand_conditional_move): Allow
1640 VOIDmode operands to conditional before canonicalization.
1642 2023-08-07 Manolis Tsamis <manolis.tsamis@vrull.eu>
1644 * regcprop.cc (maybe_copy_reg_attrs): Remove unnecessary function.
1645 (find_oldest_value_reg): Inline stack_pointer_rtx check.
1646 (copyprop_hardreg_forward_1): Inline stack_pointer_rtx check.
1648 2023-08-07 Martin Jambor <mjambor@suse.cz>
1651 * ipa-param-manipulation.h (class ipa_param_body_adjustments): New
1652 members get_ddef_if_exists_and_is_used and mark_clobbers_dead.
1653 * ipa-sra.cc (isra_track_scalar_value_uses): Ignore clobbers.
1654 (ptr_parm_has_nonarg_uses): Likewise.
1655 * ipa-param-manipulation.cc
1656 (ipa_param_body_adjustments::get_ddef_if_exists_and_is_used): New.
1657 (ipa_param_body_adjustments::mark_dead_statements): Move initial
1658 checks to get_ddef_if_exists_and_is_used.
1659 (ipa_param_body_adjustments::mark_clobbers_dead): New.
1660 (ipa_param_body_adjustments::common_initialization): Call
1661 mark_clobbers_dead when splitting.
1663 2023-08-07 Raphael Zinsly <rzinsly@ventanamicro.com>
1665 * config/riscv/riscv.cc (riscv_expand_int_scc): Add invert_ptr
1666 as an argument and pass it to riscv_emit_int_order_test.
1667 (riscv_expand_conditional_move): Handle cases where the condition
1668 is not EQ/NE or the second argument to the conditional is not
1670 * config/riscv/riscv-protos.h (riscv_expand_int_scc): Update prototype.
1671 Co-authored-by: Jeff Law <jlaw@ventanamicro.com>
1673 2023-08-07 Andrew Pinski <apinski@marvell.com>
1675 PR tree-optimization/109959
1676 * match.pd (`(a > 1) ? 0 : (cast)a`, `(a <= 1) & (cast)a`):
1679 2023-08-07 Richard Biener <rguenther@suse.de>
1681 * tree-ssa-sink.cc (pass_sink_code::execute): Do not
1682 calculate post-dominators. Calculate RPO on the inverted
1683 graph and process blocks in that order.
1685 2023-08-07 liuhongt <hongtao.liu@intel.com>
1688 * config/i386/i386-protos.h
1689 (vpternlog_redundant_operand_mask): Adjust parameter type.
1690 * config/i386/i386.cc (vpternlog_redundant_operand_mask): Use
1691 INTVAL instead of XINT, also adjust parameter type from rtx*
1692 to rtx since the function only needs operands[4] in vpternlog
1694 (substitute_vpternlog_operands): Pass operands[4] instead of
1695 operands to vpternlog_redundant_operand_mask.
1696 * config/i386/sse.md: Ditto.
1698 2023-08-07 Richard Biener <rguenther@suse.de>
1700 * tree-vect-slp.cc (vect_slp_region): Save/restore vect_location
1701 around dumping code.
1703 2023-08-07 liuhongt <hongtao.liu@intel.com>
1706 * config/i386/mmx.md (<insn><mode>3): Changed from define_insn
1707 to define_expand and break into ..
1708 (<insn>v4hf3): .. this.
1709 (divv4hf3): .. this.
1710 (<insn>v2hf3): .. this.
1711 (divv2hf3): .. this.
1712 (movd_v2hf_to_sse): New define_expand.
1713 (movq_<mode>_to_sse): Extend to V4HFmode.
1714 (mmxdoublevecmode): Ditto.
1715 (V2FI_V4HF): New mode iterator.
1716 * config/i386/sse.md (*vec_concatv4sf): Extend to hanlde V8HF
1717 by using mode iterator V4SF_V8HF, renamed to ..
1718 (*vec_concat<mode>): .. this.
1719 (*vec_concatv4sf_0): Extend to handle V8HF by using mode
1720 iterator V4SF_V8HF, renamed to ..
1721 (*vec_concat<mode>_0): .. this.
1722 (*vec_concatv8hf_movss): New define_insn.
1723 (V4SF_V8HF): New mode iterator.
1725 2023-08-07 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1727 * tree-vect-patterns.cc (vect_recog_over_widening_pattern): Add op vectype.
1729 2023-08-07 Jan Beulich <jbeulich@suse.com>
1731 * config/i386/mmx.md (*mmx_pinsrd): Drop "prefix_data16".
1732 (*mmx_pinsrb): Likewise.
1733 (*mmx_pextrb): Likewise.
1734 (*mmx_pextrb_zext): Likewise.
1735 (mmx_pshufbv8qi3): Likewise.
1736 (mmx_pshufbv4qi3): Likewise.
1737 (mmx_pswapdv2si2): Likewise.
1738 (*pinsrb): Likewise.
1739 (*pextrb): Likewise.
1740 (*pextrb_zext): Likewise.
1741 * config/i386/sse.md (*sse4_1_mulv2siv2di3<mask_name>): Likewise.
1742 (*sse2_eq<mode>3): Likewise.
1743 (*sse2_gt<mode>3): Likewise.
1744 (<sse2p4_1>_pinsr<ssemodesuffix>): Likewise.
1745 (*vec_extract<mode>): Likewise.
1746 (*vec_extract<PEXTR_MODE12:mode>_zext): Likewise.
1747 (*vec_extractv16qi_zext): Likewise.
1748 (ssse3_ph<plusminus_mnemonic>wv8hi3): Likewise.
1749 (ssse3_pmaddubsw128): Likewise.
1750 (*<ssse3_avx2>_pmulhrsw<mode>3<mask_name>): Likewise.
1751 (<ssse3_avx2>_pshufb<mode>3<mask_name>): Likewise.
1752 (<ssse3_avx2>_psign<mode>3): Likewise.
1753 (<ssse3_avx2>_palignr<mode>): Likewise.
1754 (*abs<mode>2): Likewise.
1755 (sse4_2_pcmpestr): Likewise.
1756 (sse4_2_pcmpestri): Likewise.
1757 (sse4_2_pcmpestrm): Likewise.
1758 (sse4_2_pcmpestr_cconly): Likewise.
1759 (sse4_2_pcmpistr): Likewise.
1760 (sse4_2_pcmpistri): Likewise.
1761 (sse4_2_pcmpistrm): Likewise.
1762 (sse4_2_pcmpistr_cconly): Likewise.
1763 (vgf2p8affineinvqb_<mode><mask_name>): Likewise.
1764 (vgf2p8affineqb_<mode><mask_name>): Likewise.
1765 (vgf2p8mulb_<mode><mask_name>): Likewise.
1766 (*<code>v8hi3 [smaxmin]): Drop "prefix_data16" and
1768 (*<code>v16qi3 [umaxmin]): Likewise.
1770 2023-08-07 Jan Beulich <jbeulich@suse.com>
1772 * config/i386/i386.md (sse4_1_round<mode>2): Make
1773 "length_immediate" uniformly 1.
1774 * config/i386/mmx.md (mmx_pblendvb_v8qi): Likewise.
1775 (mmx_pblendvb_<mode>): Likewise.
1777 2023-08-07 Jan Beulich <jbeulich@suse.com>
1779 * config/i386/sse.md
1780 (<avx512>_<complexopname>_<mode><maskc_name><round_name>): Add
1782 (avx512fp16_<complexopname>sh_v8hf<mask_scalarc_name><round_scalarcz_name>):
1785 2023-08-07 Jan Beulich <jbeulich@suse.com>
1787 * config/i386/sse.md (xop_phadd<u>bw): Add "prefix",
1788 "prefix_extra", and "mode" attributes.
1789 (xop_phadd<u>bd): Likewise.
1790 (xop_phadd<u>bq): Likewise.
1791 (xop_phadd<u>wd): Likewise.
1792 (xop_phadd<u>wq): Likewise.
1793 (xop_phadd<u>dq): Likewise.
1794 (xop_phsubbw): Likewise.
1795 (xop_phsubwd): Likewise.
1796 (xop_phsubdq): Likewise.
1797 (xop_rotl<mode>3): Add "prefix" and "prefix_extra" attributes.
1798 (xop_rotr<mode>3): Likewise.
1799 (xop_frcz<mode>2): Likewise.
1800 (*xop_vmfrcz<mode>2): Likewise.
1801 (xop_vrotl<mode>3): Add "prefix" attribute. Change
1802 "prefix_extra" to 1.
1803 (xop_sha<mode>3): Likewise.
1804 (xop_shl<mode>3): Likewise.
1806 2023-08-07 Jan Beulich <jbeulich@suse.com>
1808 * config/i386/sse.md
1809 (*<avx512>_eq<mode>3<mask_scalar_merge_name>_1): Drop
1811 (avx512dq_vextract<shuffletype>64x2_1_mask): Likewise.
1812 (*avx512dq_vextract<shuffletype>64x2_1): Likewise.
1813 (avx512f_vextract<shuffletype>32x4_1_mask): Likewise.
1814 (*avx512f_vextract<shuffletype>32x4_1): Likewise.
1815 (vec_extract_lo_<mode>_mask [AVX512 forms]): Likewise.
1816 (vec_extract_lo_<mode> [AVX512 forms]): Likewise.
1817 (vec_extract_hi_<mode>_mask [AVX512 forms]): Likewise.
1818 (vec_extract_hi_<mode> [AVX512 forms]): Likewise.
1819 (@vec_extract_lo_<mode> [AVX512 forms]): Likewise.
1820 (@vec_extract_hi_<mode> [AVX512 forms]): Likewise.
1821 (vec_extract_lo_v64qi): Likewise.
1822 (vec_extract_hi_v64qi): Likewise.
1823 (*vec_widen_umult_even_v16si<mask_name>): Likewise.
1824 (*vec_widen_smult_even_v16si<mask_name>): Likewise.
1825 (*avx512f_<code><mode>3<mask_name>): Likewise.
1826 (*vec_extractv4ti): Likewise.
1827 (avx512bw_<code>v32qiv32hi2<mask_name>): Likewise.
1828 (<mask_codefor>avx512dq_broadcast<mode><mask_name>_1): Likewise.
1829 Add "length_immediate".
1831 2023-08-07 Jan Beulich <jbeulich@suse.com>
1833 * config/i386/i386.md (@rdrand<mode>): Add "prefix_0f". Drop
1835 (@rdseed<mode>): Likewise.
1836 * config/i386/mmx.md (<code><mode>3 [smaxmin and umaxmin cases]):
1837 Adjust "prefix_extra".
1838 * config/i386/sse.md (@vec_set<mode>_0): Likewise.
1839 (*sse4_1_<code><mode>3<mask_name>): Likewise.
1840 (*avx2_eq<mode>3): Likewise.
1841 (avx2_gt<mode>3): Likewise.
1842 (<sse2p4_1>_pinsr<ssemodesuffix>): Likewise.
1843 (*vec_extract<mode>): Likewise.
1844 (<vi8_sse4_1_avx2_avx512>_movntdqa): Likewise.
1846 2023-08-07 Jan Beulich <jbeulich@suse.com>
1848 * config/i386/i386.md (rd<fsgs>base<mode>): Add "prefix_0f" and
1849 "prefix_rep". Drop "prefix_extra".
1850 (wr<fsgs>base<mode>): Likewise.
1851 (ptwrite<mode>): Likewise.
1853 2023-08-07 Jan Beulich <jbeulich@suse.com>
1855 * config/i386/i386.md (isa): Move up.
1856 (length_immediate): Handle "fma4".
1857 (prefix): Handle "ssemuladd".
1858 * config/i386/sse.md (*fma_fmadd_<mode>): Add "prefix" attribute.
1859 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
1861 (<avx512>_fmadd_<mode>_mask<round_name>): Likewise.
1862 (<avx512>_fmadd_<mode>_mask3<round_name>): Likewise.
1863 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>):
1865 (<avx512>_fmsub_<mode>_mask<round_name>): Likewise.
1866 (<avx512>_fmsub_<mode>_mask3<round_name>): Likewise.
1867 (*fma_fnmadd_<mode>): Likewise.
1868 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>):
1870 (<avx512>_fnmadd_<mode>_mask<round_name>): Likewise.
1871 (<avx512>_fnmadd_<mode>_mask3<round_name>): Likewise.
1872 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>):
1874 (<avx512>_fnmsub_<mode>_mask<round_name>): Likewise.
1875 (<avx512>_fnmsub_<mode>_mask3<round_name>): Likewise.
1876 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
1878 (<avx512>_fmaddsub_<mode>_mask<round_name>): Likewise.
1879 (<avx512>_fmaddsub_<mode>_mask3<round_name>): Likewise.
1880 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
1882 (<avx512>_fmsubadd_<mode>_mask<round_name>): Likewise.
1883 (<avx512>_fmsubadd_<mode>_mask3<round_name>): Likewise.
1884 (*fmai_fmadd_<mode>): Likewise.
1885 (*fmai_fmsub_<mode>): Likewise.
1886 (*fmai_fnmadd_<mode><round_name>): Likewise.
1887 (*fmai_fnmsub_<mode><round_name>): Likewise.
1888 (avx512f_vmfmadd_<mode>_mask<round_name>): Likewise.
1889 (avx512f_vmfmadd_<mode>_mask3<round_name>): Likewise.
1890 (avx512f_vmfmadd_<mode>_maskz_1<round_name>): Likewise.
1891 (*avx512f_vmfmsub_<mode>_mask<round_name>): Likewise.
1892 (avx512f_vmfmsub_<mode>_mask3<round_name>): Likewise.
1893 (*avx512f_vmfmsub_<mode>_maskz_1<round_name>): Likewise.
1894 (avx512f_vmfnmadd_<mode>_mask<round_name>): Likewise.
1895 (avx512f_vmfnmadd_<mode>_mask3<round_name>): Likewise.
1896 (avx512f_vmfnmadd_<mode>_maskz_1<round_name>): Likewise.
1897 (*avx512f_vmfnmsub_<mode>_mask<round_name>): Likewise.
1898 (*avx512f_vmfnmsub_<mode>_mask3<round_name>): Likewise.
1899 (*avx512f_vmfnmsub_<mode>_maskz_1<round_name>): Likewise.
1900 (*fma4i_vmfmadd_<mode>): Likewise.
1901 (*fma4i_vmfmsub_<mode>): Likewise.
1902 (*fma4i_vmfnmadd_<mode>): Likewise.
1903 (*fma4i_vmfnmsub_<mode>): Likewise.
1904 (fma_<complexopname>_<mode><sdc_maskz_name><round_name>): Likewise.
1905 (<avx512>_<complexopname>_<mode>_mask<round_name>): Likewise.
1906 (avx512fp16_fma_<complexopname>sh_v8hf<mask_scalarcz_name><round_scalarcz_name>):
1908 (avx512fp16_<complexopname>sh_v8hf_mask<round_name>): Likewise.
1909 (xop_p<macs><ssemodesuffix><ssemodesuffix>): Likewise.
1910 (xop_p<macs>dql): Likewise.
1911 (xop_p<macs>dqh): Likewise.
1912 (xop_p<macs>wd): Likewise.
1913 (xop_p<madcs>wd): Likewise.
1914 (fma_<complexpairopname>_<mode>_pair): Likewise. Add "mode" attribute.
1916 2023-08-07 Jan Beulich <jbeulich@suse.com>
1918 * config/i386/i386.md (length_immediate): Handle "sse4arg".
1920 (*xop_pcmov_<mode>): Add "mode" attribute.
1921 * config/i386/mmx.md (*xop_maskcmp<mode>3): Drop "prefix_data16",
1922 "prefix_rep", "prefix_extra", and "length_immediate" attributes.
1923 (*xop_maskcmp_uns<mode>3): Likewise. Switch "type" to "sse4arg".
1924 (*xop_pcmov_<mode>): Add "mode" attribute.
1925 * config/i386/sse.md (xop_pcmov_<mode><avxsizesuffix>): Add "mode"
1927 (xop_maskcmp<mode>3): Drop "prefix_data16", "prefix_rep",
1928 "prefix_extra", and "length_immediate" attributes.
1929 (xop_maskcmp_uns<mode>3): Likewise. Switch "type" to "sse4arg".
1930 (xop_maskcmp_uns2<mode>3): Drop "prefix_data16", "prefix_extra",
1931 and "length_immediate" attributes. Switch "type" to "sse4arg".
1932 (xop_pcom_tf<mode>3): Likewise.
1933 (xop_vpermil2<mode>3): Drop "length_immediate" attribute.
1935 2023-08-07 Jan Beulich <jbeulich@suse.com>
1937 * config/i386/i386.md (prefix_extra): Correct comment. Fold
1938 cases yielding 2 into ones yielding 1.
1940 2023-08-07 Jan Hubicka <jh@suse.cz>
1942 PR tree-optimization/106293
1943 * tree-vect-loop-manip.cc (vect_loop_versioning): Fix profile update.
1944 * tree-vect-loop.cc (vect_transform_loop): Likewise.
1946 2023-08-07 Andrew Pinski <apinski@marvell.com>
1948 PR tree-optimization/96695
1949 * match.pd (min_value, max_value): Extend to
1952 2023-08-06 Jan Hubicka <jh@suse.cz>
1954 * config/i386/cpuid.h (__get_cpuid_count, __get_cpuid_max): Add
1955 __builtin_expect that CPU likely supports cpuid.
1957 2023-08-06 Jan Hubicka <jh@suse.cz>
1959 * tree-loop-distribution.cc (loop_distribution::execute): Disable
1960 distribution for loops with estimated iterations 0.
1962 2023-08-06 Jan Hubicka <jh@suse.cz>
1964 * tree-vect-loop-manip.cc (vect_do_peeling): Fix profile update of peeled epilogues.
1966 2023-08-04 Xiao Zeng <zengxiao@eswincomputing.com>
1968 * config/riscv/riscv.cc (riscv_expand_conditional_move): Recognize
1969 more Zicond patterns. Fix whitespace typo.
1970 (riscv_rtx_costs): Remove accidental code duplication.
1971 Co-authored-by: Jeff Law <jlaw@ventanamicro.com>
1973 2023-08-04 Yan Simonaytes <simonaytes.yan@ispras.ru>
1976 * config/i386/i386-protos.h
1977 (vpternlog_redundant_operand_mask): Declare.
1978 (substitute_vpternlog_operands): Declare.
1979 * config/i386/i386.cc
1980 (vpternlog_redundant_operand_mask): New helper.
1981 (substitute_vpternlog_operands): New function. Use them...
1982 * config/i386/sse.md: ... here in new VPTERNLOG define_splits.
1984 2023-08-04 Roger Sayle <roger@nextmovesoftware.com>
1986 * expmed.cc (extract_bit_field_1): Document that an UNSIGNEDP
1987 value of -1 is equivalent to don't care.
1988 (extract_integral_bit_field): Indicate that we don't require
1989 the most significant word to be zero extended, if we're about
1991 (extract_fixed_bit_field_1): Document that an UNSIGNEDP value
1992 of -1 is equivalent to don't care. Don't clear the most
1993 significant bits with AND mask when UNSIGNEDP is -1.
1995 2023-08-04 Roger Sayle <roger@nextmovesoftware.com>
1997 * config/i386/sse.md (define_split): Convert highpart:DF extract
1998 from V2DFmode register into a sse2_storehpd instruction.
1999 (define_split): Likewise, convert lowpart:DF extract from V2DF
2000 register into a sse2_storelpd instruction.
2002 2023-08-04 Qing Zhao <qing.zhao@oracle.com>
2004 * doc/invoke.texi (-Wflex-array-member-not-at-end): Document
2007 2023-08-04 Vladimir N. Makarov <vmakarov@redhat.com>
2009 * lra-lives.cc (process_bb_lives): Check input insn pattern hard regs
2010 against early clobber hard regs.
2012 2023-08-04 Tamar Christina <tamar.christina@arm.com>
2014 * doc/extend.texi: Document it.
2016 2023-08-04 Tamar Christina <tamar.christina@arm.com>
2019 * config/aarch64/aarch64-simd.md (vec_widen_<sur>shiftl_lo_<mode>,
2020 vec_widen_<sur>shiftl_hi_<mode>): Remove.
2021 (aarch64_<sur>shll<mode>_internal): Renamed to...
2022 (aarch64_<su>shll<mode>): .. This.
2023 (aarch64_<sur>shll2<mode>_internal): Renamed to...
2024 (aarch64_<su>shll2<mode>): .. This.
2025 (aarch64_<sur>shll_n<mode>, aarch64_<sur>shll2_n<mode>): Re-use new
2027 * config/aarch64/constraints.md (D2, DL): New.
2028 * config/aarch64/predicates.md (aarch64_simd_shll_imm_vec): New.
2030 2023-08-04 Tamar Christina <tamar.christina@arm.com>
2032 * gensupport.cc (conlist): Support length 0 attribute.
2034 2023-08-04 Tamar Christina <tamar.christina@arm.com>
2036 * config/aarch64/aarch64.cc (aarch64_bool_compound_p): New.
2037 (aarch64_adjust_stmt_cost, aarch64_vector_costs::count_ops): Use it.
2039 2023-08-04 Tamar Christina <tamar.christina@arm.com>
2041 * config/aarch64/aarch64.cc (aarch64_multiply_add_p): Update handling
2043 (aarch64_adjust_stmt_cost): Use it.
2044 (aarch64_vector_costs::count_ops): Likewise.
2045 (aarch64_vector_costs::add_stmt_cost): Pass vinfo to
2046 aarch64_adjust_stmt_cost.
2048 2023-08-04 Richard Biener <rguenther@suse.de>
2050 PR tree-optimization/110838
2051 * tree-vect-patterns.cc (vect_recog_over_widening_pattern):
2052 Fix right-shift value sanitizing. Properly emit external
2053 def mangling in the preheader rather than in the pattern
2054 def sequence where it will fail vectorizing.
2056 2023-08-04 Matthew Malcomson <matthew.malcomson@arm.com>
2058 PR middle-end/110316
2060 * timevar.cc (NANOSEC_PER_SEC, TICKS_TO_NANOSEC,
2061 CLOCKS_TO_NANOSEC, nanosec_to_floating_sec, percent_of): New.
2062 (TICKS_TO_MSEC, CLOCKS_TO_MSEC): Remove these macros.
2063 (timer::validate_phases): Use integral arithmetic to check
2065 (timer::print_row, timer::print): Convert from integral
2066 nanoseconds to floating point seconds before printing.
2067 (timer::all_zero): Change limit to nanosec count instead of
2068 fractional count of seconds.
2069 (make_json_for_timevar_time_def): Convert from integral
2070 nanoseconds to floating point seconds before recording.
2071 * timevar.h (struct timevar_time_def): Update all measurements
2072 to use uint64_t nanoseconds rather than seconds stored in a
2075 2023-08-04 Richard Biener <rguenther@suse.de>
2077 PR tree-optimization/110838
2078 * match.pd (([rl]shift @0 out-of-bounds) -> zero): Restrict
2079 the arithmetic right-shift case to non-negative operands.
2081 2023-08-04 Pan Li <pan2.li@intel.com>
2084 2023-08-04 Pan Li <pan2.li@intel.com>
2086 * config/riscv/riscv-vector-builtins-bases.cc
2087 (class vfmacc_frm): New class for vfmacc frm.
2088 (vfmacc_frm_obj): New declaration.
2090 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
2091 * config/riscv/riscv-vector-builtins-functions.def
2092 (vfmacc_frm): New function definition.
2093 * config/riscv/riscv-vector-builtins.cc
2094 (function_expander::use_ternop_insn): Add frm operand support.
2095 * config/riscv/vector.md: Add vfmuladd to frm_mode.
2097 2023-08-04 Pan Li <pan2.li@intel.com>
2100 2023-08-04 Pan Li <pan2.li@intel.com>
2102 * config/riscv/riscv-vector-builtins-bases.cc
2103 (class vfnmacc_frm): New class for vfnmacc.
2104 (vfnmacc_frm_obj): New declaration.
2106 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
2107 * config/riscv/riscv-vector-builtins-functions.def
2108 (vfnmacc_frm): New function definition.
2110 2023-08-04 Pan Li <pan2.li@intel.com>
2113 2023-08-04 Pan Li <pan2.li@intel.com>
2115 * config/riscv/riscv-vector-builtins-bases.cc
2116 (class vfmsac_frm): New class for vfmsac frm.
2117 (vfmsac_frm_obj): New declaration.
2119 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
2120 * config/riscv/riscv-vector-builtins-functions.def
2121 (vfmsac_frm): New function definition.
2123 2023-08-04 Pan Li <pan2.li@intel.com>
2126 2023-08-04 Pan Li <pan2.li@intel.com>
2128 * config/riscv/riscv-vector-builtins-bases.cc
2129 (class vfnmsac_frm): New class for vfnmsac frm.
2130 (vfnmsac_frm_obj): New declaration.
2132 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
2133 * config/riscv/riscv-vector-builtins-functions.def
2134 (vfnmsac_frm): New function definition.
2136 2023-08-04 Georg-Johann Lay <avr@gjlay.de>
2138 * config/avr/avr-mcus.def (avr64dd14, avr64dd20, avr64dd28, avr64dd32)
2139 (avr64ea28, avr64ea32, avr64ea48, attiny424, attiny426, attiny427)
2140 (attiny824, attiny826, attiny827, attiny1624, attiny1626, attiny1627)
2141 (attiny3224, attiny3226, attiny3227, avr16dd14, avr16dd20, avr16dd28)
2142 (avr16dd32, avr32dd14, avr32dd20, avr32dd28, avr32dd32)
2143 (attiny102, attiny104): New devices.
2144 * doc/avr-mmcu.texi: Regenerate.
2146 2023-08-04 Georg-Johann Lay <avr@gjlay.de>
2148 * config/avr/avr-mcus.def (avr128d*, avr64d*): Fix their FLASH_SIZE
2149 and PM_OFFSET entries.
2151 2023-08-04 Andrew Pinski <apinski@marvell.com>
2153 PR tree-optimization/110874
2154 * gimple-match-head.cc (gimple_bit_not_with_nop): New declaration.
2155 (gimple_maybe_cmp): Likewise.
2156 (gimple_bitwise_inverted_equal_p): Rewrite to use gimple_bit_not_with_nop
2157 and gimple_maybe_cmp instead of being recursive.
2158 * match.pd (bit_not_with_nop): New match pattern.
2159 (maybe_cmp): Likewise.
2161 2023-08-04 Drew Ross <drross@redhat.com>
2163 PR middle-end/101955
2164 * match.pd ((signed x << c) >> c): New canonicalization.
2166 2023-08-04 Pan Li <pan2.li@intel.com>
2168 * config/riscv/riscv-vector-builtins-bases.cc
2169 (class vfnmsac_frm): New class for vfnmsac frm.
2170 (vfnmsac_frm_obj): New declaration.
2172 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
2173 * config/riscv/riscv-vector-builtins-functions.def
2174 (vfnmsac_frm): New function definition.
2176 2023-08-04 Pan Li <pan2.li@intel.com>
2178 * config/riscv/riscv-vector-builtins-bases.cc
2179 (class vfmsac_frm): New class for vfmsac frm.
2180 (vfmsac_frm_obj): New declaration.
2182 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
2183 * config/riscv/riscv-vector-builtins-functions.def
2184 (vfmsac_frm): New function definition.
2186 2023-08-04 Pan Li <pan2.li@intel.com>
2188 * config/riscv/riscv-vector-builtins-bases.cc
2189 (class vfnmacc_frm): New class for vfnmacc.
2190 (vfnmacc_frm_obj): New declaration.
2192 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
2193 * config/riscv/riscv-vector-builtins-functions.def
2194 (vfnmacc_frm): New function definition.
2196 2023-08-04 Hao Liu <hliu@os.amperecomputing.com>
2199 * config/aarch64/aarch64.cc (aarch64_force_single_cycle): check
2200 STMT_VINFO_REDUC_DEF to avoid failures in info_for_reduction.
2202 2023-08-04 Pan Li <pan2.li@intel.com>
2204 * config/riscv/riscv-vector-builtins-bases.cc
2205 (class vfmacc_frm): New class for vfmacc frm.
2206 (vfmacc_frm_obj): New declaration.
2208 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
2209 * config/riscv/riscv-vector-builtins-functions.def
2210 (vfmacc_frm): New function definition.
2211 * config/riscv/riscv-vector-builtins.cc
2212 (function_expander::use_ternop_insn): Add frm operand support.
2213 * config/riscv/vector.md: Add vfmuladd to frm_mode.
2215 2023-08-04 Pan Li <pan2.li@intel.com>
2217 * config/riscv/riscv-vector-builtins-bases.cc
2218 (vfwmul_frm_obj): New declaration.
2219 (vfwmul_frm): Ditto.
2220 * config/riscv/riscv-vector-builtins-bases.h:
2221 (vfwmul_frm): Ditto.
2222 * config/riscv/riscv-vector-builtins-functions.def
2223 (vfwmul_frm): New function definition.
2224 * config/riscv/vector.md: (frm_mode) Add vfwmul to frm_mode.
2226 2023-08-04 Pan Li <pan2.li@intel.com>
2228 * config/riscv/riscv-vector-builtins-bases.cc
2229 (binop_frm): New declaration.
2230 (reverse_binop_frm): Likewise.
2232 * config/riscv/riscv-vector-builtins-bases.h:
2233 (vfdiv_frm): New extern declaration.
2234 (vfrdiv_frm): Likewise.
2235 * config/riscv/riscv-vector-builtins-functions.def
2236 (vfdiv_frm): New function definition.
2237 (vfrdiv_frm): Likewise.
2238 * config/riscv/vector.md: Add vfdiv to frm_mode.
2240 2023-08-03 Jan Hubicka <jh@suse.cz>
2242 * tree-cfg.cc (print_loop_info): Print entry count.
2244 2023-08-03 Jan Hubicka <jh@suse.cz>
2246 * tree-ssa-loop-split.cc (split_loop): Update estimated iteration counts.
2248 2023-08-03 Jan Hubicka <jh@suse.cz>
2251 * cfgloopmanip.cc (scale_loop_profile): (Un)initialize
2252 unadjusted_exit_count.
2254 2023-08-03 Aldy Hernandez <aldyh@redhat.com>
2256 * ipa-prop.cc (ipa_compute_jump_functions_for_edge): Read global
2259 2023-08-03 Xiao Zeng <zengxiao@eswincomputing.com>
2261 * config/riscv/riscv.cc (riscv_expand_conditional_move): Recognize
2262 various Zicond patterns.
2263 * config/riscv/riscv.md (mov<mode>cc): Allow TARGET_ZICOND. Use
2264 sfb_alu_operand for both arms of the conditional move.
2265 Co-authored-by: Jeff Law <jlaw@ventanamicro.com>
2267 2023-08-03 Cupertino Miranda <cupertino.miranda@oracle.com>
2273 * config.gcc: Added core-builtins.cc and .o files.
2274 * config/bpf/bpf-passes.def: Removed file.
2275 * config/bpf/bpf-protos.h (bpf_add_core_reloc,
2276 bpf_replace_core_move_operands): New prototypes.
2277 * config/bpf/bpf.cc (enum bpf_builtins, is_attr_preserve_access,
2278 maybe_make_core_relo, bpf_core_field_info, bpf_core_compute,
2279 bpf_core_get_index, bpf_core_new_decl, bpf_core_walk,
2280 bpf_is_valid_preserve_field_info_arg, is_attr_preserve_access,
2281 handle_attr_preserve, pass_data_bpf_core_attr, pass_bpf_core_attr):
2283 (def_builtin, bpf_expand_builtin, bpf_resolve_overloaded_builtin): Changed.
2284 * config/bpf/bpf.md (define_expand mov<MM:mode>): Changed.
2285 (mov_reloc_core<mode>): Added.
2286 * config/bpf/core-builtins.cc (struct cr_builtin, enum
2287 cr_decision struct cr_local, struct cr_final, struct
2288 core_builtin_helpers, enum bpf_plugin_states): Added types.
2289 (builtins_data, core_builtin_helpers, core_builtin_type_defs):
2291 (allocate_builtin_data, get_builtin-data, search_builtin_data,
2292 remove_parser_plugin, compare_same_kind, compare_same_ptr_expr,
2293 compare_same_ptr_type, is_attr_preserve_access, core_field_info,
2294 bpf_core_get_index, compute_field_expr,
2295 pack_field_expr_for_access_index, pack_field_expr_for_preserve_field,
2296 process_field_expr, pack_enum_value, process_enum_value, pack_type,
2297 process_type, bpf_require_core_support, make_core_relo, read_kind,
2298 kind_access_index, kind_preserve_field_info, kind_enum_value,
2299 kind_type_id, kind_preserve_type_info, get_core_builtin_fndecl_for_type,
2300 bpf_handle_plugin_finish_type, bpf_init_core_builtins,
2301 construct_builtin_core_reloc, bpf_resolve_overloaded_core_builtin,
2302 bpf_expand_core_builtin, bpf_add_core_reloc,
2303 bpf_replace_core_move_operands): Added functions.
2304 * config/bpf/core-builtins.h (enum bpf_builtins): Added.
2305 (bpf_init_core_builtins, bpf_expand_core_builtin,
2306 bpf_resolve_overloaded_core_builtin): Added functions.
2307 * config/bpf/coreout.cc (struct bpf_core_extra): Added.
2308 (bpf_core_reloc_add, output_asm_btfext_core_reloc): Changed.
2309 * config/bpf/coreout.h (bpf_core_reloc_add) Changed prototype.
2310 * config/bpf/t-bpf: Added core-builtins.o.
2311 * doc/extend.texi: Added documentation for new BPF builtins.
2313 2023-08-03 Andrew MacLeod <amacleod@redhat.com>
2315 * gimple-range-fold.cc (fold_using_range::range_of_range_op): Add
2316 ranges to the call to relation_fold_and_or.
2317 (fold_using_range::relation_fold_and_or): Add op1 and op2 ranges.
2318 (fur_source::register_outgoing_edges): Add op1 and op2 ranges.
2319 * gimple-range-fold.h (relation_fold_and_or): Adjust params.
2320 * gimple-range-gori.cc (gori_compute::compute_operand_range): Add
2321 a varying op1 and op2 to call.
2322 * range-op-float.cc (range_operator::op1_op2_relation): New dafaults.
2323 (operator_equal::op1_op2_relation): New float version.
2324 (operator_not_equal::op1_op2_relation): Ditto.
2325 (operator_lt::op1_op2_relation): Ditto.
2326 (operator_le::op1_op2_relation): Ditto.
2327 (operator_gt::op1_op2_relation): Ditto.
2328 (operator_ge::op1_op2_relation) Ditto.
2329 * range-op-mixed.h (operator_equal::op1_op2_relation): New float
2331 (operator_not_equal::op1_op2_relation): Ditto.
2332 (operator_lt::op1_op2_relation): Ditto.
2333 (operator_le::op1_op2_relation): Ditto.
2334 (operator_gt::op1_op2_relation): Ditto.
2335 (operator_ge::op1_op2_relation): Ditto.
2336 * range-op.cc (range_op_handler::op1_op2_relation): Dispatch new
2338 (range_operator::op1_op2_relation): Add extra params.
2339 (operator_equal::op1_op2_relation): Ditto.
2340 (operator_not_equal::op1_op2_relation): Ditto.
2341 (operator_lt::op1_op2_relation): Ditto.
2342 (operator_le::op1_op2_relation): Ditto.
2343 (operator_gt::op1_op2_relation): Ditto.
2344 (operator_ge::op1_op2_relation): Ditto.
2345 * range-op.h (range_operator): New prototypes.
2346 (range_op_handler): Ditto.
2348 2023-08-03 Andrew MacLeod <amacleod@redhat.com>
2350 * gimple-range-gori.cc (gori_compute::compute_operand1_range):
2351 Use identity relation.
2352 (gori_compute::compute_operand2_range): Ditto.
2353 * value-relation.cc (get_identity_relation): New.
2354 * value-relation.h (get_identity_relation): New prototype.
2356 2023-08-03 Andrew MacLeod <amacleod@redhat.com>
2358 * value-range.h (Value_Range::set_varying): Set the type.
2359 (Value_Range::set_zero): Ditto.
2360 (Value_Range::set_nonzero): Ditto.
2362 2023-08-03 Jeff Law <jeffreyalaw@gmail.com>
2364 * config/riscv/riscv.cc (riscv_rtx_costs): Remove errant hunk from
2367 2023-08-03 Pan Li <pan2.li@intel.com>
2369 * config/riscv/riscv-vector-builtins-bases.cc: Add vfsub.
2371 2023-08-03 Richard Sandiford <richard.sandiford@arm.com>
2373 * poly-int.h (can_div_trunc_p): Succeed for more boundary conditions.
2375 2023-08-03 Richard Biener <rguenther@suse.de>
2377 PR tree-optimization/110838
2378 * tree-vect-patterns.cc (vect_recog_over_widening_pattern):
2379 Adjust the shift operand of RSHIFT_EXPRs.
2381 2023-08-03 Richard Biener <rguenther@suse.de>
2383 PR tree-optimization/110702
2384 * tree-ssa-loop-ivopts.cc (rewrite_use_address): When
2385 we created a NULL pointer based access rewrite that to
2388 2023-08-03 Richard Biener <rguenther@suse.de>
2390 * tree-ssa-sink.cc: Include tree-ssa-live.h.
2391 (pass_sink_code::execute): Instantiate virtual_operand_live
2393 (sink_code_in_bb): Pass down virtual_operand_live.
2394 (statement_sink_location): Get virtual_operand_live and
2395 verify we are not sinking loads across stores by looking up
2396 the live virtual operand at the sink location.
2398 2023-08-03 Richard Biener <rguenther@suse.de>
2400 * tree-ssa-live.h (class virtual_operand_live): New.
2401 * tree-ssa-live.cc (virtual_operand_live::init): New.
2402 (virtual_operand_live::get_live_in): Likewise.
2403 (virtual_operand_live::get_live_out): Likewise.
2405 2023-08-03 Richard Biener <rguenther@suse.de>
2407 * passes.def: Exchange loop splitting and final value
2410 2023-08-03 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
2412 * config/s390/s390.cc (expand_perm_as_a_vlbr_vstbr_candidate):
2413 New function which handles bswap patterns for vec_perm_const.
2414 (vectorize_vec_perm_const_1): Call new function.
2415 * config/s390/vector.md (*bswap<mode>): Fix operands in output
2417 (*vstbr<mode>): New insn.
2419 2023-08-03 Alexandre Oliva <oliva@adacore.com>
2421 * config/vxworks-smp.opt: New. Introduce -msmp.
2422 * config.gcc: Enable it on powerpc* vxworks prior to 7r*.
2423 * config/rs6000/vxworks.h (STARTFILE_PREFIX_SPEC): Choose
2424 lib_smp when -msmp is present in the command line.
2425 * doc/invoke.texi: Document it.
2427 2023-08-03 Yanzhang Wang <yanzhang.wang@intel.com>
2429 * config/riscv/riscv.cc (riscv_save_reg_p): Save ra for leaf
2430 when enabling -mno-omit-leaf-frame-pointer
2431 (riscv_option_override): Override omit-frame-pointer.
2432 (riscv_frame_pointer_required): Save s0 for non-leaf function
2433 (TARGET_FRAME_POINTER_REQUIRED): Override defination
2434 * config/riscv/riscv.opt: Add option support.
2436 2023-08-03 Roger Sayle <roger@nextmovesoftware.com>
2439 * config/i386/i386.md (<any_rotate>ti3): For rotations by 64 bits
2440 place operand in a register before gen_<insn>64ti2_doubleword.
2441 (<any_rotate>di3): Likewise, for rotations by 32 bits, place
2442 operand in a register before gen_<insn>32di2_doubleword.
2443 (<any_rotate>32di2_doubleword): Constrain operand to be in register.
2444 (<any_rotate>64ti2_doubleword): Likewise.
2446 2023-08-03 Pan Li <pan2.li@intel.com>
2448 * config/riscv/riscv-vector-builtins-bases.cc
2449 (vfmul_frm_obj): New declaration.
2451 * config/riscv/riscv-vector-builtins-bases.h: Likewise.
2452 * config/riscv/riscv-vector-builtins-functions.def
2453 (vfmul_frm): New function definition.
2454 * config/riscv/vector.md: Add vfmul to frm_mode.
2456 2023-08-03 Andrew Pinski <apinski@marvell.com>
2458 * match.pd (`~X & X`): Check that the types match.
2459 (`~x | x`, `~x ^ x`): Likewise.
2461 2023-08-03 Pan Li <pan2.li@intel.com>
2463 * config/riscv/riscv-vector-builtins-bases.h: Remove
2464 redudant declaration.
2466 2023-08-03 Pan Li <pan2.li@intel.com>
2468 * config/riscv/riscv-vector-builtins-bases.cc (BASE): Add
2470 * config/riscv/riscv-vector-builtins-bases.h: Add declaration.
2471 * config/riscv/riscv-vector-builtins-functions.def (vfwsub_frm):
2472 Add vfwsub function definitions.
2474 2023-08-02 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
2476 PR rtl-optimization/110867
2477 * combine.cc (simplify_compare_const): Try the optimization only
2478 in case the constant fits into the comparison mode.
2480 2023-08-02 Jeff Law <jlaw@ventanamicro.com>
2482 * config/riscv/zicond.md: Remove incorrect zicond patterns and
2483 renumber/rename them.
2484 (zero.nez.<GPR:MODE><X:mode>.opt2): Fix output string.
2486 2023-08-02 Richard Biener <rguenther@suse.de>
2488 * tree-phinodes.h (add_phi_node_to_bb): Remove.
2489 * tree-phinodes.cc (add_phi_node_to_bb): Make static.
2491 2023-08-02 Jan Beulich <jbeulich@suse.com>
2493 * config/i386/sse.md (vec_dupv2df<mask_name>): Fold the middle
2494 two of the alternatives.
2496 2023-08-02 Richard Biener <rguenther@suse.de>
2498 PR tree-optimization/92335
2499 * tree-ssa-sink.cc (select_best_block): Before loop
2500 optimizations avoid sinking unconditional loads/stores
2501 in innermost loops to conditional executed places.
2503 2023-08-02 Andrew Pinski <apinski@marvell.com>
2505 * gimple-match-head.cc (gimple_bitwise_inverted_equal_p): Valueize
2506 the comparison operands before comparing them.
2508 2023-08-02 Andrew Pinski <apinski@marvell.com>
2510 * match.pd (`~X & X`, `~X | X`): Move over to
2511 use bitwise_inverted_equal_p, removing :c as bitwise_inverted_equal_p
2512 handles that already.
2513 Remove range test simplifications to true/false as they
2514 are now handled by these patterns.
2516 2023-08-02 Andrew Pinski <apinski@marvell.com>
2518 * tree-ssa-phiopt.cc (match_simplify_replacement): Mark's cond
2519 statement's lhs and rhs to check if trivial dead.
2520 Rename inserted_exprs to exprs_maybe_dce; also move it so
2521 bitmap is not allocated if not needed.
2523 2023-08-02 Pan Li <pan2.li@intel.com>
2525 * config/riscv/riscv-vector-builtins-bases.cc
2526 (class widen_binop_frm): New class for binop frm.
2527 (BASE): Add vfwadd_frm.
2528 * config/riscv/riscv-vector-builtins-bases.h: New declaration.
2529 * config/riscv/riscv-vector-builtins-functions.def
2530 (vfwadd_frm): New function definition.
2531 * config/riscv/riscv-vector-builtins-shapes.cc
2532 (BASE_NAME_MAX_LEN): New macro.
2533 (struct alu_frm_def): Leverage new base class.
2534 (struct build_frm_base): New build base for frm.
2535 (struct widen_alu_frm_def): New struct for widen alu frm.
2536 (SHAPE): Add widen_alu_frm shape.
2537 * config/riscv/riscv-vector-builtins-shapes.h: New declaration.
2538 * config/riscv/vector.md (frm_mode): Add vfwalu type.
2540 2023-08-02 Jan Hubicka <jh@suse.cz>
2542 * cfgloop.h (loop_count_in): Declare.
2543 * cfgloopanal.cc (expected_loop_iterations_by_profile): Use count_in.
2544 (loop_count_in): Move here from ...
2545 * cfgloopmanip.cc (loop_count_in): ... here.
2546 (scale_loop_profile): Improve dumping; cast iteration bound to sreal.
2548 2023-08-02 Jan Hubicka <jh@suse.cz>
2550 * cfg.cc (scale_strictly_dominated_blocks): New function.
2551 * cfg.h (scale_strictly_dominated_blocks): Declare.
2552 * tree-cfg.cc (fold_loop_internal_call): Fixup CFG profile.
2554 2023-08-02 Richard Biener <rguenther@suse.de>
2556 PR rtl-optimization/110587
2557 * lra-spills.cc (return_regno_p): Remove.
2558 (regno_in_use_p): Likewise.
2559 (lra_final_code_change): Do not remove noop moves
2560 between hard registers.
2562 2023-08-02 liuhongt <hongtao.liu@intel.com>
2565 * config/i386/sse.md (vec_fmaddsub<mode>4): Extend to vector
2566 HFmode, use mode iterator VFH instead.
2567 (vec_fmsubadd<mode>4): Ditto.
2568 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2569 Remove scalar mode from iterator, use VFH_AVX512VL instead.
2570 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2573 2023-08-02 liuhongt <hongtao.liu@intel.com>
2575 * config/i386/sse.md (*avx2_lddqu_inserti_to_bcasti): New
2576 pre_reload define_insn_and_split.
2578 2023-08-02 Xiao Zeng <zengxiao@eswincomputing.com>
2580 * config/riscv/riscv.cc (riscv_rtx_costs): Add costing for
2581 using Zicond to implement some conditional moves.
2583 2023-08-02 Jeff Law <jlaw@ventanamicro.com>
2585 * config/riscv/zicond.md: Use the X iterator instead of ANYI
2586 on the comparison input operands.
2588 2023-08-02 Xiao Zeng <zengxiao@eswincomputing.com>
2590 * config/riscv/riscv.cc (riscv_rtx_costs, case IF_THEN_ELSE): Add
2592 (case SET): For INSNs that just set a REG, take the cost from the
2594 Co-authored-by: Jeff Law <jlaw@ventanamicro.com>
2596 2023-08-02 Hu, Lin1 <lin1.hu@intel.com>
2598 * common/config/i386/i386-common.cc (OPTION_MASK_ISA2_AMX_INT8_SET):
2599 Change OPTION_MASK_ISA2_AMX_TILE to OPTION_MASK_ISA2_AMX_TILE_SET.
2600 (OPTION_MASK_ISA2_AMX_BF16_SET): Ditto
2601 (OPTION_MASK_ISA2_AMX_FP16_SET): Ditto
2602 (OPTION_MASK_ISA2_AMX_COMPLEX_SET): Ditto
2603 (OPTION_MASK_ISA_ABM_SET):
2604 Change OPTION_MASK_ISA_POPCNT to OPTION_MASK_ISA_POPCNT_SET.
2606 2023-08-01 Andreas Krebbel <krebbel@linux.ibm.com>
2608 * config/s390/s390.cc (s390_encode_section_info): Assume external
2609 symbols without explicit alignment to be unaligned if
2610 -munaligned-symbols has been specified.
2611 * config/s390/s390.opt (-munaligned-symbols): New option.
2613 2023-08-01 Richard Ball <richard.ball@arm.com>
2615 * gimple-fold.cc (fold_ctor_reference):
2616 Add support for poly_int.
2618 2023-08-01 Georg-Johann Lay <avr@gjlay.de>
2621 * config/avr/avr.cc (avr_optimize_casesi): Set JUMP_LABEL and
2622 LABEL_NUSES of new conditional branch instruction.
2624 2023-08-01 Jan Hubicka <jh@suse.cz>
2626 * tree-vect-loop-manip.cc (vect_do_peeling): Fix profile update after
2627 constant prologue peeling.
2629 2023-08-01 Christophe Lyon <christophe.lyon@linaro.org>
2631 * doc/sourcebuild.texi (arm_v8_1m_main_cde_mve_fp): Fix spelling.
2633 2023-08-01 Pan Li <pan2.li@intel.com>
2634 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2636 * config/riscv/riscv.cc (DYNAMIC_FRM_RTL): New macro.
2637 (STATIC_FRM_P): Ditto.
2638 (struct mode_switching_info): New struct for mode switching.
2639 (struct machine_function): Add new field mode switching.
2640 (riscv_emit_frm_mode_set): Add DYN_CALL emit.
2641 (riscv_frm_adjust_mode_after_call): New function for call mode.
2642 (riscv_frm_emit_after_call_in_bb_end): New function for emit
2643 insn when call as the end of bb.
2644 (riscv_frm_mode_needed): New function for frm mode needed.
2645 (frm_unknown_dynamic_p): Remove call check.
2646 (riscv_mode_needed): Extrac function for frm.
2647 (riscv_frm_mode_after): Add DYN_CALL after.
2648 (riscv_mode_entry): Remove backup rtl initialization.
2649 * config/riscv/vector.md (frm_mode): Add dyn_call.
2650 (fsrmsi_restore_exit): Rename to _volatile.
2651 (fsrmsi_restore_volatile): Likewise.
2653 2023-08-01 Pan Li <pan2.li@intel.com>
2655 * config/riscv/riscv-vector-builtins-bases.cc
2656 (class reverse_binop_frm): Add new template for reversed frm.
2657 (vfsub_frm_obj): New obj.
2658 (vfrsub_frm_obj): Likewise.
2659 * config/riscv/riscv-vector-builtins-bases.h:
2660 (vfsub_frm): New declaration.
2661 (vfrsub_frm): Likewise.
2662 * config/riscv/riscv-vector-builtins-functions.def
2663 (vfsub_frm): New function define.
2664 (vfrsub_frm): Likewise.
2666 2023-08-01 Andrew Pinski <apinski@marvell.com>
2668 PR tree-optimization/93044
2669 * match.pd (nested int casts): A truncation (to the same size or smaller)
2670 can always remove the inner cast.
2672 2023-07-31 Hamza Mahfooz <someguy@effective-light.com>
2675 * doc/invoke.texi (-Wmissing-variable-declarations): Document
2678 2023-07-31 Andrew Pinski <apinski@marvell.com>
2680 PR tree-optimization/106164
2681 * match.pd (`a != b & a <= b`, `a != b & a >= b`,
2682 `a == b | a < b`, `a == b | a > b`): Handle these cases
2685 2023-07-31 Andrew Pinski <apinski@marvell.com>
2687 PR tree-optimization/106164
2688 * match.pd: Extend the `(X CMP1 CST1) AND/IOR (X CMP2 CST2)`
2689 patterns to support `(X CMP1 Y) AND/IOR (X CMP2 Y)`.
2691 2023-07-31 Andrew Pinski <apinski@marvell.com>
2693 PR tree-optimization/100864
2694 * generic-match-head.cc (bitwise_inverted_equal_p): New function.
2695 * gimple-match-head.cc (bitwise_inverted_equal_p): New macro.
2696 (gimple_bitwise_inverted_equal_p): New function.
2697 * match.pd ((~x | y) & x): Use bitwise_inverted_equal_p
2698 instead of direct matching bit_not.
2700 2023-07-31 Costas Argyris <costas.argyris@gmail.com>
2703 * gcc-ar.cc (main): Expand argv and use
2704 temporary response file to call ar if any
2705 expansions were made.
2707 2023-07-31 Andrew MacLeod <amacleod@redhat.com>
2709 PR tree-optimization/110582
2710 * gimple-range-fold.cc (fur_list::get_operand): Do not use the
2711 range vector for non-ssa names.
2713 2023-07-31 David Malcolm <dmalcolm@redhat.com>
2716 * diagnostic-client-data-hooks.h (class sarif_object): New forward
2718 (diagnostic_client_data_hooks::add_sarif_invocation_properties):
2720 * diagnostic-format-sarif.cc: Include "diagnostic-format-sarif.h".
2721 (class sarif_invocation): Inherit from sarif_object rather than
2723 (class sarif_result): Likewise.
2724 (class sarif_ice_notification): Likewise.
2725 (sarif_object::get_or_create_properties): New.
2726 (sarif_invocation::prepare_to_flush): Add "context" param. Use it
2727 to call the context's add_sarif_invocation_properties hook.
2728 (sarif_builder::flush_to_file): Pass m_context to
2729 sarif_invocation::prepare_to_flush.
2730 * diagnostic-format-sarif.h: New header.
2731 * doc/invoke.texi (Developer Options): Clarify that -ftime-report
2732 writes to stderr. Document that if SARIF diagnostic output is
2733 requested then any timing information is written in JSON form as
2734 part of the SARIF output, rather than to stderr.
2735 * timevar.cc: Include "json.h".
2736 (timer::named_items::m_hash_map): Split out type into...
2737 (timer::named_items::hash_map_t): ...this new typedef.
2738 (timer::named_items::make_json): New function.
2739 (timevar_diff): New function.
2740 (make_json_for_timevar_time_def): New function.
2741 (timer::timevar_def::make_json): New function.
2742 (timer::make_json): New function.
2743 * timevar.h (class json::value): New forward decl.
2744 (timer::make_json): New decl.
2745 (timer::timevar_def::make_json): New decl.
2746 * tree-diagnostic-client-data-hooks.cc: Include
2747 "diagnostic-format-sarif.h" and "timevar.h".
2748 (compiler_data_hooks::add_sarif_invocation_properties): New vfunc
2751 2023-07-31 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
2753 * combine.cc (simplify_compare_const): Narrow comparison of
2754 memory and constant.
2755 (try_combine): Adapt new function signature.
2756 (simplify_comparison): Adapt new function signature.
2758 2023-07-31 Kito Cheng <kito.cheng@sifive.com>
2760 * config/riscv/riscv-v.cc (expand_vec_series): Drop unused
2762 (expand_vector_init_insert_elems): Ditto.
2764 2023-07-31 Hao Liu <hliu@os.amperecomputing.com>
2767 * config/aarch64/aarch64.cc (count_ops): Only '* count' for
2768 single_defuse_cycle while counting reduction_latency.
2770 2023-07-31 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
2772 * internal-fn.def (DEF_INTERNAL_COND_FN): New macro.
2773 (DEF_INTERNAL_SIGNED_COND_FN): Ditto.
2794 (COND_LEN_ADD): Ditto.
2795 (COND_LEN_SUB): Ditto.
2796 (COND_LEN_MUL): Ditto.
2797 (COND_LEN_DIV): Ditto.
2798 (COND_LEN_MOD): Ditto.
2799 (COND_LEN_RDIV): Ditto.
2800 (COND_LEN_MIN): Ditto.
2801 (COND_LEN_MAX): Ditto.
2802 (COND_LEN_FMIN): Ditto.
2803 (COND_LEN_FMAX): Ditto.
2804 (COND_LEN_AND): Ditto.
2805 (COND_LEN_IOR): Ditto.
2806 (COND_LEN_XOR): Ditto.
2807 (COND_LEN_SHL): Ditto.
2808 (COND_LEN_SHR): Ditto.
2809 (COND_LEN_FMA): Ditto.
2810 (COND_LEN_FMS): Ditto.
2811 (COND_LEN_FNMA): Ditto.
2812 (COND_LEN_FNMS): Ditto.
2813 (COND_LEN_NEG): Ditto.
2814 (ADD): New macro define.
2835 2023-07-31 Roger Sayle <roger@nextmovesoftware.com>
2838 * config/i386/i386-features.cc (compute_convert_gain): Check
2839 TARGET_AVX512VL (not TARGET_AVX512F) when considering V2DImode
2840 and V4SImode rotates in STV.
2841 (general_scalar_chain::convert_rotate): Likewise.
2843 2023-07-31 Kito Cheng <kito.cheng@sifive.com>
2845 * config/riscv/autovec.md (abs<mode>2): Remove `.require ()`.
2846 * config/riscv/riscv-protos.h (get_mask_mode): Update return
2848 * config/riscv/riscv-v.cc (rvv_builder::rvv_builder): Remove
2850 (emit_vlmax_insn): Ditto.
2851 (emit_vlmax_fp_insn): Ditto.
2852 (emit_vlmax_ternary_insn): Ditto.
2853 (emit_vlmax_fp_ternary_insn): Ditto.
2854 (emit_nonvlmax_fp_ternary_tu_insn): Ditto.
2855 (emit_nonvlmax_insn): Ditto.
2856 (emit_vlmax_slide_insn): Ditto.
2857 (emit_nonvlmax_slide_tu_insn): Ditto.
2858 (emit_vlmax_merge_insn): Ditto.
2859 (emit_vlmax_masked_insn): Ditto.
2860 (emit_nonvlmax_masked_insn): Ditto.
2861 (emit_vlmax_masked_store_insn): Ditto.
2862 (emit_nonvlmax_masked_store_insn): Ditto.
2863 (emit_vlmax_masked_mu_insn): Ditto.
2864 (emit_nonvlmax_tu_insn): Ditto.
2865 (emit_nonvlmax_fp_tu_insn): Ditto.
2866 (emit_scalar_move_insn): Ditto.
2867 (emit_vlmax_compress_insn): Ditto.
2868 (emit_vlmax_reduction_insn): Ditto.
2869 (emit_vlmax_fp_reduction_insn): Ditto.
2870 (emit_nonvlmax_fp_reduction_insn): Ditto.
2871 (expand_vec_series): Ditto.
2872 (expand_vector_init_merge_repeating_sequence): Ditto.
2873 (expand_vec_perm): Ditto.
2874 (shuffle_merge_patterns): Ditto.
2875 (shuffle_compress_patterns): Ditto.
2876 (shuffle_decompress_patterns): Ditto.
2877 (expand_reduction): Ditto.
2878 (get_mask_mode): Update return type.
2879 * config/riscv/riscv.cc (riscv_get_mask_mode): Check vector type
2880 is valid, and use new get_mask_mode interface.
2882 2023-07-31 Pan Li <pan2.li@intel.com>
2884 * config/riscv/riscv-vector-builtins-shapes.cc (struct alu_frm_def):
2885 Move rm suffix before mask.
2887 2023-07-31 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2889 * config/riscv/autovec-vls.md (@vec_duplicate<mode>): New pattern.
2890 * config/riscv/riscv-v.cc (autovectorize_vector_modes): Add VLS autovec
2893 2023-07-29 Roger Sayle <roger@nextmovesoftware.com>
2896 * config/i386/i386.md (extv<mode>): Use QImode for offsets.
2897 (extzv<mode>): Likewise.
2898 (insv<mode>): Likewise.
2899 (*testqi_ext_3): Likewise.
2900 (*btr<mode>_2): Likewise.
2901 (define_split): Likewise.
2902 (*btsq_imm): Likewise.
2903 (*btrq_imm): Likewise.
2904 (*btcq_imm): Likewise.
2905 (define_peephole2 x3): Likewise.
2906 (*bt<mode>): Likewise
2907 (*bt<mode>_mask): New define_insn_and_split.
2908 (*jcc_bt<mode>): Use QImode for offsets.
2909 (*jcc_bt<mode>_1): Delete obsolete pattern.
2910 (*jcc_bt<mode>_mask): Use QImode offsets.
2911 (*jcc_bt<mode>_mask_1): Likewise.
2912 (define_split): Likewise.
2913 (*bt<mode>_setcqi): Likewise.
2914 (*bt<mode>_setncqi): Likewise.
2915 (*bt<mode>_setnc<mode>): Likewise.
2916 (*bt<mode>_setncqi_2): Likewise.
2917 (*bt<mode>_setc<mode>_mask): New define_insn_and_split.
2918 (bmi2_bzhi_<mode>3): Use QImode offsets.
2919 (*bmi2_bzhi_<mode>3): Likewise.
2920 (*bmi2_bzhi_<mode>3_1): Likewise.
2921 (*bmi2_bzhi_<mode>3_1_ccz): Likewise.
2922 (@tbm_bextri_<mode>): Likewise.
2924 2023-07-29 Jan Hubicka <jh@suse.cz>
2926 * profile-count.cc (profile_probability::sqrt): New member function.
2927 (profile_probability::pow): Likewise.
2928 * profile-count.h: (profile_probability::sqrt): Declare
2929 (profile_probability::pow): Likewise.
2930 * tree-vect-loop-manip.cc (vect_loop_versioning): Fix profile update.
2932 2023-07-28 Andrew MacLeod <amacleod@redhat.com>
2934 * gimple-range-cache.cc (ssa_cache::merge_range): New.
2935 (ssa_lazy_cache::merge_range): New.
2936 * gimple-range-cache.h (class ssa_cache): Adjust protoypes.
2937 (class ssa_lazy_cache): Ditto.
2938 * gimple-range.cc (assume_query::calculate_op): Use merge_range.
2940 2023-07-28 Andrew MacLeod <amacleod@redhat.com>
2942 * tree-ssa-propagate.cc (substitute_and_fold_engine::value_on_edge):
2943 Move from value-query.cc.
2944 (substitute_and_fold_engine::value_of_stmt): Ditto.
2945 (substitute_and_fold_engine::range_of_expr): New.
2946 * tree-ssa-propagate.h (substitute_and_fold_engine): Inherit from
2947 range_query. New prototypes.
2948 * value-query.cc (value_query::value_on_edge): Relocate.
2949 (value_query::value_of_stmt): Ditto.
2950 * value-query.h (class value_query): Remove.
2951 (class range_query): Remove base class. Adjust prototypes.
2953 2023-07-28 Andrew MacLeod <amacleod@redhat.com>
2955 PR tree-optimization/110205
2956 * gimple-range-cache.h (ranger_cache::m_estimate): Delete.
2957 * range-op-mixed.h (operator_bitwise_xor::op1_op2_relation_effect):
2959 * range-op.cc (operator_lshift): Add missing final overrides.
2960 (operator_rshift): Ditto.
2962 2023-07-28 Jose E. Marchesi <jose.marchesi@oracle.com>
2964 * config/bpf/bpf.cc (bpf_option_override): Disable tail-call
2965 optimizations in BPF target.
2967 2023-07-28 Honza <jh@ryzen4.suse.cz>
2969 * cfgloopmanip.cc (loop_count_in): Break out from ...
2970 (loop_exit_for_scaling): Break out from ...
2971 (update_loop_exit_probability_scale_dom_bbs): Break out from ...;
2972 add more sanity check and debug info.
2973 (scale_loop_profile): ... here.
2974 (create_empty_loop_on_edge): Fix whitespac.
2975 * cfgloopmanip.h (update_loop_exit_probability_scale_dom_bbs): Declare.
2976 * loop-unroll.cc (unroll_loop_constant_iterations): Use
2977 update_loop_exit_probability_scale_dom_bbs.
2978 * tree-ssa-loop-manip.cc (update_exit_probability_after_unrolling): Remove.
2979 (tree_transform_and_unroll_loop): Use
2980 update_loop_exit_probability_scale_dom_bbs.
2981 * tree-ssa-loop-split.cc (split_loop): Use
2982 update_loop_exit_probability_scale_dom_bbs.
2984 2023-07-28 Jan Hubicka <jh@suse.cz>
2987 * tree-ssa-loop-split.cc: Include value-query.h.
2988 (split_at_bb_p): Analyze cases where EQ/NE can be turned
2989 into LT/LE/GT/GE; return updated guard code.
2990 (split_loop): Use guard code.
2992 2023-07-28 Roger Sayle <roger@nextmovesoftware.com>
2993 Richard Biener <rguenther@suse.de>
2996 PR rtl-optimization/110587
2997 * expr.cc (emit_group_load_1): Simplify logic for calling
2998 force_reg on ORIG_SRC, to avoid making a copy if the source
2999 is already in a pseudo register.
3001 2023-07-28 Jan Hubicka <jh@suse.cz>
3003 PR middle-end/106923
3004 * tree-ssa-loop-split.cc (connect_loops): Change probability
3005 of the test preconditioning second loop to very_likely.
3006 (fix_loop_bb_probability): Handle correctly case where
3007 on of the arms of the conditional is empty.
3008 (split_loop): Fold the test guarding first condition to
3009 see if it is constant true; Set correct entry block
3010 probabilities of the split loops; determine correct loop
3013 2023-07-28 xuli <xuli1@eswincomputing.com>
3015 * config/riscv/riscv-vector-builtins-bases.cc: remove rounding mode of
3016 vsadd[u] and vssub[u].
3017 * config/riscv/vector.md: Ditto.
3019 2023-07-28 Jan Hubicka <jh@suse.cz>
3021 * tree-ssa-loop-split.cc (split_loop): Also support NE driven
3022 loops when IV test is not overflowing.
3024 2023-07-28 liuhongt <hongtao.liu@intel.com>
3027 * config/i386/sse.md (avx512cd_maskb_vec_dup<mode>): Add
3029 (avx512cd_maskw_vec_dup<mode>): Ditto.
3031 2023-07-27 David Faust <david.faust@oracle.com>
3035 * config/bpf/bpf.opt (msmov): New option.
3036 * config/bpf/bpf.cc (bpf_option_override): Handle it here.
3037 * config/bpf/bpf.md (*extendsidi2): New.
3043 * doc/invoke.texi (Option Summary): Add -msmov eBPF option.
3044 (eBPF Options): Add -m[no-]smov. Document that -mcpu=v4
3045 also enables -msmov.
3047 2023-07-27 David Faust <david.faust@oracle.com>
3049 * doc/invoke.texi (Option Summary): Remove -mkernel eBPF option.
3050 Add -mbswap and -msdiv eBPF options.
3051 (eBPF Options): Remove -mkernel. Add -mno-{jmpext, jmp32,
3052 alu32, v3-atomics, bswap, sdiv}. Document that -mcpu=v4 also
3055 2023-07-27 David Faust <david.faust@oracle.com>
3057 * config/bpf/bpf.md (add<AM:mode>3): Use %w2 instead of %w1
3058 in pseudo-C dialect output template.
3059 (sub<AM:mode>3): Likewise.
3061 2023-07-27 Jan Hubicka <jh@suse.cz>
3063 * tree-vect-loop.cc (optimize_mask_stores): Make store
3066 2023-07-27 Jan Hubicka <jh@suse.cz>
3068 * cfgloop.h (single_dom_exit): Declare.
3069 * cfgloopmanip.h (update_exit_probability_after_unrolling): Declare.
3070 * cfgrtl.cc (struct cfg_hooks): Fix comment.
3071 * loop-unroll.cc (unroll_loop_constant_iterations): Update exit edge.
3072 * tree-ssa-loop-ivopts.h (single_dom_exit): Do not declare it here.
3073 * tree-ssa-loop-manip.cc (update_exit_probability_after_unrolling):
3075 (tree_transform_and_unroll_loop): ... here;
3077 2023-07-27 Jan Hubicka <jh@suse.cz>
3079 * cfgloopmanip.cc (scale_dominated_blocks_in_loop): Move here from
3080 tree-ssa-loop-manip.cc and avoid recursion.
3081 (scale_loop_profile): Use scale_dominated_blocks_in_loop.
3082 (duplicate_loop_body_to_header_edge): Add DLTHE_FLAG_FLAT_PROFILE
3084 * cfgloopmanip.h (DLTHE_FLAG_FLAT_PROFILE): Define.
3085 (scale_dominated_blocks_in_loop): Declare.
3086 * predict.cc (dump_prediction): Do not ICE on uninitialized probability.
3087 (change_edge_frequency): Remove.
3088 * predict.h (change_edge_frequency): Remove.
3089 * tree-ssa-loop-manip.cc (scale_dominated_blocks_in_loop): Move to
3091 (niter_for_unrolled_loop): Remove.
3092 (tree_transform_and_unroll_loop): Fix profile update.
3094 2023-07-27 Jan Hubicka <jh@suse.cz>
3096 * tree-ssa-loop-im.cc (execute_sm_if_changed): Turn cap probability
3097 to guessed; fix count of new_bb.
3099 2023-07-27 Jan Hubicka <jh@suse.cz>
3101 * profile-count.h (profile_count::apply_probability): Fix
3102 handling of uninitialized probabilities, optimize scaling
3105 2023-07-27 Richard Biener <rguenther@suse.de>
3107 PR tree-optimization/91838
3108 * gimple-match-head.cc: Include attribs.h and asan.h.
3109 * generic-match-head.cc: Likewise.
3110 * match.pd (([rl]shift @0 out-of-bounds) -> zero): New pattern.
3112 2023-07-27 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3114 * config/riscv/riscv-modes.def (VECTOR_BOOL_MODE): Add VLS modes.
3115 (ADJUST_ALIGNMENT): Ditto.
3116 (ADJUST_PRECISION): Ditto.
3118 (VECTOR_MODE_WITH_PREFIX): Ditto.
3119 * config/riscv/riscv-opts.h (TARGET_VECTOR_VLS): New macro.
3120 * config/riscv/riscv-protos.h (riscv_v_ext_vls_mode_p): New function.
3121 * config/riscv/riscv-v.cc (INCLUDE_ALGORITHM): Add include.
3122 (legitimize_move): Enable basic VLS modes support.
3125 (get_vector_mode): Ditto.
3126 * config/riscv/riscv-vector-switch.def (VLS_ENTRY): Add vls modes.
3127 * config/riscv/riscv.cc (riscv_v_ext_vls_mode_p): New function.
3128 (VLS_ENTRY): New macro.
3129 (riscv_v_ext_mode_p): Add vls modes.
3130 (riscv_get_v_regno_alignment): New function.
3131 (riscv_print_operand): Add vls modes.
3132 (riscv_hard_regno_nregs): Ditto.
3133 (riscv_hard_regno_mode_ok): Ditto.
3134 (riscv_regmode_natural_size): Ditto.
3135 (riscv_vectorize_preferred_vector_alignment): Ditto.
3136 * config/riscv/riscv.md: Ditto.
3137 * config/riscv/vector-iterators.md: Ditto.
3138 * config/riscv/vector.md: Ditto.
3139 * config/riscv/autovec-vls.md: New file.
3141 2023-07-27 Pan Li <pan2.li@intel.com>
3143 * config/riscv/riscv_vector.h (enum RVV_CSR): Removed.
3145 (vwrite_csr): Ditto.
3147 2023-07-27 demin.han <demin.han@starfivetech.com>
3149 * config/riscv/autovec.md: Delete which_alternative use in split
3151 2023-07-27 Richard Biener <rguenther@suse.de>
3153 * tree-ssa-sink.cc (sink_code_in_bb): Remove recursion, instead
3155 (pass_sink_code::execute): ... in the caller.
3157 2023-07-27 Kewen Lin <linkw@linux.ibm.com>
3158 Richard Biener <rguenther@suse.de>
3160 PR tree-optimization/110776
3161 * tree-vect-stmts.cc (vectorizable_load): Always cost VMAT_ELEMENTWISE
3164 2023-07-26 Xiao Zeng <zengxiao@eswincomputing.com>
3166 * config/riscv/riscv.md: Include zicond.md
3167 * config/riscv/zicond.md: New file.
3169 2023-07-26 Xiao Zeng <zengxiao@eswincomputing.com>
3171 * common/config/riscv/riscv-common.cc: New extension.
3172 * config/riscv/riscv-opts.h (MASK_ZICOND): New mask.
3173 (TARGET_ZICOND): New target.
3175 2023-07-26 Carl Love <cel@us.ibm.com>
3177 * config/rs6000/rs6000-c.cc (find_instance): Add new parameter that
3178 specifies the number of built-in arguments to check.
3179 (altivec_resolve_overloaded_builtin): Update calls to find_instance
3180 to pass the number of built-in arguments to be checked.
3182 2023-07-26 David Faust <david.faust@oracle.com>
3184 * config/bpf/bpf.opt (mv3-atomics): New option.
3185 * config/bpf/bpf.cc (bpf_option_override): Handle it here.
3186 * config/bpf/bpf.h (enum_reg_class): Add R0 class.
3187 (REG_CLASS_NAMES): Likewise.
3188 (REG_CLASS_CONTENTS): Likewise.
3189 (REGNO_REG_CLASS): Handle R0.
3190 * config/bpf/bpf.md (UNSPEC_XADD): Rename to UNSPEC_AADD.
3191 (UNSPEC_AAND): New unspec.
3192 (UNSPEC_AOR): Likewise.
3193 (UNSPEC_AXOR): Likewise.
3194 (UNSPEC_AFADD): Likewise.
3195 (UNSPEC_AFAND): Likewise.
3196 (UNSPEC_AFOR): Likewise.
3197 (UNSPEC_AFXOR): Likewise.
3198 (UNSPEC_AXCHG): Likewise.
3199 (UNSPEC_ACMPX): Likewise.
3200 (atomic_add<mode>): Use UNSPEC_AADD and atomic type attribute.
3202 * config/bpf/atomic.md: ...Here. New file.
3203 * config/bpf/constraints.md (t): New constraint for R0.
3204 * doc/invoke.texi (eBPF Options): Document -mv3-atomics.
3206 2023-07-26 Matthew Malcomson <matthew.malcomson@arm.com>
3208 * tree-vect-stmts.cc (get_group_load_store_type): Reformat
3211 2023-07-26 Carl Love <cel@us.ibm.com>
3213 * config/rs6000/rs6000-builtins.def: Rename
3214 __builtin_altivec_vreplace_un_uv2di as __builtin_altivec_vreplace_un_udi
3215 __builtin_altivec_vreplace_un_uv4si as __builtin_altivec_vreplace_un_usi
3216 __builtin_altivec_vreplace_un_v2df as __builtin_altivec_vreplace_un_df
3217 __builtin_altivec_vreplace_un_v2di as __builtin_altivec_vreplace_un_di
3218 __builtin_altivec_vreplace_un_v4sf as __builtin_altivec_vreplace_un_sf
3219 __builtin_altivec_vreplace_un_v4si as __builtin_altivec_vreplace_un_si.
3220 Rename VREPLACE_UN_UV2DI as VREPLACE_UN_UDI, VREPLACE_UN_UV4SI as
3221 VREPLACE_UN_USI, VREPLACE_UN_V2DF as VREPLACE_UN_DF,
3222 VREPLACE_UN_V2DI as VREPLACE_UN_DI, VREPLACE_UN_V4SF as
3223 VREPLACE_UN_SF, VREPLACE_UN_V4SI as VREPLACE_UN_SI.
3224 Rename vreplace_un_v2di as vreplace_un_di, vreplace_un_v4si as
3225 vreplace_un_si, vreplace_un_v2df as vreplace_un_df,
3226 vreplace_un_v2di as vreplace_un_di, vreplace_un_v4sf as
3227 vreplace_un_sf, vreplace_un_v4si as vreplace_un_si.
3228 * config/rs6000/rs6000-c.cc (find_instance): Add case
3229 RS6000_OVLD_VEC_REPLACE_UN.
3230 * config/rs6000/rs6000-overload.def (__builtin_vec_replace_un):
3231 Fix first argument type. Rename VREPLACE_UN_UV4SI as
3232 VREPLACE_UN_USI, VREPLACE_UN_V4SI as VREPLACE_UN_SI,
3233 VREPLACE_UN_UV2DI as VREPLACE_UN_UDI, VREPLACE_UN_V2DI as
3234 VREPLACE_UN_DI, VREPLACE_UN_V4SF as VREPLACE_UN_SF,
3235 VREPLACE_UN_V2DF as VREPLACE_UN_DF.
3236 * config/rs6000/vsx.md (REPLACE_ELT): Rename the mode_iterator
3237 REPLACE_ELT_V for vector modes.
3238 (REPLACE_ELT): New scalar mode iterator.
3239 (REPLACE_ELT_char): Add scalar attributes.
3240 (vreplace_un_<mode>): Change iterator and mode attribute.
3242 2023-07-26 David Malcolm <dmalcolm@redhat.com>
3245 * Makefile.in (ANALYZER_OBJS): Add analyzer/symbol.o.
3247 2023-07-26 Richard Biener <rguenther@suse.de>
3249 PR tree-optimization/106081
3250 * tree-vect-slp.cc (vect_optimize_slp_pass::start_choosing_layouts):
3251 Assign layout -1 to splats.
3253 2023-07-26 Aldy Hernandez <aldyh@redhat.com>
3255 * range-op-mixed.h (class operator_cast): Add update_bitmask.
3256 * range-op.cc (operator_cast::update_bitmask): New.
3257 (operator_cast::fold_range): Call update_bitmask.
3259 2023-07-26 Li Xu <xuli1@eswincomputing.com>
3261 * config/riscv/riscv-vector-builtins.def (vfloat16mf4x2_t): Change
3262 scalar type to float16, eliminate warning.
3263 (vfloat16mf4x3_t): Ditto.
3264 (vfloat16mf4x4_t): Ditto.
3265 (vfloat16mf4x5_t): Ditto.
3266 (vfloat16mf4x6_t): Ditto.
3267 (vfloat16mf4x7_t): Ditto.
3268 (vfloat16mf4x8_t): Ditto.
3269 (vfloat16mf2x2_t): Ditto.
3270 (vfloat16mf2x3_t): Ditto.
3271 (vfloat16mf2x4_t): Ditto.
3272 (vfloat16mf2x5_t): Ditto.
3273 (vfloat16mf2x6_t): Ditto.
3274 (vfloat16mf2x7_t): Ditto.
3275 (vfloat16mf2x8_t): Ditto.
3276 (vfloat16m1x2_t): Ditto.
3277 (vfloat16m1x3_t): Ditto.
3278 (vfloat16m1x4_t): Ditto.
3279 (vfloat16m1x5_t): Ditto.
3280 (vfloat16m1x6_t): Ditto.
3281 (vfloat16m1x7_t): Ditto.
3282 (vfloat16m1x8_t): Ditto.
3283 (vfloat16m2x2_t): Ditto.
3284 (vfloat16m2x3_t): Ditto.
3285 (vfloat16m2x4_t): Ditto.
3286 (vfloat16m4x2_t): Ditto.
3287 * config/riscv/vector-iterators.md: add RVVM4x2DF in iterator V4T.
3288 * config/riscv/vector.md: add tuple mode in attr sew.
3290 2023-07-26 Uros Bizjak <ubizjak@gmail.com>
3293 * config/i386/i386.md (plusminusmult): New code iterator.
3294 * config/i386/mmx.md (mmxdoublevecmode): New mode attribute.
3295 (movq_<mode>_to_sse): New expander.
3296 (<plusminusmult:insn>v2sf3): Macroize expander from addv2sf3,
3297 subv2sf3 and mulv2sf3 using plusminusmult code iterator. Rewrite
3298 as a wrapper around V4SFmode operation.
3299 (mmx_addv2sf3): Change operand 1 and operand 2 predicates to
3300 nonimmediate_operand.
3301 (*mmx_addv2sf3): Remove SSE alternatives. Change operand 1 and
3302 operand 2 predicates to nonimmediate_operand.
3303 (mmx_subv2sf3): Change operand 2 predicate to nonimmediate_operand.
3304 (mmx_subrv2sf3): Change operand 1 predicate to nonimmediate_operand.
3305 (*mmx_subv2sf3): Remove SSE alternatives. Change operand 1 and
3306 operand 2 predicates to nonimmediate_operand.
3307 (mmx_mulv2sf3): Change operand 1 and operand 2 predicates to
3308 nonimmediate_operand.
3309 (*mmx_mulv2sf3): Remove SSE alternatives. Change operand 1 and
3310 operand 2 predicates to nonimmediate_operand.
3311 (divv2sf3): Rewrite as a wrapper around V4SFmode operation.
3312 (<smaxmin:code>v2sf3): Ditto.
3313 (mmx_<smaxmin:code>v2sf3): Change operand 1 and operand 2
3314 predicates to nonimmediate_operand.
3315 (*mmx_<smaxmin:code>v2sf3): Remove SSE alternatives. Change
3316 operand 1 and operand 2 predicates to nonimmediate_operand.
3317 (mmx_ieee_<ieee_maxmin>v2sf3): Ditto.
3318 (sqrtv2sf2): Rewrite as a wrapper around V4SFmode operation.
3319 (*mmx_haddv2sf3_low): Ditto.
3320 (*mmx_hsubv2sf3_low): Ditto.
3321 (vec_addsubv2sf3): Ditto.
3322 (*mmx_maskcmpv2sf3_comm): Remove.
3323 (*mmx_maskcmpv2sf3): Remove.
3324 (vec_cmpv2sfv2si): Rewrite as a wrapper around V4SFmode operation.
3325 (vcond<V2FI:mode>v2sf): Ditto.
3330 (fix_truncv2sfv2si2): Ditto.
3331 (fixuns_truncv2sfv2si2): Ditto.
3332 (mmx_fix_truncv2sfv2si2): Remove SSE alternatives.
3333 Change operand 1 predicate to nonimmediate_operand.
3334 (floatv2siv2sf2): Rewrite as a wrapper around V4SFmode operation.
3335 (floatunsv2siv2sf2): Ditto.
3336 (mmx_floatv2siv2sf2): Remove SSE alternatives.
3337 Change operand 1 predicate to nonimmediate_operand.
3338 (nearbyintv2sf2): Rewrite as a wrapper around V4SFmode operation.
3340 (lrintv2sfv2si2): Ditto.
3342 (lceilv2sfv2si2): Ditto.
3343 (floorv2sf2): Ditto.
3344 (lfloorv2sfv2si2): Ditto.
3345 (btruncv2sf2): Ditto.
3346 (roundv2sf2): Ditto.
3347 (lroundv2sfv2si2): Ditto.
3348 (*mmx_roundv2sf2): Remove.
3350 2023-07-26 Jose E. Marchesi <jose.marchesi@oracle.com>
3352 * config/bpf/bpf.md: Fix neg{SI,DI}2 insn.
3354 2023-07-26 Richard Biener <rguenther@suse.de>
3356 PR tree-optimization/110799
3357 * tree-ssa-pre.cc (compute_avail): More thoroughly match
3358 up TBAA behavior of redundant loads.
3360 2023-07-26 Jakub Jelinek <jakub@redhat.com>
3362 PR tree-optimization/110755
3363 * range-op-float.cc (frange_arithmetic): Change +0 result to -0
3364 for PLUS_EXPR or MINUS_EXPR if -frounding-math, inf is negative and
3365 it is exact op1 + (-op1) or op1 - op1.
3367 2023-07-26 Kewen Lin <linkw@linux.ibm.com>
3370 * config/rs6000/vsx.md (define_insn xxeval): Correct vsx
3371 operands output with "x".
3373 2023-07-26 Aldy Hernandez <aldyh@redhat.com>
3375 * range-op.cc (class operator_absu): Add update_bitmask.
3376 (operator_absu::update_bitmask): New.
3378 2023-07-26 Aldy Hernandez <aldyh@redhat.com>
3380 * range-op-mixed.h (class operator_abs): Add update_bitmask.
3381 * range-op.cc (operator_abs::update_bitmask): New.
3383 2023-07-26 Aldy Hernandez <aldyh@redhat.com>
3385 * range-op-mixed.h (class operator_bitwise_not): Add update_bitmask.
3386 * range-op.cc (operator_bitwise_not::update_bitmask): New.
3388 2023-07-26 Aldy Hernandez <aldyh@redhat.com>
3390 * range-op.cc (update_known_bitmask): Handle unary operators.
3392 2023-07-26 Aldy Hernandez <aldyh@redhat.com>
3394 * tree-ssa-ccp.cc (bit_value_unop): Initialize val when appropriate.
3396 2023-07-26 Jin Ma <jinma@linux.alibaba.com>
3398 * config/riscv/riscv.md: Likewise.
3400 2023-07-26 Jan Hubicka <jh@suse.cz>
3402 * profile-count.cc (profile_count::to_sreal_scale): Value is not know
3403 if we divide by zero.
3405 2023-07-25 David Faust <david.faust@oracle.com>
3407 * config/bpf/bpf.cc (bpf_print_operand_address): Don't print
3408 enclosing parentheses for pseudo-C dialect.
3409 * config/bpf/bpf.md (zero_exdendhidi2): Add parentheses around
3410 operands of pseudo-C dialect output templates where needed.
3411 (zero_extendqidi2): Likewise.
3412 (zero_extendsidi2): Likewise.
3413 (*mov<MM:mode>): Likewise.
3415 2023-07-25 Aldy Hernandez <aldyh@redhat.com>
3417 * tree-ssa-ccp.cc (value_mask_to_min_max): Make static.
3418 (bit_value_mult_const): Same.
3419 (get_individual_bits): Same.
3421 2023-07-25 Haochen Gui <guihaoc@gcc.gnu.org>
3424 * config/rs6000/rs6000-builtin.cc (rs6000_gimple_fold_builtin): Gimple
3425 fold RS6000_BIF_XSMINDP and RS6000_BIF_XSMAXDP when fast-math is set.
3426 * config/rs6000/rs6000.md (FMINMAX): New int iterator.
3427 (minmax_op): New int attribute.
3428 (UNSPEC_FMAX, UNSPEC_FMIN): New unspecs.
3429 (f<minmax_op><mode>3): New pattern by UNSPEC_FMAX and UNSPEC_FMIN.
3430 * config/rs6000/rs6000-builtins.def (__builtin_vsx_xsmaxdp): Set
3432 (__builtin_vsx_xsmindp): Set pattern to fmindf3.
3434 2023-07-24 David Faust <david.faust@oracle.com>
3436 * config/bpf/bpf.md (nop): Add pseudo-c asm dialect template.
3438 2023-07-24 Drew Ross <drross@redhat.com>
3439 Jakub Jelinek <jakub@redhat.com>
3441 PR middle-end/109986
3442 * generic-match-head.cc (bitwise_equal_p): New macro.
3443 * gimple-match-head.cc (bitwise_equal_p): New macro.
3444 (gimple_nop_convert): Declare.
3445 (gimple_bitwise_equal_p): Helper for bitwise_equal_p.
3446 * match.pd ((~X | Y) ^ X -> ~(X & Y)): New simplification.
3448 2023-07-24 Jeff Law <jlaw@ventanamicro.com>
3450 * common/config/riscv/riscv-common.cc (riscv_subset_list::add): Use
3451 single quote rather than backquote in diagnostic.
3453 2023-07-24 Jose E. Marchesi <jose.marchesi@oracle.com>
3456 * config/bpf/bpf.opt: New command-line option -msdiv.
3457 * config/bpf/bpf.md: Conditionalize sdiv/smod on bpf_has_sdiv.
3458 * config/bpf/bpf.cc (bpf_option_override): Initialize
3460 * doc/invoke.texi (eBPF Options): Document -msdiv.
3462 2023-07-24 Jeff Law <jlaw@ventanamicro.com>
3464 * config/riscv/riscv.cc (riscv_option_override): Spell out
3465 greater than and use cannot in diagnostic string.
3467 2023-07-24 Richard Biener <rguenther@suse.de>
3469 * tree-vectorizer.h (_slp_tree::push_vec_def): Add.
3470 (_slp_tree::vec_stmts): Remove.
3471 (SLP_TREE_VEC_STMTS): Remove.
3472 * tree-vect-slp.cc (_slp_tree::push_vec_def): Define.
3473 (_slp_tree::_slp_tree): Adjust.
3474 (_slp_tree::~_slp_tree): Likewise.
3475 (vect_get_slp_vect_def): Simplify.
3476 (vect_get_slp_defs): Likewise.
3477 (vect_transform_slp_perm_load_1): Adjust.
3478 (vect_add_slp_permutation): Likewise.
3479 (vect_schedule_slp_node): Likewise.
3480 (vectorize_slp_instance_root_stmt): Likewise.
3481 (vect_schedule_scc): Likewise.
3482 * tree-vect-stmts.cc (vectorizable_bswap): Use push_vec_def.
3483 (vectorizable_call): Likewise.
3484 (vectorizable_call): Likewise.
3485 (vect_create_vectorized_demotion_stmts): Likewise.
3486 (vectorizable_conversion): Likewise.
3487 (vectorizable_assignment): Likewise.
3488 (vectorizable_shift): Likewise.
3489 (vectorizable_operation): Likewise.
3490 (vectorizable_load): Likewise.
3491 (vectorizable_condition): Likewise.
3492 (vectorizable_comparison): Likewise.
3493 * tree-vect-loop.cc (vect_create_epilog_for_reduction): Adjust.
3494 (vectorize_fold_left_reduction): Use push_vec_def.
3495 (vect_transform_reduction): Likewise.
3496 (vect_transform_cycle_phi): Likewise.
3497 (vectorizable_lc_phi): Likewise.
3498 (vectorizable_phi): Likewise.
3499 (vectorizable_recurr): Likewise.
3500 (vectorizable_induction): Likewise.
3501 (vectorizable_live_operation): Likewise.
3503 2023-07-24 Richard Biener <rguenther@suse.de>
3505 * tree-ssa-loop.cc: Remove unused tree-vectorizer.h include.
3507 2023-07-24 Richard Biener <rguenther@suse.de>
3509 * config/i386/i386-builtins.cc: Remove tree-vectorizer.h include.
3510 * config/i386/i386-expand.cc: Likewise.
3511 * config/i386/i386-features.cc: Likewise.
3512 * config/i386/i386-options.cc: Likewise.
3514 2023-07-24 Robin Dapp <rdapp@ventanamicro.com>
3516 * tree-vect-stmts.cc (vectorizable_conversion): Handle
3517 more demotion/promotion for modifier == NONE.
3519 2023-07-24 Roger Sayle <roger@nextmovesoftware.com>
3524 * config/i386/i386.md (extv<mode>): Use QImode for offsets.
3525 (extzv<mode>): Likewise.
3526 (insv<mode>): Likewise.
3527 (*testqi_ext_3): Likewise.
3528 (*btr<mode>_2): Likewise.
3529 (define_split): Likewise.
3530 (*btsq_imm): Likewise.
3531 (*btrq_imm): Likewise.
3532 (*btcq_imm): Likewise.
3533 (define_peephole2 x3): Likewise.
3534 (*bt<mode>): Likewise
3535 (*bt<mode>_mask): New define_insn_and_split.
3536 (*jcc_bt<mode>): Use QImode for offsets.
3537 (*jcc_bt<mode>_1): Delete obsolete pattern.
3538 (*jcc_bt<mode>_mask): Use QImode offsets.
3539 (*jcc_bt<mode>_mask_1): Likewise.
3540 (define_split): Likewise.
3541 (*bt<mode>_setcqi): Likewise.
3542 (*bt<mode>_setncqi): Likewise.
3543 (*bt<mode>_setnc<mode>): Likewise.
3544 (*bt<mode>_setncqi_2): Likewise.
3545 (*bt<mode>_setc<mode>_mask): New define_insn_and_split.
3546 (bmi2_bzhi_<mode>3): Use QImode offsets.
3547 (*bmi2_bzhi_<mode>3): Likewise.
3548 (*bmi2_bzhi_<mode>3_1): Likewise.
3549 (*bmi2_bzhi_<mode>3_1_ccz): Likewise.
3550 (@tbm_bextri_<mode>): Likewise.
3552 2023-07-24 Jose E. Marchesi <jose.marchesi@oracle.com>
3554 * config/bpf/bpf-opts.h (enum bpf_kernel_version): Remove enum.
3555 * config/bpf/bpf.opt (mkernel): Remove option.
3556 * config/bpf/bpf.cc (bpf_target_macros): Do not define
3557 BPF_KERNEL_VERSION_CODE.
3559 2023-07-24 Jose E. Marchesi <jose.marchesi@oracle.com>
3562 * config/bpf/bpf.opt (mcpu): Add ISA_V4 and make it the default.
3563 (mbswap): New option.
3564 * config/bpf/bpf-opts.h (enum bpf_isa_version): New value ISA_V4.
3565 * config/bpf/bpf.cc (bpf_option_override): Set bpf_has_bswap.
3566 * config/bpf/bpf.md: Use bswap instructions if available for
3567 bswap* insn, and fix constraint.
3568 * doc/invoke.texi (eBPF Options): Document -mcpu=v4 and -mbswap.
3570 2023-07-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3572 * config/riscv/autovec.md (fold_left_plus_<mode>): New pattern.
3573 (mask_len_fold_left_plus_<mode>): Ditto.
3574 * config/riscv/riscv-protos.h (enum insn_type): New enum.
3575 (enum reduction_type): Ditto.
3576 (expand_reduction): Add in-order reduction.
3577 * config/riscv/riscv-v.cc (emit_nonvlmax_fp_reduction_insn): New function.
3578 (expand_reduction): Add in-order reduction.
3580 2023-07-24 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
3582 * tree-vect-loop.cc (get_masked_reduction_fn): Add mask_len_fold_left_plus.
3583 (vectorize_fold_left_reduction): Ditto.
3584 (vectorizable_reduction): Ditto.
3585 (vect_transform_reduction): Ditto.
3587 2023-07-24 Richard Biener <rguenther@suse.de>
3589 PR tree-optimization/110777
3590 * tree-ssa-sccvn.cc (eliminate_dom_walker::eliminate_avail):
3591 Avoid propagating abnormals.
3593 2023-07-24 Richard Biener <rguenther@suse.de>
3595 PR tree-optimization/110766
3596 * tree-scalar-evolution.cc
3597 (analyze_and_compute_bitwise_induction_effect): Check the PHI
3598 is defined in the loop header.
3600 2023-07-24 Kewen Lin <linkw@linux.ibm.com>
3602 PR tree-optimization/110740
3603 * tree-vect-loop.cc (vect_analyze_loop_costing): Do not vectorize a
3604 loop with a single scalar iteration.
3606 2023-07-24 Pan Li <pan2.li@intel.com>
3608 * config/riscv/riscv-vector-builtins-shapes.cc
3609 (struct alu_frm_def): Take range check.
3611 2023-07-22 Vineet Gupta <vineetg@rivosinc.com>
3614 * config/riscv/predicates.md (const_0_operand): Add back
3617 2023-07-22 Roger Sayle <roger@nextmovesoftware.com>
3619 * config/i386/i386-expand.cc (ix86_expand_move): Disable the
3620 64-bit insertions into TImode optimizations with -O0, unless
3621 the function has the "naked" attribute (for PR target/110533).
3623 2023-07-22 Andrew Pinski <apinski@marvell.com>
3626 * rtl.h (extended_count): Change last argument type
3629 2023-07-22 Roger Sayle <roger@nextmovesoftware.com>
3631 * config/i386/i386.md (extv<mode>): Use QImode for offsets.
3632 (extzv<mode>): Likewise.
3633 (insv<mode>): Likewise.
3634 (*testqi_ext_3): Likewise.
3635 (*btr<mode>_2): Likewise.
3636 (define_split): Likewise.
3637 (*btsq_imm): Likewise.
3638 (*btrq_imm): Likewise.
3639 (*btcq_imm): Likewise.
3640 (define_peephole2 x3): Likewise.
3641 (*bt<mode>): Likewise
3642 (*bt<mode>_mask): New define_insn_and_split.
3643 (*jcc_bt<mode>): Use QImode for offsets.
3644 (*jcc_bt<mode>_1): Delete obsolete pattern.
3645 (*jcc_bt<mode>_mask): Use QImode offsets.
3646 (*jcc_bt<mode>_mask_1): Likewise.
3647 (define_split): Likewise.
3648 (*bt<mode>_setcqi): Likewise.
3649 (*bt<mode>_setncqi): Likewise.
3650 (*bt<mode>_setnc<mode>): Likewise.
3651 (*bt<mode>_setncqi_2): Likewise.
3652 (*bt<mode>_setc<mode>_mask): New define_insn_and_split.
3653 (bmi2_bzhi_<mode>3): Use QImode offsets.
3654 (*bmi2_bzhi_<mode>3): Likewise.
3655 (*bmi2_bzhi_<mode>3_1): Likewise.
3656 (*bmi2_bzhi_<mode>3_1_ccz): Likewise.
3657 (@tbm_bextri_<mode>): Likewise.
3659 2023-07-22 Jeff Law <jlaw@ventanamicro.com>
3661 * config/bfin/bfin.md (ones): Fix length computation.
3663 2023-07-22 Vladimir N. Makarov <vmakarov@redhat.com>
3665 * lra-eliminations.cc (update_reg_eliminate): Fix the assert.
3666 (lra_update_fp2sp_elimination): Use HARD_FRAME_POINTER_REGNUM
3667 instead of FRAME_POINTER_REGNUM to spill pseudos.
3669 2023-07-21 Roger Sayle <roger@nextmovesoftware.com>
3670 Richard Biener <rguenther@suse.de>
3673 * gimplify.cc (gimplify_compound_lval): If the array's type
3674 is error_mark_node then return GS_ERROR.
3676 2023-07-21 Cupertino Miranda <cupertino.miranda@oracle.com>
3679 * config/bpf/bpf.opt: Added option -masm=<dialect>.
3680 * config/bpf/bpf-opts.h (enum bpf_asm_dialect): New type.
3681 * config/bpf/bpf.cc (bpf_print_register): New function.
3682 (bpf_print_register): Support pseudo-c syntax for registers.
3683 (bpf_print_operand_address): Likewise.
3684 * config/bpf/bpf.h (ASM_SPEC): handle -msasm.
3685 (ASSEMBLER_DIALECT): Define.
3686 * config/bpf/bpf.md: Added pseudo-c templates.
3687 * doc/invoke.texi (-masm=): New eBPF option item.
3689 2023-07-21 Cupertino Miranda <cupertino.miranda@oracle.com>
3691 * config/bpf/bpf.md: fixed template for neg instruction.
3693 2023-07-21 Jan Hubicka <jh@suse.cz>
3696 * tree-vect-loop.cc (scale_profile_for_vect_loop): Avoid scaling flat
3697 profiles by vectorization factor.
3698 (vect_transform_loop): Check for flat profiles.
3700 2023-07-21 Jan Hubicka <jh@suse.cz>
3702 * cfgloop.h (maybe_flat_loop_profile): Declare
3703 * cfgloopanal.cc (maybe_flat_loop_profile): New function.
3704 * tree-cfg.cc (print_loop_info): Print info about flat profiles.
3706 2023-07-21 Jan Hubicka <jh@suse.cz>
3708 * cfgloop.cc (get_estimated_loop_iterations): Use sreal::to_nearest_int
3709 * cfgloopanal.cc (expected_loop_iterations_unbounded): Likewise.
3710 * predict.cc (estimate_bb_frequencies): Likewise.
3711 * profile.cc (branch_prob): Likewise.
3712 * tree-ssa-loop-niter.cc (estimate_numbers_of_iterations): Likewise
3714 2023-07-21 Iain Sandoe <iain@sandoe.co.uk>
3716 * config.in: Regenerate.
3717 * config/darwin.h (DARWIN_LD_DEMANGLE): New.
3718 (LINK_COMMAND_SPEC_A): Add demangle handling.
3719 * configure: Regenerate.
3720 * configure.ac: Detect linker support for '-demangle'.
3722 2023-07-21 Jan Hubicka <jh@suse.cz>
3724 * sreal.cc (sreal::to_nearest_int): New.
3725 (sreal_verify_basics): Verify also to_nearest_int.
3726 (verify_aritmetics): Likewise.
3727 (sreal_verify_conversions): New.
3728 (sreal_cc_tests): Call sreal_verify_conversions.
3729 * sreal.h: (sreal::to_nearest_int): Declare
3731 2023-07-21 Jan Hubicka <jh@suse.cz>
3733 * tree-ssa-loop-ch.cc (enum ch_decision): New enum.
3734 (should_duplicate_loop_header_p): Return info on profitability.
3735 (do_while_loop_p): Watch for constant conditionals.
3736 (update_profile_after_ch): Do not sanity check that all
3737 static exits are taken.
3738 (ch_base::copy_headers): Run on all loops.
3739 (pass_ch::process_loop_p): Improve heuristics by handling also
3740 do_while loop and duplicating shortest sequence containing all
3743 2023-07-21 Jan Hubicka <jh@suse.cz>
3745 * tree-ssa-loop-niter.cc (finite_loop_p): Reorder to do cheap
3746 tests first; update finite_p flag.
3748 2023-07-21 Jan Hubicka <jh@suse.cz>
3750 * cfgloop.cc (flow_loop_dump): Use print_loop_info.
3751 * cfgloop.h (print_loop_info): Declare.
3752 * tree-cfg.cc (print_loop_info): Break out from ...; add
3753 printing of missing fields and profile
3754 (print_loop): ... here.
3756 2023-07-21 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3758 * config/riscv/riscv-v.cc (expand_gather_scatter): Remove redundant variables.
3760 2023-07-21 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3762 * tree-vect-stmts.cc (check_load_store_for_partial_vectors): Change condition order.
3763 (vectorizable_operation): Ditto.
3765 2023-07-21 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3767 * config/riscv/autovec.md: Align order of mask and len.
3768 * config/riscv/riscv-v.cc (expand_load_store): Ditto.
3769 (expand_gather_scatter): Ditto.
3770 * doc/md.texi: Ditto.
3771 * internal-fn.cc (add_len_and_mask_args): Ditto.
3772 (add_mask_and_len_args): Ditto.
3773 (expand_partial_load_optab_fn): Ditto.
3774 (expand_partial_store_optab_fn): Ditto.
3775 (expand_scatter_store_optab_fn): Ditto.
3776 (expand_gather_load_optab_fn): Ditto.
3777 (internal_fn_len_index): Ditto.
3778 (internal_fn_mask_index): Ditto.
3779 (internal_len_load_store_bias): Ditto.
3780 * tree-vect-stmts.cc (vectorizable_store): Ditto.
3781 (vectorizable_load): Ditto.
3783 2023-07-21 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3785 * config/riscv/autovec.md (len_maskload<mode><vm>): Change LEN_MASK into MASK_LEN.
3786 (mask_len_load<mode><vm>): Ditto.
3787 (len_maskstore<mode><vm>): Ditto.
3788 (mask_len_store<mode><vm>): Ditto.
3789 (len_mask_gather_load<RATIO64:mode><RATIO64I:mode>): Ditto.
3790 (mask_len_gather_load<RATIO64:mode><RATIO64I:mode>): Ditto.
3791 (len_mask_gather_load<RATIO32:mode><RATIO32I:mode>): Ditto.
3792 (mask_len_gather_load<RATIO32:mode><RATIO32I:mode>): Ditto.
3793 (len_mask_gather_load<RATIO16:mode><RATIO16I:mode>): Ditto.
3794 (mask_len_gather_load<RATIO16:mode><RATIO16I:mode>): Ditto.
3795 (len_mask_gather_load<RATIO8:mode><RATIO8I:mode>): Ditto.
3796 (mask_len_gather_load<RATIO8:mode><RATIO8I:mode>): Ditto.
3797 (len_mask_gather_load<RATIO4:mode><RATIO4I:mode>): Ditto.
3798 (mask_len_gather_load<RATIO4:mode><RATIO4I:mode>): Ditto.
3799 (len_mask_gather_load<RATIO2:mode><RATIO2I:mode>): Ditto.
3800 (mask_len_gather_load<RATIO2:mode><RATIO2I:mode>): Ditto.
3801 (len_mask_gather_load<RATIO1:mode><RATIO1:mode>): Ditto.
3802 (mask_len_gather_load<RATIO1:mode><RATIO1:mode>): Ditto.
3803 (len_mask_scatter_store<RATIO64:mode><RATIO64I:mode>): Ditto.
3804 (mask_len_scatter_store<RATIO64:mode><RATIO64I:mode>): Ditto.
3805 (len_mask_scatter_store<RATIO32:mode><RATIO32I:mode>): Ditto.
3806 (mask_len_scatter_store<RATIO32:mode><RATIO32I:mode>): Ditto.
3807 (len_mask_scatter_store<RATIO16:mode><RATIO16I:mode>): Ditto.
3808 (mask_len_scatter_store<RATIO16:mode><RATIO16I:mode>): Ditto.
3809 (len_mask_scatter_store<RATIO8:mode><RATIO8I:mode>): Ditto.
3810 (mask_len_scatter_store<RATIO8:mode><RATIO8I:mode>): Ditto.
3811 (len_mask_scatter_store<RATIO4:mode><RATIO4I:mode>): Ditto.
3812 (mask_len_scatter_store<RATIO4:mode><RATIO4I:mode>): Ditto.
3813 (len_mask_scatter_store<RATIO2:mode><RATIO2I:mode>): Ditto.
3814 (mask_len_scatter_store<RATIO2:mode><RATIO2I:mode>): Ditto.
3815 (len_mask_scatter_store<RATIO1:mode><RATIO1:mode>): Ditto.
3816 (mask_len_scatter_store<RATIO1:mode><RATIO1:mode>): Ditto.
3817 * doc/md.texi: Ditto.
3818 * genopinit.cc (main): Ditto.
3819 (CMP_NAME): Ditto. Ditto.
3820 * gimple-fold.cc (arith_overflowed_p): Ditto.
3821 (gimple_fold_partial_load_store_mem_ref): Ditto.
3822 (gimple_fold_call): Ditto.
3823 * internal-fn.cc (len_maskload_direct): Ditto.
3824 (mask_len_load_direct): Ditto.
3825 (len_maskstore_direct): Ditto.
3826 (mask_len_store_direct): Ditto.
3827 (expand_call_mem_ref): Ditto.
3828 (expand_len_maskload_optab_fn): Ditto.
3829 (expand_mask_len_load_optab_fn): Ditto.
3830 (expand_len_maskstore_optab_fn): Ditto.
3831 (expand_mask_len_store_optab_fn): Ditto.
3832 (direct_len_maskload_optab_supported_p): Ditto.
3833 (direct_mask_len_load_optab_supported_p): Ditto.
3834 (direct_len_maskstore_optab_supported_p): Ditto.
3835 (direct_mask_len_store_optab_supported_p): Ditto.
3836 (internal_load_fn_p): Ditto.
3837 (internal_store_fn_p): Ditto.
3838 (internal_gather_scatter_fn_p): Ditto.
3839 (internal_fn_len_index): Ditto.
3840 (internal_fn_mask_index): Ditto.
3841 (internal_fn_stored_value_index): Ditto.
3842 (internal_len_load_store_bias): Ditto.
3843 * internal-fn.def (LEN_MASK_GATHER_LOAD): Ditto.
3844 (MASK_LEN_GATHER_LOAD): Ditto.
3845 (LEN_MASK_LOAD): Ditto.
3846 (MASK_LEN_LOAD): Ditto.
3847 (LEN_MASK_SCATTER_STORE): Ditto.
3848 (MASK_LEN_SCATTER_STORE): Ditto.
3849 (LEN_MASK_STORE): Ditto.
3850 (MASK_LEN_STORE): Ditto.
3851 * optabs-query.cc (supports_vec_gather_load_p): Ditto.
3852 (supports_vec_scatter_store_p): Ditto.
3853 * optabs-tree.cc (target_supports_mask_load_store_p): Ditto.
3854 (target_supports_len_load_store_p): Ditto.
3855 * optabs.def (OPTAB_CD): Ditto.
3856 * tree-ssa-alias.cc (ref_maybe_used_by_call_p_1): Ditto.
3857 (call_may_clobber_ref_p_1): Ditto.
3858 * tree-ssa-dse.cc (initialize_ao_ref_for_dse): Ditto.
3859 (dse_optimize_stmt): Ditto.
3860 * tree-ssa-loop-ivopts.cc (get_mem_type_for_internal_fn): Ditto.
3861 (get_alias_ptr_type_for_ptr_address): Ditto.
3862 * tree-vect-data-refs.cc (vect_gather_scatter_fn_p): Ditto.
3863 * tree-vect-patterns.cc (vect_recog_gather_scatter_pattern): Ditto.
3864 * tree-vect-stmts.cc (check_load_store_for_partial_vectors): Ditto.
3865 (vect_get_strided_load_store_ops): Ditto.
3866 (vectorizable_store): Ditto.
3867 (vectorizable_load): Ditto.
3869 2023-07-21 Haochen Jiang <haochen.jiang@intel.com>
3871 * config/i386/i386.opt: Fix a typo.
3873 2023-07-21 Richard Biener <rguenther@suse.de>
3875 PR tree-optimization/88540
3876 * tree-ssa-phiopt.cc (minmax_replacement): Do not give up
3877 with NaNs but handle the simple case by if-converting to a
3880 2023-07-21 Andrew Pinski <apinski@marvell.com>
3882 * match.pd (minmax<minmax<a,b>,a>->minmax<a,b>): New
3885 2023-07-21 Richard Biener <rguenther@suse.de>
3887 PR tree-optimization/110742
3888 * tree-vect-slp.cc (vect_optimize_slp_pass::get_result_with_layout):
3889 Do not materialize an edge permutation in an external node with
3891 (vect_slp_analyze_node_operations_1): Guard purely internal
3894 2023-07-21 Jan Hubicka <jh@suse.cz>
3896 * cfgloop.cc: Include sreal.h.
3897 (flow_loop_dump): Dump sreal iteration exsitmate.
3898 (get_estimated_loop_iterations): Update.
3899 * cfgloop.h (expected_loop_iterations_by_profile): Declare.
3900 * cfgloopanal.cc (expected_loop_iterations_by_profile): New function.
3901 (expected_loop_iterations_unbounded): Use new API.
3902 * cfgloopmanip.cc (scale_loop_profile): Use
3903 expected_loop_iterations_by_profile
3904 * predict.cc (pass_profile::execute): Likewise.
3905 * profile.cc (branch_prob): Likewise.
3906 * tree-ssa-loop-niter.cc: Include sreal.h.
3907 (estimate_numbers_of_iterations): Likewise
3909 2023-07-21 Kewen Lin <linkw@linux.ibm.com>
3911 PR tree-optimization/110744
3912 * tree-ssa-sccvn.cc (vn_reference_lookup_3): Correct the index of bias
3913 operand for ifn IFN_LEN_STORE.
3915 2023-07-21 liuhongt <hongtao.liu@intel.com>
3918 * common.opt: (fcf-protection=): Add EnumSet attribute to
3919 support combination of params.
3921 2023-07-21 David Malcolm <dmalcolm@redhat.com>
3923 PR middle-end/110612
3924 * text-art/table.cc (table_geometry::table_geometry): Drop m_table
3926 (table_geometry::table_x_to_canvas_x): Add cast to comparison.
3927 (table_geometry::table_y_to_canvas_y): Likewise.
3928 * text-art/table.h (table_geometry::m_table): Drop unused field.
3929 * text-art/widget.h (wrapper_widget::update_child_alloc_rects):
3932 2023-07-20 Uros Bizjak <ubizjak@gmail.com>
3935 * config/i386/i386-features.cc
3936 (general_scalar_chain::compute_convert_gain): Calculate gain
3937 for extend higpart case.
3938 (general_scalar_chain::convert_op): Handle
3939 ASHIFTRT/ASHIFT combined RTX.
3940 (general_scalar_to_vector_candidate_p): Enable ASHIFTRT for
3941 SImode for SSE2 targets. Handle ASHIFTRT/ASHIFT combined RTX.
3942 * config/i386/i386.md (*extend<dwi>2_doubleword_highpart):
3943 New define_insn_and_split pattern.
3944 (*extendv2di2_highpart_stv): Ditto.
3946 2023-07-20 Vladimir N. Makarov <vmakarov@redhat.com>
3948 * lra-constraints.cc (simplify_operand_subreg): Check frame pointer
3951 2023-07-20 Andrew Pinski <apinski@marvell.com>
3953 * combine.cc (dump_combine_stats): Remove.
3954 (dump_combine_total_stats): Remove.
3955 (total_attempts, total_merges, total_extras,
3956 total_successes): Remove.
3957 (combine_instructions): Don't increment total stats
3958 instead use statistics_counter_event.
3959 * dumpfile.cc (print_combine_total_stats): Remove.
3960 * dumpfile.h (print_combine_total_stats): Remove.
3961 (dump_combine_total_stats): Remove.
3962 * passes.cc (finish_optimization_passes):
3963 Don't call print_combine_total_stats.
3964 * rtl.h (dump_combine_total_stats): Remove.
3965 (dump_combine_stats): Remove.
3967 2023-07-20 Jan Hubicka <jh@suse.cz>
3969 * tree-ssa-loop-ch.cc (should_duplicate_loop_header_p): Use BIT instead of TRUTH
3972 2023-07-20 Martin Jambor <mjambor@suse.cz>
3974 * doc/invoke.texi (analyzer-text-art-string-ellipsis-threshold): New.
3975 (analyzer-text-art-ideal-canvas-width): Likewise.
3976 (analyzer-text-art-string-ellipsis-head-len): Likewise.
3977 (analyzer-text-art-string-ellipsis-tail-len): Likewise.
3979 2023-07-20 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
3981 * tree-vect-stmts.cc (check_load_store_for_partial_vectors):
3982 Refine code structure.
3984 2023-07-20 Jan Hubicka <jh@suse.cz>
3986 * tree-ssa-loop-ch.cc (edge_range_query): Rename to ...
3987 (get_range_query): ... this one; do
3988 (static_loop_exit): Add query parametr, turn ranger to reference.
3989 (loop_static_stmt_p): New function.
3990 (loop_static_op_p): New function.
3991 (loop_iv_derived_p): Remove.
3992 (loop_combined_static_and_iv_p): New function.
3993 (should_duplicate_loop_header_p): Discover combined onditionals;
3994 do not track iv derived; improve dumps.
3995 (pass_ch::execute): Fix whitespace.
3997 2023-07-20 Richard Biener <rguenther@suse.de>
3999 PR tree-optimization/110204
4000 * tree-ssa-sccvn.cc (eliminate_dom_walker::eliminate_avail):
4001 Look through copies generated by PRE.
4003 2023-07-20 Matthew Malcomson <matthew.malcomson@arm.com>
4005 * tree-vect-stmts.cc (get_group_load_store_type): Account for
4006 `gap` when checking if need to peel twice.
4008 2023-07-20 Francois-Xavier Coudert <fxcoudert@gcc.gnu.org>
4011 * doc/extend.texi: Document iseqsig builtin.
4012 * builtins.cc (fold_builtin_iseqsig): New function.
4013 (fold_builtin_2): Handle BUILT_IN_ISEQSIG.
4014 (is_inexpensive_builtin): Handle BUILT_IN_ISEQSIG.
4015 * builtins.def (BUILT_IN_ISEQSIG): New built-in.
4017 2023-07-20 Pan Li <pan2.li@intel.com>
4019 * config/riscv/vector.md: Fix incorrect match_operand.
4021 2023-07-20 Roger Sayle <roger@nextmovesoftware.com>
4023 * config/i386/i386-expand.cc (ix86_expand_move): Don't call
4024 force_reg, to use SUBREG rather than create a new pseudo when
4025 inserting DFmode fields into TImode with insvti_{high,low}part.
4026 * config/i386/i386.md (*concat<mode><dwi>3_3): Split into two
4027 define_insn_and_split...
4028 (*concatditi3_3): 64-bit implementation. Provide alternative
4029 that allows register allocation to use SSE registers that is
4030 split into vec_concatv2di after reload.
4031 (*concatsidi3_3): 32-bit implementation.
4033 2023-07-20 Richard Biener <rguenther@suse.de>
4036 * internal-fn.cc (expand_vec_cond_optab_fn): When the
4037 value operands are equal to the original comparison operands
4038 preserve that equality by re-using the comparison expansion.
4039 * optabs.cc (emit_conditional_move): When the value operands
4040 are equal to the comparison operands and would be forced to
4041 a register by prepare_cmp_insn do so earlier, preserving the
4044 2023-07-20 Pan Li <pan2.li@intel.com>
4046 * config/riscv/vector.md: Align pattern format.
4048 2023-07-20 Haochen Jiang <haochen.jiang@intel.com>
4050 * doc/invoke.texi: Remove AVX512VP2INTERSECT in
4051 Granite Rapids{, D} from documentation.
4053 2023-07-20 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4055 * config/riscv/autovec.md
4056 (len_mask_gather_load<VNX16_QHSD:mode><VNX16_QHSDI:mode>):
4057 Refactor RVV machine modes.
4058 (len_mask_gather_load<VNX16_QHS:mode><VNX16_QHSI:mode>): Ditto.
4059 (len_mask_gather_load<VNX32_QHS:mode><VNX32_QHSI:mode>): Ditto.
4060 (len_mask_gather_load<VNX32_QH:mode><VNX32_QHI:mode>): Ditto.
4061 (len_mask_gather_load<VNX64_QH:mode><VNX64_QHI:mode>): Ditto.
4062 (len_mask_gather_load<mode><mode>): Ditto.
4063 (len_mask_gather_load<VNX64_Q:mode><VNX64_Q:mode>): Ditto.
4064 (len_mask_scatter_store<VNX16_QHSD:mode><VNX16_QHSDI:mode>): Ditto.
4065 (len_mask_scatter_store<VNX32_QHS:mode><VNX32_QHSI:mode>): Ditto.
4066 (len_mask_scatter_store<VNX16_QHS:mode><VNX16_QHSI:mode>): Ditto.
4067 (len_mask_scatter_store<VNX64_QH:mode><VNX64_QHI:mode>): Ditto.
4068 (len_mask_scatter_store<VNX32_QH:mode><VNX32_QHI:mode>): Ditto.
4069 (len_mask_scatter_store<mode><mode>): Ditto.
4070 (len_mask_scatter_store<VNX64_Q:mode><VNX64_Q:mode>): Ditto.
4071 * config/riscv/riscv-modes.def (VECTOR_BOOL_MODE): Ditto.
4072 (ADJUST_NUNITS): Ditto.
4073 (ADJUST_ALIGNMENT): Ditto.
4074 (ADJUST_BYTESIZE): Ditto.
4075 (ADJUST_PRECISION): Ditto.
4077 (RVV_WHOLE_MODES): Ditto.
4078 (RVV_FRACT_MODE): Ditto.
4079 (RVV_NF8_MODES): Ditto.
4080 (RVV_NF4_MODES): Ditto.
4081 (VECTOR_MODES_WITH_PREFIX): Ditto.
4082 (VECTOR_MODE_WITH_PREFIX): Ditto.
4083 (RVV_TUPLE_MODES): Ditto.
4084 (RVV_NF2_MODES): Ditto.
4085 (RVV_TUPLE_PARTIAL_MODES): Ditto.
4086 * config/riscv/riscv-v.cc (struct mode_vtype_group): Ditto.
4088 (TUPLE_ENTRY): Ditto.
4092 (preferred_simd_mode): Ditto.
4093 (autovectorize_vector_modes): Ditto.
4094 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_TYPE): Ditto.
4095 * config/riscv/riscv-vector-builtins.def (DEF_RVV_TYPE): Ditto.
4103 (vint8mf8_t): Ditto.
4104 (vuint8mf8_t): Ditto.
4105 (vint8mf4_t): Ditto.
4106 (vuint8mf4_t): Ditto.
4107 (vint8mf2_t): Ditto.
4108 (vuint8mf2_t): Ditto.
4110 (vuint8m1_t): Ditto.
4112 (vuint8m2_t): Ditto.
4114 (vuint8m4_t): Ditto.
4116 (vuint8m8_t): Ditto.
4117 (vint16mf4_t): Ditto.
4118 (vuint16mf4_t): Ditto.
4119 (vint16mf2_t): Ditto.
4120 (vuint16mf2_t): Ditto.
4121 (vint16m1_t): Ditto.
4122 (vuint16m1_t): Ditto.
4123 (vint16m2_t): Ditto.
4124 (vuint16m2_t): Ditto.
4125 (vint16m4_t): Ditto.
4126 (vuint16m4_t): Ditto.
4127 (vint16m8_t): Ditto.
4128 (vuint16m8_t): Ditto.
4129 (vint32mf2_t): Ditto.
4130 (vuint32mf2_t): Ditto.
4131 (vint32m1_t): Ditto.
4132 (vuint32m1_t): Ditto.
4133 (vint32m2_t): Ditto.
4134 (vuint32m2_t): Ditto.
4135 (vint32m4_t): Ditto.
4136 (vuint32m4_t): Ditto.
4137 (vint32m8_t): Ditto.
4138 (vuint32m8_t): Ditto.
4139 (vint64m1_t): Ditto.
4140 (vuint64m1_t): Ditto.
4141 (vint64m2_t): Ditto.
4142 (vuint64m2_t): Ditto.
4143 (vint64m4_t): Ditto.
4144 (vuint64m4_t): Ditto.
4145 (vint64m8_t): Ditto.
4146 (vuint64m8_t): Ditto.
4147 (vfloat16mf4_t): Ditto.
4148 (vfloat16mf2_t): Ditto.
4149 (vfloat16m1_t): Ditto.
4150 (vfloat16m2_t): Ditto.
4151 (vfloat16m4_t): Ditto.
4152 (vfloat16m8_t): Ditto.
4153 (vfloat32mf2_t): Ditto.
4154 (vfloat32m1_t): Ditto.
4155 (vfloat32m2_t): Ditto.
4156 (vfloat32m4_t): Ditto.
4157 (vfloat32m8_t): Ditto.
4158 (vfloat64m1_t): Ditto.
4159 (vfloat64m2_t): Ditto.
4160 (vfloat64m4_t): Ditto.
4161 (vfloat64m8_t): Ditto.
4162 * config/riscv/riscv-vector-switch.def (ENTRY): Ditto.
4163 (TUPLE_ENTRY): Ditto.
4164 * config/riscv/riscv-vsetvl.cc (change_insn): Ditto.
4165 * config/riscv/riscv.cc (riscv_valid_lo_sum_p): Ditto.
4166 (riscv_v_adjust_nunits): Ditto.
4167 (riscv_v_adjust_bytesize): Ditto.
4168 (riscv_v_adjust_precision): Ditto.
4169 (riscv_convert_vector_bits): Ditto.
4170 * config/riscv/riscv.h (riscv_v_adjust_nunits): Ditto.
4171 * config/riscv/riscv.md: Ditto.
4172 * config/riscv/vector-iterators.md: Ditto.
4173 * config/riscv/vector.md
4174 (@pred_indexed_<order>store<VNX16_QHSD:mode><VNX16_QHSDI:mode>): Ditto.
4175 (@pred_indexed_<order>store<VNX16_QHS:mode><VNX16_QHSI:mode>): Ditto.
4176 (@pred_indexed_<order>store<VNX32_QHS:mode><VNX32_QHSI:mode>): Ditto.
4177 (@pred_indexed_<order>store<VNX32_QH:mode><VNX32_QHI:mode>): Ditto.
4178 (@pred_indexed_<order>store<VNX64_QH:mode><VNX64_QHI:mode>): Ditto.
4179 (@pred_indexed_<order>store<VNX64_Q:mode><VNX64_Q:mode>): Ditto.
4180 (@pred_indexed_<order>store<VNX128_Q:mode><VNX128_Q:mode>): Ditto.
4181 (@pred_indexed_<order>load<V1T:mode><V1I:mode>): Ditto.
4182 (@pred_indexed_<order>load<V1T:mode><VNX1_QHSDI:mode>): Ditto.
4183 (@pred_indexed_<order>load<V2T:mode><V2I:mode>): Ditto.
4184 (@pred_indexed_<order>load<V2T:mode><VNX2_QHSDI:mode>): Ditto.
4185 (@pred_indexed_<order>load<V4T:mode><V4I:mode>): Ditto.
4186 (@pred_indexed_<order>load<V4T:mode><VNX4_QHSDI:mode>): Ditto.
4187 (@pred_indexed_<order>load<V8T:mode><V8I:mode>): Ditto.
4188 (@pred_indexed_<order>load<V8T:mode><VNX8_QHSDI:mode>): Ditto.
4189 (@pred_indexed_<order>load<V16T:mode><V16I:mode>): Ditto.
4190 (@pred_indexed_<order>load<V16T:mode><VNX16_QHSI:mode>): Ditto.
4191 (@pred_indexed_<order>load<V32T:mode><V32I:mode>): Ditto.
4192 (@pred_indexed_<order>load<V32T:mode><VNX32_QHI:mode>): Ditto.
4193 (@pred_indexed_<order>load<V64T:mode><V64I:mode>): Ditto.
4194 (@pred_indexed_<order>store<V1T:mode><V1I:mode>): Ditto.
4195 (@pred_indexed_<order>store<V1T:mode><VNX1_QHSDI:mode>): Ditto.
4196 (@pred_indexed_<order>store<V2T:mode><V2I:mode>): Ditto.
4197 (@pred_indexed_<order>store<V2T:mode><VNX2_QHSDI:mode>): Ditto.
4198 (@pred_indexed_<order>store<V4T:mode><V4I:mode>): Ditto.
4199 (@pred_indexed_<order>store<V4T:mode><VNX4_QHSDI:mode>): Ditto.
4200 (@pred_indexed_<order>store<V8T:mode><V8I:mode>): Ditto.
4201 (@pred_indexed_<order>store<V8T:mode><VNX8_QHSDI:mode>): Ditto.
4202 (@pred_indexed_<order>store<V16T:mode><V16I:mode>): Ditto.
4203 (@pred_indexed_<order>store<V16T:mode><VNX16_QHSI:mode>): Ditto.
4204 (@pred_indexed_<order>store<V32T:mode><V32I:mode>): Ditto.
4205 (@pred_indexed_<order>store<V32T:mode><VNX32_QHI:mode>): Ditto.
4206 (@pred_indexed_<order>store<V64T:mode><V64I:mode>): Ditto.
4208 2023-07-19 Vladimir N. Makarov <vmakarov@redhat.com>
4210 * lra-int.h (lra_update_fp2sp_elimination): New prototype.
4211 (lra_asm_insn_error): New prototype.
4212 * lra-spills.cc (remove_pseudos): Add check for pseudo slot memory
4214 (lra_spill): Call lra_update_fp2sp_elimination.
4215 * lra-eliminations.cc: Remove trailing spaces.
4216 (elimination_fp2sp_occured_p): New static flag.
4217 (lra_eliminate_regs_1): Set the flag up.
4218 (update_reg_eliminate): Modify the assert for stack to frame
4219 pointer elimination.
4220 (lra_update_fp2sp_elimination): New function.
4221 (lra_eliminate): Clear flag elimination_fp2sp_occured_p.
4223 2023-07-19 Andrew Carlotti <andrew.carlotti@arm.com>
4225 * config/aarch64/aarch64.h (TARGET_MEMTAG): Remove armv8.5
4227 * config/aarch64/arm_acle.h: Remove unnecessary armv8.x
4228 dependencies from target pragmas.
4229 * config/aarch64/arm_fp16.h (target): Likewise.
4230 * config/aarch64/arm_neon.h (target): Likewise.
4232 2023-07-19 Andrew Pinski <apinski@marvell.com>
4234 PR tree-optimization/110252
4235 * tree-ssa-phiopt.cc (class auto_flow_sensitive): New class.
4236 (auto_flow_sensitive::auto_flow_sensitive): New constructor.
4237 (auto_flow_sensitive::~auto_flow_sensitive): New deconstructor.
4238 (match_simplify_replacement): Temporarily
4239 remove the flow sensitive info on the two statements that might
4242 2023-07-19 Andrew Pinski <apinski@marvell.com>
4244 * gimple-fold.cc (fosa_unwind): Replace `vrange_storage *`
4245 with flow_sensitive_info_storage.
4246 (follow_outer_ssa_edges): Update how to save off the flow
4248 (maybe_fold_comparisons_from_match_pd): Update restoring
4249 of flow sensitive info.
4250 * tree-ssanames.cc (flow_sensitive_info_storage::save): New method.
4251 (flow_sensitive_info_storage::restore): New method.
4252 (flow_sensitive_info_storage::save_and_clear): New method.
4253 (flow_sensitive_info_storage::clear_storage): New method.
4254 * tree-ssanames.h (class flow_sensitive_info_storage): New class.
4256 2023-07-19 Andrew Pinski <apinski@marvell.com>
4258 PR tree-optimization/110726
4259 * match.pd ((a|b)&(a==b),a|(a==b),(a&b)|(a==b)):
4260 Add checks to make sure the type was one bit precision
4263 2023-07-19 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
4265 * doc/md.texi: Add mask_len_fold_left_plus.
4266 * internal-fn.cc (mask_len_fold_left_direct): Ditto.
4267 (expand_mask_len_fold_left_optab_fn): Ditto.
4268 (direct_mask_len_fold_left_optab_supported_p): Ditto.
4269 * internal-fn.def (MASK_LEN_FOLD_LEFT_PLUS): Ditto.
4270 * optabs.def (OPTAB_D): Ditto.
4272 2023-07-19 Jakub Jelinek <jakub@redhat.com>
4274 * tree-switch-conversion.h (class bit_test_cluster): Fix comment typo.
4276 2023-07-19 Jakub Jelinek <jakub@redhat.com>
4278 PR tree-optimization/110731
4279 * wide-int.cc (wi::divmod_internal): Always unpack dividend and
4280 divisor as UNSIGNED regardless of sgn.
4282 2023-07-19 Lehua Ding <lehua.ding@rivai.ai>
4284 * common/config/riscv/riscv-common.cc (riscv_supported_std_ext): Init.
4285 (standard_extensions_p): Add check.
4286 (riscv_subset_list::add): Just return NULL if it failed before.
4287 (riscv_subset_list::parse_std_ext): Continue parse when find a error
4288 (riscv_subset_list::parse): Just return NULL if it failed before.
4289 * config/riscv/riscv-subset.h (class riscv_subset_list): Add field.
4291 2023-07-19 Jan Beulich <jbeulich@suse.com>
4293 * config/i386/i386-expand.cc (ix86_expand_vector_init_duplicate):
4295 (ix86_expand_vector_extract): Use gen_vec_extract_lo /
4297 (expand_vec_perm_broadcast_1): Use gen_vec_interleave_high /
4298 gen_vec_interleave_low. Rename local variable.
4300 2023-07-19 Jan Beulich <jbeulich@suse.com>
4302 * config/i386/sse.md (vec_dupv2df<mask_name>): Add new AVX512F
4303 alternative. Move AVX512VL part of condition to new "enabled"
4306 2023-07-19 liuhongt <hongtao.liu@intel.com>
4309 * config/i386/i386-builtins.cc
4310 (ix86_register_float16_builtin_type): Remove TARGET_SSE2.
4311 (ix86_register_bf16_builtin_type): Ditto.
4312 * config/i386/i386-c.cc (ix86_target_macros): When TARGET_SSE2
4313 isn't available, undef the macros which are used to check the
4314 backend support of the _Float16/__bf16 types when building
4315 libstdc++ and libgcc.
4316 * config/i386/i386.cc (construct_container): Issue errors for
4317 HFmode/BFmode when TARGET_SSE2 is not available.
4318 (function_value_32): Ditto.
4319 (ix86_scalar_mode_supported_p): Remove TARGET_SSE2 for HFmode/BFmode.
4320 (ix86_libgcc_floating_mode_supported_p): Ditto.
4321 (ix86_emit_support_tinfos): Adjust codes.
4322 (ix86_invalid_conversion): Return diagnostic message string
4323 when there's conversion from/to BF/HFmode w/o TARGET_SSE2.
4324 (ix86_invalid_unary_op): New function.
4325 (ix86_invalid_binary_op): Ditto.
4326 (TARGET_INVALID_UNARY_OP): Define.
4327 (TARGET_INVALID_BINARY_OP): Define.
4328 * config/i386/immintrin.h [__SSE2__]: Remove for fp16/bf16
4329 related instrinsics header files.
4330 * config/i386/i386.h (VALID_SSE2_TYPE_MODE): New macro.
4332 2023-07-18 Uros Bizjak <ubizjak@gmail.com>
4334 * dwarf2asm.cc: Change FALSE to false.
4335 * dwarf2cfi.cc (execute_dwarf2_frame): Change return type to void.
4336 * dwarf2out.cc (matches_main_base): Change return type from
4337 int to bool. Change "last_match" variable to bool.
4338 (dump_struct_debug): Change return type from int to bool.
4339 Change "matches" and "result" function arguments to bool.
4340 (is_pseudo_reg): Change return type from int to bool.
4341 (is_tagged_type): Ditto.
4342 (same_loc_p): Ditto.
4343 (same_dw_val_p): Change return type from int to bool and adjust
4344 function body accordingly.
4345 (same_attr_p): Ditto.
4346 (same_die_p): Ditto.
4347 (is_type_die): Ditto.
4348 (is_declaration_die): Ditto.
4349 (should_move_die_to_comdat): Ditto.
4350 (is_base_type): Ditto.
4351 (is_based_loc): Ditto.
4352 (local_scope_p): Ditto.
4353 (class_scope_p): Ditto.
4354 (class_or_namespace_scope_p): Ditto.
4355 (is_tagged_type): Ditto.
4356 (is_rust): Use void argument.
4357 (is_nested_in_subprogram): Change return type from int to bool.
4358 (contains_subprogram_definition): Ditto.
4359 (gen_struct_or_union_type_die): Change "nested", "complete"
4360 and "ns_decl" variables to bool.
4361 (is_naming_typedef_decl): Change FALSE to false.
4363 2023-07-18 Jan Hubicka <jh@suse.cz>
4365 * tree-ssa-loop-ch.cc (edge_range_query): Take loop argument; be ready
4366 for queries not in headers.
4367 (static_loop_exit): Add basic blck parameter; update use of
4369 (should_duplicate_loop_header_p): Add ranger and static_exits
4370 parameter. Do not account statements that will be optimized
4371 out after duplicaiton in overall size. Add ranger query to
4373 (update_profile_after_ch): Take static_exits has set instead of
4374 single eliminated_edge.
4375 (ch_base::copy_headers): Do all analysis in the first pass;
4376 remember invariant_exits and static_exits.
4378 2023-07-18 Jason Merrill <jason@redhat.com>
4380 * fold-const.cc (native_interpret_aggregate): Skip empty fields.
4382 2023-07-18 Gaius Mulley <gaiusmod2@gmail.com>
4384 * doc/gm2.texi (Semantic checking): Change example testwithptr
4387 2023-07-18 Richard Biener <rguenther@suse.de>
4389 PR middle-end/105715
4390 * gimple-isel.cc (gimple_expand_vec_exprs): Merge into...
4391 (pass_gimple_isel::execute): ... this. Duplicate
4392 comparison defs of COND_EXPRs.
4394 2023-07-18 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4396 * config/riscv/riscv-selftests.cc (run_poly_int_selftests): Add more selftests.
4397 * config/riscv/riscv.cc (riscv_legitimize_poly_move): Dynamic adjust size of VLA vectors.
4398 (riscv_convert_vector_bits): Ditto.
4400 2023-07-18 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4402 * config/riscv/autovec.md (vec_shl_insert_<mode>): New patterns.
4403 * config/riscv/riscv-v.cc (shuffle_compress_patterns): Fix bugs.
4405 2023-07-18 Juergen Christ <jchrist@linux.ibm.com>
4407 * config/s390/vx-builtins.md: New vsel pattern.
4409 2023-07-18 liuhongt <hongtao.liu@intel.com>
4412 * config/i386/sse.md (<mask_codefor>one_cmpl<mode>2<mask_name>):
4413 Remove # from assemble output.
4415 2023-07-18 liuhongt <hongtao.liu@intel.com>
4418 * config/i386/sync.md (cmpccxadd_<mode>): Adjust the pattern
4419 to explicitly set FLAGS_REG like *cmp<mode>_1, also add extra
4420 3 define_peephole2 after the pattern.
4422 2023-07-18 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
4424 * rtl-ssa/internals.inl: Fix when mode1 and mode2 are not ordred.
4426 2023-07-18 Pan Li <pan2.li@intel.com>
4427 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4429 * config/riscv/riscv.cc (struct machine_function): Add new field.
4430 (riscv_static_frm_mode_p): New function.
4431 (riscv_emit_frm_mode_set): New function for emit FRM.
4432 (riscv_emit_mode_set): Extract function for FRM.
4433 (riscv_mode_needed): Fix the TODO.
4434 (riscv_mode_entry): Initial dynamic frm RTL.
4435 (riscv_mode_exit): Return DYN_EXIT.
4436 * config/riscv/riscv.md: Add rdfrm.
4437 * config/riscv/vector-iterators.md (unspecv): Add DYN_EXIT unspecv.
4438 * config/riscv/vector.md (frm_modee): Add new mode dyn_exit.
4440 (fsrmsi_backup): New pattern for swap.
4441 (fsrmsi_restore): New pattern for restore.
4442 (fsrmsi_restore_exit): New pattern for restore exit.
4443 (frrmsi): New pattern for backup.
4445 2023-07-17 Arsen Arsenović <arsen@aarsen.me>
4447 * doc/extend.texi: Add @cindex on __auto_type.
4449 2023-07-17 Uros Bizjak <ubizjak@gmail.com>
4451 * combine-stack-adj.cc (stack_memref_p): Change return type from
4452 int to bool and adjust function body accordingly.
4453 (rest_of_handle_stack_adjustments): Change return type to void.
4455 2023-07-17 Uros Bizjak <ubizjak@gmail.com>
4457 * combine.cc (struct reg_stat_type): Change last_set_invalid to bool.
4458 (cant_combine_insn_p): Change return type from int to bool and adjust
4459 function body accordingly.
4460 (can_combine_p): Ditto.
4461 (combinable_i3pat): Ditto. Change "i1_not_in_src" and "i0_not_in_src"
4462 function arguments from int to bool.
4463 (contains_muldiv): Change return type from int to bool and adjust
4464 function body accordingly.
4465 (try_combine): Ditto. Change "new_direct_jump" pointer function
4466 argument from int to bool. Change "substed_i2", "substed_i1",
4467 "substed_i0", "added_sets_0", "added_sets_1", "added_sets_2",
4468 "i2dest_in_i2src", "i1dest_in_i1src", "i2dest_in_i1src",
4469 "i0dest_in_i0src", "i1dest_in_i0src", "i2dest_in_i0src",
4470 "i2dest_killed", "i1dest_killed", "i0dest_killed", "i1_feeds_i2_n",
4471 "i0_feeds_i2_n", "i0_feeds_i1_n", "i3_subst_into_i2", "have_mult",
4472 "swap_i2i3", "split_i2i3" and "changed_i3_dest" variables
4474 (subst): Change "in_dest", "in_cond" and "unique_copy" function
4475 arguments from int to bool.
4476 (combine_simplify_rtx): Change "in_dest" and "in_cond" function
4477 arguments from int to bool.
4478 (make_extraction): Change "unsignedp", "in_dest" and "in_compare"
4479 function argument from int to bool.
4480 (force_int_to_mode): Change "just_select" function argument
4481 from int to bool. Change "next_select" variable to bool.
4482 (rtx_equal_for_field_assignment_p): Change return type from
4483 int to bool and adjust function body accordingly.
4484 (merge_outer_ops): Ditto. Change "pcomp_p" pointer function
4485 argument from int to bool.
4486 (get_last_value_validate): Change return type from int to bool
4487 and adjust function body accordingly.
4488 (reg_dead_at_p): Ditto.
4489 (reg_bitfield_target_p): Ditto.
4490 (combine_instructions): Ditto. Change "new_direct_jump"
4492 (can_combine_p): Change return type from int to bool
4493 and adjust function body accordingly.
4494 (likely_spilled_retval_p): Ditto.
4495 (can_change_dest_mode): Change "added_sets" function argument
4497 (find_split_point): Change "unsignedp" variable to bool.
4498 (simplify_if_then_else): Change "comparison_p" and "swapped"
4500 (simplify_set): Change "other_changed" variable to bool.
4501 (expand_compound_operation): Change "unsignedp" variable to bool.
4502 (force_to_mode): Change "just_select" function argument
4503 from int to bool. Change "next_select" variable to bool.
4504 (extended_count): Change "unsignedp" function argument to bool.
4505 (simplify_shift_const_1): Change "complement_p" variable to bool.
4506 (simplify_comparison): Change "changed" variable to bool.
4507 (rest_of_handle_combine): Change return type to void.
4509 2023-07-17 Andre Vieira <andre.simoesdiasvieira@arm.com>
4512 * Makefile.in (INTERNAL_FN_H): Add insn-opinit.h.
4514 2023-07-17 Senthil Kumar Selvaraj <saaadhu@gcc.gnu.org>
4516 * ira.cc (setup_reg_class_relations): Continue
4517 if regclass cl3 is hard_reg_set_empty_p.
4519 2023-07-17 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4521 * config/riscv/riscv.cc (riscv_option_override): Add sorry check.
4523 2023-07-17 Martin Jambor <mjambor@suse.cz>
4525 * tree-ssa-loop-ivcanon.cc (try_peel_loop): Remove unused variable
4528 2023-07-17 Aldy Hernandez <aldyh@redhat.com>
4530 * tree-ssa-ccp.cc (ccp_finalize): Export value/mask known bits.
4532 2023-07-17 Lehua Ding <lehua.ding@rivai.ai>
4535 * common/config/riscv/riscv-common.cc (riscv_subset_list::handle_implied_ext):
4536 recur add all implied extensions.
4537 (riscv_subset_list::check_implied_ext): Add new method.
4538 (riscv_subset_list::parse): Call checker check_implied_ext.
4539 * config/riscv/riscv-subset.h: Add new method.
4541 2023-07-17 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4543 * config/riscv/autovec.md (reduc_plus_scal_<mode>): New pattern.
4544 (reduc_smax_scal_<mode>): Ditto.
4545 (reduc_umax_scal_<mode>): Ditto.
4546 (reduc_smin_scal_<mode>): Ditto.
4547 (reduc_umin_scal_<mode>): Ditto.
4548 (reduc_and_scal_<mode>): Ditto.
4549 (reduc_ior_scal_<mode>): Ditto.
4550 (reduc_xor_scal_<mode>): Ditto.
4551 * config/riscv/riscv-protos.h (enum insn_type): Add reduction.
4552 (expand_reduction): New function.
4553 * config/riscv/riscv-v.cc (emit_vlmax_reduction_insn): Ditto.
4554 (emit_vlmax_fp_reduction_insn): Ditto.
4555 (get_m1_mode): Ditto.
4556 (expand_cond_len_binop): Fix name.
4557 (expand_reduction): New function
4558 * config/riscv/riscv-vsetvl.cc (gen_vsetvl_pat): Fix VSETVL BUG.
4559 (validate_change_or_fail): New function.
4560 (change_insn): Fix VSETVL BUG.
4561 (change_vsetvl_insn): Ditto.
4562 (pass_vsetvl::backward_demand_fusion): Ditto.
4563 (pass_vsetvl::df_post_optimization): Ditto.
4565 2023-07-17 Aldy Hernandez <aldyh@redhat.com>
4567 * ipa-prop.cc (ipcp_update_bits): Export value/mask known bits.
4569 2023-07-17 Christoph Müllner <christoph.muellner@vrull.eu>
4571 * config/riscv/riscv.cc (riscv_regno_ok_for_index_p):
4572 Remove parameter name from declaration of unused parameter.
4574 2023-07-17 Kewen Lin <linkw@linux.ibm.com>
4576 PR tree-optimization/110652
4577 * tree-vect-stmts.cc (vectorizable_load): Initialize new_temp as
4580 2023-07-17 Richard Biener <rguenther@suse.de>
4582 PR tree-optimization/110669
4583 * tree-scalar-evolution.cc (analyze_and_compute_bitop_with_inv_effect):
4584 Check we matched a header PHI.
4586 2023-07-17 Aldy Hernandez <aldyh@redhat.com>
4588 * tree-ssanames.cc (set_bitmask): New.
4589 * tree-ssanames.h (set_bitmask): New.
4591 2023-07-17 Aldy Hernandez <aldyh@redhat.com>
4593 * value-range.cc (irange_bitmask::verify_mask): Mask need not be
4595 * value-range.h (irange_bitmask::union_): Normalize beforehand.
4596 (irange_bitmask::intersect): Same.
4598 2023-07-17 Andrew Pinski <apinski@marvell.com>
4600 PR tree-optimization/95923
4601 * match.pd ((a|b)&(a==b),a|(a==b),(a&b)|(a==b)): New transformation.
4603 2023-07-17 Roger Sayle <roger@nextmovesoftware.com>
4605 * tree-if-conv.cc (predicate_scalar_phi): Make the arguments
4606 to the std::sort comparison lambda function const.
4608 2023-07-17 Andrew Pinski <apinski@marvell.com>
4610 PR tree-optimization/110666
4611 * match.pd (A NEEQ (A NEEQ CST)): Fix Outer EQ case.
4613 2023-07-17 Mo, Zewei <zewei.mo@intel.com>
4615 * common/config/i386/cpuinfo.h (get_intel_cpu): Handle Lunar Lake,
4616 Arrow Lake and Arrow Lake S.
4617 * common/config/i386/i386-common.cc:
4618 (processor_name): Add arrowlake.
4619 (processor_alias_table): Add arrow lake, arrow lake s and lunar
4621 * common/config/i386/i386-cpuinfo.h (enum processor_subtypes):
4622 Add INTEL_COREI7_ARROWLAKE and INTEL_COREI7_ARROWLAKE_S.
4623 * config.gcc: Add -march=arrowlake and -march=arrowlake-s.
4624 * config/i386/driver-i386.cc (host_detect_local_cpu): Handle
4626 * config/i386/i386-c.cc (ix86_target_macros_internal): Add
4628 * config/i386/i386-options.cc (m_ARROWLAKE): New.
4629 (processor_cost_table): Add arrowlake.
4630 * config/i386/i386.h (enum processor_type):
4631 Add PROCESSOR_ARROWLAKE.
4632 * config/i386/x86-tune.def: Add m_ARROWLAKE.
4633 * doc/extend.texi: Add arrowlake and arrowlake-s.
4634 * doc/invoke.texi: Ditto.
4636 2023-07-17 Haochen Jiang <haochen.jiang@intel.com>
4638 * config/i386/sse.md (VI2_AVX2): Delete V32HI since we actually
4639 have the same iterator. Also renaming all the occurence to
4641 (usdot_prod<mode>): New define_expand.
4642 (udot_prod<mode>): Ditto.
4644 2023-07-17 Haochen Jiang <haochen.jiang@intel.com>
4646 * common/config/i386/cpuinfo.h (get_available_features):
4648 * common/config/i386/i386-common.cc (OPTION_MASK_ISA2_SM4_SET,
4649 OPTION_MASK_ISA2_SM4_UNSET): New.
4650 (OPTION_MASK_ISA2_AVX_UNSET): Add SM4.
4651 (ix86_handle_option): Handle -msm4.
4652 * common/config/i386/i386-cpuinfo.h (enum processor_features):
4654 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
4656 * config.gcc: Add sm4intrin.h.
4657 * config/i386/cpuid.h (bit_SM4): New.
4658 * config/i386/i386-builtin.def (BDESC): Add new builtins.
4659 * config/i386/i386-c.cc (ix86_target_macros_internal): Define
4661 * config/i386/i386-isa.def (SM4): Add DEF_PTA(SM4).
4662 * config/i386/i386-options.cc (isa2_opts): Add -msm4.
4663 (ix86_valid_target_attribute_inner_p): Handle sm4.
4664 * config/i386/i386.opt: Add option -msm4.
4665 * config/i386/immintrin.h: Include sm4intrin.h
4666 * config/i386/sse.md (vsm4key4_<mode>): New define insn.
4667 (vsm4rnds4_<mode>): Ditto.
4668 * doc/extend.texi: Document sm4.
4669 * doc/invoke.texi: Document -msm4.
4670 * doc/sourcebuild.texi: Document target sm4.
4671 * config/i386/sm4intrin.h: New file.
4673 2023-07-17 Haochen Jiang <haochen.jiang@intel.com>
4675 * common/config/i386/cpuinfo.h (get_available_features):
4677 * common/config/i386/i386-common.cc (OPTION_MASK_ISA2_SHA512_SET,
4678 OPTION_MASK_ISA2_SHA512_UNSET): New.
4679 (OPTION_MASK_ISA2_AVX_UNSET): Add SHA512.
4680 (ix86_handle_option): Handle -msha512.
4681 * common/config/i386/i386-cpuinfo.h (enum processor_features):
4683 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
4685 * config.gcc: Add sha512intrin.h.
4686 * config/i386/cpuid.h (bit_SHA512): New.
4687 * config/i386/i386-builtin-types.def:
4688 Add DEF_FUNCTION_TYPE (V4DI, V4DI, V4DI, V2DI).
4689 * config/i386/i386-builtin.def (BDESC): Add new builtins.
4690 * config/i386/i386-c.cc (ix86_target_macros_internal): Define
4692 * config/i386/i386-expand.cc (ix86_expand_args_builtin): Handle
4693 V4DI_FTYPE_V4DI_V4DI_V2DI and V4DI_FTYPE_V4DI_V2DI.
4694 * config/i386/i386-isa.def (SHA512): Add DEF_PTA(SHA512).
4695 * config/i386/i386-options.cc (isa2_opts): Add -msha512.
4696 (ix86_valid_target_attribute_inner_p): Handle sha512.
4697 * config/i386/i386.opt: Add option -msha512.
4698 * config/i386/immintrin.h: Include sha512intrin.h.
4699 * config/i386/sse.md (vsha512msg1): New define insn.
4700 (vsha512msg2): Ditto.
4701 (vsha512rnds2): Ditto.
4702 * doc/extend.texi: Document sha512.
4703 * doc/invoke.texi: Document -msha512.
4704 * doc/sourcebuild.texi: Document target sha512.
4705 * config/i386/sha512intrin.h: New file.
4707 2023-07-17 Haochen Jiang <haochen.jiang@intel.com>
4709 * common/config/i386/cpuinfo.h (get_available_features):
4711 * common/config/i386/i386-common.cc (OPTION_MASK_ISA2_SM3_SET,
4712 OPTION_MASK_ISA2_SM3_UNSET): New.
4713 (OPTION_MASK_ISA2_AVX_UNSET): Add SM3.
4714 (ix86_handle_option): Handle -msm3.
4715 * common/config/i386/i386-cpuinfo.h (enum processor_features):
4717 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
4719 * config.gcc: Add sm3intrin.h
4720 * config/i386/cpuid.h (bit_SM3): New.
4721 * config/i386/i386-builtin-types.def:
4722 Add DEF_FUNCTION_TYPE (V4SI, V4SI, V4SI, V4SI, INT).
4723 * config/i386/i386-builtin.def (BDESC): Add new builtins.
4724 * config/i386/i386-c.cc (ix86_target_macros_internal): Define
4726 * config/i386/i386-expand.cc (ix86_expand_args_builtin): Handle
4727 V4SI_FTYPE_V4SI_V4SI_V4SI_INT.
4728 * config/i386/i386-isa.def (SM3): Add DEF_PTA(SM3).
4729 * config/i386/i386-options.cc (isa2_opts): Add -msm3.
4730 (ix86_valid_target_attribute_inner_p): Handle sm3.
4731 * config/i386/i386.opt: Add option -msm3.
4732 * config/i386/immintrin.h: Include sm3intrin.h.
4733 * config/i386/sse.md (vsm3msg1): New define insn.
4736 * doc/extend.texi: Document sm3.
4737 * doc/invoke.texi: Document -msm3.
4738 * doc/sourcebuild.texi: Document target sm3.
4739 * config/i386/sm3intrin.h: New file.
4741 2023-07-17 Kong Lingling <lingling.kong@intel.com>
4742 Haochen Jiang <haochen.jiang@intel.com>
4744 * common/config/i386/cpuinfo.h (get_available_features): Detect
4746 * common/config/i386/i386-common.cc
4747 (OPTION_MASK_ISA2_AVXVNNIINT16_SET): New.
4748 (OPTION_MASK_ISA2_AVXVNNIINT16_UNSET): Ditto.
4749 (ix86_handle_option): Handle -mavxvnniint16.
4750 * common/config/i386/i386-cpuinfo.h (enum processor_features):
4751 Add FEATURE_AVXVNNIINT16.
4752 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
4754 * config.gcc: Add avxvnniint16.h.
4755 * config/i386/avxvnniint16intrin.h: New file.
4756 * config/i386/cpuid.h (bit_AVXVNNIINT16): New.
4757 * config/i386/i386-builtin.def: Add new builtins.
4758 * config/i386/i386-c.cc (ix86_target_macros_internal): Define
4760 * config/i386/i386-options.cc (isa2_opts): Add -mavxvnniint16.
4761 (ix86_valid_target_attribute_inner_p): Handle avxvnniint16intrin.h.
4762 * config/i386/i386-isa.def: Add DEF_PTA(AVXVNNIINT16).
4763 * config/i386/i386.opt: Add option -mavxvnniint16.
4764 * config/i386/immintrin.h: Include avxvnniint16.h.
4765 * config/i386/sse.md
4766 (vpdp<vpdpwprodtype>_<mode>): New define_insn.
4767 * doc/extend.texi: Document avxvnniint16.
4768 * doc/invoke.texi: Document -mavxvnniint16.
4769 * doc/sourcebuild.texi: Document target avxvnniint16.
4771 2023-07-16 Jan Hubicka <jh@suse.cz>
4773 PR middle-end/110649
4774 * tree-vect-loop.cc (scale_profile_for_vect_loop): Rewrite.
4775 (vect_transform_loop): Move scale_profile_for_vect_loop after
4776 upper bound updates.
4778 2023-07-16 Jan Hubicka <jh@suse.cz>
4780 PR tree-optimization/110649
4781 * tree-vect-loop.cc (optimize_mask_stores): Set correctly
4782 probability of the if-then-else construct.
4784 2023-07-16 Jan Hubicka <jh@suse.cz>
4786 PR middle-end/110649
4787 * tree-ssa-loop-ivcanon.cc (try_peel_loop): Avoid double profile update.
4789 2023-07-15 Andrew Pinski <apinski@marvell.com>
4791 * doc/contrib.texi: Update my entry.
4793 2023-07-15 John David Anglin <danglin@gcc.gnu.org>
4795 * config/pa/pa.md: Define constants R1_REGNUM, R19_REGNUM and
4797 (tgd_load): Restrict to !TARGET_64BIT. Use register constants.
4798 (tld_load): Likewise.
4799 (tgd_load_pic): Change to expander.
4800 (tld_load_pic, tld_offset_load, tp_load): Likewise.
4801 (tie_load_pic, tle_load): Likewise.
4802 (tgd_load_picsi, tgd_load_picdi): New.
4803 (tld_load_picsi, tld_load_picdi): New.
4804 (tld_offset_load<P:mode>): New.
4805 (tp_load<P:mode>): New.
4806 (tie_load_picsi, tie_load_picdi): New.
4807 (tle_load<P:mode>): New.
4809 2023-07-14 Christophe Lyon <christophe.lyon@linaro.org>
4811 * config/arm/arm-mve-builtins-base.cc (vcmlaq, vcmlaq_rot90)
4812 (vcmlaq_rot180, vcmlaq_rot270): New.
4813 * config/arm/arm-mve-builtins-base.def (vcmlaq, vcmlaq_rot90)
4814 (vcmlaq_rot180, vcmlaq_rot270): New.
4815 * config/arm/arm-mve-builtins-base.h: (vcmlaq, vcmlaq_rot90)
4816 (vcmlaq_rot180, vcmlaq_rot270): New.
4817 * config/arm/arm-mve-builtins.cc
4818 (function_instance::has_inactive_argument): Handle vcmlaq,
4819 vcmlaq_rot90, vcmlaq_rot180, vcmlaq_rot270.
4820 * config/arm/arm_mve.h (vcmlaq): Delete.
4821 (vcmlaq_rot180): Delete.
4822 (vcmlaq_rot270): Delete.
4823 (vcmlaq_rot90): Delete.
4825 (vcmlaq_rot180_m): Delete.
4826 (vcmlaq_rot270_m): Delete.
4827 (vcmlaq_rot90_m): Delete.
4828 (vcmlaq_f16): Delete.
4829 (vcmlaq_rot180_f16): Delete.
4830 (vcmlaq_rot270_f16): Delete.
4831 (vcmlaq_rot90_f16): Delete.
4832 (vcmlaq_f32): Delete.
4833 (vcmlaq_rot180_f32): Delete.
4834 (vcmlaq_rot270_f32): Delete.
4835 (vcmlaq_rot90_f32): Delete.
4836 (vcmlaq_m_f32): Delete.
4837 (vcmlaq_m_f16): Delete.
4838 (vcmlaq_rot180_m_f32): Delete.
4839 (vcmlaq_rot180_m_f16): Delete.
4840 (vcmlaq_rot270_m_f32): Delete.
4841 (vcmlaq_rot270_m_f16): Delete.
4842 (vcmlaq_rot90_m_f32): Delete.
4843 (vcmlaq_rot90_m_f16): Delete.
4844 (__arm_vcmlaq_f16): Delete.
4845 (__arm_vcmlaq_rot180_f16): Delete.
4846 (__arm_vcmlaq_rot270_f16): Delete.
4847 (__arm_vcmlaq_rot90_f16): Delete.
4848 (__arm_vcmlaq_f32): Delete.
4849 (__arm_vcmlaq_rot180_f32): Delete.
4850 (__arm_vcmlaq_rot270_f32): Delete.
4851 (__arm_vcmlaq_rot90_f32): Delete.
4852 (__arm_vcmlaq_m_f32): Delete.
4853 (__arm_vcmlaq_m_f16): Delete.
4854 (__arm_vcmlaq_rot180_m_f32): Delete.
4855 (__arm_vcmlaq_rot180_m_f16): Delete.
4856 (__arm_vcmlaq_rot270_m_f32): Delete.
4857 (__arm_vcmlaq_rot270_m_f16): Delete.
4858 (__arm_vcmlaq_rot90_m_f32): Delete.
4859 (__arm_vcmlaq_rot90_m_f16): Delete.
4860 (__arm_vcmlaq): Delete.
4861 (__arm_vcmlaq_rot180): Delete.
4862 (__arm_vcmlaq_rot270): Delete.
4863 (__arm_vcmlaq_rot90): Delete.
4864 (__arm_vcmlaq_m): Delete.
4865 (__arm_vcmlaq_rot180_m): Delete.
4866 (__arm_vcmlaq_rot270_m): Delete.
4867 (__arm_vcmlaq_rot90_m): Delete.
4869 2023-07-14 Christophe Lyon <christophe.lyon@linaro.org>
4871 * config/arm/arm_mve_builtins.def (vcmlaq_rot90_f)
4872 (vcmlaq_rot270_f, vcmlaq_rot180_f, vcmlaq_f): Add "_f" suffix.
4873 * config/arm/iterators.md (MVE_VCMLAQ_M): New.
4874 (mve_insn): Add vcmla.
4875 (rot): Add VCMLAQ_M_F, VCMLAQ_ROT90_M_F, VCMLAQ_ROT180_M_F,
4877 (mve_rot): Add VCMLAQ_M_F, VCMLAQ_ROT90_M_F, VCMLAQ_ROT180_M_F,
4879 * config/arm/mve.md (mve_vcmlaq<mve_rot><mode>): Rename into ...
4880 (@mve_<mve_insn>q<mve_rot>_f<mode>): ... this.
4881 (mve_vcmlaq_m_f<mode>, mve_vcmlaq_rot180_m_f<mode>)
4882 (mve_vcmlaq_rot270_m_f<mode>, mve_vcmlaq_rot90_m_f<mode>): Merge
4884 (@mve_<mve_insn>q<mve_rot>_m_f<mode>): ... this.
4886 2023-07-14 Christophe Lyon <christophe.lyon@linaro.org>
4888 * config/arm/arm-mve-builtins-base.cc (vcmulq, vcmulq_rot90)
4889 (vcmulq_rot180, vcmulq_rot270): New.
4890 * config/arm/arm-mve-builtins-base.def (vcmulq, vcmulq_rot90)
4891 (vcmulq_rot180, vcmulq_rot270): New.
4892 * config/arm/arm-mve-builtins-base.h: (vcmulq, vcmulq_rot90)
4893 (vcmulq_rot180, vcmulq_rot270): New.
4894 * config/arm/arm_mve.h (vcmulq_rot90): Delete.
4895 (vcmulq_rot270): Delete.
4896 (vcmulq_rot180): Delete.
4899 (vcmulq_rot180_m): Delete.
4900 (vcmulq_rot270_m): Delete.
4901 (vcmulq_rot90_m): Delete.
4903 (vcmulq_rot90_x): Delete.
4904 (vcmulq_rot180_x): Delete.
4905 (vcmulq_rot270_x): Delete.
4906 (vcmulq_rot90_f16): Delete.
4907 (vcmulq_rot270_f16): Delete.
4908 (vcmulq_rot180_f16): Delete.
4909 (vcmulq_f16): Delete.
4910 (vcmulq_rot90_f32): Delete.
4911 (vcmulq_rot270_f32): Delete.
4912 (vcmulq_rot180_f32): Delete.
4913 (vcmulq_f32): Delete.
4914 (vcmulq_m_f32): Delete.
4915 (vcmulq_m_f16): Delete.
4916 (vcmulq_rot180_m_f32): Delete.
4917 (vcmulq_rot180_m_f16): Delete.
4918 (vcmulq_rot270_m_f32): Delete.
4919 (vcmulq_rot270_m_f16): Delete.
4920 (vcmulq_rot90_m_f32): Delete.
4921 (vcmulq_rot90_m_f16): Delete.
4922 (vcmulq_x_f16): Delete.
4923 (vcmulq_x_f32): Delete.
4924 (vcmulq_rot90_x_f16): Delete.
4925 (vcmulq_rot90_x_f32): Delete.
4926 (vcmulq_rot180_x_f16): Delete.
4927 (vcmulq_rot180_x_f32): Delete.
4928 (vcmulq_rot270_x_f16): Delete.
4929 (vcmulq_rot270_x_f32): Delete.
4930 (__arm_vcmulq_rot90_f16): Delete.
4931 (__arm_vcmulq_rot270_f16): Delete.
4932 (__arm_vcmulq_rot180_f16): Delete.
4933 (__arm_vcmulq_f16): Delete.
4934 (__arm_vcmulq_rot90_f32): Delete.
4935 (__arm_vcmulq_rot270_f32): Delete.
4936 (__arm_vcmulq_rot180_f32): Delete.
4937 (__arm_vcmulq_f32): Delete.
4938 (__arm_vcmulq_m_f32): Delete.
4939 (__arm_vcmulq_m_f16): Delete.
4940 (__arm_vcmulq_rot180_m_f32): Delete.
4941 (__arm_vcmulq_rot180_m_f16): Delete.
4942 (__arm_vcmulq_rot270_m_f32): Delete.
4943 (__arm_vcmulq_rot270_m_f16): Delete.
4944 (__arm_vcmulq_rot90_m_f32): Delete.
4945 (__arm_vcmulq_rot90_m_f16): Delete.
4946 (__arm_vcmulq_x_f16): Delete.
4947 (__arm_vcmulq_x_f32): Delete.
4948 (__arm_vcmulq_rot90_x_f16): Delete.
4949 (__arm_vcmulq_rot90_x_f32): Delete.
4950 (__arm_vcmulq_rot180_x_f16): Delete.
4951 (__arm_vcmulq_rot180_x_f32): Delete.
4952 (__arm_vcmulq_rot270_x_f16): Delete.
4953 (__arm_vcmulq_rot270_x_f32): Delete.
4954 (__arm_vcmulq_rot90): Delete.
4955 (__arm_vcmulq_rot270): Delete.
4956 (__arm_vcmulq_rot180): Delete.
4957 (__arm_vcmulq): Delete.
4958 (__arm_vcmulq_m): Delete.
4959 (__arm_vcmulq_rot180_m): Delete.
4960 (__arm_vcmulq_rot270_m): Delete.
4961 (__arm_vcmulq_rot90_m): Delete.
4962 (__arm_vcmulq_x): Delete.
4963 (__arm_vcmulq_rot90_x): Delete.
4964 (__arm_vcmulq_rot180_x): Delete.
4965 (__arm_vcmulq_rot270_x): Delete.
4967 2023-07-14 Christophe Lyon <christophe.lyon@linaro.org>
4969 * config/arm/arm_mve_builtins.def (vcmulq_rot90_f)
4970 (vcmulq_rot270_f, vcmulq_rot180_f, vcmulq_f): Add "_f" suffix.
4971 * config/arm/iterators.md (MVE_VCADDQ_VCMULQ)
4972 (MVE_VCADDQ_VCMULQ_M): New.
4973 (mve_insn): Add vcmul.
4974 (rot): Add VCMULQ_M_F, VCMULQ_ROT90_M_F, VCMULQ_ROT180_M_F,
4977 (mve_rot): Add VCMULQ_M_F, VCMULQ_ROT90_M_F, VCMULQ_ROT180_M_F,
4979 * config/arm/mve.md (mve_vcmulq<mve_rot><mode>): Merge into
4980 @mve_<mve_insn>q<mve_rot>_f<mode>.
4981 (mve_vcmulq_m_f<mode>, mve_vcmulq_rot180_m_f<mode>)
4982 (mve_vcmulq_rot270_m_f<mode>, mve_vcmulq_rot90_m_f<mode>): Merge
4983 into @mve_<mve_insn>q<mve_rot>_m_f<mode>.
4985 2023-07-14 Christophe Lyon <christophe.lyon@linaro.org>
4987 * config/arm/arm-mve-builtins-base.cc (vcaddq_rot90)
4988 (vcaddq_rot270, vhcaddq_rot90, vhcaddq_rot270): New.
4989 * config/arm/arm-mve-builtins-base.def (vcaddq_rot90)
4990 (vcaddq_rot270, vhcaddq_rot90, vhcaddq_rot270): New.
4991 * config/arm/arm-mve-builtins-base.h: (vcaddq_rot90)
4992 (vcaddq_rot270, vhcaddq_rot90, vhcaddq_rot270): New.
4993 * config/arm/arm-mve-builtins-functions.h (class
4994 unspec_mve_function_exact_insn_rot): New.
4995 * config/arm/arm_mve.h (vcaddq_rot90): Delete.
4996 (vcaddq_rot270): Delete.
4997 (vhcaddq_rot90): Delete.
4998 (vhcaddq_rot270): Delete.
4999 (vcaddq_rot270_m): Delete.
5000 (vcaddq_rot90_m): Delete.
5001 (vhcaddq_rot270_m): Delete.
5002 (vhcaddq_rot90_m): Delete.
5003 (vcaddq_rot90_x): Delete.
5004 (vcaddq_rot270_x): Delete.
5005 (vhcaddq_rot90_x): Delete.
5006 (vhcaddq_rot270_x): Delete.
5007 (vcaddq_rot90_u8): Delete.
5008 (vcaddq_rot270_u8): Delete.
5009 (vhcaddq_rot90_s8): Delete.
5010 (vhcaddq_rot270_s8): Delete.
5011 (vcaddq_rot90_s8): Delete.
5012 (vcaddq_rot270_s8): Delete.
5013 (vcaddq_rot90_u16): Delete.
5014 (vcaddq_rot270_u16): Delete.
5015 (vhcaddq_rot90_s16): Delete.
5016 (vhcaddq_rot270_s16): Delete.
5017 (vcaddq_rot90_s16): Delete.
5018 (vcaddq_rot270_s16): Delete.
5019 (vcaddq_rot90_u32): Delete.
5020 (vcaddq_rot270_u32): Delete.
5021 (vhcaddq_rot90_s32): Delete.
5022 (vhcaddq_rot270_s32): Delete.
5023 (vcaddq_rot90_s32): Delete.
5024 (vcaddq_rot270_s32): Delete.
5025 (vcaddq_rot90_f16): Delete.
5026 (vcaddq_rot270_f16): Delete.
5027 (vcaddq_rot90_f32): Delete.
5028 (vcaddq_rot270_f32): Delete.
5029 (vcaddq_rot270_m_s8): Delete.
5030 (vcaddq_rot270_m_s32): Delete.
5031 (vcaddq_rot270_m_s16): Delete.
5032 (vcaddq_rot270_m_u8): Delete.
5033 (vcaddq_rot270_m_u32): Delete.
5034 (vcaddq_rot270_m_u16): Delete.
5035 (vcaddq_rot90_m_s8): Delete.
5036 (vcaddq_rot90_m_s32): Delete.
5037 (vcaddq_rot90_m_s16): Delete.
5038 (vcaddq_rot90_m_u8): Delete.
5039 (vcaddq_rot90_m_u32): Delete.
5040 (vcaddq_rot90_m_u16): Delete.
5041 (vhcaddq_rot270_m_s8): Delete.
5042 (vhcaddq_rot270_m_s32): Delete.
5043 (vhcaddq_rot270_m_s16): Delete.
5044 (vhcaddq_rot90_m_s8): Delete.
5045 (vhcaddq_rot90_m_s32): Delete.
5046 (vhcaddq_rot90_m_s16): Delete.
5047 (vcaddq_rot270_m_f32): Delete.
5048 (vcaddq_rot270_m_f16): Delete.
5049 (vcaddq_rot90_m_f32): Delete.
5050 (vcaddq_rot90_m_f16): Delete.
5051 (vcaddq_rot90_x_s8): Delete.
5052 (vcaddq_rot90_x_s16): Delete.
5053 (vcaddq_rot90_x_s32): Delete.
5054 (vcaddq_rot90_x_u8): Delete.
5055 (vcaddq_rot90_x_u16): Delete.
5056 (vcaddq_rot90_x_u32): Delete.
5057 (vcaddq_rot270_x_s8): Delete.
5058 (vcaddq_rot270_x_s16): Delete.
5059 (vcaddq_rot270_x_s32): Delete.
5060 (vcaddq_rot270_x_u8): Delete.
5061 (vcaddq_rot270_x_u16): Delete.
5062 (vcaddq_rot270_x_u32): Delete.
5063 (vhcaddq_rot90_x_s8): Delete.
5064 (vhcaddq_rot90_x_s16): Delete.
5065 (vhcaddq_rot90_x_s32): Delete.
5066 (vhcaddq_rot270_x_s8): Delete.
5067 (vhcaddq_rot270_x_s16): Delete.
5068 (vhcaddq_rot270_x_s32): Delete.
5069 (vcaddq_rot90_x_f16): Delete.
5070 (vcaddq_rot90_x_f32): Delete.
5071 (vcaddq_rot270_x_f16): Delete.
5072 (vcaddq_rot270_x_f32): Delete.
5073 (__arm_vcaddq_rot90_u8): Delete.
5074 (__arm_vcaddq_rot270_u8): Delete.
5075 (__arm_vhcaddq_rot90_s8): Delete.
5076 (__arm_vhcaddq_rot270_s8): Delete.
5077 (__arm_vcaddq_rot90_s8): Delete.
5078 (__arm_vcaddq_rot270_s8): Delete.
5079 (__arm_vcaddq_rot90_u16): Delete.
5080 (__arm_vcaddq_rot270_u16): Delete.
5081 (__arm_vhcaddq_rot90_s16): Delete.
5082 (__arm_vhcaddq_rot270_s16): Delete.
5083 (__arm_vcaddq_rot90_s16): Delete.
5084 (__arm_vcaddq_rot270_s16): Delete.
5085 (__arm_vcaddq_rot90_u32): Delete.
5086 (__arm_vcaddq_rot270_u32): Delete.
5087 (__arm_vhcaddq_rot90_s32): Delete.
5088 (__arm_vhcaddq_rot270_s32): Delete.
5089 (__arm_vcaddq_rot90_s32): Delete.
5090 (__arm_vcaddq_rot270_s32): Delete.
5091 (__arm_vcaddq_rot270_m_s8): Delete.
5092 (__arm_vcaddq_rot270_m_s32): Delete.
5093 (__arm_vcaddq_rot270_m_s16): Delete.
5094 (__arm_vcaddq_rot270_m_u8): Delete.
5095 (__arm_vcaddq_rot270_m_u32): Delete.
5096 (__arm_vcaddq_rot270_m_u16): Delete.
5097 (__arm_vcaddq_rot90_m_s8): Delete.
5098 (__arm_vcaddq_rot90_m_s32): Delete.
5099 (__arm_vcaddq_rot90_m_s16): Delete.
5100 (__arm_vcaddq_rot90_m_u8): Delete.
5101 (__arm_vcaddq_rot90_m_u32): Delete.
5102 (__arm_vcaddq_rot90_m_u16): Delete.
5103 (__arm_vhcaddq_rot270_m_s8): Delete.
5104 (__arm_vhcaddq_rot270_m_s32): Delete.
5105 (__arm_vhcaddq_rot270_m_s16): Delete.
5106 (__arm_vhcaddq_rot90_m_s8): Delete.
5107 (__arm_vhcaddq_rot90_m_s32): Delete.
5108 (__arm_vhcaddq_rot90_m_s16): Delete.
5109 (__arm_vcaddq_rot90_x_s8): Delete.
5110 (__arm_vcaddq_rot90_x_s16): Delete.
5111 (__arm_vcaddq_rot90_x_s32): Delete.
5112 (__arm_vcaddq_rot90_x_u8): Delete.
5113 (__arm_vcaddq_rot90_x_u16): Delete.
5114 (__arm_vcaddq_rot90_x_u32): Delete.
5115 (__arm_vcaddq_rot270_x_s8): Delete.
5116 (__arm_vcaddq_rot270_x_s16): Delete.
5117 (__arm_vcaddq_rot270_x_s32): Delete.
5118 (__arm_vcaddq_rot270_x_u8): Delete.
5119 (__arm_vcaddq_rot270_x_u16): Delete.
5120 (__arm_vcaddq_rot270_x_u32): Delete.
5121 (__arm_vhcaddq_rot90_x_s8): Delete.
5122 (__arm_vhcaddq_rot90_x_s16): Delete.
5123 (__arm_vhcaddq_rot90_x_s32): Delete.
5124 (__arm_vhcaddq_rot270_x_s8): Delete.
5125 (__arm_vhcaddq_rot270_x_s16): Delete.
5126 (__arm_vhcaddq_rot270_x_s32): Delete.
5127 (__arm_vcaddq_rot90_f16): Delete.
5128 (__arm_vcaddq_rot270_f16): Delete.
5129 (__arm_vcaddq_rot90_f32): Delete.
5130 (__arm_vcaddq_rot270_f32): Delete.
5131 (__arm_vcaddq_rot270_m_f32): Delete.
5132 (__arm_vcaddq_rot270_m_f16): Delete.
5133 (__arm_vcaddq_rot90_m_f32): Delete.
5134 (__arm_vcaddq_rot90_m_f16): Delete.
5135 (__arm_vcaddq_rot90_x_f16): Delete.
5136 (__arm_vcaddq_rot90_x_f32): Delete.
5137 (__arm_vcaddq_rot270_x_f16): Delete.
5138 (__arm_vcaddq_rot270_x_f32): Delete.
5139 (__arm_vcaddq_rot90): Delete.
5140 (__arm_vcaddq_rot270): Delete.
5141 (__arm_vhcaddq_rot90): Delete.
5142 (__arm_vhcaddq_rot270): Delete.
5143 (__arm_vcaddq_rot270_m): Delete.
5144 (__arm_vcaddq_rot90_m): Delete.
5145 (__arm_vhcaddq_rot270_m): Delete.
5146 (__arm_vhcaddq_rot90_m): Delete.
5147 (__arm_vcaddq_rot90_x): Delete.
5148 (__arm_vcaddq_rot270_x): Delete.
5149 (__arm_vhcaddq_rot90_x): Delete.
5150 (__arm_vhcaddq_rot270_x): Delete.
5152 2023-07-14 Christophe Lyon <christophe.lyon@linaro.org>
5154 * config/arm/arm_mve_builtins.def (vcaddq_rot90_, vcaddq_rot270_)
5155 (vcaddq_rot90_f, vcaddq_rot90_f): Add "_" or "_f" suffix.
5156 * config/arm/iterators.md (mve_insn): Add vcadd, vhcadd.
5157 (isu): Add UNSPEC_VCADD90, UNSPEC_VCADD270, VCADDQ_ROT270_M_U,
5158 VCADDQ_ROT270_M_S, VCADDQ_ROT90_M_U, VCADDQ_ROT90_M_S,
5159 VHCADDQ_ROT90_M_S, VHCADDQ_ROT270_M_S, VHCADDQ_ROT90_S,
5161 (rot): Add VCADDQ_ROT90_M_F, VCADDQ_ROT90_M_S, VCADDQ_ROT90_M_U,
5162 VCADDQ_ROT270_M_F, VCADDQ_ROT270_M_S, VCADDQ_ROT270_M_U,
5163 VHCADDQ_ROT90_S, VHCADDQ_ROT270_S, VHCADDQ_ROT90_M_S,
5165 (mve_rot): Add VCADDQ_ROT90_M_F, VCADDQ_ROT90_M_S,
5166 VCADDQ_ROT90_M_U, VCADDQ_ROT270_M_F, VCADDQ_ROT270_M_S,
5167 VCADDQ_ROT270_M_U, VHCADDQ_ROT90_S, VHCADDQ_ROT270_S,
5168 VHCADDQ_ROT90_M_S, VHCADDQ_ROT270_M_S.
5169 (supf): Add VHCADDQ_ROT90_M_S, VHCADDQ_ROT270_M_S,
5170 VHCADDQ_ROT90_S, VHCADDQ_ROT270_S, UNSPEC_VCADD90,
5172 (VCADDQ_ROT270_M): Delete.
5173 (VCADDQ_M_F VxCADDQ VxCADDQ_M): New.
5174 (VCADDQ_ROT90_M): Delete.
5175 * config/arm/mve.md (mve_vcaddq<mve_rot><mode>)
5176 (mve_vhcaddq_rot270_s<mode>, mve_vhcaddq_rot90_s<mode>): Merge
5178 (@mve_<mve_insn>q<mve_rot>_<supf><mode>): ... this.
5179 (mve_vcaddq<mve_rot><mode>): Rename into ...
5180 (@mve_<mve_insn>q<mve_rot>_f<mode>): ... this
5181 (mve_vcaddq_rot270_m_<supf><mode>)
5182 (mve_vcaddq_rot90_m_<supf><mode>, mve_vhcaddq_rot270_m_s<mode>)
5183 (mve_vhcaddq_rot90_m_s<mode>): Merge into ...
5184 (@mve_<mve_insn>q<mve_rot>_m_<supf><mode>): ... this.
5185 (mve_vcaddq_rot270_m_f<mode>, mve_vcaddq_rot90_m_f<mode>): Merge
5187 (@mve_<mve_insn>q<mve_rot>_m_f<mode>): ... this.
5189 2023-07-14 Roger Sayle <roger@nextmovesoftware.com>
5192 * config/i386/i386.md (*bt<mode>_setcqi): Prefer string form
5193 preparation statement over braces for a single statement.
5194 (*bt<mode>_setncqi): Likewise.
5195 (*bt<mode>_setncqi_2): New define_insn_and_split.
5197 2023-07-14 Roger Sayle <roger@nextmovesoftware.com>
5199 * config/i386/i386-expand.cc (ix86_expand_move): Generalize special
5200 case inserting of 64-bit values into a TImode register, to handle
5201 both DImode and DFmode using either *insvti_lowpart_1
5202 or *isnvti_highpart_1.
5204 2023-07-14 Uros Bizjak <ubizjak@gmail.com>
5207 * fwprop.cc (contains_paradoxical_subreg_p): Move to ...
5208 * rtlanal.cc (contains_paradoxical_subreg_p): ... here.
5209 * rtlanal.h (contains_paradoxical_subreg_p): Add prototype.
5210 * cprop.cc (try_replace_reg): Do not set REG_EQUAL note
5211 when the original source contains a paradoxical subreg.
5213 2023-07-14 Jan Hubicka <jh@suse.cz>
5215 * passes.cc (execute_function_todo): Remove
5216 TODO_rebuild_frequencies
5217 * passes.def: Add rebuild_frequencies pass.
5218 * predict.cc (estimate_bb_frequencies): Drop
5220 (tree_estimate_probability): Update call of
5221 estimate_bb_frequencies.
5222 (rebuild_frequencies): Turn into a pass; verify CFG profile consistency
5223 first and do not rebuild if not necessary.
5224 (class pass_rebuild_frequencies): New.
5225 (make_pass_rebuild_frequencies): New.
5226 * profile-count.h: Add profile_count::very_large_p.
5227 * tree-inline.cc (optimize_inline_calls): Do not return
5228 TODO_rebuild_frequencies
5229 * tree-pass.h (TODO_rebuild_frequencies): Remove.
5230 (make_pass_rebuild_frequencies): Declare.
5232 2023-07-14 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5234 * config/riscv/autovec.md (cond_len_fma<mode>): New pattern.
5235 * config/riscv/riscv-protos.h (enum insn_type): New enum.
5236 (expand_cond_len_ternop): New function.
5237 * config/riscv/riscv-v.cc (emit_nonvlmax_fp_ternary_tu_insn): Ditto.
5238 (expand_cond_len_ternop): Ditto.
5240 2023-07-14 Jose E. Marchesi <jose.marchesi@oracle.com>
5243 * config/bpf/bpf.md: Enable instruction scheduling.
5245 2023-07-14 Tamar Christina <tamar.christina@arm.com>
5247 PR tree-optimization/109154
5248 * tree-if-conv.cc (INCLUDE_ALGORITHM): Include.
5249 (struct bb_predicate): Add no_predicate_stmts.
5250 (set_bb_predicate): Increase predicate count.
5251 (set_bb_predicate_gimplified_stmts): Conditionally initialize
5253 (get_bb_num_predicate_stmts): New.
5254 (init_bb_predicate): Initialzie no_predicate_stmts.
5255 (release_bb_predicate): Cleanup no_predicate_stmts.
5256 (insert_gimplified_predicates): Preserve no_predicate_stmts.
5258 2023-07-14 Tamar Christina <tamar.christina@arm.com>
5260 PR tree-optimization/109154
5261 * tree-if-conv.cc (gen_simplified_condition,
5262 gen_phi_nest_statement): New.
5263 (gen_phi_arg_condition, predicate_scalar_phi): Use it.
5265 2023-07-14 Richard Biener <rguenther@suse.de>
5267 * gimple.h (gimple_phi_arg): New const overload.
5268 (gimple_phi_arg_def): Make gimple arg const.
5269 (gimple_phi_arg_def_from_edge): New inline function.
5270 * tree-phinodes.h (gimple_phi_arg_imm_use_ptr_from_edge):
5272 * tree-ssa-operands.h (PHI_ARG_DEF_FROM_EDGE): Direct to
5273 new inline function.
5274 (PHI_ARG_DEF_PTR_FROM_EDGE): Likewise.
5276 2023-07-14 Monk Chiang <monk.chiang@sifive.com>
5278 * common/config/riscv/riscv-common.cc:
5279 (riscv_implied_info): Add zihintntl item.
5280 (riscv_ext_version_table): Ditto.
5281 (riscv_ext_flag_table): Ditto.
5282 * config/riscv/riscv-opts.h (MASK_ZIHINTNTL): New macro.
5283 (TARGET_ZIHINTNTL): Ditto.
5285 2023-07-14 Die Li <lidie@eswincomputing.com>
5287 * config/riscv/riscv.md: Remove redundant portion in and<mode>3.
5289 2023-07-14 Oleg Endo <olegendo@gcc.gnu.org>
5292 * config/sh/sh.md (peephole2): Handle case where eliminated reg is also
5293 used by the address of the following memory operand.
5295 2023-07-13 Mikael Pettersson <mikpelinux@gmail.com>
5298 * config/pdp11/pdp11.cc (pdp11_expand_epilogue): Also
5299 deallocate alloca-only frame.
5301 2023-07-13 Iain Sandoe <iain@sandoe.co.uk>
5304 * config/darwin.h (DARWIN_PLATFORM_ID): New.
5305 (LINK_COMMAND_A): Use DARWIN_PLATFORM_ID to pass OS, OS version
5306 and SDK data to the static linker.
5308 2023-07-13 Carl Love <cel@us.ibm.com>
5310 * config/rs6000/rs6000-builtins.def (__builtin_set_fpscr_rn): Update
5311 built-in definition return type.
5312 * config/rs6000/rs6000-c.cc (rs6000_target_modify_macros): Add check,
5313 define __SET_FPSCR_RN_RETURNS_FPSCR__ macro.
5314 * config/rs6000/rs6000.md (rs6000_set_fpscr_rn): Add return
5315 argument to return FPSCR fields.
5316 * doc/extend.texi (__builtin_set_fpscr_rn): Update description for
5317 the return value. Add description for
5318 __SET_FPSCR_RN_RETURNS_FPSCR__ macro.
5320 2023-07-13 Uros Bizjak <ubizjak@gmail.com>
5323 * config/alpha/alpha.cc (alpha_emit_set_long_const):
5324 Always use DImode when constructing long const.
5326 2023-07-13 Uros Bizjak <ubizjak@gmail.com>
5328 * haifa-sched.cc: Change TRUE/FALSE to true/false.
5330 * lra-assigns.cc: Ditto.
5331 * lra-constraints.cc: Ditto.
5332 * sel-sched.cc: Ditto.
5334 2023-07-13 Andrew Pinski <apinski@marvell.com>
5336 PR tree-optimization/110293
5337 PR tree-optimization/110539
5338 * match.pd: Expand the `x != (typeof x)(x == 0)`
5339 pattern to handle where the inner and outer comparsions
5340 are either `!=` or `==` and handle other constants
5343 2023-07-13 Vladimir N. Makarov <vmakarov@redhat.com>
5345 PR middle-end/109520
5346 * lra-int.h (lra_insn_recog_data): Add member asm_reloads_num.
5347 (lra_asm_insn_error): New prototype.
5348 * lra.cc: Include rtl_error.h.
5349 (lra_set_insn_recog_data): Initialize asm_reloads_num.
5350 (lra_asm_insn_error): New func whose code is taken from ...
5351 * lra-assigns.cc (lra_split_hard_reg_for): ... here. Use lra_asm_insn_error.
5352 * lra-constraints.cc (curr_insn_transform): Check reloads nummber for asm.
5354 2023-07-13 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
5356 * genmatch.cc (commutative_op): Add COND_LEN_*
5357 * internal-fn.cc (first_commutative_argument): Ditto.
5359 (get_unconditional_internal_fn): Ditto.
5360 (can_interpret_as_conditional_op_p): Ditto.
5361 (internal_fn_len_index): Ditto.
5362 * internal-fn.h (can_interpret_as_conditional_op_p): Ditt.
5363 * tree-ssa-math-opts.cc (convert_mult_to_fma_1): Ditto.
5364 (convert_mult_to_fma): Ditto.
5365 (math_opts_dom_walker::after_dom_children): Ditto.
5367 2023-07-13 Pan Li <pan2.li@intel.com>
5369 * config/riscv/riscv.cc (vxrm_rtx): New static var.
5371 (global_state_unknown_p): Removed.
5372 (riscv_entity_mode_after): Removed.
5373 (asm_insn_p): New function.
5374 (vxrm_unknown_p): New function for fixed-point.
5375 (riscv_vxrm_mode_after): Ditto.
5376 (frm_unknown_dynamic_p): New function for floating-point.
5377 (riscv_frm_mode_after): Ditto.
5378 (riscv_mode_after): Leverage new functions.
5380 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
5382 * tree-vect-stmts.cc (vect_model_load_cost): Remove.
5383 (vectorizable_load): Adjust the cost handling on VMAT_CONTIGUOUS without
5384 calling vect_model_load_cost.
5386 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
5388 * tree-vect-stmts.cc (vect_model_load_cost): Assert this function only
5389 handle memory_access_type VMAT_CONTIGUOUS, remove some
5390 VMAT_CONTIGUOUS_PERMUTE related handlings.
5391 (vectorizable_load): Adjust the cost handling on VMAT_CONTIGUOUS_PERMUTE
5392 without calling vect_model_load_cost.
5394 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
5396 * tree-vect-stmts.cc (vect_model_load_cost): Assert it won't get
5397 VMAT_CONTIGUOUS_REVERSE any more.
5398 (vectorizable_load): Adjust the costing handling on
5399 VMAT_CONTIGUOUS_REVERSE without calling vect_model_load_cost.
5401 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
5403 * tree-vect-stmts.cc (vectorizable_load): Adjust the cost handling on
5404 VMAT_LOAD_STORE_LANES without calling vect_model_load_cost.
5405 (vectorizable_load): Remove VMAT_LOAD_STORE_LANES related handling and
5406 assert it will never get VMAT_LOAD_STORE_LANES.
5408 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
5410 * tree-vect-stmts.cc (vectorizable_load): Adjust the cost handling on
5411 VMAT_GATHER_SCATTER without calling vect_model_load_cost.
5412 (vect_model_load_cost): Adjut the assertion on VMAT_GATHER_SCATTER,
5413 remove VMAT_GATHER_SCATTER related handlings and the related parameter
5416 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
5418 * tree-vect-stmts.cc (vectorizable_load): Adjust the cost handling
5419 on VMAT_ELEMENTWISE and VMAT_STRIDED_SLP without calling
5420 vect_model_load_cost.
5421 (vect_model_load_cost): Assert it won't get VMAT_ELEMENTWISE and
5422 VMAT_STRIDED_SLP any more, and remove their related handlings.
5424 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
5426 * tree-vect-stmts.cc (hoist_defs_of_uses): Add one argument HOIST_P.
5427 (vectorizable_load): Adjust the handling on VMAT_INVARIANT to respect
5428 hoisting decision and without calling vect_model_load_cost.
5429 (vect_model_load_cost): Assert it won't get VMAT_INVARIANT any more
5430 and remove VMAT_INVARIANT related handlings.
5432 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
5434 * tree-vect-stmts.cc (vect_build_gather_load_calls): Add the handlings
5435 on costing with one extra argument cost_vec.
5436 (vectorizable_load): Adjust the call to vect_build_gather_load_calls.
5437 (vect_model_load_cost): Assert it won't get VMAT_GATHER_SCATTER with
5438 gs_info.decl set any more.
5440 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
5442 * tree-vect-stmts.cc (vectorizable_load): Move and duplicate the call
5443 to vect_model_load_cost down to some different transform paths
5444 according to the handlings of different vect_memory_access_types.
5446 2023-07-13 Kewen Lin <linkw@linux.ibm.com>
5448 * tree.h (wi::from_mpz): Hide from GENERATOR_FILE.
5450 2023-07-13 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
5452 * config/riscv/autovec.md
5453 (len_mask_gather_load<VNX1_QHSD:mode><VNX1_QHSDI:mode>): New pattern.
5454 (len_mask_gather_load<VNX2_QHSD:mode><VNX2_QHSDI:mode>): Ditto.
5455 (len_mask_gather_load<VNX4_QHSD:mode><VNX4_QHSDI:mode>): Ditto.
5456 (len_mask_gather_load<VNX8_QHSD:mode><VNX8_QHSDI:mode>): Ditto.
5457 (len_mask_gather_load<VNX16_QHSD:mode><VNX16_QHSDI:mode>): Ditto.
5458 (len_mask_gather_load<VNX32_QHS:mode><VNX32_QHSI:mode>): Ditto.
5459 (len_mask_gather_load<VNX64_QH:mode><VNX64_QHI:mode>): Ditto.
5460 (len_mask_gather_load<mode><mode>): Ditto.
5461 (len_mask_scatter_store<VNX1_QHSD:mode><VNX1_QHSDI:mode>): Ditto.
5462 (len_mask_scatter_store<VNX2_QHSD:mode><VNX2_QHSDI:mode>): Ditto.
5463 (len_mask_scatter_store<VNX4_QHSD:mode><VNX4_QHSDI:mode>): Ditto.
5464 (len_mask_scatter_store<VNX8_QHSD:mode><VNX8_QHSDI:mode>): Ditto.
5465 (len_mask_scatter_store<VNX16_QHSD:mode><VNX16_QHSDI:mode>): Ditto.
5466 (len_mask_scatter_store<VNX32_QHS:mode><VNX32_QHSI:mode>): Ditto.
5467 (len_mask_scatter_store<VNX64_QH:mode><VNX64_QHI:mode>): Ditto.
5468 (len_mask_scatter_store<mode><mode>): Ditto.
5469 * config/riscv/predicates.md (const_1_operand): New predicate.
5470 (vector_gs_scale_operand_16): Ditto.
5471 (vector_gs_scale_operand_32): Ditto.
5472 (vector_gs_scale_operand_64): Ditto.
5473 (vector_gs_extension_operand): Ditto.
5474 (vector_gs_scale_operand_16_rv32): Ditto.
5475 (vector_gs_scale_operand_32_rv32): Ditto.
5476 * config/riscv/riscv-protos.h (enum insn_type): Add gather/scatter.
5477 (expand_gather_scatter): New function.
5478 * config/riscv/riscv-v.cc (gen_const_vector_dup): Add gather/scatter.
5479 (emit_vlmax_masked_store_insn): New function.
5480 (emit_nonvlmax_masked_store_insn): Ditto.
5481 (modulo_sel_indices): Ditto.
5482 (expand_vec_perm): Fix SLP for gather/scatter.
5483 (prepare_gather_scatter): New function.
5484 (expand_gather_scatter): Ditto.
5485 * config/riscv/riscv.cc (riscv_legitimize_move): Fix bug of
5486 (subreg:SI (DI CONST_POLY_INT)).
5487 * config/riscv/vector-iterators.md: Add gather/scatter.
5488 * config/riscv/vector.md (vec_duplicate<mode>): Use "@" instead.
5489 (@vec_duplicate<mode>): Ditto.
5490 (@pred_indexed_<order>store<VNX16_QHS:mode><VNX16_QHSDI:mode>):
5492 (@pred_indexed_<order>store<VNX16_QHSD:mode><VNX16_QHSDI:mode>): Ditto.
5494 2023-07-12 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5496 * config/riscv/autovec.md (cond_len_<optab><mode>): New pattern.
5497 * config/riscv/riscv-protos.h (enum insn_type): New enum.
5498 (expand_cond_len_binop): New function.
5499 * config/riscv/riscv-v.cc (emit_nonvlmax_tu_insn): Ditto.
5500 (emit_nonvlmax_fp_tu_insn): Ditto.
5501 (need_fp_rounding_p): Ditto.
5502 (expand_cond_len_binop): Ditto.
5503 * config/riscv/riscv.cc (riscv_preferred_else_value): Ditto.
5504 (TARGET_PREFERRED_ELSE_VALUE): New target hook.
5506 2023-07-12 Jan Hubicka <jh@suse.cz>
5508 * tree-cfg.cc (gimple_duplicate_sese_region): Rename to ...
5509 (gimple_duplicate_seme_region): ... this; break out profile updating
5511 * tree-ssa-loop-ch.cc (update_profile_after_ch): ... here.
5512 (ch_base::copy_headers): Update.
5513 * tree-cfg.h (gimple_duplicate_sese_region): Rename to ...
5514 (gimple_duplicate_seme_region): ... this.
5516 2023-07-12 Aldy Hernandez <aldyh@redhat.com>
5518 PR tree-optimization/107043
5519 * range-op.cc (operator_bitwise_and::op1_range): Update bitmask.
5521 2023-07-12 Aldy Hernandez <aldyh@redhat.com>
5523 PR tree-optimization/107053
5524 * gimple-range-op.cc (cfn_popcount): Use known set bits.
5526 2023-07-12 Uros Bizjak <ubizjak@gmail.com>
5528 * ira.cc (equiv_init_varies_p): Change return type from int to bool
5529 and adjust function body accordingly.
5530 (equiv_init_movable_p): Ditto.
5531 (memref_used_between_p): Ditto.
5532 * lra-constraints.cc (valid_address_p): Ditto.
5534 2023-07-12 Aldy Hernandez <aldyh@redhat.com>
5536 * range-op.cc (irange_to_masked_value): Remove.
5537 (update_known_bitmask): Update irange value/mask pair instead of
5538 only updating nonzero bits.
5540 2023-07-12 Jan Hubicka <jh@suse.cz>
5542 * tree-cfg.cc (gimple_duplicate_sese_region): Add ORIG_ELIMINATED_EDGES
5543 parameter and rewrite profile updating code to handle edges elimination.
5544 * tree-cfg.h (gimple_duplicate_sese_region): Update prototpe.
5545 * tree-ssa-loop-ch.cc (loop_invariant_op_p): New function.
5546 (loop_iv_derived_p): New function.
5547 (should_duplicate_loop_header_p): Track invariant exit edges; fix handling
5548 of PHIs and propagation of IV derived variables.
5549 (ch_base::copy_headers): Pass around the invariant edges hash set.
5551 2023-07-12 Uros Bizjak <ubizjak@gmail.com>
5553 * ifcvt.cc (cond_exec_changed_p): Change variable to bool.
5554 (last_active_insn): Change "skip_use_p" function argument to bool.
5555 (noce_operand_ok): Change return type from int to bool.
5556 (find_cond_trap): Ditto.
5557 (block_jumps_and_fallthru_p): Change "fallthru_p" and
5558 "jump_p" variables to bool.
5559 (noce_find_if_block): Change return type from int to bool.
5560 (cond_exec_find_if_block): Ditto.
5561 (find_if_case_1): Ditto.
5562 (find_if_case_2): Ditto.
5563 (dead_or_predicable): Ditto. Change "reversep" function arg to bool.
5564 (block_jumps_and_fallthru): Rename from block_jumps_and_fallthru_p.
5565 (cond_exec_process_insns): Change return type from int to bool.
5566 Change "mod_ok" function arg to bool.
5567 (cond_exec_process_if_block): Change return type from int to bool.
5568 Change "do_multiple_p" function arg to bool. Change "then_mod_ok"
5570 (noce_emit_store_flag): Change return type from int to bool.
5571 Change "reversep" function arg to bool. Change "cond_complex"
5573 (noce_try_move): Change return type from int to bool.
5574 (noce_try_ifelse_collapse): Ditto.
5575 (noce_try_store_flag): Ditto. Change "reversep" variable to bool.
5576 (noce_try_addcc): Change return type from int to bool. Change
5577 "subtract" variable to bool.
5578 (noce_try_store_flag_constants): Change return type from int to bool.
5579 (noce_try_store_flag_mask): Ditto. Change "reversep" variable to bool.
5580 (noce_try_cmove): Change return type from int to bool.
5581 (noce_try_cmove_arith): Ditto. Change "is_mem" variable to bool.
5582 (noce_try_minmax): Change return type from int to bool. Change
5583 "unsignedp" variable to bool.
5584 (noce_try_abs): Change return type from int to bool. Change
5585 "negate" variable to bool.
5586 (noce_try_sign_mask): Change return type from int to bool.
5587 (noce_try_move): Ditto.
5588 (noce_try_store_flag_constants): Ditto.
5589 (noce_try_cmove): Ditto.
5590 (noce_try_cmove_arith): Ditto.
5591 (noce_try_minmax): Ditto. Change "unsignedp" variable to bool.
5592 (noce_try_bitop): Change return type from int to bool.
5593 (noce_operand_ok): Ditto.
5594 (noce_convert_multiple_sets): Ditto.
5595 (noce_convert_multiple_sets_1): Ditto.
5596 (noce_process_if_block): Ditto.
5597 (check_cond_move_block): Ditto.
5598 (cond_move_process_if_block): Ditto. Change "success_p"
5600 (rest_of_handle_if_conversion): Change return type to void.
5602 2023-07-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
5604 * internal-fn.cc (FOR_EACH_CODE_MAPPING): Adapt for COND_LEN_* support.
5606 (get_conditional_len_internal_fn): New function.
5607 * internal-fn.h (get_conditional_len_internal_fn): Ditto.
5608 * tree-vect-stmts.cc (vectorizable_operation): Adapt for COND_LEN_*
5611 2023-07-12 Roger Sayle <roger@nextmovesoftware.com>
5614 * config/i386/i386.md (*add<dwi>3_doubleword_concat_zext): Typo.
5616 2023-07-12 Roger Sayle <roger@nextmovesoftware.com>
5619 * config/i386/i386.md (*add<dwi>3_doubleword_concat_zext): New
5620 define_insn_and_split derived from *add<dwi>3_doubleword_concat
5621 and *add<dwi>3_doubleword_zext.
5623 2023-07-12 Roger Sayle <roger@nextmovesoftware.com>
5626 * config/i386/i386.md (peephole2): Check !reg_mentioned_p when
5627 optimizing rega = 0; rega op= regb for op in [XOR,IOR,PLUS].
5628 (peephole2): Simplify rega = 0; rega op= rega cases.
5630 2023-07-12 Roger Sayle <roger@nextmovesoftware.com>
5632 * config/i386/i386-expand.cc (ix86_expand_int_compare): If
5633 testing a TImode SUBREG of a 128-bit vector register against
5634 zero, use a PTEST instruction instead of first moving it to
5635 a pair of scalar registers.
5637 2023-07-12 Robin Dapp <rdapp@ventanamicro.com>
5639 * genopinit.cc (main): Adjust maximal number of optabs and
5641 * gensupport.cc (find_optab): Shift optab by 20 and mode by
5643 * optabs-query.h (optab_handler): Ditto.
5644 (convert_optab_handler): Ditto.
5646 2023-07-12 Richard Biener <rguenther@suse.de>
5648 PR tree-optimization/110630
5649 * tree-vect-slp.cc (vect_add_slp_permutation): New
5650 offset parameter, honor that for the extract code generation.
5651 (vectorizable_slp_permutation_1): Handle offsetted identities.
5653 2023-07-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
5655 * config/riscv/autovec.md (smul<mode>3_highpart): New pattern.
5656 (umul<mode>3_highpart): Ditto.
5658 2023-07-12 Jan Beulich <jbeulich@suse.com>
5660 * config/i386/i386.md (extendbfsf2_1): Add new AVX512F
5661 alternative. Adjust original last alternative's "prefix"
5662 attribute to maybe_evex.
5664 2023-07-12 Jan Beulich <jbeulich@suse.com>
5666 * config/i386/sse.md (vec_dupv4sf): Make first alternative use
5667 vbroadcastss for AVX2. New AVX512F alternative.
5668 (*vec_dupv4si): New AVX2 and AVX512F alternatives using
5669 vpbroadcastd. Replace sselog1 by sseshuf1 in "type" attribute.
5671 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
5673 * config/riscv/peephole.md: Remove XThead* peephole passes.
5674 * config/riscv/thead.md: Include thead-peephole.md.
5675 * config/riscv/thead-peephole.md: New file.
5677 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
5679 * config/riscv/riscv-protos.h (riscv_regno_ok_for_index_p):
5681 (riscv_index_reg_class): Likewise.
5682 * config/riscv/riscv.cc (riscv_regno_ok_for_index_p): New function.
5683 (riscv_index_reg_class): New function.
5684 * config/riscv/riscv.h (INDEX_REG_CLASS): Call new function
5685 riscv_index_reg_class().
5686 (REGNO_OK_FOR_INDEX_P): Call new function
5687 riscv_regno_ok_for_index_p().
5689 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
5691 * config/riscv/riscv-protos.h (enum riscv_address_type):
5692 New location of type definition.
5693 (struct riscv_address_info): Likewise.
5694 * config/riscv/riscv.cc (enum riscv_address_type):
5695 Old location of type definition.
5696 (struct riscv_address_info): Likewise.
5698 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
5700 * config/riscv/riscv.h (Xmode): New macro.
5702 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
5704 * config/riscv/riscv.cc (riscv_print_operand_address): Use
5705 output_addr_const rather than riscv_print_operand.
5707 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
5709 * config/riscv/thead.md: Adjust constraints of th_addsl.
5711 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
5713 * config/riscv/thead.cc (th_mempair_operands_p):
5714 Fix documentation of th_mempair_order_operands().
5716 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
5718 * config/riscv/thead.cc (th_mempair_save_regs):
5719 Emit REG_FRAME_RELATED_EXPR notes in prologue.
5721 2023-07-12 Christoph Müllner <christoph.muellner@vrull.eu>
5723 * config/riscv/riscv.md: No base-ISA extension splitter for XThead*.
5724 * config/riscv/thead.md (*extend<SHORT:mode><SUPERQI:mode>2_th_ext):
5725 New XThead extension INSN.
5726 (*zero_extendsidi2_th_extu): New XThead extension INSN.
5727 (*zero_extendhi<GPR:mode>2_th_extu): New XThead extension INSN.
5729 2023-07-12 liuhongt <hongtao.liu@intel.com>
5733 * config/i386/predicates.md
5734 (int_float_vector_all_ones_operand): New predicate.
5735 * config/i386/sse.md (*vmov<mode>_constm1_pternlog_false_dep): New
5737 (*<avx512>_cvtmask2<ssemodesuffix><mode>_pternlog_false_dep):
5739 (*<avx512>_cvtmask2<ssemodesuffix><mode>_pternlog_false_dep):
5741 (*<avx512>_cvtmask2<ssemodesuffix><mode>): Adjust to
5742 define_insn_and_split to avoid false dependence.
5743 (*<avx512>_cvtmask2<ssemodesuffix><mode>): Ditto.
5744 (<mask_codefor>one_cmpl<mode>2<mask_name>): Adjust constraint
5745 of operands 1 to '0' to avoid false dependence.
5746 (*andnot<mode>3): Ditto.
5747 (iornot<mode>3): Ditto.
5748 (*<nlogic><mode>3): Ditto.
5750 2023-07-12 Mo, Zewei <zewei.mo@intel.com>
5752 * common/config/i386/cpuinfo.h
5753 (get_intel_cpu): Handle Granite Rapids D.
5754 * common/config/i386/i386-common.cc:
5755 (processor_alias_table): Add graniterapids-d.
5756 * common/config/i386/i386-cpuinfo.h
5757 (enum processor_subtypes): Add INTEL_COREI7_GRANITERAPIDS_D.
5758 * config.gcc: Add -march=graniterapids-d.
5759 * config/i386/driver-i386.cc (host_detect_local_cpu):
5760 Handle graniterapids-d.
5761 * config/i386/i386.h: (PTA_GRANITERAPIDS_D): New.
5762 * doc/extend.texi: Add graniterapids-d.
5763 * doc/invoke.texi: Ditto.
5765 2023-07-12 Haochen Jiang <haochen.jiang@intel.com>
5767 * config/i386/i386-builtins.cc (ix86_init_mmx_sse_builtins):
5768 Add OPTION_MASK_ISA_AVX512VL.
5769 * config/i386/i386-expand.cc (ix86_check_builtin_isa_match):
5772 2023-07-11 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
5774 * config/riscv/riscv-protos.h (enum insn_type): Add vcompress optimization.
5775 * config/riscv/riscv-v.cc (emit_vlmax_compress_insn): Ditto.
5776 (shuffle_compress_patterns): Ditto.
5777 (expand_vec_perm_const_1): Ditto.
5779 2023-07-11 Uros Bizjak <ubizjak@gmail.com>
5781 * cfghooks.cc (verify_flow_info): Change "err" variable to bool.
5782 * cfghooks.h (struct cfg_hooks): Change return type of
5783 verify_flow_info from integer to bool.
5784 * cfgrtl.cc (can_delete_note_p): Change return type from int to bool.
5785 (can_delete_label_p): Ditto.
5786 (rtl_verify_flow_info): Change return type from int to bool
5787 and adjust function body accordingly. Change "err" variable to bool.
5788 (rtl_verify_flow_info_1): Ditto.
5789 (free_bb_for_insn): Change return type to void.
5790 (rtl_merge_blocks): Change "b_empty" variable to bool.
5791 (try_redirect_by_replacing_jump): Change "fallthru" variable to bool.
5792 (verify_hot_cold_block_grouping): Change return type from int to bool.
5793 Change "err" variable to bool.
5794 (rtl_verify_edges): Ditto.
5795 (rtl_verify_bb_insns): Ditto.
5796 (rtl_verify_bb_pointers): Ditto.
5797 (rtl_verify_bb_insn_chain): Ditto.
5798 (rtl_verify_fallthru): Ditto.
5799 (rtl_verify_bb_layout): Ditto.
5800 (purge_all_dead_edges): Change "purged" variable to bool.
5801 * cfgrtl.h (free_bb_for_insn): Change return type from int to void.
5802 * postreload-gcse.cc (expr_hasher::equal): Change "equiv_p" to bool.
5803 (load_killed_in_block_p): Change return type from int to bool
5804 and adjust function body accordingly.
5805 (oprs_unchanged_p): Return true/false.
5806 (rest_of_handle_gcse2): Change return type to void.
5807 * tree-cfg.cc (gimple_verify_flow_info): Change return type from
5808 int to bool. Change "err" variable to bool.
5810 2023-07-11 Gaius Mulley <gaiusmod2@gmail.com>
5812 * doc/gm2.texi (-Wuninit-variable-checking=) New item.
5814 2023-07-11 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
5816 * doc/md.texi: Add COND_LEN_* operations for loop control with length.
5817 * internal-fn.cc (cond_len_unary_direct): Ditto.
5818 (cond_len_binary_direct): Ditto.
5819 (cond_len_ternary_direct): Ditto.
5820 (expand_cond_len_unary_optab_fn): Ditto.
5821 (expand_cond_len_binary_optab_fn): Ditto.
5822 (expand_cond_len_ternary_optab_fn): Ditto.
5823 (direct_cond_len_unary_optab_supported_p): Ditto.
5824 (direct_cond_len_binary_optab_supported_p): Ditto.
5825 (direct_cond_len_ternary_optab_supported_p): Ditto.
5826 * internal-fn.def (COND_LEN_ADD): Ditto.
5827 (COND_LEN_SUB): Ditto.
5828 (COND_LEN_MUL): Ditto.
5829 (COND_LEN_DIV): Ditto.
5830 (COND_LEN_MOD): Ditto.
5831 (COND_LEN_RDIV): Ditto.
5832 (COND_LEN_MIN): Ditto.
5833 (COND_LEN_MAX): Ditto.
5834 (COND_LEN_FMIN): Ditto.
5835 (COND_LEN_FMAX): Ditto.
5836 (COND_LEN_AND): Ditto.
5837 (COND_LEN_IOR): Ditto.
5838 (COND_LEN_XOR): Ditto.
5839 (COND_LEN_SHL): Ditto.
5840 (COND_LEN_SHR): Ditto.
5841 (COND_LEN_FMA): Ditto.
5842 (COND_LEN_FMS): Ditto.
5843 (COND_LEN_FNMA): Ditto.
5844 (COND_LEN_FNMS): Ditto.
5845 (COND_LEN_NEG): Ditto.
5846 * optabs.def (OPTAB_D): Ditto.
5848 2023-07-11 Richard Biener <rguenther@suse.de>
5850 PR tree-optimization/110614
5851 * tree-vect-data-refs.cc (vect_supportable_dr_alignment):
5852 SLP splats are not suitable for re-align ops.
5854 2023-07-10 Peter Bergner <bergner@linux.ibm.com>
5856 * config/rs6000/predicates.md (quad_memory_operand): Remove redundant
5858 (vsx_quad_dform_memory_operand): Likewise.
5860 2023-07-10 Uros Bizjak <ubizjak@gmail.com>
5862 * reorg.cc (stop_search_p): Change return type from int to bool
5863 and adjust function body accordingly.
5864 (resource_conflicts_p): Ditto.
5865 (insn_references_resource_p): Change return type from int to bool.
5866 (insn_sets_resource_p): Ditto.
5867 (redirect_with_delay_slots_safe_p): Ditto.
5868 (condition_dominates_p): Change return type from int to bool
5869 and adjust function body accordingly.
5870 (redirect_with_delay_list_safe_p): Ditto.
5871 (check_annul_list_true_false): Ditto. Change "annul_true_p"
5872 function argument to bool.
5873 (steal_delay_list_from_target): Change "pannul_p" function
5874 argument to bool pointer. Change "must_annul" and "used_annul"
5875 variables from int to bool.
5876 (steal_delay_list_from_fallthrough): Ditto.
5877 (own_thread_p): Change return type from int to bool and adjust
5878 function body accordingly. Change "allow_fallthrough" function
5880 (reorg_redirect_jump): Change return type from int to bool.
5881 (fill_simple_delay_slots): Change "non_jumps_p" function
5882 argument from int to bool. Change "maybe_never" varible to bool.
5883 (fill_slots_from_thread): Change "likely", "thread_if_true" and
5884 "own_thread" function arguments to bool. Change "lose" and
5885 "must_annul" variables to bool.
5886 (delete_from_delay_slot): Change "had_barrier" variable to bool.
5887 (try_merge_delay_insns): Change "annul_p" variable to bool.
5888 (fill_eager_delay_slots): Change "own_target" and "own_fallthrouhg"
5890 (rest_of_handle_delay_slots): Change return type from int to void
5891 and adjust function body accordingly.
5893 2023-07-10 Kito Cheng <kito.cheng@sifive.com>
5895 * doc/extend.texi (RISC-V Operand Modifiers): New.
5897 2023-07-10 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
5899 * config/riscv/riscv-vsetvl.cc (add_label_notes): Remove it.
5900 (insert_insn_end_basic_block): Ditto.
5901 (pass_vsetvl::commit_vsetvls): Adapt for new helper function.
5902 * gcse.cc (insert_insn_end_basic_block): Export as global function.
5903 * gcse.h (insert_insn_end_basic_block): Ditto.
5905 2023-07-10 Christophe Lyon <christophe.lyon@linaro.org>
5908 * config/arm/arm-builtins.cc (arm_init_mve_builtins): Handle LTO.
5909 (arm_builtin_decl): Hahndle MVE builtins.
5910 * config/arm/arm-mve-builtins.cc (builtin_decl): New function.
5911 (add_unique_function): Fix handling of
5912 __ARM_MVE_PRESERVE_USER_NAMESPACE.
5913 (add_overloaded_function): Likewise.
5914 * config/arm/arm-protos.h (builtin_decl): New declaration.
5916 2023-07-10 Christophe Lyon <christophe.lyon@linaro.org>
5918 * doc/sourcebuild.texi (arm_v8_1m_main_cde_mve_fp): Document.
5920 2023-07-10 Xi Ruoyao <xry111@xry111.site>
5922 PR tree-optimization/110557
5923 * tree-vect-patterns.cc (vect_recog_bitfield_ref_pattern):
5924 Ensure the output sign-extended if necessary.
5926 2023-07-10 Roger Sayle <roger@nextmovesoftware.com>
5928 * config/i386/i386.md (peephole2): Transform xchg insn with a
5929 REG_UNUSED note to a (simple) move.
5930 (*insvti_lowpart_1): New define_insn_and_split.
5931 (*insvdi_lowpart_1): Likewise.
5933 2023-07-10 Roger Sayle <roger@nextmovesoftware.com>
5935 * config/i386/i386-features.cc (compute_convert_gain): Tweak
5936 gains/costs for ROTATE/ROTATERT by integer constant on AVX512VL.
5937 (general_scalar_chain::convert_rotate): On TARGET_AVX512F generate
5938 avx512vl_rolv2di or avx412vl_rolv4si when appropriate.
5940 2023-07-10 liuhongt <hongtao.liu@intel.com>
5943 * config/i386/i386.md (*ieee_max<mode>3_1): New pre_reload
5944 splitter to detect fp max pattern.
5945 (*ieee_min<mode>3_1): Ditto, but for fp min pattern.
5947 2023-07-09 Jan Hubicka <jh@suse.cz>
5949 * cfg.cc (check_bb_profile): Dump counts with relative frequency.
5950 (dump_edge_info): Likewise.
5951 (dump_bb_info): Likewise.
5952 * profile-count.cc (profile_count::dump): Add comma between quality and
5955 2023-07-08 Jan Hubicka <jh@suse.cz>
5957 PR tree-optimization/110600
5958 * cfgloopmanip.cc (scale_loop_profile): Add mising profile_dump check.
5960 2023-07-08 Jan Hubicka <jh@suse.cz>
5962 PR middle-end/110590
5963 * cfgloopmanip.cc (scale_loop_profile): Avoid scaling exits within
5964 inner loops and be more careful about inconsistent profiles.
5965 (duplicate_loop_body_to_header_edge): Fix profile update when eliminated
5966 exit is followed by other exit.
5968 2023-07-08 Uros Bizjak <ubizjak@gmail.com>
5970 * cprop.cc (reg_available_p): Change return type from int to bool.
5971 (reg_not_set_p): Ditto.
5972 (try_replace_reg): Ditto. Change "success" variable to bool.
5973 (cprop_jump): Change return type from int to void
5974 and adjust function body accordingly.
5975 (constprop_register): Ditto.
5976 (cprop_insn): Ditto. Change "changed" variable to bool.
5977 (local_cprop_pass): Change return type from int to void
5978 and adjust function body accordingly.
5979 (bypass_block): Ditto. Change "change", "may_be_loop_header"
5980 and "removed_p" variables to bool.
5981 (bypass_conditional_jumps): Change return type from int to void
5982 and adjust function body accordingly. Change "changed"
5984 (one_cprop_pass): Ditto.
5986 2023-07-08 Uros Bizjak <ubizjak@gmail.com>
5988 * gcse.cc (expr_equiv_p): Change return type from int to bool.
5989 (oprs_unchanged_p): Change return type from int to void
5990 and adjust function body accordingly.
5991 (oprs_anticipatable_p): Ditto.
5992 (oprs_available_p): Ditto.
5993 (insert_expr_in_table): Ditto. Change "antic_p" and "avail_p"
5994 arguments to bool. Change "found" variable to bool.
5995 (load_killed_in_block_p): Change return type from int to void and
5996 adjust function body accordingly. Change "avail_p" argument to bool.
5997 (pre_expr_reaches_here_p): Change return type from int to void
5998 and adjust function body accordingly.
5999 (pre_delete): Ditto. Change "changed" variable to bool.
6000 (pre_gcse): Change return type from int to void
6001 and adjust function body accordingly. Change "did_insert" and
6002 "changed" variables to bool.
6003 (one_pre_gcse_pass): Change return type from int to void
6004 and adjust function body accordingly. Change "changed" variable
6006 (should_hoist_expr_to_dom): Change return type from int to void
6007 and adjust function body accordingly. Change
6008 "visited_allocated_locally" variable to bool.
6009 (hoist_code): Change return type from int to void and adjust
6010 function body accordingly. Change "changed" variable to bool.
6011 (one_code_hoisting_pass): Ditto.
6012 (pre_edge_insert): Change return type from int to void and adjust
6013 function body accordingly. Change "did_insert" variable to bool.
6014 (pre_expr_reaches_here_p_work): Change return type from int to void
6015 and adjust function body accordingly.
6016 (simple_mem): Ditto.
6017 (want_to_gcse_p): Change return type from int to void
6018 and adjust function body accordingly.
6019 (can_assign_to_reg_without_clobbers_p): Update function body
6020 for bool return type.
6021 (hash_scan_set): Change "antic_p" and "avail_p" variables to bool.
6022 (pre_insert_copies): Change "added_copy" variable to bool.
6024 2023-07-08 Jonathan Wakely <jwakely@redhat.com>
6028 * doc/invoke.texi (Warning Options): Fix typos.
6030 2023-07-07 Jan Hubicka <jh@suse.cz>
6032 * profile-count.cc (profile_count::dump): Add FUN
6033 parameter; print relative frequency.
6034 (profile_count::debug): Update.
6035 * profile-count.h (profile_count::dump): Update
6038 2023-07-07 Roger Sayle <roger@nextmovesoftware.com>
6042 * config/i386/i386-expand.cc (ix86_expand_move): Convert SETs of
6043 TImode destinations from paradoxical SUBREGs (setting the lowpart)
6044 into explicit zero extensions. Use *insvti_highpart_1 instruction
6045 to set the highpart of a TImode destination.
6047 2023-07-07 Jan Hubicka <jh@suse.cz>
6049 * predict.cc (force_edge_cold): Use
6050 set_edge_probability_and_rescale_others; improve dumps.
6052 2023-07-07 Jan Hubicka <jh@suse.cz>
6054 * cfgloopmanip.cc (scale_loop_profile): Fix computation of count_in and scaling blocks
6056 * tree-vect-loop-manip.cc (vect_do_peeling): Scale loop profile of the epilogue if bound
6059 2023-07-07 Juergen Christ <jchrist@linux.ibm.com>
6061 * config/s390/s390.cc (vec_init): Fix default case
6063 2023-07-07 Vladimir N. Makarov <vmakarov@redhat.com>
6065 * lra-assigns.cc (assign_by_spills): Add reload insns involving
6066 reload pseudos with non-refined class to be processed on the next
6068 * lra-constraints.cc (enough_allocatable_hard_regs_p): New func.
6069 (in_class_p): Use it.
6070 (print_curr_insn_alt): New func.
6071 (process_alt_operands): Use it. Improve debug info.
6072 (curr_insn_transform): Use print_curr_insn_alt. Refine reload
6073 pseudo class if it is not refined yet.
6075 2023-07-07 Aldy Hernandez <aldyh@redhat.com>
6077 * value-range.cc (irange::get_bitmask_from_range): Return all the
6078 known bits for a singleton.
6079 (irange::set_range_from_bitmask): Set a range of a singleton when
6082 2023-07-07 Aldy Hernandez <aldyh@redhat.com>
6084 * value-range.cc (irange::intersect): Leave normalization to
6087 2023-07-07 Aldy Hernandez <aldyh@redhat.com>
6089 * data-streamer-in.cc (streamer_read_value_range): Adjust for
6091 * data-streamer-out.cc (streamer_write_vrange): Same.
6092 * range-op.cc (operator_cast::fold_range): Same.
6093 * value-range-pretty-print.cc
6094 (vrange_printer::print_irange_bitmasks): Same.
6095 * value-range-storage.cc (irange_storage::write_lengths_address):
6097 (irange_storage::set_irange): Same.
6098 (irange_storage::get_irange): Same.
6099 (irange_storage::size): Same.
6100 (irange_storage::dump): Same.
6101 * value-range-storage.h: Same.
6102 * value-range.cc (debug): New.
6103 (irange_bitmask::dump): New.
6104 (add_vrange): Adjust for value/mask.
6105 (irange::operator=): Same.
6106 (irange::set): Same.
6107 (irange::verify_range): Same.
6108 (irange::operator==): Same.
6109 (irange::contains_p): Same.
6110 (irange::irange_single_pair_union): Same.
6111 (irange::union_): Same.
6112 (irange::intersect): Same.
6113 (irange::invert): Same.
6114 (irange::get_nonzero_bits_from_range): Rename to...
6115 (irange::get_bitmask_from_range): ...this.
6116 (irange::set_range_from_nonzero_bits): Rename to...
6117 (irange::set_range_from_bitmask): ...this.
6118 (irange::set_nonzero_bits): Rename to...
6119 (irange::update_bitmask): ...this.
6120 (irange::get_nonzero_bits): Rename to...
6121 (irange::get_bitmask): ...this.
6122 (irange::intersect_nonzero_bits): Rename to...
6123 (irange::intersect_bitmask): ...this.
6124 (irange::union_nonzero_bits): Rename to...
6125 (irange::union_bitmask): ...this.
6126 (irange_bitmask::verify_mask): New.
6127 * value-range.h (class irange_bitmask): New.
6128 (irange_bitmask::set_unknown): New.
6129 (irange_bitmask::unknown_p): New.
6130 (irange_bitmask::irange_bitmask): New.
6131 (irange_bitmask::get_precision): New.
6132 (irange_bitmask::get_nonzero_bits): New.
6133 (irange_bitmask::set_nonzero_bits): New.
6134 (irange_bitmask::operator==): New.
6135 (irange_bitmask::union_): New.
6136 (irange_bitmask::intersect): New.
6137 (class irange): Friend vrange_printer.
6138 (irange::varying_compatible_p): Adjust for bitmask.
6139 (irange::set_varying): Same.
6140 (irange::set_nonzero): Same.
6142 2023-07-07 Jan Beulich <jbeulich@suse.com>
6144 * config/i386/sse.md (*vec_extractv2ti): Drop g modifiers.
6146 2023-07-07 Jan Beulich <jbeulich@suse.com>
6148 * config/i386/sse.md (@vec_extract_hi_<mode>): Drop last
6149 alternative. Switch new last alternative's "isa" attribute to
6151 (vec_extract_hi_v32qi): Likewise.
6153 2023-07-07 Pan Li <pan2.li@intel.com>
6154 Robin Dapp <rdapp@ventanamicro.com>
6156 * config/riscv/riscv.cc (riscv_emit_mode_set): Avoid emit insn
6158 (riscv_mode_entry): Take FRM_MODE_DYN as entry mode.
6159 (riscv_mode_exit): Likewise for exit mode.
6160 (riscv_mode_needed): Likewise for needed mode.
6161 (riscv_mode_after): Likewise for after mode.
6163 2023-07-07 Pan Li <pan2.li@intel.com>
6165 * config/riscv/vector.md: Fix typo.
6167 2023-07-06 Jan Hubicka <jh@suse.cz>
6170 * tree-ssa-loop-ch.cc (ch_base::copy_headers): Scale loop frequency to maximal number
6171 of iterations determined.
6172 * tree-ssa-loop-ivcanon.cc (try_unroll_loop_completely): Likewise.
6174 2023-07-06 Jan Hubicka <jh@suse.cz>
6176 * cfgloopmanip.cc (scale_loop_profile): Rewrite exit edge
6177 probability update to be safe on loops with subloops.
6178 Make bound parameter to be iteration bound.
6179 * tree-ssa-loop-ivcanon.cc (try_peel_loop): Update call
6180 of scale_loop_profile.
6181 * tree-vect-loop-manip.cc (vect_do_peeling): Likewise.
6183 2023-07-06 Hao Liu OS <hliu@os.amperecomputing.com>
6185 PR tree-optimization/110449
6186 * tree-vect-loop.cc (vectorizable_induction): use vec_n to replace
6187 vec_loop for the unrolled loop.
6189 2023-07-06 Jan Hubicka <jh@suse.cz>
6191 * cfg.cc (set_edge_probability_and_rescale_others): New function.
6192 (update_bb_profile_for_threading): Use it; simplify the rest.
6193 * cfg.h (set_edge_probability_and_rescale_others): Declare.
6194 * profile-count.h (profile_probability::apply_scale): New.
6196 2023-07-06 Claudiu Zissulescu <claziss@gmail.com>
6198 * doc/extend.texi (ARC Built-in Functions): Update documentation
6199 with missing builtins.
6201 2023-07-06 Richard Biener <rguenther@suse.de>
6203 PR tree-optimization/110556
6204 * tree-ssa-tail-merge.cc (gimple_equal_p): Check
6205 assign code and all operands of non-stores.
6207 2023-07-06 Richard Biener <rguenther@suse.de>
6209 PR tree-optimization/110563
6210 * tree-vectorizer.h (vect_determine_partial_vectors_and_peeling):
6211 Remove second argument.
6212 * tree-vect-loop.cc (vect_determine_partial_vectors_and_peeling):
6213 Remove for_epilogue_p argument. Merge assert ...
6214 (vect_analyze_loop_2): ... with check done before determining
6215 partial vectors by moving it after.
6216 * tree-vect-loop-manip.cc (vect_do_peeling): Adjust.
6218 2023-07-06 Thomas Schwinge <thomas@codesourcery.com>
6220 * ggc-common.cc (gt_pch_note_reorder, gt_pch_save): Tighten up a
6221 few things re 'reorder' option and strings.
6222 * stringpool.cc (gt_pch_p_S): This is now 'gcc_unreachable'.
6224 2023-07-06 Thomas Schwinge <thomas@codesourcery.com>
6226 * gengtype-parse.cc: Clean up obsolete parametrized structs
6228 * gengtype.cc: Likewise.
6229 * gengtype.h: Likewise.
6231 2023-07-06 Thomas Schwinge <thomas@codesourcery.com>
6233 * gengtype.cc (struct walk_type_data): Remove 'needs_cast_p'.
6236 2023-07-06 Thomas Schwinge <thomas@codesourcery.com>
6238 * gengtype-parse.cc (token_names): Add '"user"'.
6239 * gengtype.h (gty_token): Add 'UNUSED_PARAM_IS' for use with
6240 'FIRST_TOKEN_WITH_VALUE'.
6242 2023-07-06 Thomas Schwinge <thomas@codesourcery.com>
6244 * doc/gty.texi (GTY Options) <string_length>: Enhance.
6246 2023-07-06 Thomas Schwinge <thomas@codesourcery.com>
6248 * gengtype.cc (write_root, write_roots): Explicitly reject
6249 'string_length' option.
6250 * doc/gty.texi (GTY Options) <string_length>: Document.
6252 2023-07-06 Thomas Schwinge <thomas@codesourcery.com>
6254 * ggc-internal.h (ggc_pch_count_object, ggc_pch_alloc_object)
6255 (ggc_pch_write_object): Remove 'bool is_string' argument.
6256 * ggc-common.cc: Adjust.
6257 * ggc-page.cc: Likewise.
6259 2023-07-06 Roger Sayle <roger@nextmovesoftware.com>
6261 * dwarf2out.cc (mem_loc_descriptor): Handle COPYSIGN.
6263 2023-07-06 Hongyu Wang <hongyu.wang@intel.com>
6265 * doc/extend.texi: Move x86 inlining rule to a new subsubsection
6266 and add description for inling of function with arch and tune
6269 2023-07-06 Richard Biener <rguenther@suse.de>
6271 PR tree-optimization/110515
6272 * tree-ssa-pre.cc (compute_avail): Make code dealing
6273 with hoisting loads with different alias-sets more
6276 2023-07-06 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
6278 * tree-vect-stmts.cc (vect_get_strided_load_store_ops): Fix ICE.
6280 2023-07-06 Hongyu Wang <hongyu.wang@intel.com>
6282 * config/i386/i386.cc (ix86_can_inline_p): If callee has
6283 default arch=x86-64 and tune=generic, do not block the
6284 inlining to its caller. Also allow callee with different
6285 arch= to be inlined if it has always_inline attribute and
6286 it's ISA is subset of caller's.
6288 2023-07-06 liuhongt <hongtao.liu@intel.com>
6290 * config/i386/i386.cc (ix86_rtx_costs): Adjust rtx_cost for
6291 DF/SFmode AND/IOR/XOR/ANDN operations.
6293 2023-07-06 Andrew Pinski <apinski@marvell.com>
6295 PR middle-end/110554
6296 * tree-vect-generic.cc (expand_vector_condition): For comparisons,
6297 just build using boolean_type_node instead of the cond_type.
6298 For non-comparisons/non-scalar-bitmask, build a ` != 0` gimple
6299 that will feed into the COND_EXPR.
6301 2023-07-06 liuhongt <hongtao.liu@intel.com>
6304 * config/i386/i386.md (movdf_internal): Disparage slightly for
6305 2 alternatives (r,v) and (v,r) by adding constraint modifier
6308 2023-07-06 Jeevitha Palanisamy <jeevitha@linux.ibm.com>
6311 * config/rs6000/rs6000.cc (rs6000_expand_vector_extract): Remove redundant
6312 initialization of new_addr.
6314 2023-07-06 Hao Liu <hliu@os.amperecomputing.com>
6316 PR tree-optimization/110474
6317 * tree-vect-loop.cc (vect_analyze_loop_2): unscale the VF by suggested
6318 unroll factor while selecting the epilog vect loop VF.
6320 2023-07-05 Andrew MacLeod <amacleod@redhat.com>
6322 * gimple-range-gori.cc (compute_operand_range): Convert to a tail
6325 2023-07-05 Andrew MacLeod <amacleod@redhat.com>
6327 * gimple-range-gori.cc (compute_operand_range): After calling
6328 compute_operand2_range, recursively call self if needed.
6329 (compute_operand2_range): Turn into a leaf function.
6330 (gori_compute::compute_operand1_and_operand2_range): Finish
6331 operand2 calculation.
6332 * gimple-range-gori.h (compute_operand2_range): Remove name param.
6334 2023-07-05 Andrew MacLeod <amacleod@redhat.com>
6336 * gimple-range-gori.cc (compute_operand_range): After calling
6337 compute_operand1_range, recursively call self if needed.
6338 (compute_operand1_range): Turn into a leaf function.
6339 (gori_compute::compute_operand1_and_operand2_range): Finish
6340 operand1 calculation.
6341 * gimple-range-gori.h (compute_operand1_range): Remove name param.
6343 2023-07-05 Andrew MacLeod <amacleod@redhat.com>
6345 * gimple-range-gori.cc (compute_operand_range): Check for
6346 operand interdependence when both op1 and op2 are computed.
6347 (compute_operand1_and_operand2_range): No checks required now.
6349 2023-07-05 Andrew MacLeod <amacleod@redhat.com>
6351 * gimple-range-gori.cc (compute_operand_range): Check for
6352 a relation between op1 and op2 and use that instead.
6353 (compute_operand1_range): Don't look for a relation override.
6354 (compute_operand2_range): Ditto.
6356 2023-07-05 Jonathan Wakely <jwakely@redhat.com>
6358 * doc/contrib.texi (Contributors): Update my entry.
6360 2023-07-05 Filip Kastl <filip.kastl@gmail.com>
6362 * value-prof.cc (gimple_mod_subtract_transform): Correct edge
6365 2023-07-05 Uros Bizjak <ubizjak@gmail.com>
6367 * sched-int.h (struct haifa_sched_info): Change can_schedule_ready_p,
6368 scehdule_more_p and contributes_to_priority indirect frunction
6369 type from int to bool.
6370 (no_real_insns_p): Change return type from int to bool.
6371 (contributes_to_priority): Ditto.
6372 * haifa-sched.cc (no_real_insns_p): Change return type from
6373 int to bool and adjust function body accordingly.
6374 * modulo-sched.cc (try_scheduling_node_in_cycle): Change "success"
6375 variable type from int to bool.
6376 (ps_insn_advance_column): Change return type from int to bool.
6377 (ps_has_conflicts): Ditto. Change "has_conflicts"
6378 variable type from int to bool.
6379 * sched-deps.cc (deps_may_trap_p): Change return type from int to bool.
6380 (conditions_mutex_p): Ditto.
6381 * sched-ebb.cc (schedule_more_p): Ditto.
6382 (ebb_contributes_to_priority): Change return type from
6383 int to bool and adjust function body accordingly.
6384 * sched-rgn.cc (is_cfg_nonregular): Ditto.
6385 (check_live_1): Ditto.
6387 (find_conditional_protection): Ditto.
6388 (is_conditionally_protected): Ditto.
6390 (is_exception_free): Ditto.
6391 (haifa_find_rgns): Change "unreachable" and "too_large_failure"
6392 variables from int to bool.
6393 (extend_rgns): Change "rescan" variable from int to bool.
6394 (check_live): Change return type from
6395 int to bool and adjust function body accordingly.
6396 (can_schedule_ready_p): Ditto.
6397 (schedule_more_p): Ditto.
6398 (contributes_to_priority): Ditto.
6400 2023-07-05 Robin Dapp <rdapp@ventanamicro.com>
6402 * doc/md.texi: Document that vec_set and vec_extract must not
6404 * gimple-isel.cc (gimple_expand_vec_set_expr): Rename this...
6405 (gimple_expand_vec_set_extract_expr): ...to this.
6406 (gimple_expand_vec_exprs): Call renamed function.
6407 * internal-fn.cc (vec_extract_direct): Add.
6408 (expand_vec_extract_optab_fn): New function to expand
6410 (direct_vec_extract_optab_supported_p): Add.
6411 * internal-fn.def (VEC_EXTRACT): Add.
6412 * optabs.cc (can_vec_extract_var_idx_p): New function.
6413 * optabs.h (can_vec_extract_var_idx_p): Declare.
6415 2023-07-05 Robin Dapp <rdapp@ventanamicro.com>
6417 * config/riscv/autovec.md: Add gen_lowpart.
6419 2023-07-05 Robin Dapp <rdapp@ventanamicro.com>
6421 * config/riscv/autovec.md: Allow register index operand.
6423 2023-07-05 Pan Li <pan2.li@intel.com>
6425 * config/riscv/riscv-vector-builtins.cc
6426 (function_expander::use_exact_insn): Use FRM_DYN instead of const0.
6428 2023-07-05 Robin Dapp <rdapp@ventanamicro.com>
6430 * config/riscv/autovec.md: Use float_truncate.
6432 2023-07-05 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
6434 * internal-fn.cc (internal_fn_len_index): Apply
6435 LEN_MASK_GATHER_LOAD/SCATTER_STORE into vectorizer.
6436 (internal_fn_mask_index): Ditto.
6437 * optabs-query.cc (supports_vec_gather_load_p): Ditto.
6438 (supports_vec_scatter_store_p): Ditto.
6439 * tree-vect-data-refs.cc (vect_gather_scatter_fn_p): Ditto.
6440 * tree-vect-patterns.cc (vect_recog_gather_scatter_pattern): Ditto.
6441 * tree-vect-stmts.cc (check_load_store_for_partial_vectors): Ditto.
6442 (vect_get_strided_load_store_ops): Ditto.
6443 (vectorizable_store): Ditto.
6444 (vectorizable_load): Ditto.
6446 2023-07-05 Robin Dapp <rdapp@ventanamicro.com>
6447 Juzhe-Zhong <juzhe.zhong@rivai.ai>
6449 * simplify-rtx.cc (native_encode_rtx): Ditto.
6450 (native_decode_vector_rtx): Ditto.
6451 (simplify_const_vector_byte_offset): Ditto.
6452 (simplify_const_vector_subreg): Ditto.
6453 * tree.cc (build_truth_vector_type_for_mode): Ditto.
6454 * varasm.cc (output_constant_pool_2): Ditto.
6456 2023-07-05 YunQiang Su <yunqiang.su@cipunited.com>
6458 * config/mips/mips.cc (mips_expand_block_move): don't expand for
6459 r6 with -mno-unaligned-access option if one or both of src and
6460 dest are unaligned. restruct: return directly if length is not const.
6461 (mips_block_move_straight): emit_move if ISA_HAS_UNALIGNED_ACCESS.
6463 2023-07-05 Jan Beulich <jbeulich@suse.com>
6466 * config/i386/sse.md: New splitters to simplify
6467 not;vec_duplicate as a singular vpternlog.
6468 (one_cmpl<mode>2): Allow broadcast for operand 1.
6469 (<mask_codefor>one_cmpl<mode>2<mask_name>): Likewise.
6471 2023-07-05 Jan Beulich <jbeulich@suse.com>
6474 * config/i386/sse.md: New splitters to simplify
6475 not;vec_duplicate;{ior,xor} as vec_duplicate;{iornot,xnor}.
6477 2023-07-05 Jan Beulich <jbeulich@suse.com>
6480 * config/i386/sse.md: Permit non-immediate operand 1 in AVX2
6481 form of splitter for PR target/100711.
6483 2023-07-05 Richard Biener <rguenther@suse.de>
6485 PR middle-end/110541
6486 * tree.def (VEC_PERM_EXPR): Adjust documentation to reflect
6489 2023-07-05 Jan Beulich <jbeulich@suse.com>
6492 * config/i386/sse.md (*andnot<mode>3): Add new alternatives
6493 for memory form operand 1.
6495 2023-07-05 Jan Beulich <jbeulich@suse.com>
6498 * config/i386/i386.cc (ix86_rtx_costs): Further special-case
6499 bitwise vector operations.
6500 * config/i386/sse.md (*iornot<mode>3): New insn.
6501 (*xnor<mode>3): Likewise.
6502 (*<nlogic><mode>3): Likewise.
6503 (andor): New code iterator.
6504 (nlogic): New code attribute.
6505 (ternlog_nlogic): Likewise.
6507 2023-07-05 Richard Biener <rguenther@suse.de>
6509 * tree-vect-stmts.cc (vect_mark_relevant): Fix typo.
6511 2023-07-05 yulong <shiyulong@iscas.ac.cn>
6513 * config/riscv/vector.md: Add float16 attr at sew、vlmul and ratio.
6515 2023-07-05 yulong <shiyulong@iscas.ac.cn>
6517 * config/riscv/genrvv-type-indexer.cc (valid_type): Enable FP16 tuple.
6518 * config/riscv/riscv-modes.def (RVV_TUPLE_MODES): New macro.
6519 (ADJUST_ALIGNMENT): Ditto.
6520 (RVV_TUPLE_PARTIAL_MODES): Ditto.
6521 (ADJUST_NUNITS): Ditto.
6522 * config/riscv/riscv-vector-builtins-types.def (vfloat16mf4x2_t):
6524 (vfloat16mf4x3_t): Ditto.
6525 (vfloat16mf4x4_t): Ditto.
6526 (vfloat16mf4x5_t): Ditto.
6527 (vfloat16mf4x6_t): Ditto.
6528 (vfloat16mf4x7_t): Ditto.
6529 (vfloat16mf4x8_t): Ditto.
6530 (vfloat16mf2x2_t): Ditto.
6531 (vfloat16mf2x3_t): Ditto.
6532 (vfloat16mf2x4_t): Ditto.
6533 (vfloat16mf2x5_t): Ditto.
6534 (vfloat16mf2x6_t): Ditto.
6535 (vfloat16mf2x7_t): Ditto.
6536 (vfloat16mf2x8_t): Ditto.
6537 (vfloat16m1x2_t): Ditto.
6538 (vfloat16m1x3_t): Ditto.
6539 (vfloat16m1x4_t): Ditto.
6540 (vfloat16m1x5_t): Ditto.
6541 (vfloat16m1x6_t): Ditto.
6542 (vfloat16m1x7_t): Ditto.
6543 (vfloat16m1x8_t): Ditto.
6544 (vfloat16m2x2_t): Ditto.
6545 (vfloat16m2x3_t): Ditto.
6546 (vfloat16m2x4_t): Ditto.
6547 (vfloat16m4x2_t): Ditto.
6548 * config/riscv/riscv-vector-builtins.def (vfloat16mf4x2_t): New macro.
6549 (vfloat16mf4x3_t): Ditto.
6550 (vfloat16mf4x4_t): Ditto.
6551 (vfloat16mf4x5_t): Ditto.
6552 (vfloat16mf4x6_t): Ditto.
6553 (vfloat16mf4x7_t): Ditto.
6554 (vfloat16mf4x8_t): Ditto.
6555 (vfloat16mf2x2_t): Ditto.
6556 (vfloat16mf2x3_t): Ditto.
6557 (vfloat16mf2x4_t): Ditto.
6558 (vfloat16mf2x5_t): Ditto.
6559 (vfloat16mf2x6_t): Ditto.
6560 (vfloat16mf2x7_t): Ditto.
6561 (vfloat16mf2x8_t): Ditto.
6562 (vfloat16m1x2_t): Ditto.
6563 (vfloat16m1x3_t): Ditto.
6564 (vfloat16m1x4_t): Ditto.
6565 (vfloat16m1x5_t): Ditto.
6566 (vfloat16m1x6_t): Ditto.
6567 (vfloat16m1x7_t): Ditto.
6568 (vfloat16m1x8_t): Ditto.
6569 (vfloat16m2x2_t): Ditto.
6570 (vfloat16m2x3_t): Ditto.
6571 (vfloat16m2x4_t): Ditto.
6572 (vfloat16m4x2_t): Ditto.
6573 * config/riscv/riscv-vector-switch.def (TUPLE_ENTRY): New.
6574 * config/riscv/riscv.md: New.
6575 * config/riscv/vector-iterators.md: New.
6577 2023-07-04 Andrew Pinski <apinski@marvell.com>
6579 PR tree-optimization/110487
6580 * match.pd (a !=/== CST1 ? CST2 : CST3): Always
6581 build a nonstandard integer and use that.
6583 2023-07-04 Andrew Pinski <apinski@marvell.com>
6585 * match.pd (a?-1:0): Cast type an integer type
6586 rather the type before the negative.
6589 2023-07-04 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
6591 * config/xtensa/xtensa.cc (machine_function, xtensa_expand_prologue):
6592 Change to use HARD_REG_BIT and its macros.
6593 * config/xtensa/xtensa.md
6594 (peephole2: regmove elimination during DFmode input reload):
6597 2023-07-04 Richard Biener <rguenther@suse.de>
6599 PR tree-optimization/110491
6600 * tree-ssa-phiopt.cc (match_simplify_replacement): Check
6601 whether the PHI args are possibly undefined before folding
6604 2023-07-04 Pan Li <pan2.li@intel.com>
6605 Thomas Schwinge <thomas@codesourcery.com>
6607 * lto-streamer-in.cc (lto_input_mode_table): Stream in the mode
6608 bits for machine mode table.
6609 * lto-streamer-out.cc (lto_write_mode_table): Stream out the
6610 HOST machine mode bits.
6611 * lto-streamer.h (struct lto_file_decl_data): New fields mode_bits.
6612 * tree-streamer.cc (streamer_mode_table): Take MAX_MACHINE_MODE
6614 * tree-streamer.h (streamer_mode_table): Ditto.
6615 (bp_pack_machine_mode): Take 1 << ceil_log2 (MAX_MACHINE_MODE)
6616 as the packing limit.
6617 (bp_unpack_machine_mode): Ditto with 'file_data->mode_bits'.
6619 2023-07-04 Thomas Schwinge <thomas@codesourcery.com>
6621 * lto-streamer.h (class lto_input_block): Capture
6622 'lto_file_decl_data *file_data' instead of just
6623 'unsigned char *mode_table'.
6624 * ipa-devirt.cc (ipa_odr_read_section): Adjust.
6625 * ipa-fnsummary.cc (inline_read_section): Likewise.
6626 * ipa-icf.cc (sem_item_optimizer::read_section): Likewise.
6627 * ipa-modref.cc (read_section): Likewise.
6628 * ipa-prop.cc (ipa_prop_read_section, read_replacements_section):
6630 * ipa-sra.cc (isra_read_summary_section): Likewise.
6631 * lto-cgraph.cc (input_cgraph_opt_section): Likewise.
6632 * lto-section-in.cc (lto_create_simple_input_block): Likewise.
6633 * lto-streamer-in.cc (lto_read_body_or_constructor)
6634 (lto_input_toplevel_asms): Likewise.
6635 * tree-streamer.h (bp_unpack_machine_mode): Likewise.
6637 2023-07-04 Richard Biener <rguenther@suse.de>
6639 * tree-ssa-phiopt.cc (pass_phiopt::execute): Mark SSA undefs.
6640 (empty_bb_or_one_feeding_into_p): Check for them.
6641 * tree-ssa.h (gimple_uses_undefined_value_p): Remove.
6642 * tree-ssa.cc (gimple_uses_undefined_value_p): Likewise.
6644 2023-07-04 Richard Biener <rguenther@suse.de>
6646 * tree-vect-loop.cc (vect_analyze_loop_costing): Remove
6647 check guarding scalar_niter underflow.
6649 2023-07-04 Hao Liu <hliu@os.amperecomputing.com>
6651 PR tree-optimization/110531
6652 * tree-vect-loop.cc (vect_analyze_loop_1): initialize
6653 slp_done_for_suggested_uf to false.
6655 2023-07-04 Richard Biener <rguenther@suse.de>
6657 PR tree-optimization/110228
6658 * tree-ssa-ifcombine.cc (pass_tree_ifcombine::execute):
6659 Mark SSA may-undefs.
6660 (bb_no_side_effects_p): Check stmt uses for undefs.
6662 2023-07-04 Richard Biener <rguenther@suse.de>
6664 PR tree-optimization/110436
6665 * tree-vect-stmts.cc (vect_mark_relevant): Expand dumping,
6666 force live but not relevant pattern stmts relevant.
6668 2023-07-04 Lili Cui <lili.cui@intel.com>
6670 * config/i386/i386.h: Add PTA_ENQCMD and PTA_UINTR to PTA_SIERRAFOREST.
6671 * doc/invoke.texi: Update new isa to march=sierraforest and grandridge.
6673 2023-07-04 Richard Biener <rguenther@suse.de>
6675 PR middle-end/110495
6676 * tree.h (TREE_OVERFLOW): Do not mention VECTOR_CSTs
6677 since we do not set TREE_OVERFLOW on those since the
6678 introduction of VL vectors.
6679 * match.pd (x +- CST +- CST): For VECTOR_CST do not look
6680 at TREE_OVERFLOW to determine validity of association.
6682 2023-07-04 Richard Biener <rguenther@suse.de>
6684 PR tree-optimization/110310
6685 * tree-vect-loop.cc (vect_determine_partial_vectors_and_peeling):
6686 Move costing part ...
6687 (vect_analyze_loop_costing): ... here. Integrate better
6688 estimate for epilogues from ...
6689 (vect_analyze_loop_2): Call vect_determine_partial_vectors_and_peeling
6690 with actual epilogue status.
6691 * tree-vect-loop-manip.cc (vect_do_peeling): ... here and
6692 avoid cancelling epilogue vectorization.
6693 (vect_update_epilogue_niters): Remove. No longer update
6694 epilogue LOOP_VINFO_NITERS.
6696 2023-07-04 Pan Li <pan2.li@intel.com>
6699 2023-07-03 Pan Li <pan2.li@intel.com>
6701 * config/riscv/vector.md: Fix typo.
6703 2023-07-04 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
6705 * doc/md.texi: Add len_mask_gather_load/len_mask_scatter_store.
6706 * internal-fn.cc (expand_scatter_store_optab_fn): Ditto.
6707 (expand_gather_load_optab_fn): Ditto.
6708 (internal_load_fn_p): Ditto.
6709 (internal_store_fn_p): Ditto.
6710 (internal_gather_scatter_fn_p): Ditto.
6711 (internal_fn_len_index): Ditto.
6712 (internal_fn_mask_index): Ditto.
6713 (internal_fn_stored_value_index): Ditto.
6714 * internal-fn.def (LEN_MASK_GATHER_LOAD): Ditto.
6715 (LEN_MASK_SCATTER_STORE): Ditto.
6716 * optabs.def (OPTAB_CD): Ditto.
6718 2023-07-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
6720 * config/riscv/riscv-vsetvl.cc
6721 (vector_insn_info::parse_insn): Add early break.
6723 2023-07-04 Hans-Peter Nilsson <hp@axis.com>
6725 * config/cris/cris.md (CRIS_UNSPEC_SWAP_BITS): Remove.
6726 ("cris_swap_bits", "ctzsi2"): Use bitreverse instead.
6728 2023-07-04 Hans-Peter Nilsson <hp@axis.com>
6730 * dwarf2out.cc (mem_loc_descriptor): Handle BITREVERSE.
6732 2023-07-03 Christoph Müllner <christoph.muellner@vrull.eu>
6734 * common/config/riscv/riscv-common.cc: Add support for zvbb,
6735 zvbc, zvkg, zvkned, zvknha, zvknhb, zvksed, zvksh, zvkn,
6736 zvknc, zvkng, zvks, zvksc, zvksg, zvkt and the implied subsets.
6737 * config/riscv/arch-canonicalize: Add canonicalization info for
6738 zvkn, zvknc, zvkng, zvks, zvksc, zvksg.
6739 * config/riscv/riscv-opts.h (MASK_ZVBB): New macro.
6740 (MASK_ZVBC): Likewise.
6741 (TARGET_ZVBB): Likewise.
6742 (TARGET_ZVBC): Likewise.
6743 (MASK_ZVKG): Likewise.
6744 (MASK_ZVKNED): Likewise.
6745 (MASK_ZVKNHA): Likewise.
6746 (MASK_ZVKNHB): Likewise.
6747 (MASK_ZVKSED): Likewise.
6748 (MASK_ZVKSH): Likewise.
6749 (MASK_ZVKN): Likewise.
6750 (MASK_ZVKNC): Likewise.
6751 (MASK_ZVKNG): Likewise.
6752 (MASK_ZVKS): Likewise.
6753 (MASK_ZVKSC): Likewise.
6754 (MASK_ZVKSG): Likewise.
6755 (MASK_ZVKT): Likewise.
6756 (TARGET_ZVKG): Likewise.
6757 (TARGET_ZVKNED): Likewise.
6758 (TARGET_ZVKNHA): Likewise.
6759 (TARGET_ZVKNHB): Likewise.
6760 (TARGET_ZVKSED): Likewise.
6761 (TARGET_ZVKSH): Likewise.
6762 (TARGET_ZVKN): Likewise.
6763 (TARGET_ZVKNC): Likewise.
6764 (TARGET_ZVKNG): Likewise.
6765 (TARGET_ZVKS): Likewise.
6766 (TARGET_ZVKSC): Likewise.
6767 (TARGET_ZVKSG): Likewise.
6768 (TARGET_ZVKT): Likewise.
6769 * config/riscv/riscv.opt: Introduction of riscv_zv{b,k}_subext.
6771 2023-07-03 Andrew Pinski <apinski@marvell.com>
6773 PR middle-end/110510
6774 * except.h (struct eh_landing_pad_d): Add chain_next GTY.
6776 2023-07-03 Iain Sandoe <iain@sandoe.co.uk>
6778 * config/darwin.h: Avoid duplicate multiply_defined specs on
6779 earlier Darwin versions with shared libgcc.
6781 2023-07-03 Uros Bizjak <ubizjak@gmail.com>
6783 * tree.h (tree_int_cst_equal): Change return type from int to bool.
6784 (operand_equal_for_phi_arg_p): Ditto.
6785 (tree_map_base_marked_p): Ditto.
6786 * tree.cc (contains_placeholder_p): Update function body
6787 for bool return type.
6788 (type_cache_hasher::equal): Ditto.
6789 (tree_map_base_hash): Change return type
6790 from int to void and adjust function body accordingly.
6791 (tree_int_cst_equal): Ditto.
6792 (operand_equal_for_phi_arg_p): Ditto.
6793 (get_narrower): Change "first" variable to bool.
6794 (cl_option_hasher::equal): Update function body for bool return type.
6795 * ggc.h (ggc_set_mark): Change return type from int to bool.
6796 (ggc_marked_p): Ditto.
6797 * ggc-page.cc (gt_ggc_mx): Change return type
6798 from int to void and adjust function body accordingly.
6799 (ggc_set_mark): Ditto.
6801 2023-07-03 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
6803 * config/riscv/autovec.md: Change order of
6804 LEN_MASK_LOAD/LEN_MASK_STORE/LEN_LOAD/LEN_STORE arguments.
6805 * config/riscv/riscv-v.cc (expand_load_store): Ditto.
6806 * doc/md.texi: Ditto.
6807 * gimple-fold.cc (gimple_fold_partial_load_store_mem_ref): Ditto.
6808 * internal-fn.cc (len_maskload_direct): Ditto.
6809 (len_maskstore_direct): Ditto.
6810 (add_len_and_mask_args): New function.
6811 (expand_partial_load_optab_fn): Change order of
6812 LEN_MASK_LOAD/LEN_MASK_STORE/LEN_LOAD/LEN_STORE arguments.
6813 (expand_partial_store_optab_fn): Ditto.
6814 (internal_fn_len_index): New function.
6815 (internal_fn_mask_index): Change order of
6816 LEN_MASK_LOAD/LEN_MASK_STORE/LEN_LOAD/LEN_STORE arguments.
6817 (internal_fn_stored_value_index): Ditto.
6818 (internal_len_load_store_bias): Ditto.
6819 * internal-fn.h (internal_fn_len_index): New function.
6820 * tree-ssa-dse.cc (initialize_ao_ref_for_dse): Change order of
6821 LEN_MASK_LOAD/LEN_MASK_STORE/LEN_LOAD/LEN_STORE arguments.
6822 * tree-vect-stmts.cc (vectorizable_store): Ditto.
6823 (vectorizable_load): Ditto.
6825 2023-07-03 Gaius Mulley <gaiusmod2@gmail.com>
6828 * doc/gm2.texi (Semantic checking): Include examples using
6829 -Wuninit-variable-checking.
6831 2023-07-03 Juzhe-Zhong <juzhe.zhong@rivai.ai>
6833 * config/riscv/autovec-opt.md (*double_widen_fnma<mode>): New pattern.
6834 (*single_widen_fnma<mode>): Ditto.
6835 (*double_widen_fms<mode>): Ditto.
6836 (*single_widen_fms<mode>): Ditto.
6837 (*double_widen_fnms<mode>): Ditto.
6838 (*single_widen_fnms<mode>): Ditto.
6840 2023-07-03 Juzhe-Zhong <juzhe.zhong@rivai.ai>
6842 * config/riscv/autovec-opt.md (@pred_single_widen_mul<any_extend:su><mode>): Change "@"
6843 into "*" in pattern name which simplifies build files.
6844 (*pred_single_widen_mul<any_extend:su><mode>): Ditto.
6845 (*pred_single_widen_mul<mode>): New pattern.
6847 2023-07-03 Richard Sandiford <richard.sandiford@arm.com>
6849 * config/aarch64/aarch64-simd.md (vec_extract<mode><Vhalf>): Expect
6850 the index to be 0 or 1.
6852 2023-07-03 Lehua Ding <lehua.ding@rivai.ai>
6855 2023-07-03 Juzhe-Zhong <juzhe.zhong@rivai.ai>
6857 * config/riscv/autovec-opt.md (*double_widen_fnma<mode>): New pattern.
6858 (*single_widen_fnma<mode>): Ditto.
6859 (*double_widen_fms<mode>): Ditto.
6860 (*single_widen_fms<mode>): Ditto.
6861 (*double_widen_fnms<mode>): Ditto.
6862 (*single_widen_fnms<mode>): Ditto.
6864 2023-07-03 Juzhe-Zhong <juzhe.zhong@rivai.ai>
6866 * config/riscv/autovec-opt.md (*double_widen_fnma<mode>): New pattern.
6867 (*single_widen_fnma<mode>): Ditto.
6868 (*double_widen_fms<mode>): Ditto.
6869 (*single_widen_fms<mode>): Ditto.
6870 (*double_widen_fnms<mode>): Ditto.
6871 (*single_widen_fnms<mode>): Ditto.
6873 2023-07-03 Pan Li <pan2.li@intel.com>
6875 * config/riscv/vector.md: Fix typo.
6877 2023-07-03 Richard Biener <rguenther@suse.de>
6879 PR tree-optimization/110506
6880 * tree-vect-patterns.cc (vect_recog_rotate_pattern): Re-order
6881 TYPE_PRECISION access with INTEGRAL_TYPE_P check.
6883 2023-07-03 Richard Biener <rguenther@suse.de>
6885 PR tree-optimization/110506
6886 * tree-ssa-ccp.cc (get_value_for_expr): Check for integral
6887 type before relying on TYPE_PRECISION to produce a nonzero mask.
6889 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
6891 * config/mips/mips.md(*and<mode>3_mips16): Generates
6892 ZEB/ZEH instructions.
6894 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
6896 * config/mips/mips.cc(mips_9bit_offset_address_p): Restrict the
6897 address register to M16_REGS for MIPS16.
6898 (BUILTIN_AVAIL_MIPS16E2): Defined a new macro.
6899 (AVAIL_MIPS16E2_OR_NON_MIPS16): Same as above.
6900 (AVAIL_NON_MIPS16 (cache..)): Update to
6901 AVAIL_MIPS16E2_OR_NON_MIPS16.
6902 * config/mips/mips.h (ISA_HAS_CACHE): Add clause for ISA_HAS_MIPS16E2.
6903 * config/mips/mips.md (mips_cache): Mark as extended MIPS16.
6905 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
6907 * config/mips/mips.h(ISA_HAS_9BIT_DISPLACEMENT): Add clause
6908 for ISA_HAS_MIPS16E2.
6909 (ISA_HAS_SYNC): Same as above.
6910 (ISA_HAS_LL_SC): Same as above.
6912 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
6914 * config/mips/mips.cc(mips_expand_ins_as_unaligned_store):
6915 Add logics for generating instruction.
6916 * config/mips/mips.h(ISA_HAS_LWL_LWR): Add clause for ISA_HAS_MIPS16E2.
6917 * config/mips/mips.md(mov_<load>l): Generates instructions.
6918 (mov_<load>r): Same as above.
6919 (mov_<store>l): Adjusted for the conditions above.
6920 (mov_<store>r): Same as above.
6921 (mov_<store>l_mips16e2): Add machine description for `define_insn mov_<store>l_mips16e2`.
6922 (mov_<store>r_mips16e2): Add machine description for `define_insn mov_<store>r_mips16e2`.
6924 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
6926 * config/mips/mips.cc(mips_symbol_insns_1): Generates LUI instruction.
6927 (mips_const_insns): Same as above.
6928 (mips_output_move): Same as above.
6929 (mips_output_function_prologue): Same as above.
6930 * config/mips/mips.md: Same as above
6932 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
6934 * config/mips/constraints.md(Yz): New constraints for mips16e2.
6935 * config/mips/mips-protos.h(mips_bit_clear_p): Declared new function.
6936 (mips_bit_clear_info): Same as above.
6937 * config/mips/mips.cc(mips_bit_clear_info): New function for
6938 generating instructions.
6939 (mips_bit_clear_p): Same as above.
6940 * config/mips/mips.h(ISA_HAS_EXT_INS): Add clause for ISA_HAS_MIPS16E2.
6941 * config/mips/mips.md(extended_mips16): Generates EXT and INS instructions.
6942 (*and<mode>3): Generates INS instruction.
6943 (*and<mode>3_mips16): Generates EXT, INS and ANDI instructions.
6944 (ior<mode>3): Add logics for ORI instruction.
6945 (*ior<mode>3_mips16_asmacro): Generates ORI instrucion.
6946 (*ior<mode>3_mips16): Add logics for XORI instruction.
6947 (*xor<mode>3_mips16): Generates XORI instrucion.
6948 (*extzv<mode>): Add logics for EXT instruction.
6949 (*insv<mode>): Add logics for INS instruction.
6950 * config/mips/predicates.md(bit_clear_operand): New predicate for
6951 generating bitwise instructions.
6952 (and_reg_operand): Add logics for generating bitwise instructions.
6954 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
6956 * config/mips/mips.cc(mips_regno_mode_ok_for_base_p): Generate instructions
6957 that uses global pointer register.
6958 (mips16_unextended_reference_p): Same as above.
6959 (mips_pic_base_register): Same as above.
6960 (mips_init_relocs): Same as above.
6961 * config/mips/mips.h(MIPS16_GP_LOADS): Defined a new macro.
6962 (GLOBAL_POINTER_REGNUM): Moved to machine description `mips.md`.
6963 * config/mips/mips.md(GLOBAL_POINTER_REGNUM): Moved to here from above.
6964 (*lowsi_mips16_gp):New `define_insn *low<mode>_mips16`.
6966 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
6968 * config/mips/mips.h(ISA_HAS_CONDMOVE): Add condition for ISA_HAS_MIPS16E2.
6969 * config/mips/mips.md(*mov<GPR:mode>_on_<MOVECC:mode>): Add logics for MOVx insts.
6970 (*mov<GPR:mode>_on_<MOVECC:mode>_mips16e2): Generate MOVx instruction.
6971 (*mov<GPR:mode>_on_<GPR2:mode>_ne): Add logics for MOVx insts.
6972 (*mov<GPR:mode>_on_<GPR2:mode>_ne_mips16e2): Generate MOVx instruction.
6973 * config/mips/predicates.md(reg_or_0_operand_mips16e2): New predicate for MOVx insts.
6975 2023-07-03 Jie Mei <jie.mei@oss.cipunited.com>
6977 * config/mips/mips.cc(mips_file_start): Add mips16e2 info
6979 * config/mips/mips.h(__mips_mips16e2): Defined a new
6981 (ISA_HAS_MIPS16E2): Defined a new macro.
6982 (ASM_SPEC): Pass mmips16e2 to the assembler.
6983 * config/mips/mips.opt: Add -m(no-)mips16e2 option.
6984 * config/mips/predicates.md: Add clause for TARGET_MIPS16E2.
6985 * doc/invoke.texi: Add -m(no-)mips16e2 option..
6987 2023-07-02 Jakub Jelinek <jakub@redhat.com>
6989 PR tree-optimization/110508
6990 * tree-ssa-math-opts.cc (match_uaddc_usubc): Only replace re2 with
6991 REALPART_EXPR opf nlhs if re2 is non-NULL.
6993 2023-07-02 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
6995 * config/xtensa/xtensa.cc (xtensa_match_CLAMPS_imms_p):
6997 * config/xtensa/xtensa.md (*xtensa_clamps):
6998 Add TARGET_MINMAX to the condition.
7000 2023-07-02 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
7002 * config/xtensa/xtensa.md (*eqne_INT_MIN):
7003 Add missing ":SI" to the match_operator.
7005 2023-07-02 Iain Sandoe <iain@sandoe.co.uk>
7008 * config/darwin.opt: Add fconstant-cfstrings alias to
7009 mconstant-cfstrings.
7010 * doc/invoke.texi: Amend invocation descriptions to reflect
7011 that the fconstant-cfstrings is a target-option alias and to
7012 add the missing mconstant-cfstrings option description to the
7015 2023-07-01 Jan Hubicka <jh@suse.cz>
7017 * tree-cfg.cc (gimple_duplicate_sese_region): Add elliminated_edge
7018 parmaeter; update profile.
7019 * tree-cfg.h (gimple_duplicate_sese_region): Update prototype.
7020 * tree-ssa-loop-ch.cc (entry_loop_condition_is_static): Rename to ...
7021 (static_loop_exit): ... this; return the edge to be elliminated.
7022 (ch_base::copy_headers): Handle profile updating for eliminated exits.
7024 2023-07-01 Roger Sayle <roger@nextmovesoftware.com>
7026 * config/i386/i386-features.cc (compute_convert_gain): Provide
7027 gains/costs for ROTATE and ROTATERT (by an integer constant).
7028 (general_scalar_chain::convert_rotate): New helper function to
7029 convert a DImode or SImode rotation by an integer constant into
7031 (general_scalar_chain::convert_insn): Call the new convert_rotate
7032 for ROTATE and ROTATERT.
7033 (general_scalar_to_vector_candidate_p): Consider ROTATE and
7034 ROTATERT to be candidates if the second operand is an integer
7035 constant, valid for a rotation (or shift) in the given mode.
7036 * config/i386/i386-features.h (general_scalar_chain): Add new
7037 helper method convert_rotate.
7039 2023-07-01 Jan Hubicka <jh@suse.cz>
7041 PR tree-optimization/103680
7042 * cfg.cc (update_bb_profile_for_threading): Fix profile update;
7043 make message clearer.
7045 2023-06-30 Qing Zhao <qing.zhao@oracle.com>
7047 PR tree-optimization/101832
7048 * tree-object-size.cc (addr_object_size): Handle structure/union type
7049 when it has flexible size.
7051 2023-06-30 Eric Botcazou <ebotcazou@adacore.com>
7053 * gimple-fold.cc (fold_array_ctor_reference): Fix head comment.
7054 (fold_nonarray_ctor_reference): Likewise. Specifically deal
7055 with integral bit-fields.
7056 (fold_ctor_reference): Make sure that the constructor uses the
7057 native storage order.
7059 2023-06-30 Jan Hubicka <jh@suse.cz>
7061 PR middle-end/109849
7062 * predict.cc (estimate_bb_frequencies): Turn to static function.
7063 (expr_expected_value_1): Fix handling of binary expressions with
7065 * predict.def (PRED_MALLOC_NONNULL): Move later in the priority queue.
7066 (PRED_BUILTIN_EXPECT_WITH_PROBABILITY): Move to almost top of the priority
7068 * predict.h (estimate_bb_frequencies): No longer declare it.
7070 2023-06-30 Uros Bizjak <ubizjak@gmail.com>
7072 * fold-const.h (multiple_of_p): Change return type from int to bool.
7073 * fold-const.cc (split_tree): Change negl_p, neg_litp_p,
7074 neg_conp_p and neg_var_p variables to bool.
7075 (const_binop): Change sat_p variable to bool.
7076 (merge_ranges): Change no_overlap variable to bool.
7077 (extract_muldiv_1): Change same_p variable to bool.
7078 (tree_swap_operands_p): Update function body for bool return type.
7079 (fold_truth_andor): Change commutative variable to bool.
7080 (multiple_of_p): Change return type
7081 from int to void and adjust function body accordingly.
7082 * optabs.h (expand_twoval_unop): Change return type from int to bool.
7083 (expand_twoval_binop): Ditto.
7084 (can_compare_p): Ditto.
7085 (have_add2_insn): Ditto.
7086 (have_addptr3_insn): Ditto.
7087 (have_sub2_insn): Ditto.
7088 (have_insn_for): Ditto.
7089 * optabs.cc (add_equal_note): Ditto.
7090 (widen_operand): Change no_extend argument from int to bool.
7091 (expand_binop): Ditto.
7092 (expand_twoval_unop): Change return type
7093 from int to void and adjust function body accordingly.
7094 (expand_twoval_binop): Ditto.
7095 (can_compare_p): Ditto.
7096 (have_add2_insn): Ditto.
7097 (have_addptr3_insn): Ditto.
7098 (have_sub2_insn): Ditto.
7099 (have_insn_for): Ditto.
7101 2023-06-30 Oluwatamilore Adebayo <oluwatamilore.adebayo@arm.com>
7103 * config/aarch64/aarch64-simd.md
7104 (vec_widen_<su>abdl_lo_<mode>, vec_widen_<su>abdl_hi_<mode>):
7105 Expansions for abd vec widen optabs.
7106 (aarch64_<su>abdl<mode>_insn): VQW based abdl RTL.
7107 * config/aarch64/iterators.md (USMAX_EXT): Code attributes
7108 that give the appropriate extend RTL for the max RTL.
7110 2023-06-30 Oluwatamilore Adebayo <oluwatamilore.adebayo@arm.com>
7112 * internal-fn.def (VEC_WIDEN_ABD): New internal hilo optab.
7113 * optabs.def (vec_widen_sabd_optab,
7114 vec_widen_sabd_hi_optab, vec_widen_sabd_lo_optab,
7115 vec_widen_sabd_odd_even, vec_widen_sabd_even_optab,
7116 vec_widen_uabd_optab,
7117 vec_widen_uabd_hi_optab, vec_widen_uabd_lo_optab,
7118 vec_widen_uabd_odd_even, vec_widen_uabd_even_optab):
7120 * doc/md.texi: Document them.
7121 * tree-vect-patterns.cc (vect_recog_abd_pattern): Update to
7122 to build a VEC_WIDEN_ABD call if the input precision is smaller
7123 than the precision of the output.
7124 (vect_recog_widen_abd_pattern): Should an ABD expression be
7125 found preceeding an extension, replace the two with a
7128 2023-06-30 Pan Li <pan2.li@intel.com>
7130 * config/riscv/vector.md: Refactor the common condition.
7132 2023-06-30 Richard Biener <rguenther@suse.de>
7134 PR tree-optimization/110496
7135 * gimple-ssa-store-merging.cc (find_bswap_or_nop_1): Re-order
7136 verifying and TYPE_PRECISION query for the BIT_FIELD_REF case.
7138 2023-06-30 Richard Biener <rguenther@suse.de>
7140 PR middle-end/110489
7141 * statistics.cc (curr_statistics_hash): Add argument
7142 indicating whether we should allocate the hash.
7143 (statistics_fini_pass): If the hash isn't allocated
7144 only print the summary header.
7146 2023-06-30 Segher Boessenkool <segher@kernel.crashing.org>
7147 Thomas Schwinge <thomas@codesourcery.com>
7149 * config/nvptx/nvptx.cc (TARGET_LRA_P): Remove.
7151 2023-06-30 Jovan Dmitrović <jovan.dmitrovic@syrmia.com>
7154 * config/mips/mips.cc (mips_function_arg_alignment): Returns
7155 the alignment of function argument. In case of typedef type,
7156 it returns the aligment of the aliased type.
7157 (mips_function_arg_boundary): Relocated calculation of the
7158 aligment of function arguments.
7160 2023-06-29 Jan Hubicka <jh@suse.cz>
7162 PR tree-optimization/109849
7163 * ipa-fnsummary.cc (decompose_param_expr): Skip
7164 functions returning its parameter.
7165 (set_cond_stmt_execution_predicate): Return early
7166 if predicate was constructed.
7168 2023-06-29 Qing Zhao <qing.zhao@oracle.com>
7171 * doc/extend.texi: Document GCC extension on a structure containing
7172 a flexible array member to be a member of another structure.
7174 2023-06-29 Qing Zhao <qing.zhao@oracle.com>
7176 * print-tree.cc (print_node): Print new bit type_include_flexarray.
7177 * tree-core.h (struct tree_type_common): Use bit no_named_args_stdarg_p
7178 as type_include_flexarray for RECORD_TYPE or UNION_TYPE.
7179 * tree-streamer-in.cc (unpack_ts_type_common_value_fields): Stream
7180 in bit no_named_args_stdarg_p properly for its corresponding type.
7181 * tree-streamer-out.cc (pack_ts_type_common_value_fields): Stream
7182 out bit no_named_args_stdarg_p properly for its corresponding type.
7183 * tree.h (TYPE_INCLUDES_FLEXARRAY): New macro TYPE_INCLUDES_FLEXARRAY.
7185 2023-06-29 Aldy Hernandez <aldyh@redhat.com>
7187 * tree-vrp.cc (maybe_set_nonzero_bits): Move from here...
7188 * tree-ssa-dom.cc (maybe_set_nonzero_bits): ...to here.
7189 * tree-vrp.h (maybe_set_nonzero_bits): Remove.
7191 2023-06-29 Aldy Hernandez <aldyh@redhat.com>
7193 * value-range.cc (frange::set): Do not call verify_range.
7194 (frange::normalize_kind): Verify range.
7195 (frange::union_nans): Do not call verify_range.
7196 (frange::union_): Same.
7197 (frange::intersect): Same.
7198 (irange::irange_single_pair_union): Call normalize_kind if
7200 (irange::union_): Same.
7201 (irange::intersect): Same.
7202 (irange::set_range_from_nonzero_bits): Verify range.
7203 (irange::set_nonzero_bits): Call normalize_kind if necessary.
7204 (irange::get_nonzero_bits): Tweak comment.
7205 (irange::intersect_nonzero_bits): Call normalize_kind if
7207 (irange::union_nonzero_bits): Same.
7208 * value-range.h (irange::normalize_kind): Verify range.
7210 2023-06-29 Uros Bizjak <ubizjak@gmail.com>
7212 * cselib.h (rtx_equal_for_cselib_1):
7213 Change return type from int to bool.
7214 (references_value_p): Ditto.
7215 (rtx_equal_for_cselib_p): Ditto.
7216 * expr.h (can_store_by_pieces): Ditto.
7217 (try_casesi): Ditto.
7218 (try_tablejump): Ditto.
7219 (safe_from_p): Ditto.
7220 * sbitmap.h (bitmap_equal_p): Ditto.
7221 * cselib.cc (references_value_p): Change return type
7222 from int to void and adjust function body accordingly.
7223 (rtx_equal_for_cselib_1): Ditto.
7224 * expr.cc (is_aligning_offset): Ditto.
7225 (can_store_by_pieces): Ditto.
7226 (mostly_zeros_p): Ditto.
7227 (all_zeros_p): Ditto.
7228 (safe_from_p): Ditto.
7229 (is_aligning_offset): Ditto.
7230 (try_casesi): Ditto.
7231 (try_tablejump): Ditto.
7232 (store_constructor): Change "need_to_clear" and
7233 "const_bounds_p" variables to bool.
7234 * sbitmap.cc (bitmap_equal_p): Change return type from int to bool.
7236 2023-06-29 Robin Dapp <rdapp@ventanamicro.com>
7238 * tree-ssa-math-opts.cc (divmod_candidate_p): Use
7241 2023-06-29 Richard Biener <rguenther@suse.de>
7243 PR tree-optimization/110460
7244 * tree-vect-stmts.cc (get_related_vectype_for_scalar_type):
7245 Only allow integral, pointer and scalar float type scalar_type.
7247 2023-06-29 Lili Cui <lili.cui@intel.com>
7249 PR tree-optimization/110148
7250 * tree-ssa-reassoc.cc (rewrite_expr_tree_parallel): Handle loop-carried
7251 ops in this function.
7253 2023-06-29 Richard Biener <rguenther@suse.de>
7255 PR middle-end/110452
7256 * expr.cc (store_constructor): Handle uniform boolean
7257 vectors with integer mode specially.
7259 2023-06-29 Richard Biener <rguenther@suse.de>
7261 PR middle-end/110461
7262 * match.pd (bitop (convert@2 @0) (convert?@3 @1)): Disable
7265 2023-06-29 Richard Sandiford <richard.sandiford@arm.com>
7267 * vec.h (gt_pch_nx): Add overloads for va_gc_atomic.
7268 (array_slice): Relax va_gc constructor to handle all vectors
7269 with a vl_embed layout.
7271 2023-06-29 Pan Li <pan2.li@intel.com>
7273 * config/riscv/riscv.cc (riscv_emit_mode_set): Add emit for FRM.
7274 (riscv_mode_needed): Likewise.
7275 (riscv_entity_mode_after): Likewise.
7276 (riscv_mode_after): Likewise.
7277 (riscv_mode_entry): Likewise.
7278 (riscv_mode_exit): Likewise.
7279 * config/riscv/riscv.h (NUM_MODES_FOR_MODE_SWITCHING): Add number
7281 * config/riscv/riscv.md: Add FRM register.
7282 * config/riscv/vector-iterators.md: Add FRM type.
7283 * config/riscv/vector.md (frm_mode): Define new attr for FRM mode.
7284 (fsrm): Define new insn for fsrm instruction.
7286 2023-06-29 Pan Li <pan2.li@intel.com>
7288 * config/riscv/riscv-protos.h (enum floating_point_rounding_mode):
7289 Add macro for static frm min and max.
7290 * config/riscv/riscv-vector-builtins-bases.cc
7291 (class binop_frm): New class for floating-point with frm.
7292 (BASE): Add vfadd for frm.
7293 * config/riscv/riscv-vector-builtins-bases.h: Likewise.
7294 * config/riscv/riscv-vector-builtins-functions.def
7295 (vfadd_frm): Likewise.
7296 * config/riscv/riscv-vector-builtins-shapes.cc
7297 (struct alu_frm_def): New struct for alu with frm.
7298 (SHAPE): Add alu with frm.
7299 * config/riscv/riscv-vector-builtins-shapes.h: Likewise.
7300 * config/riscv/riscv-vector-builtins.cc
7301 (function_checker::report_out_of_range_and_not): New function
7302 for report out of range and not val.
7303 (function_checker::require_immediate_range_or): New function
7304 for checking in range or one val.
7305 * config/riscv/riscv-vector-builtins.h: Add function decl.
7307 2023-06-29 Cui, Lili <lili.cui@intel.com>
7309 * common/config/i386/cpuinfo.h (get_intel_cpu): Remove model value 0xa8
7310 from Rocketlake, move model value 0xbf from Alderlake to Raptorlake.
7312 2023-06-28 Hans-Peter Nilsson <hp@axis.com>
7315 * config/cris/cris.cc (cris_postdbr_cmpelim): Don't apply PATTERN
7316 to insn before validating it.
7318 2023-06-28 Jan Hubicka <jh@suse.cz>
7320 PR middle-end/110334
7321 * ipa-fnsummary.h (ipa_fn_summary): Add
7322 safe_to_inline_to_always_inline.
7323 * ipa-inline.cc (can_early_inline_edge_p): ICE
7324 if SSA is not built; do cycle checking for
7325 always_inline functions.
7326 (inline_always_inline_functions): Be recrusive;
7327 watch for cycles; do not updat overall summary.
7328 (early_inliner): Do not give up on always_inlines.
7329 * ipa-utils.cc (ipa_reverse_postorder): Do not skip
7332 2023-06-28 Uros Bizjak <ubizjak@gmail.com>
7334 * output.h (leaf_function_p): Change return type from int to bool.
7335 (final_forward_branch_p): Ditto.
7336 (only_leaf_regs_used): Ditto.
7337 (maybe_assemble_visibility): Ditto.
7338 * varasm.h (supports_one_only): Ditto.
7339 * rtl.h (compute_alignments): Change return type from int to void.
7340 * final.cc (app_on): Change return type from int to bool.
7341 (compute_alignments): Change return type from int to void
7342 and adjust function body accordingly.
7343 (shorten_branches): Change "something_changed" variable
7344 type from int to bool.
7345 (leaf_function_p): Change return type from int to bool
7346 and adjust function body accordingly.
7347 (final_forward_branch_p): Ditto.
7348 (only_leaf_regs_used): Ditto.
7349 * varasm.cc (contains_pointers_p): Change return type from
7350 int to bool and adjust function body accordingly.
7351 (compare_constant): Ditto.
7352 (maybe_assemble_visibility): Ditto.
7353 (supports_one_only): Ditto.
7355 2023-06-28 Manolis Tsamis <manolis.tsamis@vrull.eu>
7358 * regcprop.cc (maybe_mode_change): Check stack_pointer_rtx mode.
7359 (maybe_copy_reg_attrs): New function.
7360 (find_oldest_value_reg): Use maybe_copy_reg_attrs.
7361 (copyprop_hardreg_forward_1): Ditto.
7363 2023-06-28 Richard Biener <rguenther@suse.de>
7365 PR tree-optimization/110434
7366 * tree-nrv.cc (pass_nrv::execute): Remove CLOBBERs of
7367 VAR we replace with <retval>.
7369 2023-06-28 Richard Biener <rguenther@suse.de>
7371 PR tree-optimization/110451
7372 * tree-ssa-loop-im.cc (stmt_cost): [VEC_]COND_EXPR and
7373 tcc_comparison are expensive.
7375 2023-06-28 Roger Sayle <roger@nextmovesoftware.com>
7377 * config/i386/i386-expand.cc (ix86_expand_branch): Also use ptest
7378 for TImode comparisons on 32-bit architectures.
7379 * config/i386/i386.md (cbranch<mode>4): Change from SDWIM to
7380 SWIM1248x to exclude/avoid TImode being conditional on -m64.
7381 (cbranchti4): New define_expand for TImode on both TARGET_64BIT
7382 and/or with TARGET_SSE4_1.
7383 * config/i386/predicates.md (ix86_timode_comparison_operator):
7384 New predicate that depends upon TARGET_64BIT.
7385 (ix86_timode_comparison_operand): Likewise.
7387 2023-06-28 Roger Sayle <roger@nextmovesoftware.com>
7390 * config/i386/i386-features.cc (compute_convert_gain): Provide
7391 more accurate gains for conversion of scalar comparisons to
7394 2023-06-28 Richard Biener <rguenther@suse.de>
7396 PR tree-optimization/110443
7397 * tree-vect-slp.cc (vect_build_slp_tree_1): Reject non-grouped
7400 2023-06-28 Haochen Gui <guihaoc@gcc.gnu.org>
7402 * config/rs6000/rs6000.md (peephole2 for compare_and_move): New.
7403 (peephole2 for move_and_compare): New.
7404 (mode_iterator WORD): New. Set the mode to SI/DImode by
7406 (*mov<mode>_internal2): Change the mode iterator from P to WORD.
7407 (split pattern for compare_and_move): Likewise.
7409 2023-06-28 Juzhe-Zhong <juzhe.zhong@rivai.ai>
7411 * config/riscv/autovec-opt.md (*double_widen_fma<mode>): New pattern.
7412 (*single_widen_fma<mode>): Ditto.
7414 2023-06-28 Haochen Gui <guihaoc@gcc.gnu.org>
7417 * config/rs6000/altivec.md (*altivec_vupkhs<VU_char>_direct): Rename
7419 (altivec_vupkhs<VU_char>_direct): ...this.
7420 * config/rs6000/predicates.md (vspltisw_vupkhsw_constant_split): New
7421 predicate to test if a constant can be loaded with vspltisw and
7423 (easy_vector_constant): Call vspltisw_vupkhsw_constant_p to Check if
7424 a vector constant can be synthesized with a vspltisw and a vupkhsw.
7425 * config/rs6000/rs6000-protos.h (vspltisw_vupkhsw_constant_p):
7427 * config/rs6000/rs6000.cc (vspltisw_vupkhsw_constant_p): New
7428 function to return true if OP mode is V2DI and can be synthesized
7429 with vupkhsw and vspltisw.
7430 * config/rs6000/vsx.md (*vspltisw_v2di_split): New insn to load up
7431 constants with vspltisw and vupkhsw.
7433 2023-06-28 Jan Hubicka <jh@suse.cz>
7435 PR tree-optimization/110377
7436 * ipa-prop.cc (ipa_compute_jump_functions_for_edge): Pass statement to
7438 (ipa_analyze_node): Enable ranger.
7440 2023-06-28 Richard Biener <rguenther@suse.de>
7442 * tree.h (TYPE_PRECISION): Check for non-VECTOR_TYPE.
7443 (TYPE_PRECISION_RAW): Provide raw access to the precision
7445 * tree.cc (verify_type_variant): Compare TYPE_PRECISION_RAW.
7446 (gimple_canonical_types_compatible_p): Likewise.
7447 * tree-streamer-out.cc (pack_ts_type_common_value_fields):
7448 Stream TYPE_PRECISION_RAW.
7449 * tree-streamer-in.cc (unpack_ts_type_common_value_fields):
7451 * lto-streamer-out.cc (hash_tree): Hash TYPE_PRECISION_RAW.
7453 2023-06-28 Alexandre Oliva <oliva@adacore.com>
7455 * doc/extend.texi (zero-call-used-regs): Document leafy and
7457 * flag-types.h (zero_regs_flags): Add LEAFY_MODE, as well as
7459 * function.cc (gen_call_ued_regs_seq): Set only_used for leaf
7460 functions in leafy mode.
7461 * opts.cc (zero_call_used_regs_opts): Add leafy and variants.
7463 2023-06-28 Juzhe-Zhong <juzhe.zhong@rivai.ai>
7465 * config/riscv/riscv-vector-builtins-bases.cc: Adapt expand.
7466 * config/riscv/vector.md (@pred_single_widen_<plus_minus:optab><mode>):
7468 (@pred_single_widen_add<mode>): New pattern.
7469 (@pred_single_widen_sub<mode>): New pattern.
7471 2023-06-28 liuhongt <hongtao.liu@intel.com>
7473 * config/i386/i386.cc (ix86_invalid_conversion): New function.
7474 (TARGET_INVALID_CONVERSION): Define as
7475 ix86_invalid_conversion.
7477 2023-06-27 Robin Dapp <rdapp@ventanamicro.com>
7479 * config/riscv/autovec.md (<optab><vnconvert><mode>2): New
7481 (<float_cvt><vnconvert><mode>2): Ditto.
7482 (<optab><mode><vnconvert>2): Ditto.
7483 (<float_cvt><mode><vnconvert>2): Ditto.
7484 * config/riscv/vector-iterators.md: Add vnconvert.
7486 2023-06-27 Robin Dapp <rdapp@ventanamicro.com>
7488 * config/riscv/autovec.md (extend<v_double_trunc><mode>2): New
7490 (extend<v_quad_trunc><mode>2): Ditto.
7491 (trunc<mode><v_double_trunc>2): Ditto.
7492 (trunc<mode><v_quad_trunc>2): Ditto.
7493 * config/riscv/vector-iterators.md: Add VQEXTF and HF to
7494 V_QUAD_TRUNC and v_quad_trunc.
7496 2023-06-27 Robin Dapp <rdapp@ventanamicro.com>
7498 * config/riscv/autovec.md (<float_cvt><vconvert><mode>2): New
7501 2023-06-27 Robin Dapp <rdapp@ventanamicro.com>
7503 * config/riscv/autovec.md (copysign<mode>3): Add expander.
7504 (xorsign<mode>3): Ditto.
7505 * config/riscv/riscv-vector-builtins-bases.cc (class vfsgnjn):
7507 * config/riscv/vector-iterators.md (copysign): Remove ncopysign.
7511 * config/riscv/vector.md (@pred_ncopysign<mode>): Split off.
7512 (@pred_ncopysign<mode>_scalar): Ditto.
7514 2023-06-27 Robin Dapp <rdapp@ventanamicro.com>
7516 * config/riscv/autovec.md: VF_AUTO -> VF.
7517 * config/riscv/vector-iterators.md: Introduce VF_ZVFHMIN,
7518 VWEXTF_ZVFHMIN and use TARGET_ZVFH in VWCONVERTI, VHF and
7520 * config/riscv/vector.md: Use new iterators.
7522 2023-06-27 Robin Dapp <rdapp@ventanamicro.com>
7524 * match.pd: Use element_mode and check if target supports
7525 operation with new type.
7527 2023-06-27 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
7529 * config/aarch64/aarch64-sve-builtins-base.cc
7530 (svdupq_impl::fold_nonconst_dupq): New method.
7531 (svdupq_impl::fold): Call fold_nonconst_dupq.
7533 2023-06-27 Andrew Pinski <apinski@marvell.com>
7535 PR middle-end/110420
7536 PR middle-end/103979
7538 * gimplify.cc (gimplify_asm_expr): Mark asm with labels as volatile.
7540 2023-06-27 Aldy Hernandez <aldyh@redhat.com>
7542 * ipa-cp.cc (decide_whether_version_node): Adjust comment.
7543 * ipa-fnsummary.cc (evaluate_conditions_for_known_args): Adjust
7545 (set_switch_stmt_execution_predicate): Same.
7546 * ipa-prop.cc (ipa_compute_jump_functions_for_edge): Same.
7548 2023-06-27 Aldy Hernandez <aldyh@redhat.com>
7550 * ipa-prop.cc (struct ipa_vr_ggc_hash_traits): Adjust for use with
7551 ipa_vr instead of value_range.
7554 (ipa_get_value_range): Same.
7555 * value-range.cc (gt_pch_nx): Move to ipa-prop.cc and adjust for
7559 2023-06-27 Aldy Hernandez <aldyh@redhat.com>
7561 * ipa-cp.cc (ipa_vr_operation_and_type_effects): New.
7562 * ipa-prop.cc (ipa_get_value_range): Adjust for ipa_vr.
7563 (ipa_set_jfunc_vr): Take a range.
7564 (ipa_compute_jump_functions_for_edge): Pass range to
7566 (ipa_write_jump_function): Call streamer write helper.
7567 (ipa_read_jump_function): Call streamer read helper.
7568 * ipa-prop.h (class ipa_vr): Change m_vr to an ipa_vr.
7570 2023-06-27 Richard Sandiford <richard.sandiford@arm.com>
7572 * gengtype-parse.cc (consume_until_comma_or_eos): Parse "= { ... }"
7573 as a probable initializer rather than a probable complete statement.
7575 2023-06-27 Richard Biener <rguenther@suse.de>
7577 PR tree-optimization/96208
7578 * tree-vect-slp.cc (vect_build_slp_tree_1): Allow
7579 a non-grouped load if it is the same for all lanes.
7580 (vect_build_slp_tree_2): Handle not grouped loads.
7581 (vect_optimize_slp_pass::remove_redundant_permutations):
7583 (vect_transform_slp_perm_load_1): Likewise.
7584 * tree-vect-stmts.cc (vect_model_load_cost): Likewise.
7585 (get_group_load_store_type): Likewise. Handle
7587 (vectorizable_load): Likewise.
7589 2023-06-27 liuhongt <hongtao.liu@intel.com>
7591 PR rtl-optimization/110237
7592 * config/i386/sse.md (<avx512>_store<mode>_mask): Refine with
7594 (maskstore<mode><avx512fmaskmodelower): Ditto.
7595 (*<avx512>_store<mode>_mask): New define_insn, it's renamed
7596 from original <avx512>_store<mode>_mask.
7598 2023-06-27 liuhongt <hongtao.liu@intel.com>
7600 * config/i386/i386-features.cc (pass_insert_vzeroupper:gate):
7601 Move flag_expensive_optimizations && !optimize_size to ..
7602 * config/i386/i386-options.cc (ix86_option_override_internal):
7603 .. this, it makes -mvzeroupper independent of optimization
7604 level, but still keeps the behavior of architecture
7605 tuning(emit_vzeroupper) unchanged.
7607 2023-06-27 liuhongt <hongtao.liu@intel.com>
7610 * config/i386/i386.cc (ix86_avx_u127_mode_needed): Don't emit
7611 vzeroupper for vzeroupper call_insn.
7613 2023-06-27 Andrew Pinski <apinski@marvell.com>
7615 * doc/extend.texi (__builtin_alloca_with_align_and_max): Fix
7618 2023-06-27 Juzhe-Zhong <juzhe.zhong@rivai.ai>
7620 * config/riscv/riscv-v.cc (expand_const_vector): Fix stepped vector
7623 2023-06-26 Andrew Pinski <apinski@marvell.com>
7625 * doc/extend.texi (access attribute): Add
7627 (interrupt/interrupt_handler attribute):
7630 2023-06-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7632 * config/aarch64/aarch64-simd.md (aarch64_sqrshrun_n<mode>_insn):
7633 Use <DWI> instead of <V2XWIDE>.
7634 (aarch64_sqrshrun_n<mode>): Likewise.
7636 2023-06-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7638 * config/aarch64/aarch64-protos.h (aarch64_const_vec_rsra_rnd_imm_p):
7640 (aarch64_rnd_imm_p): ... This.
7641 * config/aarch64/predicates.md (aarch64_simd_rsra_rnd_imm_vec):
7643 (aarch64_int_rnd_operand): ... This.
7644 (aarch64_simd_rshrn_imm_vec): Delete.
7645 * config/aarch64/aarch64-simd.md (aarch64_<sra_op>rsra_n<mode>_insn):
7646 Adjust for the above.
7647 (aarch64_<sra_op>rshr_n<mode><vczle><vczbe>_insn): Likewise.
7648 (*aarch64_<shrn_op>rshrn_n<mode>_insn): Likewise.
7649 (*aarch64_sqrshrun_n<mode>_insn<vczle><vczbe>): Likewise.
7650 (aarch64_sqrshrun_n<mode>_insn): Likewise.
7651 (aarch64_<shrn_op>rshrn2_n<mode>_insn_le): Likewise.
7652 (aarch64_<shrn_op>rshrn2_n<mode>_insn_be): Likewise.
7653 (aarch64_sqrshrun2_n<mode>_insn_le): Likewise.
7654 (aarch64_sqrshrun2_n<mode>_insn_be): Likewise.
7655 * config/aarch64/aarch64.cc (aarch64_const_vec_rsra_rnd_imm_p):
7657 (aarch64_rnd_imm_p): ... This.
7659 2023-06-26 Andreas Krebbel <krebbel@linux.ibm.com>
7661 * config/s390/s390.cc (s390_encode_section_info): Set
7662 SYMBOL_FLAG_SET_NOTALIGN2 only if the symbol has explicitely been
7665 2023-06-26 Jan Hubicka <jh@suse.cz>
7667 PR tree-optimization/109849
7668 * tree-ssa-dce.cc (make_forwarders_with_degenerate_phis): Fix profile
7669 count of newly constructed forwarder block.
7671 2023-06-26 Andrew Carlotti <andrew.carlotti@arm.com>
7673 * doc/optinfo.texi: Fix "steam" -> "stream".
7675 2023-06-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
7677 * tree-ssa-dse.cc (initialize_ao_ref_for_dse): Add LEN_MASK_STORE and
7679 (dse_optimize_stmt): Add LEN_MASK_STORE.
7681 2023-06-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
7683 * gimple-fold.cc (gimple_fold_partial_load_store_mem_ref): Fix gimple
7684 fold of LOAD/STORE with length.
7686 2023-06-26 Andrew MacLeod <amacleod@redhat.com>
7688 * gimple-range-gori.cc (compute_operand1_and_operand2_range):
7689 Check for interdependence between operands 1 and 2.
7691 2023-06-26 Richard Sandiford <richard.sandiford@arm.com>
7693 * tree-vect-stmts.cc (vectorizable_conversion): Take multi_step_cvt
7694 into account when costing non-widening/truncating conversions.
7696 2023-06-26 Richard Biener <rguenther@suse.de>
7698 PR tree-optimization/110381
7699 * tree-vect-slp.cc (vect_optimize_slp_pass::start_choosing_layouts):
7700 Materialize permutes before fold-left reductions.
7702 2023-06-26 Pan Li <pan2.li@intel.com>
7704 * config/riscv/riscv-vector-builtins-bases.h: Remove duplicated decl.
7706 2023-06-26 Richard Biener <rguenther@suse.de>
7708 * varasm.cc (initializer_constant_valid_p_1): Also
7709 constrain the type of value to be scalar integral
7710 before dispatching to narrowing_initializer_constant_valid_p.
7712 2023-06-26 Richard Biener <rguenther@suse.de>
7714 * tree-ssa-scopedtables.cc (hashable_expr_equal_p):
7715 Use element_precision.
7717 2023-06-26 Juzhe-Zhong <juzhe.zhong@rivai.ai>
7719 * config/riscv/autovec.md (vcond<V:mode><VI:mode>): Remove redundant
7721 (vcondu<V:mode><VI:mode>): Ditto.
7722 * config/riscv/riscv-protos.h (expand_vcond): Ditto.
7723 * config/riscv/riscv-v.cc (expand_vcond): Ditto.
7725 2023-06-26 Richard Biener <rguenther@suse.de>
7727 PR tree-optimization/110392
7728 * gimple-predicate-analysis.cc (uninit_analysis::is_use_guarded):
7729 Do early exits on true/false predicate only after normalization.
7731 2023-06-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
7733 * tree-ssa-sccvn.cc (vn_reference_lookup_3): Change name "len" into
7736 2023-06-26 Roger Sayle <roger@nextmovesoftware.com>
7738 * config/i386/i386.md (peephole2): Simplify zeroing a register
7739 followed by an IOR, XOR or PLUS operation on it, into a move.
7740 (*ashl<dwi>3_doubleword_highpart): New define_insn_and_split to
7741 eliminate (and hide from reload) unnecessary word to doubleword
7742 extensions that are followed by left shifts by sufficiently large,
7743 but valid, bit counts.
7745 2023-06-26 liuhongt <hongtao.liu@intel.com>
7747 PR tree-optimization/110371
7748 PR tree-optimization/110018
7749 * tree-vect-stmts.cc (vectorizable_conversion): Use cvt_op to
7750 save intermediate type operand instead of "subtle" vec_dest
7753 2023-06-26 liuhongt <hongtao.liu@intel.com>
7755 PR tree-optimization/110371
7756 PR tree-optimization/110018
7757 * tree-vect-stmts.cc (vectorizable_conversion): Don't use
7758 intermiediate type for FIX_TRUNC_EXPR when ftrapping-math.
7760 2023-06-26 Hongyu Wang <hongyu.wang@intel.com>
7762 * config/i386/i386-options.cc (ix86_valid_target_attribute_tree):
7763 Override tune_string with arch_string if tune_string is not
7764 explicitly specified.
7766 2023-06-25 Juzhe-Zhong <juzhe.zhong@rivai.ai>
7768 * config/riscv/riscv-vsetvl.cc (vector_insn_info::parse_insn): Ehance
7770 * config/riscv/riscv-vsetvl.h: New function.
7772 2023-06-25 Li Xu <xuli1@eswincomputing.com>
7774 * config/riscv/riscv-vector-builtins-bases.cc: change emit_insn to
7777 2023-06-25 Juzhe-Zhong <juzhe.zhong@rivai.ai>
7779 * config/riscv/autovec.md (len_load_<mode>): Remove.
7780 (len_maskload<mode><vm>): Remove.
7781 (len_store_<mode>): New pattern.
7782 (len_maskstore<mode><vm>): New pattern.
7783 * config/riscv/predicates.md (autovec_length_operand): New predicate.
7784 * config/riscv/riscv-protos.h (enum insn_type): New enum.
7785 (expand_load_store): New function.
7786 * config/riscv/riscv-v.cc (emit_vlmax_masked_insn): Ditto.
7787 (emit_nonvlmax_masked_insn): Ditto.
7788 (expand_load_store): Ditto.
7789 * config/riscv/riscv-vector-builtins.cc
7790 (function_expander::use_contiguous_store_insn): Add avl_type operand
7792 * config/riscv/vector.md: Ditto.
7794 2023-06-25 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
7796 * internal-fn.cc (expand_partial_store_optab_fn): Fix bug of BIAS
7799 2023-06-25 Pan Li <pan2.li@intel.com>
7801 * config/riscv/vector.md: Revert.
7803 2023-06-25 Pan Li <pan2.li@intel.com>
7805 * config/riscv/genrvv-type-indexer.cc (valid_type): Revert changes.
7806 * config/riscv/riscv-modes.def (RVV_TUPLE_MODES): Ditto.
7807 (ADJUST_ALIGNMENT): Ditto.
7808 (RVV_TUPLE_PARTIAL_MODES): Ditto.
7809 (ADJUST_NUNITS): Ditto.
7810 * config/riscv/riscv-vector-builtins-types.def (vfloat16mf4x2_t): Ditto.
7811 (vfloat16mf4x3_t): Ditto.
7812 (vfloat16mf4x4_t): Ditto.
7813 (vfloat16mf4x5_t): Ditto.
7814 (vfloat16mf4x6_t): Ditto.
7815 (vfloat16mf4x7_t): Ditto.
7816 (vfloat16mf4x8_t): Ditto.
7817 (vfloat16mf2x2_t): Ditto.
7818 (vfloat16mf2x3_t): Ditto.
7819 (vfloat16mf2x4_t): Ditto.
7820 (vfloat16mf2x5_t): Ditto.
7821 (vfloat16mf2x6_t): Ditto.
7822 (vfloat16mf2x7_t): Ditto.
7823 (vfloat16mf2x8_t): Ditto.
7824 (vfloat16m1x2_t): Ditto.
7825 (vfloat16m1x3_t): Ditto.
7826 (vfloat16m1x4_t): Ditto.
7827 (vfloat16m1x5_t): Ditto.
7828 (vfloat16m1x6_t): Ditto.
7829 (vfloat16m1x7_t): Ditto.
7830 (vfloat16m1x8_t): Ditto.
7831 (vfloat16m2x2_t): Ditto.
7832 (vfloat16m2x3_t): Diito.
7833 (vfloat16m2x4_t): Diito.
7834 (vfloat16m4x2_t): Diito.
7835 * config/riscv/riscv-vector-builtins.def (vfloat16mf4x2_t): Ditto.
7836 (vfloat16mf4x3_t): Ditto.
7837 (vfloat16mf4x4_t): Ditto.
7838 (vfloat16mf4x5_t): Ditto.
7839 (vfloat16mf4x6_t): Ditto.
7840 (vfloat16mf4x7_t): Ditto.
7841 (vfloat16mf4x8_t): Ditto.
7842 (vfloat16mf2x2_t): Ditto.
7843 (vfloat16mf2x3_t): Ditto.
7844 (vfloat16mf2x4_t): Ditto.
7845 (vfloat16mf2x5_t): Ditto.
7846 (vfloat16mf2x6_t): Ditto.
7847 (vfloat16mf2x7_t): Ditto.
7848 (vfloat16mf2x8_t): Ditto.
7849 (vfloat16m1x2_t): Ditto.
7850 (vfloat16m1x3_t): Ditto.
7851 (vfloat16m1x4_t): Ditto.
7852 (vfloat16m1x5_t): Ditto.
7853 (vfloat16m1x6_t): Ditto.
7854 (vfloat16m1x7_t): Ditto.
7855 (vfloat16m1x8_t): Ditto.
7856 (vfloat16m2x2_t): Ditto.
7857 (vfloat16m2x3_t): Ditto.
7858 (vfloat16m2x4_t): Ditto.
7859 (vfloat16m4x2_t): Ditto.
7860 * config/riscv/riscv-vector-switch.def (TUPLE_ENTRY): Ditto.
7861 * config/riscv/riscv.md: Ditto.
7862 * config/riscv/vector-iterators.md: Ditto.
7864 2023-06-25 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
7866 * gimple-fold.cc (arith_overflowed_p): Apply LEN_MASK_{LOAD,STORE}.
7867 (gimple_fold_partial_load_store_mem_ref): Ditto.
7868 (gimple_fold_partial_store): Ditto.
7869 (gimple_fold_call): Ditto.
7871 2023-06-25 liuhongt <hongtao.liu@intel.com>
7874 * config/i386/sse.md (maskload<mode><avx512fmaskmodelower>):
7875 Refine pattern with UNSPEC_MASKLOAD.
7876 (maskload<mode><avx512fmaskmodelower>): Ditto.
7877 (*<avx512>_load<mode>_mask): Extend mode iterator to
7879 (*<avx512>_load<mode>): Ditto.
7881 2023-06-25 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
7883 * tree-ssa-alias.cc (call_may_clobber_ref_p_1): Add LEN_MASK_STORE.
7885 2023-06-25 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
7887 * tree-ssa-alias.cc (ref_maybe_used_by_call_p_1): Apply
7888 LEN_MASK_{LOAD,STORE}
7890 2023-06-25 yulong <shiyulong@iscas.ac.cn>
7892 * config/riscv/vector.md: Add float16 attr at sew、vlmul and ratio.
7894 2023-06-24 Roger Sayle <roger@nextmovesoftware.com>
7896 * config/i386/i386.md (*<code>qi_ext<mode>_3): New define_insn.
7898 2023-06-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
7900 * config/riscv/autovec.md (*fma<mode>): set clobber to Pmode in expand stage.
7901 (*fma<VI:mode><P:mode>): Ditto.
7902 (*fnma<mode>): Ditto.
7903 (*fnma<VI:mode><P:mode>): Ditto.
7905 2023-06-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
7907 * config/riscv/autovec.md (fma<mode>4): New pattern.
7908 (*fma<mode>): Ditto.
7909 (fnma<mode>4): Ditto.
7910 (*fnma<mode>): Ditto.
7911 (fms<mode>4): Ditto.
7912 (*fms<mode>): Ditto.
7913 (fnms<mode>4): Ditto.
7914 (*fnms<mode>): Ditto.
7915 * config/riscv/riscv-protos.h (emit_vlmax_fp_ternary_insn):
7917 * config/riscv/riscv-v.cc (emit_vlmax_fp_ternary_insn): Ditto.
7918 * config/riscv/vector.md: Fix attribute bug.
7920 2023-06-24 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
7922 * tree-ssa-loop-ivopts.cc (get_mem_type_for_internal_fn):
7923 Apply LEN_MASK_{LOAD,STORE}.
7925 2023-06-24 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
7927 * tree-ssa-loop-ivopts.cc (get_alias_ptr_type_for_ptr_address):
7928 Add LEN_MASK_{LOAD,STORE}.
7930 2023-06-24 David Malcolm <dmalcolm@redhat.com>
7932 * diagnostic-format-sarif.cc: Add #define INCLUDE_VECTOR.
7933 * diagnostic.cc: Likewise.
7934 * text-art/box-drawing.cc: Likewise.
7935 * text-art/canvas.cc: Likewise.
7936 * text-art/ruler.cc: Likewise.
7937 * text-art/selftests.cc: Likewise.
7938 * text-art/selftests.h (text_art::canvas): New forward decl.
7939 * text-art/style.cc: Add #define INCLUDE_VECTOR.
7940 * text-art/styled-string.cc: Likewise.
7941 * text-art/table.cc: Likewise.
7942 * text-art/table.h: Remove #include <vector>.
7943 * text-art/theme.cc: Add #define INCLUDE_VECTOR.
7944 * text-art/types.h: Check that INCLUDE_VECTOR is defined.
7945 Remove #include of <vector> and <string>.
7946 * text-art/widget.cc: Add #define INCLUDE_VECTOR.
7947 * text-art/widget.h: Remove #include <vector>.
7949 2023-06-24 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
7951 * internal-fn.cc (expand_partial_store_optab_fn): Adapt for LEN_MASK_STORE.
7952 (internal_load_fn_p): Add LEN_MASK_LOAD.
7953 (internal_store_fn_p): Add LEN_MASK_STORE.
7954 (internal_fn_mask_index): Add LEN_MASK_{LOAD,STORE}.
7955 (internal_fn_stored_value_index): Add LEN_MASK_STORE.
7956 (internal_len_load_store_bias): Add LEN_MASK_{LOAD,STORE}.
7957 * optabs-tree.cc (can_vec_mask_load_store_p): Adapt for LEN_MASK_{LOAD,STORE}.
7958 (get_len_load_store_mode): Ditto.
7959 * optabs-tree.h (can_vec_mask_load_store_p): Ditto.
7960 (get_len_load_store_mode): Ditto.
7961 * tree-vect-stmts.cc (check_load_store_for_partial_vectors): Ditto.
7962 (get_all_ones_mask): New function.
7963 (vectorizable_store): Apply LEN_MASK_{LOAD,STORE} into vectorizer.
7964 (vectorizable_load): Ditto.
7966 2023-06-23 Marek Polacek <polacek@redhat.com>
7968 * doc/cpp.texi (__cplusplus): Document value for -std=c++26 and
7969 -std=gnu++26. Document that for C++23, its value is 202302L.
7970 * doc/invoke.texi: Document -std=c++26 and -std=gnu++26.
7971 * dwarf2out.cc (highest_c_language): Handle GNU C++26.
7972 (gen_compile_unit_die): Likewise.
7974 2023-06-23 Jan Hubicka <jh@suse.cz>
7976 * tree-ssa-phiprop.cc (propagate_with_phi): Compute post dominators on
7978 (pass_phiprop::execute): Do not compute it here; return
7979 update_ssa_only_virtuals if something changed.
7980 (pass_data_phiprop): Remove TODO_update_ssa from todos.
7982 2023-06-23 Michael Meissner <meissner@linux.ibm.com>
7983 Aaron Sawdey <acsawdey@linux.ibm.com>
7986 * config/rs6000/genfusion.pl (gen_ld_cmpi_p10_one): Fix problems that
7987 allowed prefixed lwa to be generated.
7988 * config/rs6000/fusion.md: Regenerate.
7989 * config/rs6000/predicates.md (ds_form_mem_operand): Delete.
7990 * config/rs6000/rs6000.md (prefixed attribute): Add support for load
7991 plus compare immediate fused insns.
7992 (maybe_prefixed): Likewise.
7994 2023-06-23 Roger Sayle <roger@nextmovesoftware.com>
7996 * simplify-rtx.cc (simplify_subreg): Optimize lowpart SUBREGs
7997 of ASHIFT to const0_rtx with sufficiently large shift count.
7998 Optimize highpart SUBREGs of ASHIFT as the shift operand when
7999 the shift count is the correct offset. Optimize SUBREGs of
8000 multi-word logic operations if the SUBREGs of both operands
8003 2023-06-23 Richard Biener <rguenther@suse.de>
8005 * varasm.cc (initializer_constant_valid_p_1): Only
8006 allow conversions between scalar floating point types.
8008 2023-06-23 Richard Biener <rguenther@suse.de>
8010 * tree-vect-stmts.cc (vectorizable_assignment):
8011 Properly handle non-integral operands when analyzing
8014 2023-06-23 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
8016 PR tree-optimization/110280
8017 * match.pd (vec_perm_expr(v, v, mask) -> v): Explicitly build vector
8018 using build_vector_from_val with the element of input operand, and
8019 mask's type if operand and mask's types don't match.
8021 2023-06-23 Richard Biener <rguenther@suse.de>
8023 * fold-const.cc (tree_simple_nonnegative_warnv_p): Guard
8024 the truth_value_p case with !VECTOR_TYPE_P.
8026 2023-06-23 Richard Biener <rguenther@suse.de>
8028 * tree-vect-patterns.cc (vect_look_through_possible_promotion):
8029 Exit early when the type isn't scalar integral.
8031 2023-06-23 Richard Biener <rguenther@suse.de>
8033 * match.pd ((outertype)((innertype0)a+(innertype1)b)
8034 -> ((newtype)a+(newtype)b)): Use element_precision
8037 2023-06-23 Richard Biener <rguenther@suse.de>
8039 * fold-const.cc (fold_binary_loc): Use element_precision
8040 when trying (double)float1 CMP (double)float2 to
8041 float1 CMP float2 simplification.
8042 * match.pd: Likewise.
8044 2023-06-23 Richard Biener <rguenther@suse.de>
8046 * tree-vect-stmts.cc (vectorizable_load): Avoid useless
8047 copies of VMAT_INVARIANT vectorized stmts, fix SLP support.
8049 2023-06-23 Richard Biener <rguenther@suse.de>
8051 * tree-vect-stmts.cc (vector_vector_composition_type):
8052 Handle composition of a vector from a number of elements that
8053 happens to match its number of lanes.
8055 2023-06-22 Marek Polacek <polacek@redhat.com>
8057 * configure.ac (--enable-host-bind-now): New check. Add
8058 -Wl,-z,now to LD_PICFLAG if --enable-host-bind-now.
8059 * configure: Regenerate.
8060 * doc/install.texi: Document --enable-host-bind-now.
8062 2023-06-22 Di Zhao OS <dizhao@os.amperecomputing.com>
8064 * config/aarch64/aarch64.cc: Change fma_reassoc_width for ampere1.
8066 2023-06-22 Richard Biener <rguenther@suse.de>
8068 PR tree-optimization/110332
8069 * tree-ssa-phiprop.cc (propagate_with_phi): Always
8070 check aliasing with edge inserted loads.
8072 2023-06-22 Roger Sayle <roger@nextmovesoftware.com>
8073 Uros Bizjak <ubizjak@gmail.com>
8075 * config/i386/i386-expand.cc (ix86_expand_sse_ptest): Recognize
8076 expansion of ptestc with equal operands as producing const1_rtx.
8077 * config/i386/i386.cc (ix86_rtx_costs): Provide accurate cost
8078 estimates of UNSPEC_PTEST, where the ptest performs the PAND
8079 or PAND of its operands.
8080 * config/i386/sse.md (define_split): Transform CCCmode UNSPEC_PTEST
8081 of reg_equal_p operands into an x86_stc instruction.
8082 (define_split): Split pandn/ptestz/set{n?}e into ptestc/set{n?}c.
8083 (define_split): Similar to above for strict_low_part destinations.
8084 (define_split): Split pandn/ptestz/j{n?}e into ptestc/j{n?}c.
8086 2023-06-22 David Malcolm <dmalcolm@redhat.com>
8089 * Makefile.in (ANALYZER_OBJS): Add analyzer/access-diagram.o.
8090 * doc/invoke.texi (Wanalyzer-out-of-bounds): Add description of
8092 (fanalyzer-debug-text-art): New.
8094 2023-06-22 David Malcolm <dmalcolm@redhat.com>
8096 * Makefile.in (OBJS-libcommon): Add text-art/box-drawing.o,
8097 text-art/canvas.o, text-art/ruler.o, text-art/selftests.o,
8098 text-art/style.o, text-art/styled-string.o, text-art/table.o,
8099 text-art/theme.o, and text-art/widget.o.
8100 * color-macros.h (COLOR_FG_BRIGHT_BLACK): New.
8101 (COLOR_FG_BRIGHT_RED): New.
8102 (COLOR_FG_BRIGHT_GREEN): New.
8103 (COLOR_FG_BRIGHT_YELLOW): New.
8104 (COLOR_FG_BRIGHT_BLUE): New.
8105 (COLOR_FG_BRIGHT_MAGENTA): New.
8106 (COLOR_FG_BRIGHT_CYAN): New.
8107 (COLOR_FG_BRIGHT_WHITE): New.
8108 (COLOR_BG_BRIGHT_BLACK): New.
8109 (COLOR_BG_BRIGHT_RED): New.
8110 (COLOR_BG_BRIGHT_GREEN): New.
8111 (COLOR_BG_BRIGHT_YELLOW): New.
8112 (COLOR_BG_BRIGHT_BLUE): New.
8113 (COLOR_BG_BRIGHT_MAGENTA): New.
8114 (COLOR_BG_BRIGHT_CYAN): New.
8115 (COLOR_BG_BRIGHT_WHITE): New.
8116 * common.opt (fdiagnostics-text-art-charset=): New option.
8117 (diagnostic-text-art.h): New SourceInclude.
8118 (diagnostic_text_art_charset) New Enum and EnumValues.
8119 * configure: Regenerate.
8120 * configure.ac (gccdepdir): Add text-art to loop.
8121 * diagnostic-diagram.h: New file.
8122 * diagnostic-format-json.cc (json_emit_diagram): New.
8123 (diagnostic_output_format_init_json): Wire it up to
8124 context->m_diagrams.m_emission_cb.
8125 * diagnostic-format-sarif.cc: Include "diagnostic-diagram.h" and
8126 "text-art/canvas.h".
8127 (sarif_result::on_nested_diagnostic): Move code to...
8128 (sarif_result::add_related_location): ...this new function.
8129 (sarif_result::on_diagram): New.
8130 (sarif_builder::emit_diagram): New.
8131 (sarif_builder::make_message_object_for_diagram): New.
8132 (sarif_emit_diagram): New.
8133 (diagnostic_output_format_init_sarif): Set
8134 context->m_diagrams.m_emission_cb to sarif_emit_diagram.
8135 * diagnostic-text-art.h: New file.
8136 * diagnostic.cc: Include "diagnostic-text-art.h",
8137 "diagnostic-diagram.h", and "text-art/theme.h".
8138 (diagnostic_initialize): Initialize context->m_diagrams and
8139 call diagnostics_text_art_charset_init.
8140 (diagnostic_finish): Clean up context->m_diagrams.m_theme.
8141 (diagnostic_emit_diagram): New.
8142 (diagnostics_text_art_charset_init): New.
8143 * diagnostic.h (text_art::theme): New forward decl.
8144 (class diagnostic_diagram): Likewise.
8145 (diagnostic_context::m_diagrams): New field.
8146 (diagnostic_emit_diagram): New decl.
8147 * doc/invoke.texi (Diagnostic Message Formatting Options): Add
8148 -fdiagnostics-text-art-charset=.
8149 (-fdiagnostics-plain-output): Add
8150 -fdiagnostics-text-art-charset=none.
8151 * gcc.cc: Include "diagnostic-text-art.h".
8152 (driver_handle_option): Handle OPT_fdiagnostics_text_art_charset_.
8153 * opts-common.cc (decode_cmdline_options_to_array): Add
8154 "-fdiagnostics-text-art-charset=none" to expanded_args for
8155 -fdiagnostics-plain-output.
8156 * opts.cc: Include "diagnostic-text-art.h".
8157 (common_handle_option): Handle OPT_fdiagnostics_text_art_charset_.
8158 * pretty-print.cc (pp_unicode_character): New.
8159 * pretty-print.h (pp_unicode_character): New decl.
8160 * selftest-run-tests.cc: Include "text-art/selftests.h".
8161 (selftest::run_tests): Call text_art_tests.
8162 * text-art/box-drawing-chars.inc: New file, generated by
8163 contrib/unicode/gen-box-drawing-chars.py.
8164 * text-art/box-drawing.cc: New file.
8165 * text-art/box-drawing.h: New file.
8166 * text-art/canvas.cc: New file.
8167 * text-art/canvas.h: New file.
8168 * text-art/ruler.cc: New file.
8169 * text-art/ruler.h: New file.
8170 * text-art/selftests.cc: New file.
8171 * text-art/selftests.h: New file.
8172 * text-art/style.cc: New file.
8173 * text-art/styled-string.cc: New file.
8174 * text-art/table.cc: New file.
8175 * text-art/table.h: New file.
8176 * text-art/theme.cc: New file.
8177 * text-art/theme.h: New file.
8178 * text-art/types.h: New file.
8179 * text-art/widget.cc: New file.
8180 * text-art/widget.h: New file.
8182 2023-06-21 Uros Bizjak <ubizjak@gmail.com>
8184 * function.h (emit_initial_value_sets):
8185 Change return type from int to void.
8186 (aggregate_value_p): Change return type from int to bool.
8187 (prologue_contains): Ditto.
8188 (epilogue_contains): Ditto.
8189 (prologue_epilogue_contains): Ditto.
8190 * function.cc (temp_slot): Make "in_use" variable bool.
8191 (make_slot_available): Update for changed "in_use" variable.
8192 (assign_stack_temp_for_type): Ditto.
8193 (emit_initial_value_sets): Change return type from int to void
8194 and update function body accordingly.
8195 (instantiate_virtual_regs): Ditto.
8196 (rest_of_handle_thread_prologue_and_epilogue): Ditto.
8197 (safe_insn_predicate): Change return type from int to bool.
8198 (aggregate_value_p): Change return type from int to bool
8199 and update function body accordingly.
8200 (prologue_contains): Change return type from int to bool.
8201 (prologue_epilogue_contains): Ditto.
8203 2023-06-21 Alexander Monakov <amonakov@ispras.ru>
8205 * common.opt (fp_contract_mode) [on]: Remove fallback.
8206 * config/sh/sh.md (*fmasf4): Correct flag_fp_contract_mode test.
8207 * doc/invoke.texi (-ffp-contract): Update.
8208 * trans-mem.cc (diagnose_tm_1): Skip internal function calls.
8210 2023-06-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8212 * config/aarch64/aarch64-sve.md (mask_gather_load<mode><v_int_container>):
8213 Add alternatives to prefer to avoid same input and output Z register.
8214 (mask_gather_load<mode><v_int_container>): Likewise.
8215 (*mask_gather_load<mode><v_int_container>_<su>xtw_unpacked): Likewise.
8216 (*mask_gather_load<mode><v_int_container>_sxtw): Likewise.
8217 (*mask_gather_load<mode><v_int_container>_uxtw): Likewise.
8218 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_4HSI:mode><SVE_4BHI:mode>):
8220 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode><SVE_2BHSI:mode>):
8222 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
8223 <SVE_2BHSI:mode>_<ANY_EXTEND2:su>xtw_unpacked): Likewise.
8224 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
8225 <SVE_2BHSI:mode>_sxtw): Likewise.
8226 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
8227 <SVE_2BHSI:mode>_uxtw): Likewise.
8228 (@aarch64_ldff1_gather<mode>): Likewise.
8229 (@aarch64_ldff1_gather<mode>): Likewise.
8230 (*aarch64_ldff1_gather<mode>_sxtw): Likewise.
8231 (*aarch64_ldff1_gather<mode>_uxtw): Likewise.
8232 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx4_WIDE:mode>
8233 <VNx4_NARROW:mode>): Likewise.
8234 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
8235 <VNx2_NARROW:mode>): Likewise.
8236 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
8237 <VNx2_NARROW:mode>_sxtw): Likewise.
8238 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
8239 <VNx2_NARROW:mode>_uxtw): Likewise.
8240 * config/aarch64/aarch64-sve2.md (@aarch64_gather_ldnt<mode>): Likewise.
8241 (@aarch64_gather_ldnt_<ANY_EXTEND:optab><SVE_FULL_SDI:mode>
8242 <SVE_PARTIAL_I:mode>): Likewise.
8244 2023-06-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8246 * config/aarch64/aarch64-sve.md (mask_gather_load<mode><v_int_container>):
8247 Convert to compact alternatives syntax.
8248 (mask_gather_load<mode><v_int_container>): Likewise.
8249 (*mask_gather_load<mode><v_int_container>_<su>xtw_unpacked): Likewise.
8250 (*mask_gather_load<mode><v_int_container>_sxtw): Likewise.
8251 (*mask_gather_load<mode><v_int_container>_uxtw): Likewise.
8252 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_4HSI:mode><SVE_4BHI:mode>):
8254 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode><SVE_2BHSI:mode>):
8256 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
8257 <SVE_2BHSI:mode>_<ANY_EXTEND2:su>xtw_unpacked): Likewise.
8258 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
8259 <SVE_2BHSI:mode>_sxtw): Likewise.
8260 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
8261 <SVE_2BHSI:mode>_uxtw): Likewise.
8262 (@aarch64_ldff1_gather<mode>): Likewise.
8263 (@aarch64_ldff1_gather<mode>): Likewise.
8264 (*aarch64_ldff1_gather<mode>_sxtw): Likewise.
8265 (*aarch64_ldff1_gather<mode>_uxtw): Likewise.
8266 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx4_WIDE:mode>
8267 <VNx4_NARROW:mode>): Likewise.
8268 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
8269 <VNx2_NARROW:mode>): Likewise.
8270 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
8271 <VNx2_NARROW:mode>_sxtw): Likewise.
8272 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
8273 <VNx2_NARROW:mode>_uxtw): Likewise.
8274 * config/aarch64/aarch64-sve2.md (@aarch64_gather_ldnt<mode>): Likewise.
8275 (@aarch64_gather_ldnt_<ANY_EXTEND:optab><SVE_FULL_SDI:mode>
8276 <SVE_PARTIAL_I:mode>): Likewise.
8278 2023-06-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8281 2023-06-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8283 * config/aarch64/aarch64-sve.md (mask_gather_load<mode><v_int_container>):
8284 Convert to compact alternatives syntax.
8285 (mask_gather_load<mode><v_int_container>): Likewise.
8286 (*mask_gather_load<mode><v_int_container>_<su>xtw_unpacked): Likewise.
8287 (*mask_gather_load<mode><v_int_container>_sxtw): Likewise.
8288 (*mask_gather_load<mode><v_int_container>_uxtw): Likewise.
8289 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_4HSI:mode><SVE_4BHI:mode>):
8291 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode><SVE_2BHSI:mode>):
8293 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
8294 <SVE_2BHSI:mode>_<ANY_EXTEND2:su>xtw_unpacked): Likewise.
8295 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
8296 <SVE_2BHSI:mode>_sxtw): Likewise.
8297 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
8298 <SVE_2BHSI:mode>_uxtw): Likewise.
8299 (@aarch64_ldff1_gather<mode>): Likewise.
8300 (@aarch64_ldff1_gather<mode>): Likewise.
8301 (*aarch64_ldff1_gather<mode>_sxtw): Likewise.
8302 (*aarch64_ldff1_gather<mode>_uxtw): Likewise.
8303 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx4_WIDE:mode>
8304 <VNx4_NARROW:mode>): Likewise.
8305 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
8306 <VNx2_NARROW:mode>): Likewise.
8307 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
8308 <VNx2_NARROW:mode>_sxtw): Likewise.
8309 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
8310 <VNx2_NARROW:mode>_uxtw): Likewise.
8311 * config/aarch64/aarch64-sve2.md (@aarch64_gather_ldnt<mode>): Likewise.
8312 (@aarch64_gather_ldnt_<ANY_EXTEND:optab><SVE_FULL_SDI:mode>
8313 <SVE_PARTIAL_I:mode>): Likewise.
8315 2023-06-21 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
8317 * optabs-query.cc (can_vec_mask_load_store_p): Move to optabs-tree.cc.
8318 (get_len_load_store_mode): Ditto.
8319 * optabs-query.h (can_vec_mask_load_store_p): Move to optabs-tree.h.
8320 (get_len_load_store_mode): Ditto.
8321 * optabs-tree.cc (can_vec_mask_load_store_p): New function.
8322 (get_len_load_store_mode): Ditto.
8323 * optabs-tree.h (can_vec_mask_load_store_p): Ditto.
8324 (get_len_load_store_mode): Ditto.
8325 * tree-if-conv.cc: include optabs-tree instead of optabs-query
8327 2023-06-21 Richard Biener <rguenther@suse.de>
8329 * tree-ssa-loop-ivopts.cc (add_iv_candidate_for_use): Use
8330 split_constant_offset for the POINTER_PLUS_EXPR case.
8332 2023-06-21 Richard Biener <rguenther@suse.de>
8334 * tree-ssa-loop-ivopts.cc (record_group_use): Use
8335 split_constant_offset.
8337 2023-06-21 Richard Biener <rguenther@suse.de>
8339 * tree-loop-distribution.cc (classify_builtin_st): Use
8340 split_constant_offset.
8341 * tree-ssa-loop-ivopts.h (strip_offset): Remove.
8342 * tree-ssa-loop-ivopts.cc (strip_offset): Make static.
8344 2023-06-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8346 * config/aarch64/aarch64-sve.md (mask_gather_load<mode><v_int_container>):
8347 Convert to compact alternatives syntax.
8348 (mask_gather_load<mode><v_int_container>): Likewise.
8349 (*mask_gather_load<mode><v_int_container>_<su>xtw_unpacked): Likewise.
8350 (*mask_gather_load<mode><v_int_container>_sxtw): Likewise.
8351 (*mask_gather_load<mode><v_int_container>_uxtw): Likewise.
8352 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_4HSI:mode><SVE_4BHI:mode>):
8354 (@aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode><SVE_2BHSI:mode>):
8356 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
8357 <SVE_2BHSI:mode>_<ANY_EXTEND2:su>xtw_unpacked): Likewise.
8358 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
8359 <SVE_2BHSI:mode>_sxtw): Likewise.
8360 (*aarch64_gather_load_<ANY_EXTEND:optab><SVE_2HSDI:mode>
8361 <SVE_2BHSI:mode>_uxtw): Likewise.
8362 (@aarch64_ldff1_gather<mode>): Likewise.
8363 (@aarch64_ldff1_gather<mode>): Likewise.
8364 (*aarch64_ldff1_gather<mode>_sxtw): Likewise.
8365 (*aarch64_ldff1_gather<mode>_uxtw): Likewise.
8366 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx4_WIDE:mode>
8367 <VNx4_NARROW:mode>): Likewise.
8368 (@aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
8369 <VNx2_NARROW:mode>): Likewise.
8370 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
8371 <VNx2_NARROW:mode>_sxtw): Likewise.
8372 (*aarch64_ldff1_gather_<ANY_EXTEND:optab><VNx2_WIDE:mode>
8373 <VNx2_NARROW:mode>_uxtw): Likewise.
8374 * config/aarch64/aarch64-sve2.md (@aarch64_gather_ldnt<mode>): Likewise.
8375 (@aarch64_gather_ldnt_<ANY_EXTEND:optab><SVE_FULL_SDI:mode>
8376 <SVE_PARTIAL_I:mode>): Likewise.
8378 2023-06-21 Tamar Christina <tamar.christina@arm.com>
8381 * doc/md.texi: Replace backslashchar.
8383 2023-06-21 Richard Biener <rguenther@suse.de>
8385 * config/i386/i386.cc (ix86_vector_costs::finish_cost):
8386 Overload. For masked main loops make sure the vectorization
8387 factor isn't more than double the number of iterations.
8389 2023-06-21 Jan Beulich <jbeulich@suse.com>
8391 * config/i386/i386-expand.cc (ix86_expand_copysign): Request
8392 value duplication by ix86_build_signbit_mask() when AVX512F and
8394 * config/i386/sse.md (*<avx512>_vternlog<mode>_all): Convert to
8395 2-alternative form. Adjust "mode" attribute. Add "enabled"
8397 (*<avx512>_vpternlog<mode>_1): Also permit when TARGET_AVX512F
8398 && !TARGET_PREFER_AVX256.
8399 (*<avx512>_vpternlog<mode>_2): Likewise.
8400 (*<avx512>_vpternlog<mode>_3): Likewise.
8402 2023-06-21 liuhongt <hongtao.liu@intel.com>
8405 * tree-vect-stmts.cc (vectorizable_conversion): Use
8406 intermiediate integer type for float_expr/fix_trunc_expr when
8407 direct optab is not existed.
8409 2023-06-20 Tamar Christina <tamar.christina@arm.com>
8412 * gensupport.cc (convert_syntax): Explicitly check for RTX code.
8414 2023-06-20 Richard Sandiford <richard.sandiford@arm.com>
8416 * config/aarch64/aarch64.md (stack_tie): Hard-code the first
8417 register operand to the stack pointer. Require the second register
8418 operand to have the number specified in a separate const_int operand.
8419 * config/aarch64/aarch64.cc (aarch64_emit_stack_tie): New function.
8420 (aarch64_allocate_and_probe_stack_space): Use it.
8421 (aarch64_expand_prologue, aarch64_expand_epilogue): Likewise.
8422 (aarch64_expand_epilogue): Likewise.
8424 2023-06-20 Jakub Jelinek <jakub@redhat.com>
8427 * tree-ssa-math-opts.cc (match_uaddc_usubc): Remember lhs of
8428 IMAGPART_EXPR of arg2/arg3 and use that as arg3 if it has the right
8431 2023-06-20 Uros Bizjak <ubizjak@gmail.com>
8433 * calls.h (setjmp_call_p): Change return type from int to bool.
8434 * calls.cc (struct arg_data): Change "pass_on_stack" to bool.
8435 (store_one_arg): Change return type from int to bool
8436 and adjust function body accordingly. Change "sibcall_failure"
8438 (finalize_must_preallocate): Ditto. Change *must_preallocate pointer
8439 argument to bool. Change "partial_seen" variable to bool.
8440 (load_register_parameters): Change *sibcall_failure
8441 pointer argument to bool.
8442 (check_sibcall_argument_overlap_1): Change return type from int to bool
8443 and adjust function body accordingly.
8444 (check_sibcall_argument_overlap): Ditto. Change
8445 "mark_stored_args_map" argument to bool.
8446 (emit_call_1): Change "already_popped" variable to bool.
8447 (setjmp_call_p): Change return type from int to bool
8448 and adjust function body accordingly.
8449 (initialize_argument_information): Change *must_preallocate
8450 pointer argument to bool.
8451 (expand_call): Change "pcc_struct_value", "must_preallocate"
8452 and "sibcall_failure" variables to bool.
8453 (emit_library_call_value_1): Change "pcc_struct_value"
8456 2023-06-20 Martin Jambor <mjambor@suse.cz>
8459 * ipa-sra.cc (struct caller_issues): New field there_is_one.
8460 (check_for_caller_issues): Set it.
8461 (check_all_callers_for_issues): Check it.
8463 2023-06-20 Martin Jambor <mjambor@suse.cz>
8465 * ipa-prop.h (ipa_uid_to_idx_map_elt): New type.
8466 (struct ipcp_transformation): Rearrange members according to
8467 C++ class coding convention, add m_uid_to_idx,
8468 get_param_index and maybe_create_parm_idx_map.
8469 * ipa-cp.cc (ipcp_transformation::get_param_index): New function.
8470 (compare_uids): Likewise.
8471 (ipcp_transformation::maype_create_parm_idx_map): Likewise.
8472 * ipa-prop.cc (ipcp_get_parm_bits): Use get_param_index.
8473 (ipcp_update_bits): Accept TS as a parameter, assume it is not NULL.
8474 (ipcp_update_vr): Likewise.
8475 (ipcp_transform_function): Call, maybe_create_parm_idx_map of TS, bail
8476 out quickly if empty, pass it to ipcp_update_bits and ipcp_update_vr.
8478 2023-06-20 Carl Love <cel@us.ibm.com>
8480 * config/rs6000/rs6000-builtin.cc (rs6000_expand_builtin):
8481 Rename CODE_FOR_xsxsigqp_tf to CODE_FOR_xsxsigqp_tf_ti.
8482 Rename CODE_FOR_xsxsigqp_kf to CODE_FOR_xsxsigqp_kf_ti.
8483 Rename CCDE_FOR_xsxexpqp_tf to CODE_FOR_xsxexpqp_tf_di.
8484 Rename CODE_FOR_xsxexpqp_kf to CODE_FOR_xsxexpqp_kf_di.
8485 (CODE_FOR_xsxexpqp_kf_v2di, CODE_FOR_xsxsigqp_kf_v1ti,
8486 CODE_FOR_xsiexpqp_kf_v2di): Add case statements.
8487 * config/rs6000/rs6000-builtins.def
8488 (__builtin_vsx_scalar_extract_exp_to_vec,
8489 __builtin_vsx_scalar_extract_sig_to_vec,
8490 __builtin_vsx_scalar_insert_exp_vqp): Add new builtin definitions.
8491 Rename xsxexpqp_kf, xsxsigqp_kf, xsiexpqp_kf to xsexpqp_kf_di,
8492 xsxsigqp_kf_ti, xsiexpqp_kf_di respectively.
8493 * config/rs6000/rs6000-c.cc (altivec_resolve_overloaded_builtin):
8494 Update case RS6000_OVLD_VEC_VSIE to handle MODE_VECTOR_INT for new
8495 overloaded instance. Update comments.
8496 * config/rs6000/rs6000-overload.def
8497 (__builtin_vec_scalar_insert_exp): Add new overload definition with
8499 (scalar_extract_exp_to_vec, scalar_extract_sig_to_vec): New
8500 overloaded definitions.
8501 * config/rs6000/vsx.md (V2DI_DI): New mode iterator.
8502 (DI_to_TI): New mode attribute.
8503 Rename xsxexpqp_<mode> to sxexpqp_<IEEE128:mode>_<V2DI_DI:mode>.
8504 Rename xsxsigqp_<mode> to xsxsigqp_<IEEE128:mode>_<VEC_TI:mode>.
8505 Rename xsiexpqp_<mode> to xsiexpqp_<IEEE128:mode>_<V2DI_DI:mode>.
8506 * doc/extend.texi (scalar_extract_exp_to_vec,
8507 scalar_extract_sig_to_vec): Add documentation for new builtins.
8508 (scalar_insert_exp): Add new overloaded builtin definition.
8510 2023-06-20 Li Xu <xuli1@eswincomputing.com>
8512 * config/riscv/riscv.cc (riscv_regmode_natural_size): set the natural
8513 size of vector mask mode to one rvv register.
8515 2023-06-20 Juzhe-Zhong <juzhe.zhong@rivai.ai>
8517 * config/riscv/riscv-v.cc (expand_const_vector): Optimize codegen.
8519 2023-06-20 Lehua Ding <lehua.ding@rivai.ai>
8521 * config/riscv/riscv.cc (riscv_arg_has_vector): Add default
8524 2023-06-20 Richard Biener <rguenther@suse.de>
8526 * tree-ssa-dse.cc (dse_classify_store): When we found
8527 no defs and the basic-block with the original definition
8528 ends in __builtin_unreachable[_trap] the store is dead.
8530 2023-06-20 Richard Biener <rguenther@suse.de>
8532 * tree-ssa-phiprop.cc (phiprop_insert_phi): For simple loads
8533 keep the virtual SSA form up-to-date.
8535 2023-06-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8537 * config/aarch64/aarch64-simd.md (*aarch64_addp_same_reg<mode>):
8538 New define_insn_and_split.
8540 2023-06-20 Tamar Christina <tamar.christina@arm.com>
8542 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Drop test comment.
8544 2023-06-20 Jan Beulich <jbeulich@suse.com>
8546 * config/i386/sse.md (vec_dupv2di): Correct %vmovddup input
8547 constraint. Add new AVX512F alternative.
8549 2023-06-20 Richard Biener <rguenther@suse.de>
8552 * dwarf2out.cc (process_scope_var): Continue processing
8553 the decl after setting a parent in case the existing DIE
8556 2023-06-20 Lehua Ding <lehua.ding@rivai.ai>
8558 * config/riscv/riscv.cc (riscv_scalable_vector_type_p): Delete.
8559 (riscv_arg_has_vector): Simplify.
8560 (riscv_pass_in_vector_p): Adjust warning message.
8562 2023-06-19 Jin Ma <jinma@linux.alibaba.com>
8564 * config/riscv/riscv.cc (riscv_compute_frame_info): Allocate frame for FCSR.
8565 (riscv_for_each_saved_reg): Save and restore FCSR in interrupt functions.
8566 * config/riscv/riscv.md (riscv_frcsr): New patterns.
8567 (riscv_fscsr): Likewise.
8569 2023-06-19 Toru Kisuki <tkisuki@tachyum.com>
8571 PR rtl-optimization/110305
8572 * simplify-rtx.cc (simplify_context::simplify_binary_operation_1):
8573 Handle HONOR_SNANS for x + 0.0.
8575 2023-06-19 Jan Hubicka <jh@suse.cz>
8577 PR tree-optimization/109811
8578 PR tree-optimization/109849
8579 * passes.def: Add phiprop to early optimization passes.
8580 * tree-ssa-phiprop.cc: Allow clonning.
8582 2023-06-19 Tamar Christina <tamar.christina@arm.com>
8584 * config/aarch64/aarch64.md (arches): Add nosimd.
8585 (*mov<mode>_aarch64, *movsi_aarch64, *movdi_aarch64): Rewrite to
8588 2023-06-19 Tamar Christina <tamar.christina@arm.com>
8589 Omar Tahir <Omar.Tahir2@arm.com>
8591 * gensupport.cc (class conlist, add_constraints, add_attributes,
8592 skip_spaces, expect_char, preprocess_compact_syntax,
8593 parse_section_layout, parse_section, convert_syntax): New.
8594 (process_rtx): Check for conversion.
8595 * genoutput.cc (process_template): Check for unresolved iterators.
8596 (class data): Add compact_syntax_p.
8598 * gensupport.h (compact_syntax): New.
8599 (hash-set.h): Include.
8600 * doc/md.texi: Document it.
8602 2023-06-19 Uros Bizjak <ubizjak@gmail.com>
8604 * recog.h (check_asm_operands): Change return type from int to bool.
8605 (insn_invalid_p): Ditto.
8606 (verify_changes): Ditto.
8607 (apply_change_group): Ditto.
8608 (constrain_operands): Ditto.
8609 (constrain_operands_cached): Ditto.
8610 (validate_replace_rtx_subexp): Ditto.
8611 (validate_replace_rtx): Ditto.
8612 (validate_replace_rtx_part): Ditto.
8613 (validate_replace_rtx_part_nosimplify): Ditto.
8614 (added_clobbers_hard_reg_p): Ditto.
8615 (peep2_regno_dead_p): Ditto.
8616 (peep2_reg_dead_p): Ditto.
8617 (store_data_bypass_p): Ditto.
8618 (if_test_bypass_p): Ditto.
8619 * rtl.h (split_all_insns_noflow): Change
8620 return type from unsigned int to void.
8621 * genemit.cc (output_added_clobbers_hard_reg_p): Change return type
8622 of generated added_clobbers_hard_reg_p from int to bool and adjust
8623 function body accordingly. Change "used" variable type from
8625 * recog.cc (check_asm_operands): Change return type
8626 from int to bool and adjust function body accordingly.
8627 (insn_invalid_p): Ditto. Change "is_asm" variable to bool.
8628 (verify_changes): Change return type from int to bool.
8629 (apply_change_group): Change return type from int to bool
8630 and adjust function body accordingly.
8631 (validate_replace_rtx_subexp): Change return type from int to bool.
8632 (validate_replace_rtx): Ditto.
8633 (validate_replace_rtx_part): Ditto.
8634 (validate_replace_rtx_part_nosimplify): Ditto.
8635 (constrain_operands_cached): Ditto.
8636 (constrain_operands): Ditto. Change "lose" and "win"
8637 variables type from int to bool.
8638 (split_all_insns_noflow): Change return type from unsigned int
8639 to void and adjust function body accordingly.
8640 (peep2_regno_dead_p): Change return type from int to bool.
8641 (peep2_reg_dead_p): Ditto.
8642 (peep2_find_free_register): Change "success"
8643 variable type from int to bool
8644 (store_data_bypass_p_1): Change return type from int to bool.
8645 (store_data_bypass_p): Ditto.
8647 2023-06-19 Li Xu <xuli1@eswincomputing.com>
8649 * config/riscv/vector-iterators.md: zvfh/zvfhmin depends on the
8652 2023-06-19 Pan Li <pan2.li@intel.com>
8655 * config/riscv/riscv-vector-builtins-bases.cc: Adjust expand for
8657 * config/riscv/vector-iterators.md: Remove VWLMUL1, VWLMUL1_ZVE64,
8658 VWLMUL1_ZVE32, VI_ZVE64, VI_ZVE32, VWI, VWI_ZVE64, VWI_ZVE32,
8659 VF_ZVE63 and VF_ZVE32.
8660 * config/riscv/vector.md
8661 (@pred_widen_reduc_plus<v_su><mode><vwlmul1>): Removed.
8662 (@pred_widen_reduc_plus<v_su><mode><vwlmul1_zve64>): Ditto.
8663 (@pred_widen_reduc_plus<v_su><mode><vwlmul1_zve32>): Ditto.
8664 (@pred_widen_reduc_plus<order><mode><vwlmul1>): Ditto.
8665 (@pred_widen_reduc_plus<order><mode><vwlmul1_zve64>): Ditto.
8666 (@pred_widen_reduc_plus<v_su><VQI:mode><VHI_LMUL1:mode>): New pattern.
8667 (@pred_widen_reduc_plus<v_su><VHI:mode><VSI_LMUL1:mode>): Ditto.
8668 (@pred_widen_reduc_plus<v_su><VSI:mode><VDI_LMUL1:mode>): Ditto.
8669 (@pred_widen_reduc_plus<order><VHF:mode><VSF_LMUL1:mode>): Ditto.
8670 (@pred_widen_reduc_plus<order><VSF:mode><VDF_LMUL1:mode>): Ditto.
8672 2023-06-19 Pan Li <pan2.li@intel.com>
8675 * config/riscv/riscv-vector-builtins-bases.cc: Adjust expand for
8677 * config/riscv/vector-iterators.md: Add VHF, VSF, VDF,
8678 VHF_LMUL1, VSF_LMUL1, VDF_LMUL1, and remove unused attr.
8679 * config/riscv/vector.md (@pred_reduc_<reduc><mode><vlmul1>): Removed.
8680 (@pred_reduc_<reduc><mode><vlmul1_zve64>): Ditto.
8681 (@pred_reduc_<reduc><mode><vlmul1_zve32>): Ditto.
8682 (@pred_reduc_plus<order><mode><vlmul1>): Ditto.
8683 (@pred_reduc_plus<order><mode><vlmul1_zve32>): Ditto.
8684 (@pred_reduc_plus<order><mode><vlmul1_zve64>): Ditto.
8685 (@pred_reduc_<reduc><VHF:mode><VHF_LMUL1:mode>): New pattern.
8686 (@pred_reduc_<reduc><VSF:mode><VSF_LMUL1:mode>): Ditto.
8687 (@pred_reduc_<reduc><VDF:mode><VDF_LMUL1:mode>): Ditto.
8688 (@pred_reduc_plus<order><VHF:mode><VHF_LMUL1:mode>): Ditto.
8689 (@pred_reduc_plus<order><VSF:mode><VSF_LMUL1:mode>): Ditto.
8690 (@pred_reduc_plus<order><VDF:mode><VDF_LMUL1:mode>): Ditto.
8692 2023-06-19 Andrew Stubbs <ams@codesourcery.com>
8694 * config/gcn/gcn.cc (gcn_expand_divmod_libfunc): New function.
8695 (gcn_init_libfuncs): Add div and mod functions for all modes.
8696 Add placeholders for divmod functions.
8697 (TARGET_EXPAND_DIVMOD_LIBFUNC): Define.
8699 2023-06-19 Andrew Stubbs <ams@codesourcery.com>
8701 * tree-vect-generic.cc: Include optabs-libfuncs.h.
8702 (get_compute_type): Check optab_libfunc.
8703 * tree-vect-stmts.cc: Include optabs-libfuncs.h.
8704 (vectorizable_operation): Check optab_libfunc.
8706 2023-06-19 Andrew Stubbs <ams@codesourcery.com>
8708 * config/gcn/gcn-protos.h (vgpr_4reg_mode_p): New function.
8709 * config/gcn/gcn-valu.md (V_4REG, V_4REG_ALT): New iterators.
8710 (V_MOV, V_MOV_ALT): Likewise.
8711 (scalar_mode, SCALAR_MODE): Add TImode.
8712 (vnsi, VnSI, vndi, VnDI): Likewise.
8713 (vec_merge, vec_merge_with_clobber, vec_merge_with_vcc): Use V_MOV.
8714 (mov<mode>, mov<mode>_unspec): Use V_MOV.
8715 (*mov<mode>_4reg): New insn.
8716 (mov<mode>_exec): New 4reg variant.
8717 (mov<mode>_sgprbase): Likewise.
8718 (reload_in<mode>, reload_out<mode>): Use V_MOV.
8719 (vec_set<mode>): Likewise.
8720 (vec_duplicate<mode><exec>): New 4reg variant.
8721 (vec_extract<mode><scalar_mode>): Likewise.
8722 (vec_extract<V_ALL:mode><V_ALL_ALT:mode>): Rename to ...
8723 (vec_extract<V_MOV:mode><V_MOV_ALT:mode>): ... this, and use V_MOV.
8724 (vec_extract<V_4REG:mode><V_4REG_ALT:mode>_nop): New 4reg variant.
8725 (fold_extract_last_<mode>): Use V_MOV.
8726 (vec_init<V_ALL:mode><V_ALL_ALT:mode>): Rename to ...
8727 (vec_init<V_MOV:mode><V_MOV_ALT:mode>): ... this, and use V_MOV.
8728 (gather_load<mode><vnsi>, gather<mode>_expr<exec>,
8729 gather<mode>_insn_1offset<exec>, gather<mode>_insn_1offset_ds<exec>,
8730 gather<mode>_insn_2offsets<exec>): Use V_MOV.
8731 (scatter_store<mode><vnsi>, scatter<mode>_expr<exec_scatter>,
8732 scatter<mode>_insn_1offset<exec_scatter>,
8733 scatter<mode>_insn_1offset_ds<exec_scatter>,
8734 scatter<mode>_insn_2offsets<exec_scatter>): Likewise.
8735 (maskload<mode>di, maskstore<mode>di, mask_gather_load<mode><vnsi>,
8736 mask_scatter_store<mode><vnsi>): Likewise.
8737 * config/gcn/gcn.cc (gcn_class_max_nregs): Use vgpr_4reg_mode_p.
8738 (gcn_hard_regno_mode_ok): Likewise.
8739 (GEN_VNM): Add TImode support.
8740 (USE_TI): New macro. Separate TImode operations from non-TImode ones.
8741 (gcn_vector_mode_supported_p): Add V64TImode, V32TImode, V16TImode,
8742 V8TImode, and V2TImode.
8743 (print_operand): Add 'J' and 'K' print codes.
8745 2023-06-19 Richard Biener <rguenther@suse.de>
8747 PR tree-optimization/110298
8748 * tree-ssa-loop-ivcanon.cc (tree_unroll_loops_completely):
8749 Clear number of iterations info before cleaning up the CFG.
8751 2023-06-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8753 * simplify-rtx.cc (simplify_context::simplify_binary_operation_1):
8754 Simplify vec_concat of lowpart subreg and high part vec_select.
8756 2023-06-19 Tobias Burnus <tobias@codesourcery.com>
8758 * doc/invoke.texi (-foffload-options): Remove '-O3' from the examples.
8760 2023-06-19 Richard Sandiford <richard.sandiford@arm.com>
8762 * tree-vect-loop-manip.cc (vect_set_loop_condition_partial_vectors):
8763 Handle null niters_skip.
8765 2023-06-19 Richard Biener <rguenther@suse.de>
8767 * config/aarch64/aarch64.cc
8768 (aarch64_vector_costs::analyze_loop_vinfo): Fix reference
8769 to LOOP_VINFO_MASKS.
8771 2023-06-19 Senthil Kumar Selvaraj <saaadhu@gcc.gnu.org>
8774 * common/config/avr/avr-common.cc: Remove setting
8775 of OPT_fdelete_null_pointer_checks.
8776 * config/avr/avr.cc (avr_option_override): Clear
8777 flag_delete_null_pointer_checks if zero_address_valid.
8778 (avr_addr_space_zero_address_valid): New function.
8779 (TARGET_ADDR_SPACE_ZERO_ADDRESS_VALID): Provide target
8782 2023-06-19 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
8783 Robin Dapp <rdapp.gcc@gmail.com>
8785 * doc/md.texi: Add len_mask{load,store}.
8786 * genopinit.cc (main): Ditto.
8788 * internal-fn.cc (len_maskload_direct): Ditto.
8789 (len_maskstore_direct): Ditto.
8790 (expand_call_mem_ref): Ditto.
8791 (expand_partial_load_optab_fn): Ditto.
8792 (expand_len_maskload_optab_fn): Ditto.
8793 (expand_partial_store_optab_fn): Ditto.
8794 (expand_len_maskstore_optab_fn): Ditto.
8795 (direct_len_maskload_optab_supported_p): Ditto.
8796 (direct_len_maskstore_optab_supported_p): Ditto.
8797 * internal-fn.def (LEN_MASK_LOAD): Ditto.
8798 (LEN_MASK_STORE): Ditto.
8799 * optabs.def (OPTAB_CD): Ditto.
8801 2023-06-19 Robin Dapp <rdapp@ventanamicro.com>
8803 * config/riscv/autovec.md (<optab><mode>2): Add unop expanders.
8805 2023-06-19 Robin Dapp <rdapp@ventanamicro.com>
8807 * config/riscv/autovec.md (<optab><mode>3): Implement binop
8809 * config/riscv/riscv-protos.h (emit_vlmax_fp_insn): Declare.
8810 (enum vxrm_field_enum): Rename this...
8811 (enum fixed_point_rounding_mode): ...to this.
8812 (enum frm_field_enum): Rename this...
8813 (enum floating_point_rounding_mode): ...to this.
8814 * config/riscv/riscv-v.cc (emit_vlmax_fp_insn): New function
8815 * config/riscv/riscv.cc (riscv_const_insns): Clarify const
8817 (riscv_libgcc_floating_mode_supported_p): Adjust comment.
8818 (riscv_excess_precision): Do not convert to float for ZVFH.
8819 * config/riscv/vector-iterators.md: Add VF_AUTO iterator.
8821 2023-06-19 Robin Dapp <rdapp@ventanamicro.com>
8823 * config/riscv/vector-iterators.md: Add VI_QH iterator.
8824 * config/riscv/autovec-opt.md
8825 (@pred_extract_first_sextdi<mode>): New vmv.x.s pattern
8826 that includes sign extension.
8827 (@pred_extract_first_sextsi<mode>): Dito for SImode.
8829 2023-06-19 Robin Dapp <rdapp@ventanamicro.com>
8831 * config/riscv/autovec.md (vec_set<mode>): Implement.
8832 (vec_extract<mode><vel>): Implement.
8833 * config/riscv/riscv-protos.h (enum insn_type): Add slide insn.
8834 (emit_vlmax_slide_insn): Declare.
8835 (emit_nonvlmax_slide_tu_insn): Declare.
8836 (emit_scalar_move_insn): Export.
8837 (emit_nonvlmax_integer_move_insn): Export.
8838 * config/riscv/riscv-v.cc (emit_vlmax_slide_insn): New function.
8839 (emit_nonvlmax_slide_tu_insn): New function.
8840 (emit_vlmax_masked_mu_insn): No change.
8841 (emit_vlmax_integer_move_insn): Export.
8843 2023-06-19 Richard Biener <rguenther@suse.de>
8845 * tree-vectorizer.h (enum vect_partial_vector_style): New.
8846 (_loop_vec_info::partial_vector_style): Likewise.
8847 (LOOP_VINFO_PARTIAL_VECTORS_STYLE): Likewise.
8848 (rgroup_controls::compare_type): Add.
8849 (vec_loop_masks): Change from a typedef to auto_vec<>
8851 * tree-vect-loop-manip.cc (vect_set_loop_condition_partial_vectors):
8852 Adjust. Convert niters_skip to compare_type.
8853 (vect_set_loop_condition_partial_vectors_avx512): New function
8854 implementing the AVX512 partial vector codegen.
8855 (vect_set_loop_condition): Dispatch to the correct
8856 vect_set_loop_condition_partial_vectors_* function based on
8857 LOOP_VINFO_PARTIAL_VECTORS_STYLE.
8858 (vect_prepare_for_masked_peels): Compute LOOP_VINFO_MASK_SKIP_NITERS
8859 in the original niter type.
8860 * tree-vect-loop.cc (_loop_vec_info::_loop_vec_info): Initialize
8861 partial_vector_style.
8862 (can_produce_all_loop_masks_p): Adjust.
8863 (vect_verify_full_masking): Produce the rgroup_controls vector
8864 here. Set LOOP_VINFO_PARTIAL_VECTORS_STYLE on success.
8865 (vect_verify_full_masking_avx512): New function implementing
8866 verification of AVX512 style masking.
8867 (vect_verify_loop_lens): Set LOOP_VINFO_PARTIAL_VECTORS_STYLE.
8868 (vect_analyze_loop_2): Also try AVX512 style masking.
8870 (vect_estimate_min_profitable_iters): Implement AVX512 style
8871 mask producing cost.
8872 (vect_record_loop_mask): Do not build the rgroup_controls
8873 vector here but record masks in a hash-set.
8874 (vect_get_loop_mask): Implement AVX512 style mask query,
8875 complementing the existing while_ult style.
8877 2023-06-19 Richard Biener <rguenther@suse.de>
8879 * tree-vectorizer.h (vect_get_loop_mask): Add loop_vec_info
8881 * tree-vect-loop.cc (vect_get_loop_mask): Likewise.
8882 (vectorize_fold_left_reduction): Adjust.
8883 (vect_transform_reduction): Likewise.
8884 (vectorizable_live_operation): Likewise.
8885 * tree-vect-stmts.cc (vectorizable_call): Likewise.
8886 (vectorizable_operation): Likewise.
8887 (vectorizable_store): Likewise.
8888 (vectorizable_load): Likewise.
8889 (vectorizable_condition): Likewise.
8891 2023-06-19 Senthil Kumar Selvaraj <saaadhu@gcc.gnu.org>
8894 * config/avr/avr.opt (mgas-isr-prologues, mmain-is-OS_task):
8895 Add Optimization option property.
8897 2023-06-19 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
8899 * config/xtensa/xtensa.cc (xtensa_constantsynth_2insn):
8900 Add new pattern for the abovementioned case.
8902 2023-06-19 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
8904 * config/xtensa/xtensa.cc
8905 (TARGET_MEMORY_MOVE_COST, xtensa_memory_move_cost): Remove.
8907 2023-06-19 Jiufu Guo <guojiufu@linux.ibm.com>
8909 * config/rs6000/rs6000.cc (TARGET_CONST_ANCHOR): New define.
8911 2023-06-19 Jiufu Guo <guojiufu@linux.ibm.com>
8913 * cse.cc (try_const_anchors): Check SCALAR_INT_MODE.
8915 2023-06-19 liuhongt <hongtao.liu@intel.com>
8918 * config/i386/sse.md (<sse2_avx2>_packsswb<mask_name>):
8920 (sse2_packsswb<mask_name>): .. this, ..
8921 (avx2_packsswb<mask_name>): .. this and ..
8922 (avx512bw_packsswb<mask_name>): .. this.
8923 (<sse2_avx2>_packssdw<mask_name>): Substitute with ..
8924 (sse2_packssdw<mask_name>): .. this, ..
8925 (avx2_packssdw<mask_name>): .. this and ..
8926 (avx512bw_packssdw<mask_name>): .. this.
8928 2023-06-19 liuhongt <hongtao.liu@intel.com>
8931 * config/i386/i386-expand.cc (ix86_split_mmx_pack): Use
8932 UNSPEC_US_TRUNCATE instead of original us_truncate for
8934 * config/i386/mmx.md (mmx_pack<s_trunsuffix>swb): Substitute
8936 (mmx_packsswb): .. this and ..
8937 (mmx_packuswb): .. this.
8938 (mmx_packusdw): Use UNSPEC_US_TRUNCATE instead of original
8940 (s_trunsuffix): Removed code iterator.
8941 (any_s_truncate): Ditto.
8942 * config/i386/sse.md (<sse2_avx2>_packuswb<mask_name>): Use
8943 UNSPEC_US_TRUNCATE instead of original us_truncate.
8944 (<sse4_1_avx2>_packusdw<mask_name>): Ditto.
8945 * config/i386/i386.md (UNSPEC_US_TRUNCATE): New unspec_c_enum.
8947 2023-06-18 Pan Li <pan2.li@intel.com>
8949 * config/riscv/riscv-vector-builtins-bases.cc: Fix one typo.
8951 2023-06-18 Uros Bizjak <ubizjak@gmail.com>
8953 * rtl.h (*rtx_equal_p_callback_function):
8954 Change return type from int to bool.
8955 (rtx_equal_p): Ditto.
8956 (*hash_rtx_callback_function): Ditto.
8957 * rtl.cc (rtx_equal_p): Change return type from int to bool
8958 and adjust function body accordingly.
8959 * early-remat.cc (scratch_equal): Ditto.
8960 * sel-sched-ir.cc (skip_unspecs_callback): Ditto.
8961 (hash_with_unspec_callback): Ditto.
8963 2023-06-18 Jeff Law <jlaw@ventanamicro.com>
8965 * config/arc/arc.md (movqi_insn): Allow certain constants to
8966 be stored into memory in the pattern's condition.
8967 (movsf_insn): Similarly.
8969 2023-06-18 Honza <jh@ryzen3.suse.cz>
8971 PR tree-optimization/109849
8972 * ipa-fnsummary.cc (evaluate_conditions_for_known_args): Add new parameter
8973 ES; handle ipa_predicate::not_sra_candidate.
8974 (evaluate_properties_for_edge): Pass es to
8975 evaluate_conditions_for_known_args.
8976 (ipa_fn_summary_t::duplicate): Handle sra candidates.
8977 (dump_ipa_call_summary): Dump points_to_possible_sra_candidate.
8978 (load_or_store_of_ptr_parameter): New function.
8979 (points_to_possible_sra_candidate_p): New function.
8980 (analyze_function_body): Initialize points_to_possible_sra_candidate;
8981 determine sra predicates.
8982 (estimate_ipcp_clone_size_and_time): Update call of
8983 evaluate_conditions_for_known_args.
8984 (remap_edge_params): Update points_to_possible_sra_candidate.
8985 (read_ipa_call_summary): Stream points_to_possible_sra_candidate
8986 (write_ipa_call_summary): Likewise.
8987 * ipa-predicate.cc (ipa_predicate::add_clause): Handle not_sra_candidate.
8988 (dump_condition): Dump it.
8989 * ipa-predicate.h (struct inline_param_summary): Add
8990 points_to_possible_sra_candidate.
8992 2023-06-18 Roger Sayle <roger@nextmovesoftware.com>
8994 * config/i386/i386-expand.cc (ix86_expand_carry): New helper
8995 function for setting the carry flag.
8996 (ix86_expand_builtin) <handlecarry>: Use it here.
8997 * config/i386/i386-protos.h (ix86_expand_carry): Prototype here.
8998 * config/i386/i386.md (uaddc<mode>5): Use ix86_expand_carry.
8999 (usubc<mode>5): Likewise.
9001 2023-06-18 Roger Sayle <roger@nextmovesoftware.com>
9003 * config/i386/i386.md (*concat<mode><dwi>3_1): Use QImode
9004 for the immediate constant shift count.
9005 (*concat<mode><dwi>3_2): Likewise.
9006 (*concat<mode><dwi>3_3): Likewise.
9007 (*concat<mode><dwi>3_4): Likewise.
9008 (*concat<mode><dwi>3_5): Likewise.
9009 (*concat<mode><dwi>3_6): Likewise.
9011 2023-06-18 Uros Bizjak <ubizjak@gmail.com>
9013 * cse.cc (hash_rtx_cb): Rename to hash_rtx.
9015 * early-remat.cc (remat_candidate_hasher::equal): Update
9016 to call rtx_equal_p with rtx_equal_p_callback_function argument.
9017 * rtl.cc (rtx_equal_p_cb): Rename to rtx_equal_p.
9018 (rtx_equal_p): Remove.
9019 * rtl.h (rtx_equal_p): Add rtx_equal_p_callback_function
9020 argument with NULL default value.
9021 (rtx_equal_p_cb): Remove function declaration.
9022 (hash_rtx_cb): Ditto.
9023 (hash_rtx): Add hash_rtx_callback_function argument
9024 with NULL default value.
9025 * sel-sched-ir.cc (free_nop_pool): Update function comment.
9026 (skip_unspecs_callback): Ditto.
9027 (vinsn_init): Update to call hash_rtx with
9028 hash_rtx_callback_function argument.
9029 (vinsn_equal_p): Ditto.
9031 2023-06-18 yulong <shiyulong@iscas.ac.cn>
9033 * config/riscv/genrvv-type-indexer.cc (valid_type): Enable FP16 tuple.
9034 * config/riscv/riscv-modes.def (RVV_TUPLE_MODES): New macro.
9035 (ADJUST_ALIGNMENT): Ditto.
9036 (RVV_TUPLE_PARTIAL_MODES): Ditto.
9037 (ADJUST_NUNITS): Ditto.
9038 * config/riscv/riscv-vector-builtins-types.def (vfloat16mf4x2_t):
9040 (vfloat16mf4x3_t): Ditto.
9041 (vfloat16mf4x4_t): Ditto.
9042 (vfloat16mf4x5_t): Ditto.
9043 (vfloat16mf4x6_t): Ditto.
9044 (vfloat16mf4x7_t): Ditto.
9045 (vfloat16mf4x8_t): Ditto.
9046 (vfloat16mf2x2_t): Ditto.
9047 (vfloat16mf2x3_t): Ditto.
9048 (vfloat16mf2x4_t): Ditto.
9049 (vfloat16mf2x5_t): Ditto.
9050 (vfloat16mf2x6_t): Ditto.
9051 (vfloat16mf2x7_t): Ditto.
9052 (vfloat16mf2x8_t): Ditto.
9053 (vfloat16m1x2_t): Ditto.
9054 (vfloat16m1x3_t): Ditto.
9055 (vfloat16m1x4_t): Ditto.
9056 (vfloat16m1x5_t): Ditto.
9057 (vfloat16m1x6_t): Ditto.
9058 (vfloat16m1x7_t): Ditto.
9059 (vfloat16m1x8_t): Ditto.
9060 (vfloat16m2x2_t): Ditto.
9061 (vfloat16m2x3_t): Ditto.
9062 (vfloat16m2x4_t): Ditto.
9063 (vfloat16m4x2_t): Ditto.
9064 * config/riscv/riscv-vector-builtins.def (vfloat16mf4x2_t): New macro.
9065 (vfloat16mf4x3_t): Ditto.
9066 (vfloat16mf4x4_t): Ditto.
9067 (vfloat16mf4x5_t): Ditto.
9068 (vfloat16mf4x6_t): Ditto.
9069 (vfloat16mf4x7_t): Ditto.
9070 (vfloat16mf4x8_t): Ditto.
9071 (vfloat16mf2x2_t): Ditto.
9072 (vfloat16mf2x3_t): Ditto.
9073 (vfloat16mf2x4_t): Ditto.
9074 (vfloat16mf2x5_t): Ditto.
9075 (vfloat16mf2x6_t): Ditto.
9076 (vfloat16mf2x7_t): Ditto.
9077 (vfloat16mf2x8_t): Ditto.
9078 (vfloat16m1x2_t): Ditto.
9079 (vfloat16m1x3_t): Ditto.
9080 (vfloat16m1x4_t): Ditto.
9081 (vfloat16m1x5_t): Ditto.
9082 (vfloat16m1x6_t): Ditto.
9083 (vfloat16m1x7_t): Ditto.
9084 (vfloat16m1x8_t): Ditto.
9085 (vfloat16m2x2_t): Ditto.
9086 (vfloat16m2x3_t): Ditto.
9087 (vfloat16m2x4_t): Ditto.
9088 (vfloat16m4x2_t): Ditto.
9089 * config/riscv/riscv-vector-switch.def (TUPLE_ENTRY): New.
9090 * config/riscv/riscv.md: New.
9091 * config/riscv/vector-iterators.md: New.
9093 2023-06-17 Roger Sayle <roger@nextmovesoftware.com>
9095 * config/i386/i386-expand.cc (ix86_expand_move): Check that OP1 is
9096 CONST_WIDE_INT_P before calling ix86_convert_wide_int_to_broadcast.
9097 Generalize special case for converting TImode to V1TImode to handle
9098 all 128-bit vector conversions.
9100 2023-06-17 Costas Argyris <costas.argyris@gmail.com>
9102 * gcc-ar.cc (main): Refactor to slightly reduce code
9103 duplication. Avoid unnecessary elements in nargv.
9105 2023-06-16 Pan Li <pan2.li@intel.com>
9108 * config/riscv/riscv-vector-builtins-bases.cc: Add ret_mode for
9109 integer reduction expand.
9110 * config/riscv/vector-iterators.md: Add VQI, VHI, VSI and VDI,
9111 and the LMUL1 attr respectively.
9112 * config/riscv/vector.md
9113 (@pred_reduc_<reduc><mode><vlmul1>): Removed.
9114 (@pred_reduc_<reduc><mode><vlmul1_zve64>): Likewise.
9115 (@pred_reduc_<reduc><mode><vlmul1_zve32>): Likewise.
9116 (@pred_reduc_<reduc><VQI:mode><VQI_LMUL1:mode>): New pattern.
9117 (@pred_reduc_<reduc><VHI:mode><VHI_LMUL1:mode>): Likewise.
9118 (@pred_reduc_<reduc><VSI:mode><VSI_LMUL1:mode>): Likewise.
9119 (@pred_reduc_<reduc><VDI:mode><VDI_LMUL1:mode>): Likewise.
9121 2023-06-16 Juzhe-Zhong <juzhe.zhong@rivai.ai>
9124 * config/riscv/riscv-vsetvl.cc (insert_vsetvl): Fix bug.
9126 2023-06-16 Jakub Jelinek <jakub@redhat.com>
9129 * builtin-types.def (BT_FN_UINT_UINT_UINT_UINT_UINTPTR,
9130 BT_FN_ULONG_ULONG_ULONG_ULONG_ULONGPTR,
9131 BT_FN_ULONGLONG_ULONGLONG_ULONGLONG_ULONGLONG_ULONGLONGPTR): New
9133 * builtins.def (BUILT_IN_ADDC, BUILT_IN_ADDCL, BUILT_IN_ADDCLL,
9134 BUILT_IN_SUBC, BUILT_IN_SUBCL, BUILT_IN_SUBCLL): New builtins.
9135 * builtins.cc (fold_builtin_addc_subc): New function.
9136 (fold_builtin_varargs): Handle BUILT_IN_{ADD,SUB}C{,L,LL}.
9137 * doc/extend.texi (__builtin_addc, __builtin_subc): Document.
9139 2023-06-16 Jakub Jelinek <jakub@redhat.com>
9141 PR tree-optimization/110271
9142 * tree-ssa-math-opts.cc (math_opts_dom_walker::after_dom_children)
9143 <case PLUS_EXPR>: Ignore return value from match_arith_overflow,
9144 instead call match_uaddc_usubc only if gsi_stmt (gsi) is still stmt.
9146 2023-06-16 Martin Jambor <mjambor@suse.cz>
9148 * configure: Regenerate.
9150 2023-06-16 Roger Sayle <roger@nextmovesoftware.com>
9151 Uros Bizjak <ubizjak@gmail.com>
9154 * config/i386/i386.md (*add<dwi>3_doubleword_concat): New
9155 define_insn_and_split combine *add<dwi>3_doubleword with
9156 a *concat<mode><dwi>3 for more efficient lowering after reload.
9158 2023-06-16 Vladimir N. Makarov <vmakarov@redhat.com>
9160 * ira-lives.cc: Include except.h.
9161 (process_bb_node_lives): Ignore conflicts from cleanup exceptions
9162 when the pseudo does not live at the exception landing pad.
9164 2023-06-16 Alex Coplan <alex.coplan@arm.com>
9166 * doc/invoke.texi: Document -Welaborated-enum-base.
9168 2023-06-16 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9170 * config/aarch64/aarch64-simd-builtins.def (shrn2_n): Rename builtins to...
9171 (ushrn2_n): ... This.
9172 (sqshrn2_n): Rename builtins to...
9173 (ssqshrn2_n): ... This.
9174 (uqshrn2_n): Rename builtins to...
9175 (uqushrn2_n): ... This.
9176 * config/aarch64/arm_neon.h (vqshrn_high_n_s16): Adjust for the above.
9177 (vqshrn_high_n_s32): Likewise.
9178 (vqshrn_high_n_s64): Likewise.
9179 (vqshrn_high_n_u16): Likewise.
9180 (vqshrn_high_n_u32): Likewise.
9181 (vqshrn_high_n_u64): Likewise.
9182 (vshrn_high_n_s16): Likewise.
9183 (vshrn_high_n_s32): Likewise.
9184 (vshrn_high_n_s64): Likewise.
9185 (vshrn_high_n_u16): Likewise.
9186 (vshrn_high_n_u32): Likewise.
9187 (vshrn_high_n_u64): Likewise.
9188 * config/aarch64/aarch64-simd.md (aarch64_<shrn_op>shrn2_n<mode>_insn_le):
9190 (aarch64_<shrn_op><sra_op>shrn2_n<mode>_insn_le): ... This.
9191 Use SHIFTRT iterator and AARCH64_VALID_SHRN_OP check.
9192 (aarch64_<shrn_op>shrn2_n<mode>_insn_be): Rename to...
9193 (aarch64_<shrn_op><sra_op>shrn2_n<mode>_insn_be): ... This.
9194 Use SHIFTRT iterator and AARCH64_VALID_SHRN_OP check.
9195 (aarch64_<shrn_op>shrn2_n<mode>): Rename to...
9196 (aarch64_<shrn_op><sra_op>shrn2_n<mode>): ... This.
9197 Update expander for the above.
9199 2023-06-16 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9201 * config/aarch64/aarch64-simd-builtins.def (shrn2): Rename builtins to...
9202 (shrn2_n): ... This.
9203 (rshrn2): Rename builtins to...
9204 (rshrn2_n): ... This.
9205 * config/aarch64/arm_neon.h (vrshrn_high_n_s16): Adjust for the above.
9206 (vrshrn_high_n_s32): Likewise.
9207 (vrshrn_high_n_s64): Likewise.
9208 (vrshrn_high_n_u16): Likewise.
9209 (vrshrn_high_n_u32): Likewise.
9210 (vrshrn_high_n_u64): Likewise.
9211 (vshrn_high_n_s16): Likewise.
9212 (vshrn_high_n_s32): Likewise.
9213 (vshrn_high_n_s64): Likewise.
9214 (vshrn_high_n_u16): Likewise.
9215 (vshrn_high_n_u32): Likewise.
9216 (vshrn_high_n_u64): Likewise.
9217 * config/aarch64/aarch64-simd.md (*aarch64_<srn_op>shrn<mode>2_vect_le):
9219 (*aarch64_<srn_op>shrn<mode>2_vect_be): Likewise.
9220 (aarch64_shrn2<mode>_insn_le): Likewise.
9221 (aarch64_shrn2<mode>_insn_be): Likewise.
9222 (aarch64_shrn2<mode>): Likewise.
9223 (aarch64_rshrn2<mode>_insn_le): Likewise.
9224 (aarch64_rshrn2<mode>_insn_be): Likewise.
9225 (aarch64_rshrn2<mode>): Likewise.
9226 (aarch64_<sur>q<r>shr<u>n2_n<mode>_insn_le): Likewise.
9227 (aarch64_<shrn_op>shrn2_n<mode>_insn_le): New define_insn.
9228 (aarch64_<sur>q<r>shr<u>n2_n<mode>_insn_be): Delete.
9229 (aarch64_<shrn_op>shrn2_n<mode>_insn_be): New define_insn.
9230 (aarch64_<sur>q<r>shr<u>n2_n<mode>): Delete.
9231 (aarch64_<shrn_op>shrn2_n<mode>): New define_expand.
9232 (aarch64_<shrn_op>rshrn2_n<mode>_insn_le): New define_insn.
9233 (aarch64_<shrn_op>rshrn2_n<mode>_insn_be): New define_insn.
9234 (aarch64_<shrn_op>rshrn2_n<mode>): New define_expand.
9235 (aarch64_sqshrun2_n<mode>_insn_le): New define_insn.
9236 (aarch64_sqshrun2_n<mode>_insn_be): New define_insn.
9237 (aarch64_sqshrun2_n<mode>): New define_expand.
9238 (aarch64_sqrshrun2_n<mode>_insn_le): New define_insn.
9239 (aarch64_sqrshrun2_n<mode>_insn_be): New define_insn.
9240 (aarch64_sqrshrun2_n<mode>): New define_expand.
9241 * config/aarch64/iterators.md (UNSPEC_SQSHRUN, UNSPEC_SQRSHRUN,
9242 UNSPEC_SQSHRN, UNSPEC_UQSHRN, UNSPEC_SQRSHRN, UNSPEC_UQRSHRN):
9243 Delete unspec values.
9244 (VQSHRN_N): Delete int iterator.
9246 2023-06-16 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9248 * config/aarch64/aarch64.h (AARCH64_VALID_SHRN_OP): Define.
9249 * config/aarch64/aarch64-simd.md
9250 (*aarch64_<shrn_op>shrn_n<mode>_insn<vczle><vczbe>): Rename to...
9251 (*aarch64_<shrn_op><shrn_s>shrn_n<mode>_insn<vczle><vczbe>): ... This.
9252 Use SHIFTRT iterator and add AARCH64_VALID_SHRN_OP to condition.
9253 * config/aarch64/iterators.md (shrn_s): New code attribute.
9255 2023-06-16 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9257 * config/aarch64/aarch64-simd.md (aarch64_<sur>q<r>shr<u>n_n<mode>):
9259 (aarch64_<shrn_op>shrn_n<mode>): ... This. Reimplement with RTL codes.
9260 (*aarch64_<shrn_op>rshrn_n<mode>_insn): New define_insn.
9261 (aarch64_sqrshrun_n<mode>_insn): Likewise.
9262 (aarch64_sqshrun_n<mode>_insn): Likewise.
9263 (aarch64_<shrn_op>rshrn_n<mode>): New define_expand.
9264 (aarch64_sqshrun_n<mode>): Likewise.
9265 (aarch64_sqrshrun_n<mode>): Likewise.
9266 * config/aarch64/iterators.md (V2XWIDE): Add HI and SI modes.
9268 2023-06-16 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9270 * config/aarch64/aarch64-simd-builtins.def (shrn): Rename builtins to...
9272 (rshrn): Rename builtins to...
9273 (rshrn_n): ... This.
9274 * config/aarch64/arm_neon.h (vshrn_n_s16): Adjust for the above.
9275 (vshrn_n_s32): Likewise.
9276 (vshrn_n_s64): Likewise.
9277 (vshrn_n_u16): Likewise.
9278 (vshrn_n_u32): Likewise.
9279 (vshrn_n_u64): Likewise.
9280 (vrshrn_n_s16): Likewise.
9281 (vrshrn_n_s32): Likewise.
9282 (vrshrn_n_s64): Likewise.
9283 (vrshrn_n_u16): Likewise.
9284 (vrshrn_n_u32): Likewise.
9285 (vrshrn_n_u64): Likewise.
9286 * config/aarch64/aarch64-simd.md
9287 (*aarch64_<srn_op>shrn<mode><vczle><vczbe>): Delete.
9288 (aarch64_shrn<mode>): Likewise.
9289 (aarch64_rshrn<mode><vczle><vczbe>_insn): Likewise.
9290 (aarch64_rshrn<mode>): Likewise.
9291 (aarch64_<sur>q<r>shr<u>n_n<mode>_insn<vczle><vczbe>): Likewise.
9292 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
9293 (*aarch64_<shrn_op>shrn_n<mode>_insn<vczle><vczbe>): New define_insn.
9294 (*aarch64_<shrn_op>rshrn_n<mode>_insn<vczle><vczbe>): Likewise.
9295 (*aarch64_sqshrun_n<mode>_insn<vczle><vczbe>): Likewise.
9296 (*aarch64_sqrshrun_n<mode>_insn<vczle><vczbe>): Likewise.
9297 (aarch64_<shrn_op>shrn_n<mode>): New define_expand.
9298 (aarch64_<shrn_op>rshrn_n<mode>): Likewise.
9299 (aarch64_sqshrun_n<mode>): Likewise.
9300 (aarch64_sqrshrun_n<mode>): Likewise.
9301 * config/aarch64/iterators.md (ALL_TRUNC): New code iterator.
9302 (TRUNCEXTEND): New code attribute.
9303 (TRUNC_SHIFT): Likewise.
9304 (shrn_op): Likewise.
9305 * config/aarch64/predicates.md (aarch64_simd_umax_quarter_mode):
9308 2023-06-16 Pan Li <pan2.li@intel.com>
9310 * config/riscv/riscv-vsetvl.cc
9311 (pass_vsetvl::global_eliminate_vsetvl_insn): Initialize var by NULL.
9313 2023-06-16 Richard Biener <rguenther@suse.de>
9315 PR tree-optimization/110278
9316 * match.pd (uns < (typeof uns)(uns != 0) -> false): New.
9317 (x != (typeof x)(x == 0) -> true): Likewise.
9319 2023-06-16 Pali Rohár <pali@kernel.org>
9321 * config/i386/mingw-w64.h (CPP_SPEC): Adjust for -mcrtdll=.
9322 (REAL_LIBGCC_SPEC): New define.
9323 * config/i386/mingw.opt: Add mcrtdll=
9324 * config/i386/mingw32.h (CPP_SPEC): Adjust for -mcrtdll=.
9325 (REAL_LIBGCC_SPEC): Adjust for -mcrtdll=.
9326 (STARTFILE_SPEC): Adjust for -mcrtdll=.
9327 * doc/invoke.texi: Add mcrtdll= documentation.
9329 2023-06-16 Simon Dardis <simon.dardis@imgtec.com>
9331 * config/mips/mips.cc (enum mips_code_readable_setting):New enmu.
9332 (mips_handle_code_readable_attr):New static function.
9333 (mips_get_code_readable_attr):New static enum function.
9334 (mips_set_current_function):Set the code_readable mode.
9335 (mips_option_override):Same as above.
9336 * doc/extend.texi:Document code_readable.
9338 2023-06-16 Richard Biener <rguenther@suse.de>
9340 PR tree-optimization/110269
9341 * fold-const.cc (fold_binary_loc): Merge x != 0 folding
9342 with tree_expr_nonzero_p ...
9343 * match.pd (cmp (convert? addr@0) integer_zerop): With this
9346 2023-06-15 Marek Polacek <polacek@redhat.com>
9348 * Makefile.in: Set LD_PICFLAG. Use it. Set enable_host_pie.
9349 Remove NO_PIE_CFLAGS and NO_PIE_FLAG. Pass LD_PICFLAG to
9350 ALL_LINKERFLAGS. Use the "pic" build of libiberty if --enable-host-pie.
9351 * configure.ac (--enable-host-shared): Don't set PICFLAG here.
9352 (--enable-host-pie): New check. Set PICFLAG and LD_PICFLAG after this
9354 * configure: Regenerate.
9355 * doc/install.texi: Document --enable-host-pie.
9357 2023-06-15 Manolis Tsamis <manolis.tsamis@vrull.eu>
9359 * regcprop.cc (maybe_mode_change): Enable stack pointer
9362 2023-06-15 Andrew MacLeod <amacleod@redhat.com>
9364 PR tree-optimization/110266
9365 * gimple-range-fold.cc (adjust_imagpart_expr): Check for integer
9367 (adjust_realpart_expr): Ditto.
9369 2023-06-15 Jan Beulich <jbeulich@suse.com>
9371 * config/i386/sse.md (<avx512>_vec_dup<mode><mask_name>): Use
9374 2023-06-15 Jan Beulich <jbeulich@suse.com>
9376 * config/i386/constraints.md: Mention k and r for B.
9378 2023-06-15 Lulu Cheng <chenglulu@loongson.cn>
9379 Andrew Pinski <apinski@marvell.com>
9382 * config/loongarch/loongarch.md: Modify the register constraints for template
9383 "jumptable" and "indirect_jump" from "r" to "e".
9385 2023-06-15 Xi Ruoyao <xry111@xry111.site>
9387 * config/loongarch/loongarch-tune.h (loongarch_align): New
9389 * config/loongarch/loongarch-def.h (loongarch_cpu_align): New
9391 * config/loongarch/loongarch-def.c (loongarch_cpu_align): Define
9393 * config/loongarch/loongarch.cc
9394 (loongarch_option_override_internal): Set the value of
9395 -falign-functions= if -falign-functions is enabled but no value
9396 is given. Likewise for -falign-labels=.
9398 2023-06-15 Jakub Jelinek <jakub@redhat.com>
9401 * internal-fn.def (UADDC, USUBC): New internal functions.
9402 * internal-fn.cc (expand_UADDC, expand_USUBC): New functions.
9403 (commutative_ternary_fn_p): Return true also for IFN_UADDC.
9404 * optabs.def (uaddc5_optab, usubc5_optab): New optabs.
9405 * tree-ssa-math-opts.cc (uaddc_cast, uaddc_ne0, uaddc_is_cplxpart,
9406 match_uaddc_usubc): New functions.
9407 (math_opts_dom_walker::after_dom_children): Call match_uaddc_usubc
9408 for PLUS_EXPR, MINUS_EXPR, BIT_IOR_EXPR and BIT_XOR_EXPR unless
9409 other optimizations have been successful for those.
9410 * gimple-fold.cc (gimple_fold_call): Handle IFN_UADDC and IFN_USUBC.
9411 * fold-const-call.cc (fold_const_call): Likewise.
9412 * gimple-range-fold.cc (adjust_imagpart_expr): Likewise.
9413 * tree-ssa-dce.cc (eliminate_unnecessary_stmts): Likewise.
9414 * doc/md.texi (uaddc<mode>5, usubc<mode>5): Document new named
9416 * config/i386/i386.md (uaddc<mode>5, usubc<mode>5): New
9417 define_expand patterns.
9418 (*setcc_qi_addqi3_cconly_overflow_1_<mode>, *setccc): Split
9419 into NOTE_INSN_DELETED note rather than nop instruction.
9420 (*setcc_qi_negqi_ccc_1_<mode>, *setcc_qi_negqi_ccc_2_<mode>):
9423 2023-06-15 Jakub Jelinek <jakub@redhat.com>
9426 * config/i386/i386.md (subborrow<mode>): Add alternative with
9427 memory destination and add for it define_peephole2
9428 TARGET_READ_MODIFY_WRITE/-Os patterns to prefer using memory
9429 destination in these patterns.
9431 2023-06-15 Jakub Jelinek <jakub@redhat.com>
9434 * config/i386/i386.md (*sub<mode>_3, @add<mode>3_carry,
9435 addcarry<mode>, @sub<mode>3_carry, *add<mode>3_cc_overflow_1): Add
9436 define_peephole2 TARGET_READ_MODIFY_WRITE/-Os patterns to prefer
9437 using memory destination in these patterns.
9439 2023-06-15 Jakub Jelinek <jakub@redhat.com>
9441 * gimple-fold.cc (gimple_fold_call): Move handling of arg0
9442 as well as arg1 INTEGER_CSTs for .UBSAN_CHECK_{ADD,SUB,MUL}
9443 and .{ADD,SUB,MUL}_OVERFLOW calls from here...
9444 * fold-const-call.cc (fold_const_call): ... here.
9446 2023-06-15 Oluwatamilore Adebayo <oluwatamilore.adebayo@arm.com>
9448 * config/aarch64/aarch64-simd.md (aarch64_<su>abd<mode>):
9449 Rename to <su>abd<mode>3.
9450 * config/aarch64/aarch64-sve.md (<su>abd<mode>_3): Rename
9453 2023-06-15 Oluwatamilore Adebayo <oluwatamilore.adebayo@arm.com>
9455 * doc/md.texi (sabd, uabd): Document them.
9456 * internal-fn.def (ABD): Use new optab.
9457 * optabs.def (sabd_optab, uabd_optab): New optabs,
9458 * tree-vect-patterns.cc (vect_recog_absolute_difference):
9459 Recognize the following idiom abs (a - b).
9460 (vect_recog_sad_pattern): Refactor to use
9461 vect_recog_absolute_difference.
9462 (vect_recog_abd_pattern): Use patterns found by
9463 vect_recog_absolute_difference to build a new ABD
9466 2023-06-15 chenxiaolong <chenxl04200420@163.com>
9468 * config/loongarch/loongarch.h (LARCH_CALL_RATIO): Modify the value
9469 of macro LARCH_CALL_RATIO on LoongArch to make it perform optimally.
9471 2023-06-15 Juzhe-Zhong <juzhe.zhong@rivai.ai>
9473 * config/riscv/riscv-v.cc (shuffle_merge_patterns): New pattern.
9474 (expand_vec_perm_const_1): Add merge optmization.
9476 2023-06-15 Lehua Ding <lehua.ding@rivai.ai>
9479 * config/riscv/riscv.cc (riscv_get_arg_info): Return NULL_RTX for vector mode
9480 (riscv_pass_by_reference): Return true for vector mode
9482 2023-06-15 Pan Li <pan2.li@intel.com>
9484 * config/riscv/autovec-opt.md: Align the predictor sytle.
9485 * config/riscv/autovec.md: Ditto.
9487 2023-06-15 Pan Li <pan2.li@intel.com>
9489 * config/riscv/riscv-v.cc (rvv_builder::get_merge_scalar_mask):
9490 Take elen instead of scalar BITS_PER_WORD.
9491 (expand_vector_init_merge_repeating_sequence): Use inner_bits_size
9492 instead of scaler BITS_PER_WORD.
9494 2023-06-14 Jivan Hakobyan <jivanhakobyan9@gmail.com>
9496 * config/moxie/uclinux.h (MFWRAP_SPEC): Remove
9498 2023-06-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9500 * config/aarch64/aarch64-sve-builtins-base.cc (svlast_impl::fold):
9501 Fix signed comparison warning in loop from npats to enelts.
9503 2023-06-14 Thomas Schwinge <thomas@codesourcery.com>
9505 * gcc.cc (driver_handle_option): Forward host '-lgfortran', '-lm'
9506 to offloading compilation.
9507 * config/gcn/mkoffload.cc (main): Adjust.
9508 * config/nvptx/mkoffload.cc (main): Likewise.
9509 * doc/invoke.texi (foffload-options): Update example.
9511 2023-06-14 liuhongt <hongtao.liu@intel.com>
9514 * config/i386/sse.md (mov<mode>_internal>): Use x instead of v
9515 for alternative 2 since there's no evex version for vpcmpeqd
9518 2023-06-13 Jeff Law <jlaw@ventanamicro.com>
9520 * gcc.cc (LINK_COMMAND_SPEC): Remove mudflap spec handling.
9522 2023-06-13 Jeff Law <jlaw@ventanamicro.com>
9524 * config/sh/divtab.cc: Remove.
9526 2023-06-13 Jakub Jelinek <jakub@redhat.com>
9528 * config/i386/i386.cc (standard_sse_constant_opcode): Remove
9529 superfluous spaces around \t for vpcmpeqd.
9531 2023-06-13 Roger Sayle <roger@nextmovesoftware.com>
9533 * expr.cc (store_constructor) <case VECTOR_TYPE>: Don't bother
9534 clearing vectors with only a single element. Set CLEARED if the
9535 vector was initialized to zero.
9537 2023-06-13 Lehua Ding <lehua.ding@rivai.ai>
9539 * config/riscv/riscv-v.cc (struct mode_vtype_group): Remove duplicate
9542 (TUPLE_ENTRY): Undef.
9544 2023-06-13 Juzhe-Zhong <juzhe.zhong@rivai.ai>
9546 * config/riscv/riscv-v.cc (rvv_builder::single_step_npatterns_p): Add comment.
9547 (shuffle_generic_patterns): Ditto.
9548 (expand_vec_perm_const_1): Ditto.
9550 2023-06-13 Juzhe-Zhong <juzhe.zhong@rivai.ai>
9552 * config/riscv/riscv-v.cc (emit_vlmax_decompress_insn): Fix bug.
9553 (shuffle_decompress_patterns): Ditto.
9555 2023-06-13 Richard Biener <rguenther@suse.de>
9557 * tree-ssa-loop-ch.cc (ch_base::copy_headers): Free loop BBs.
9559 2023-06-13 Yanzhang Wang <yanzhang.wang@intel.com>
9560 Kito Cheng <kito.cheng@sifive.com>
9562 * config/riscv/riscv-protos.h (riscv_init_cumulative_args): Set
9563 warning flag if func is not builtin
9564 * config/riscv/riscv.cc
9565 (riscv_scalable_vector_type_p): Determine whether the type is scalable vector.
9566 (riscv_arg_has_vector): Determine whether the arg is vector type.
9567 (riscv_pass_in_vector_p): Check the vector type param is passed by value.
9568 (riscv_init_cumulative_args): The same as header.
9569 (riscv_get_arg_info): Add the checking.
9570 (riscv_function_value): Check the func return and set warning flag
9571 * config/riscv/riscv.h (INIT_CUMULATIVE_ARGS): Add a flag to
9572 determine whether warning psabi or not.
9574 2023-06-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9576 * config/arm/arm-opts.h (enum arm_tp_type): Remove TP_CP15.
9577 Add TP_TPIDRURW, TP_TPIDRURO, TP_TPIDRPRW values.
9578 * config/arm/arm-protos.h (arm_output_load_tpidr): Declare prototype.
9579 * config/arm/arm.cc (arm_option_reconfigure_globals): Replace TP_CP15
9581 (arm_output_load_tpidr): Define.
9582 * config/arm/arm.h (TARGET_HARD_TP): Define in terms of TARGET_SOFT_TP.
9583 * config/arm/arm.md (load_tp_hard): Call arm_output_load_tpidr to output
9585 (reload_tp_hard): Likewise.
9586 * config/arm/arm.opt (tpidrurw, tpidruro, tpidrprw): New values for
9588 * doc/invoke.texi (Arm Options, mtp): Document new values.
9590 2023-06-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9593 * config/aarch64/aarch64-opts.h (enum aarch64_tp_reg): Add
9594 AARCH64_TPIDRRO_EL0 value.
9595 * config/aarch64/aarch64.cc (aarch64_output_load_tp): Define.
9596 * config/aarch64/aarch64.opt (tpidr_el0, tpidr_el1, tpidr_el2,
9597 tpidr_el3, tpidrro_el3): New accepted values to -mtp=.
9598 * doc/invoke.texi (AArch64 Options): Document new -mtp= options.
9600 2023-06-13 Alexandre Oliva <oliva@adacore.com>
9602 * range-op-float.cc (frange_nextafter): Drop inline.
9603 (frelop_early_resolve): Add static.
9604 (frange_float): Likewise.
9606 2023-06-13 Richard Biener <rguenther@suse.de>
9608 PR middle-end/110232
9609 * fold-const.cc (native_interpret_vector): Use TYPE_SIZE_UNIT
9610 to check whether the buffer covers the whole vector.
9612 2023-06-13 Richard Biener <rguenther@suse.de>
9614 * tree-ssa-alias.cc (ref_maybe_used_by_call_p_1): For
9615 .MASK_LOAD and friends set the size of the access to unknown.
9617 2023-06-13 Tejas Belagod <tbelagod@arm.com>
9620 * config/aarch64/aarch64-sve-builtins-base.cc (svlast_impl::fold): Fold sve
9621 calls that have a constant input predicate vector.
9622 (svlast_impl::is_lasta): Query to check if intrinsic is svlasta.
9623 (svlast_impl::is_lastb): Query to check if intrinsic is svlastb.
9624 (svlast_impl::vect_all_same): Check if all vector elements are equal.
9626 2023-06-13 Andi Kleen <ak@linux.intel.com>
9628 * config/i386/gcc-auto-profile: Regenerate.
9630 2023-06-13 Juzhe-Zhong <juzhe.zhong@rivai.ai>
9632 * config/riscv/vector-iterators.md: Fix requirement.
9634 2023-06-13 Juzhe-Zhong <juzhe.zhong@rivai.ai>
9636 * config/riscv/riscv-v.cc (emit_vlmax_decompress_insn): New function.
9637 (shuffle_decompress_patterns): New function.
9638 (expand_vec_perm_const_1): Add decompress optimization.
9640 2023-06-12 Jeff Law <jlaw@ventanamicro.com>
9642 PR rtl-optimization/101188
9643 * postreload.cc (reload_cse_move2add_invalidate): New function,
9645 (reload_cse_move2add): Call reload_cse_move2add_invalidate.
9647 2023-06-12 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
9649 * config/aarch64/aarch64.cc (aarch64_expand_vector_init): Tweak condition
9650 if (n_var == n_elts && n_elts <= 16) to allow a single constant,
9651 and if maxv == 1, use constant element for duplicating into register.
9653 2023-06-12 Tobias Burnus <tobias@codesourcery.com>
9655 * gimplify.cc (gimplify_adjust_omp_clauses_1): Use
9656 GOMP_MAP_FORCE_PRESENT for 'present alloc' implicit mapping.
9657 (gimplify_adjust_omp_clauses): Change
9658 GOMP_MAP_PRESENT_{TO,TOFROM,FROM,ALLOC} to the equivalent
9659 GOMP_MAP_FORCE_PRESENT.
9660 * omp-low.cc (lower_omp_target): Remove handling of no-longer valid
9661 GOMP_MAP_PRESENT_{TO,TOFROM,FROM,ALLOC}; update map kinds used for
9662 to/from clauses with present modifier.
9664 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9666 PR tree-optimization/110205
9667 * range-op-float.cc (range_operator::fold_range): Add default FII
9669 * range-op-mixed.h (class operator_gt): Add missing final overrides.
9670 * range-op.cc (range_op_handler::fold_range): Add RO_FII case.
9671 (operator_lshift ::update_bitmask): Add final override.
9672 (operator_rshift ::update_bitmask): Add final override.
9673 * range-op.h (range_operator::fold_range): Add FII prototype.
9675 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9677 * gimple-range-op.cc (gimple_range_op_handler::maybe_non_standard):
9678 Use range_op_handler directly.
9679 * range-op.cc (range_op_handler::range_op_handler): Unsigned
9680 param instead of tree-code.
9681 (ptr_op_widen_plus_signed): Delete.
9682 (ptr_op_widen_plus_unsigned): Delete.
9683 (ptr_op_widen_mult_signed): Delete.
9684 (ptr_op_widen_mult_unsigned): Delete.
9685 (range_op_table::initialize_integral_ops): Add new opcodes.
9686 * range-op.h (range_op_handler): Use unsigned.
9687 (OP_WIDEN_MULT_SIGNED): New.
9688 (OP_WIDEN_MULT_UNSIGNED): New.
9689 (OP_WIDEN_PLUS_SIGNED): New.
9690 (OP_WIDEN_PLUS_UNSIGNED): New.
9691 (RANGE_OP_TABLE_SIZE): New.
9692 (range_op_table::operator []): Use unsigned.
9693 (range_op_table::set): Use unsigned.
9694 (m_range_tree): Make unsigned.
9695 (ptr_op_widen_mult_signed): Remove.
9696 (ptr_op_widen_mult_unsigned): Remove.
9697 (ptr_op_widen_plus_signed): Remove.
9698 (ptr_op_widen_plus_unsigned): Remove.
9700 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9702 * gimple-range-op.cc (gimple_range_op_handler): Set m_operator
9703 manually as there is no access to the default operator.
9704 (cfn_copysign::fold_range): Don't check for validity.
9705 (cfn_ubsan::fold_range): Ditto.
9706 (gimple_range_op_handler::maybe_builtin_call): Don't set to NULL.
9707 * range-op.cc (default_operator): New.
9708 (range_op_handler::range_op_handler): Use default_operator
9710 (range_op_handler::operator bool): Move from header, compare
9711 against default operator.
9712 (range_op_handler::range_op): New.
9713 * range-op.h (range_op_handler::operator bool): Move.
9715 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9717 * range-op.cc (unified_table): Delete.
9718 (range_op_table operator_table): Instantiate.
9719 (range_op_table::range_op_table): Rename from unified_table.
9720 (range_op_handler::range_op_handler): Use range_op_table.
9721 * range-op.h (range_op_table::operator []): Inline.
9722 (range_op_table::set): Inline.
9724 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9726 * gimple-range-gori.cc (gori_compute::condexpr_adjust): Do not
9728 * gimple-range-op.cc (get_code): Rename from get_code_and_type
9730 (gimple_range_op_handler::supported_p): No need for type.
9731 (gimple_range_op_handler::gimple_range_op_handler): Ditto.
9732 (cfn_copysign::fold_range): Ditto.
9733 (cfn_ubsan::fold_range): Ditto.
9734 * ipa-cp.cc (ipa_vr_operation_and_type_effects): Ditto.
9735 * ipa-fnsummary.cc (evaluate_conditions_for_known_args): Ditto.
9736 * range-op-float.cc (operator_plus::op1_range): Ditto.
9737 (operator_mult::op1_range): Ditto.
9738 (range_op_float_tests): Ditto.
9739 * range-op.cc (get_op_handler): Remove.
9740 (range_op_handler::set_op_handler): Remove.
9741 (operator_plus::op1_range): No need for type.
9742 (operator_minus::op1_range): Ditto.
9743 (operator_mult::op1_range): Ditto.
9744 (operator_exact_divide::op1_range): Ditto.
9745 (operator_cast::op1_range): Ditto.
9746 (perator_bitwise_not::fold_range): Ditto.
9747 (operator_negate::fold_range): Ditto.
9748 * range-op.h (range_op_handler::range_op_handler): Remove type param.
9749 (range_cast): No need for type.
9750 (range_op_table::operator[]): Check for enum_code >= 0.
9751 * tree-data-ref.cc (compute_distributive_range): No need for type.
9752 * tree-ssa-loop-unswitch.cc (unswitch_predicate): Ditto.
9753 * value-query.cc (range_query::get_tree_range): Ditto.
9754 * value-relation.cc (relation_oracle::validate_relation): Ditto.
9755 * vr-values.cc (range_of_var_in_loop): Ditto.
9756 (simplify_using_ranges::fold_cond_with_ops): Ditto.
9758 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9760 * range-op-mixed.h (operator_max): Remove final.
9761 * range-op-ptr.cc (pointer_table::pointer_table): Remove MAX_EXPR.
9762 (pointer_table::pointer_table): Remove.
9763 (class hybrid_max_operator): New.
9764 (range_op_table::initialize_pointer_ops): Add hybrid_max_operator.
9765 * range-op.cc (pointer_tree_table): Remove.
9766 (unified_table::unified_table): Comment out MAX_EXPR.
9767 (get_op_handler): Remove check of pointer table.
9768 * range-op.h (class pointer_table): Remove.
9770 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9772 * range-op-mixed.h (operator_min): Remove final.
9773 * range-op-ptr.cc (pointer_table::pointer_table): Remove MIN_EXPR.
9774 (class hybrid_min_operator): New.
9775 (range_op_table::initialize_pointer_ops): Add hybrid_min_operator.
9776 * range-op.cc (unified_table::unified_table): Comment out MIN_EXPR.
9778 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9780 * range-op-mixed.h (operator_bitwise_or): Remove final.
9781 * range-op-ptr.cc (pointer_table::pointer_table): Remove BIT_IOR_EXPR.
9782 (class hybrid_or_operator): New.
9783 (range_op_table::initialize_pointer_ops): Add hybrid_or_operator.
9784 * range-op.cc (unified_table::unified_table): Comment out BIT_IOR_EXPR.
9786 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9788 * range-op-mixed.h (operator_bitwise_and): Remove final.
9789 * range-op-ptr.cc (pointer_table::pointer_table): Remove BIT_AND_EXPR.
9790 (class hybrid_and_operator): New.
9791 (range_op_table::initialize_pointer_ops): Add hybrid_and_operator.
9792 * range-op.cc (unified_table::unified_table): Comment out BIT_AND_EXPR.
9794 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9796 * Makefile.in (OBJS): Add range-op-ptr.o.
9797 * range-op-mixed.h (update_known_bitmask): Move prototype here.
9798 (minus_op1_op2_relation_effect): Move prototype here.
9799 (wi_includes_zero_p): Move function to here.
9801 * range-op.cc (update_known_bitmask): Remove static.
9802 (wi_includes_zero_p): Move to header.
9803 (wi_zero_p): Move to header.
9804 (minus_op1_op2_relation_effect): Remove static.
9805 (operator_pointer_diff): Move class and routines to range-op-ptr.cc.
9806 (pointer_plus_operator): Ditto.
9807 (pointer_min_max_operator): Ditto.
9808 (pointer_and_operator): Ditto.
9809 (pointer_or_operator): Ditto.
9810 (pointer_table): Ditto.
9811 (range_op_table::initialize_pointer_ops): Ditto.
9812 * range-op-ptr.cc: New.
9814 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9816 * range-op-mixed.h (class operator_max): Move from...
9817 * range-op.cc (unified_table::unified_table): Add MAX_EXPR.
9818 (get_op_handler): Remove the integral table.
9819 (class operator_max): Move from here.
9820 (integral_table::integral_table): Delete.
9821 * range-op.h (class integral_table): Delete.
9823 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9825 * range-op-mixed.h (class operator_min): Move from...
9826 * range-op.cc (unified_table::unified_table): Add MIN_EXPR.
9827 (class operator_min): Move from here.
9828 (integral_table::integral_table): Remove MIN_EXPR.
9830 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9832 * range-op-mixed.h (class operator_bitwise_or): Move from...
9833 * range-op.cc (unified_table::unified_table): Add BIT_IOR_EXPR.
9834 (class operator_bitwise_or): Move from here.
9835 (integral_table::integral_table): Remove BIT_IOR_EXPR.
9837 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9839 * range-op-mixed.h (class operator_bitwise_and): Move from...
9840 * range-op.cc (unified_table::unified_table): Add BIT_AND_EXPR.
9841 (get_op_handler): Check for a pointer table entry first.
9842 (class operator_bitwise_and): Move from here.
9843 (integral_table::integral_table): Remove BIT_AND_EXPR.
9845 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9847 * range-op-mixed.h (class operator_bitwise_xor): Move from...
9848 * range-op.cc (unified_table::unified_table): Add BIT_XOR_EXPR.
9849 (class operator_bitwise_xor): Move from here.
9850 (integral_table::integral_table): Remove BIT_XOR_EXPR.
9851 (pointer_table::pointer_table): Remove BIT_XOR_EXPR.
9853 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9855 * range-op-mixed.h (class operator_bitwise_not): Move from...
9856 * range-op.cc (unified_table::unified_table): Add BIT_NOT_EXPR.
9857 (class operator_bitwise_not): Move from here.
9858 (integral_table::integral_table): Remove BIT_NOT_EXPR.
9859 (pointer_table::pointer_table): Remove BIT_NOT_EXPR.
9861 2023-06-12 Andrew MacLeod <amacleod@redhat.com>
9863 * range-op-mixed.h (class operator_addr_expr): Move from...
9864 * range-op.cc (unified_table::unified_table): Add ADDR_EXPR.
9865 (class operator_addr_expr): Move from here.
9866 (integral_table::integral_table): Remove ADDR_EXPR.
9867 (pointer_table::pointer_table): Remove ADDR_EXPR.
9869 2023-06-12 Pan Li <pan2.li@intel.com>
9871 * config/riscv/riscv-vector-builtins-types.def
9872 (vfloat16m1_t): Add type to lmul1 ops.
9873 (vfloat16m2_t): Likewise.
9874 (vfloat16m4_t): Likewise.
9876 2023-06-12 Richard Biener <rguenther@suse.de>
9878 * tree-ssa-alias.cc (call_may_clobber_ref_p_1): For
9879 .MASK_STORE and friend set the size of the access to
9882 2023-06-12 Tamar Christina <tamar.christina@arm.com>
9884 * config.in: Regenerate.
9885 * configure: Regenerate.
9886 * configure.ac: Remove DEFAULT_MATCHPD_PARTITIONS.
9888 2023-06-12 Juzhe-Zhong <juzhe.zhong@rivai.ai>
9890 * config/riscv/autovec-opt.md
9891 (*v<any_shiftrt:optab><any_extend:optab>trunc<mode>): New pattern.
9892 (*<any_shiftrt:optab>trunc<mode>): Ditto.
9893 * config/riscv/autovec.md (<optab><mode>3): Change to
9894 define_insn_and_split.
9895 (v<optab><mode>3): Ditto.
9896 (trunc<mode><v_double_trunc>2): Ditto.
9898 2023-06-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9900 * simplify-rtx.cc (simplify_const_unary_operation):
9901 Handle US_TRUNCATE, SS_TRUNCATE.
9903 2023-06-12 Eric Botcazou <ebotcazou@adacore.com>
9906 * doc/gm2.texi (Standard procedures): Fix Next link.
9908 2023-06-12 Tamar Christina <tamar.christina@arm.com>
9910 * config.in: Regenerate.
9912 2023-06-12 Andre Vieira <andre.simoesdiasvieira@arm.com>
9914 PR middle-end/110142
9915 * tree-vect-patterns.cc (vect_recog_widen_op_pattern): Don't pass
9916 subtype to vect_widened_op_tree and remove subtype parameter, also
9917 remove superfluous overloaded function definition.
9918 (vect_recog_widen_plus_pattern): Remove subtype parameter and dont pass
9919 to call to vect_recog_widen_op_pattern.
9920 (vect_recog_widen_minus_pattern): Likewise.
9922 2023-06-12 liuhongt <hongtao.liu@intel.com>
9924 * config/i386/sse.md (vec_pack<floatprefix>_float_<mode>): New expander.
9925 (vec_unpack_<fixprefix>fix_trunc_lo_<mode>): Ditto.
9926 (vec_unpack_<fixprefix>fix_trunc_hi_<mode>): Ditto.
9927 (vec_unpacks_lo_<mode>): Ditto.
9928 (vec_unpacks_hi_<mode>): Ditto.
9929 (sse_movlhps_<mode>): New define_insn.
9930 (ssse3_palignr<mode>_perm): Extend to V_128H.
9931 (V_128H): New mode iterator.
9932 (ssepackPHmode): New mode attribute.
9933 (vunpck_extract_mode): Ditto.
9934 (vpckfloat_concat_mode): Extend to VxSI/VxSF for _Float16.
9935 (vpckfloat_temp_mode): Ditto.
9936 (vpckfloat_op_mode): Ditto.
9937 (vunpckfixt_mode): Extend to VxHF.
9938 (vunpckfixt_model): Ditto.
9939 (vunpckfixt_extract_mode): Ditto.
9941 2023-06-12 Richard Biener <rguenther@suse.de>
9943 PR middle-end/110200
9944 * genmatch.cc (expr::gen_transform): Put braces around
9945 the if arm for the (convert ...) short-cut.
9947 2023-06-12 Kewen Lin <linkw@linux.ibm.com>
9950 * config/rs6000/rs6000-builtins.def (__builtin_pack_vector_int128,
9951 __builtin_unpack_vector_int128): Move from stanza power7 to vsx.
9953 2023-06-12 Kewen Lin <linkw@linux.ibm.com>
9956 * config/rs6000/rs6000.cc (output_toc): Use the mode of the 128-bit
9957 floating constant itself for real_to_target call.
9959 2023-06-12 Pan Li <pan2.li@intel.com>
9961 * config/riscv/riscv-vector-builtins-types.def
9962 (vfloat16mf4_t): Add type to X2/X4/X8/X16/X32 vlmul ext ops.
9963 (vfloat16mf2_t): Ditto.
9964 (vfloat16m1_t): Ditto.
9965 (vfloat16m2_t): Ditto.
9966 (vfloat16m4_t): Ditto.
9968 2023-06-12 David Edelsohn <dje.gcc@gmail.com>
9970 * config/rs6000/rs6000-logue.cc (rs6000_stack_info):
9971 Do not require a stack frame when debugging is enabled for AIX.
9973 2023-06-11 Georg-Johann Lay <avr@gjlay.de>
9975 * config/avr/avr.md (adjust_len) [insv_notbit_0, insv_notbit_7]:
9976 Remove attribute values.
9977 (insv_notbit): New post-reload insn.
9978 (*insv.not-shiftrt_split, *insv.xor1-bit.0_split)
9979 (*insv.not-bit.0_split, *insv.not-bit.7_split)
9980 (*insv.xor-extract_split): Split to insv_notbit.
9981 (*insv.not-shiftrt, *insv.xor1-bit.0, *insv.not-bit.0, *insv.not-bit.7)
9982 (*insv.xor-extract): Remove post-reload insns.
9983 * config/avr/avr.cc (avr_out_insert_notbit) [bitno]: Remove parameter.
9984 (avr_adjust_insn_length): Adjust call of avr_out_insert_notbit.
9985 [ADJUST_LEN_INSV_NOTBIT_0, ADJUST_LEN_INSV_NOTBIT_7]: Remove cases.
9986 * config/avr/avr-protos.h (avr_out_insert_notbit): Adjust prototype.
9988 2023-06-11 Georg-Johann Lay <avr@gjlay.de>
9991 * config/avr/avr.md (adjust_len) [extr, extr_not]: New elements.
9992 (MSB, SIZE): New mode attributes.
9993 (any_shift): New code iterator.
9994 (*lshr<mode>3_split, *lshr<mode>3, lshr<mode>3)
9995 (*lshr<mode>3_const_split): Add constraint alternative for
9996 the case of shift-offset = MSB. Ditch "length" attribute.
9997 (extzv<mode): New. replaces extzv. Adjust following patterns.
9998 Use avr_out_extr, avr_out_extr_not to print asm.
9999 (*extzv.subreg.<mode>, *extzv.<mode>.subreg, *extzv.xor)
10000 (*extzv<mode>.ge, *neg.ashiftrt<mode>.msb, *extzv.io.lsr7): New.
10001 * config/avr/constraints.md (C15, C23, C31, Yil): New
10002 * config/avr/predicates.md (reg_or_low_io_operand)
10003 (const7_operand, reg_or_low_io_operand)
10004 (const15_operand, const_0_to_15_operand)
10005 (const23_operand, const_0_to_23_operand)
10006 (const31_operand, const_0_to_31_operand): New.
10007 * config/avr/avr-protos.h (avr_out_extr, avr_out_extr_not): New.
10008 * config/avr/avr.cc (avr_out_extr, avr_out_extr_not): New funcs.
10009 (lshrqi3_out, lshrhi3_out, lshrpsi3_out, lshrsi3_out): Adjust
10010 MSB case to new insn constraint "r" for operands[1].
10011 (avr_adjust_insn_length) [ADJUST_LEN_EXTR_NOT, ADJUST_LEN_EXTR]:
10012 Handle these cases.
10013 (avr_rtx_costs_1): Adjust cost for a new pattern.
10015 2023-06-11 Juzhe-Zhong <juzhe.zhong@rivai.ai>
10017 * config/riscv/riscv-vsetvl.cc (available_occurrence_p): Enhance user vsetvl optimization.
10018 (vector_insn_info::parse_insn): Add rtx_insn parse.
10019 (pass_vsetvl::local_eliminate_vsetvl_insn): Enhance user vsetvl optimization.
10020 (get_first_vsetvl): New function.
10021 (pass_vsetvl::global_eliminate_vsetvl_insn): Ditto.
10022 (pass_vsetvl::cleanup_insns): Remove it.
10023 (pass_vsetvl::ssa_post_optimization): New function.
10024 (has_no_uses): Ditto.
10025 (pass_vsetvl::propagate_avl): Remove it.
10026 (pass_vsetvl::df_post_optimization): New function.
10027 (pass_vsetvl::lazy_vsetvl): Rework Phase 5 && Phase 6.
10028 * config/riscv/riscv-vsetvl.h: Adapt declaration.
10030 2023-06-10 Aldy Hernandez <aldyh@redhat.com>
10032 * ipa-cp.cc (ipcp_vr_lattice::init): Take type argument.
10033 (ipcp_vr_lattice::print): Call dump method.
10034 (ipcp_vr_lattice::meet_with): Adjust for m_vr being a
10036 (ipcp_vr_lattice::meet_with_1): Make argument a reference.
10037 (ipcp_vr_lattice::set_to_bottom): Set varying for an unsupported
10039 (initialize_node_lattices): Pass type when appropriate.
10040 (ipa_vr_operation_and_type_effects): Make type agnostic.
10041 (ipa_value_range_from_jfunc): Same.
10042 (propagate_vr_across_jump_function): Same.
10043 * ipa-fnsummary.cc (evaluate_conditions_for_known_args): Same.
10044 (evaluate_properties_for_edge): Same.
10045 * ipa-prop.cc (ipa_vr::get_vrange): Same.
10046 (ipcp_update_vr): Same.
10047 * ipa-prop.h (ipa_value_range_from_jfunc): Same.
10048 (ipa_range_set_and_normalize): Same.
10050 2023-06-10 Georg-Johann Lay <avr@gjlay.de>
10054 * config/avr/avr-passes.def (avr_pass_ifelse): Insert new pass.
10055 * config/avr/avr.cc (avr_pass_ifelse): New RTL pass.
10056 (avr_pass_data_ifelse): New pass_data for it.
10057 (make_avr_pass_ifelse, avr_redundant_compare, avr_cbranch_cost)
10058 (avr_canonicalize_comparison, avr_out_plus_set_ZN)
10059 (avr_out_cmp_ext): New functions.
10060 (compare_condtition): Make sure REG_CC dies in the branch insn.
10061 (avr_rtx_costs_1): Add computation of cbranch costs.
10062 (avr_adjust_insn_length) [ADJUST_LEN_ADD_SET_ZN, ADJUST_LEN_CMP_ZEXT]:
10063 [ADJUST_LEN_CMP_SEXT]Handle them.
10064 (TARGET_CANONICALIZE_COMPARISON): New define.
10065 (avr_simplify_comparison_p, compare_diff_p, avr_compare_pattern)
10066 (avr_reorg_remove_redundant_compare, avr_reorg): Remove functions.
10067 (TARGET_MACHINE_DEPENDENT_REORG): Remove define.
10068 * config/avr/avr-protos.h (avr_simplify_comparison_p): Remove proto.
10069 (make_avr_pass_ifelse, avr_out_plus_set_ZN, cc_reg_rtx)
10070 (avr_out_cmp_zext): New Protos
10071 * config/avr/avr.md (branch, difficult_branch): Don't split insns.
10072 (*cbranchhi.zero-extend.0", *cbranchhi.zero-extend.1")
10073 (*swapped_tst<mode>, *add.for.eqne.<mode>): New insns.
10074 (*cbranch<mode>4): Rename to cbranch<mode>4_insn.
10075 (define_peephole): Add dead_or_set_regno_p(insn,REG_CC) as needed.
10076 (define_deephole2): Add peep2_regno_dead_p(*,REG_CC) as needed.
10077 Add new RTL peepholes for decrement-and-branch and *swapped_tst<mode>.
10078 Rework signtest-and-branch peepholes for *sbrx_branch<mode>.
10079 (adjust_len) [add_set_ZN, cmp_zext]: New.
10080 (QIPSI): New mode iterator.
10081 (ALLs1, ALLs2, ALLs4, ALLs234): New mode iterators.
10082 (gelt): New code iterator.
10083 (gelt_eqne): New code attribute.
10084 (rvbranch, *rvbranch, difficult_rvbranch, *difficult_rvbranch)
10085 (branch_unspec, *negated_tst<mode>, *reversed_tst<mode>)
10086 (*cmpqi_sign_extend): Remove insns.
10087 (define_c_enum "unspec") [UNSPEC_IDENTITY]: Remove.
10088 * config/avr/avr-dimode.md (cbranch<mode>4): Canonicalize comparisons.
10089 * config/avr/predicates.md (scratch_or_d_register_operand): New.
10090 * config/avr/constraints.md (Yxx): New constraint.
10092 2023-06-10 Juzhe-Zhong <juzhe.zhong@rivai.ai>
10094 * config/riscv/autovec.md (select_vl<mode>): New pattern.
10095 * config/riscv/riscv-protos.h (expand_select_vl): New function.
10096 * config/riscv/riscv-v.cc (expand_select_vl): Ditto.
10098 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
10100 * range-op-float.cc (foperator_mult_div_base): Delete.
10101 (foperator_mult_div_base::find_range): Make static local function.
10102 (foperator_mult): Remove. Move prototypes to range-op-mixed.h
10103 (operator_mult::op1_range): Rename from foperator_mult.
10104 (operator_mult::op2_range): Ditto.
10105 (operator_mult::rv_fold): Ditto.
10106 (float_table::float_table): Remove MULT_EXPR.
10107 (class foperator_div): Inherit from range_operator.
10108 (float_table::float_table): Delete.
10109 * range-op-mixed.h (class operator_mult): Combined from integer
10111 * range-op.cc (float_tree_table): Delete.
10112 (op_mult): New object.
10113 (unified_table::unified_table): Add MULT_EXPR.
10114 (get_op_handler): Do not check float table any longer.
10115 (class cross_product_operator): Move to range-op-mixed.h.
10116 (class operator_mult): Move to range-op-mixed.h.
10117 (integral_table::integral_table): Remove MULT_EXPR.
10118 (pointer_table::pointer_table): Remove MULT_EXPR.
10119 * range-op.h (float_table): Remove.
10121 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
10123 * range-op-float.cc (foperator_negate): Remove. Move prototypes
10124 to range-op-mixed.h
10125 (operator_negate::fold_range): Rename from foperator_negate.
10126 (operator_negate::op1_range): Ditto.
10127 (float_table::float_table): Remove NEGATE_EXPR.
10128 * range-op-mixed.h (class operator_negate): Combined from integer
10130 * range-op.cc (op_negate): New object.
10131 (unified_table::unified_table): Add NEGATE_EXPR.
10132 (class operator_negate): Move to range-op-mixed.h.
10133 (integral_table::integral_table): Remove NEGATE_EXPR.
10134 (pointer_table::pointer_table): Remove NEGATE_EXPR.
10136 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
10138 * range-op-float.cc (foperator_minus): Remove. Move prototypes
10139 to range-op-mixed.h
10140 (operator_minus::fold_range): Rename from foperator_minus.
10141 (operator_minus::op1_range): Ditto.
10142 (operator_minus::op2_range): Ditto.
10143 (operator_minus::rv_fold): Ditto.
10144 (float_table::float_table): Remove MINUS_EXPR.
10145 * range-op-mixed.h (class operator_minus): Combined from integer
10147 * range-op.cc (op_minus): New object.
10148 (unified_table::unified_table): Add MINUS_EXPR.
10149 (class operator_minus): Move to range-op-mixed.h.
10150 (integral_table::integral_table): Remove MINUS_EXPR.
10151 (pointer_table::pointer_table): Remove MINUS_EXPR.
10153 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
10155 * range-op-float.cc (foperator_abs): Remove. Move prototypes
10156 to range-op-mixed.h
10157 (operator_abs::fold_range): Rename from foperator_abs.
10158 (operator_abs::op1_range): Ditto.
10159 (float_table::float_table): Remove ABS_EXPR.
10160 * range-op-mixed.h (class operator_abs): Combined from integer
10162 * range-op.cc (op_abs): New object.
10163 (unified_table::unified_table): Add ABS_EXPR.
10164 (class operator_abs): Move to range-op-mixed.h.
10165 (integral_table::integral_table): Remove ABS_EXPR.
10166 (pointer_table::pointer_table): Remove ABS_EXPR.
10168 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
10170 * range-op-float.cc (foperator_plus): Remove. Move prototypes
10171 to range-op-mixed.h
10172 (operator_plus::fold_range): Rename from foperator_plus.
10173 (operator_plus::op1_range): Ditto.
10174 (operator_plus::op2_range): Ditto.
10175 (operator_plus::rv_fold): Ditto.
10176 (float_table::float_table): Remove PLUS_EXPR.
10177 * range-op-mixed.h (class operator_plus): Combined from integer
10179 * range-op.cc (op_plus): New object.
10180 (unified_table::unified_table): Add PLUS_EXPR.
10181 (class operator_plus): Move to range-op-mixed.h.
10182 (integral_table::integral_table): Remove PLUS_EXPR.
10183 (pointer_table::pointer_table): Remove PLUS_EXPR.
10185 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
10187 * range-op-mixed.h (class operator_cast): Combined from integer
10189 * range-op.cc (op_cast): New object.
10190 (unified_table::unified_table): Add op_cast
10191 (class operator_cast): Move to range-op-mixed.h.
10192 (integral_table::integral_table): Remove op_cast
10193 (pointer_table::pointer_table): Remove op_cast.
10195 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
10197 * range-op-float.cc (operator_cst::fold_range): New.
10198 * range-op-mixed.h (class operator_cst): Move from integer file.
10199 * range-op.cc (op_cst): New object.
10200 (unified_table::unified_table): Add op_cst. Also use for REAL_CST.
10201 (class operator_cst): Move to range-op-mixed.h.
10202 (integral_table::integral_table): Remove op_cst.
10203 (pointer_table::pointer_table): Remove op_cst.
10205 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
10207 * range-op-float.cc (foperator_identity): Remove. Move prototypes
10208 to range-op-mixed.h
10209 (operator_identity::fold_range): Rename from foperator_identity.
10210 (operator_identity::op1_range): Ditto.
10211 (float_table::float_table): Remove fop_identity.
10212 * range-op-mixed.h (class operator_identity): Combined from integer
10214 * range-op.cc (op_identity): New object.
10215 (unified_table::unified_table): Add op_identity.
10216 (class operator_identity): Move to range-op-mixed.h.
10217 (integral_table::integral_table): Remove identity.
10218 (pointer_table::pointer_table): Remove identity.
10220 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
10222 * range-op-float.cc (foperator_ge): Remove. Move prototypes
10223 to range-op-mixed.h
10224 (operator_ge::fold_range): Rename from foperator_ge.
10225 (operator_ge::op1_range): Ditto.
10226 (float_table::float_table): Remove GE_EXPR.
10227 * range-op-mixed.h (class operator_ge): Combined from integer
10229 * range-op.cc (op_ge): New object.
10230 (unified_table::unified_table): Add GE_EXPR.
10231 (class operator_ge): Move to range-op-mixed.h.
10232 (ge_op1_op2_relation): Fold into
10233 operator_ge::op1_op2_relation.
10234 (integral_table::integral_table): Remove GE_EXPR.
10235 (pointer_table::pointer_table): Remove GE_EXPR.
10236 * range-op.h (ge_op1_op2_relation): Delete.
10238 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
10240 * range-op-float.cc (foperator_gt): Remove. Move prototypes
10241 to range-op-mixed.h
10242 (operator_gt::fold_range): Rename from foperator_gt.
10243 (operator_gt::op1_range): Ditto.
10244 (float_table::float_table): Remove GT_EXPR.
10245 * range-op-mixed.h (class operator_gt): Combined from integer
10247 * range-op.cc (op_gt): New object.
10248 (unified_table::unified_table): Add GT_EXPR.
10249 (class operator_gt): Move to range-op-mixed.h.
10250 (gt_op1_op2_relation): Fold into
10251 operator_gt::op1_op2_relation.
10252 (integral_table::integral_table): Remove GT_EXPR.
10253 (pointer_table::pointer_table): Remove GT_EXPR.
10254 * range-op.h (gt_op1_op2_relation): Delete.
10256 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
10258 * range-op-float.cc (foperator_le): Remove. Move prototypes
10259 to range-op-mixed.h
10260 (operator_le::fold_range): Rename from foperator_le.
10261 (operator_le::op1_range): Ditto.
10262 (float_table::float_table): Remove LE_EXPR.
10263 * range-op-mixed.h (class operator_le): Combined from integer
10265 * range-op.cc (op_le): New object.
10266 (unified_table::unified_table): Add LE_EXPR.
10267 (class operator_le): Move to range-op-mixed.h.
10268 (le_op1_op2_relation): Fold into
10269 operator_le::op1_op2_relation.
10270 (integral_table::integral_table): Remove LE_EXPR.
10271 (pointer_table::pointer_table): Remove LE_EXPR.
10272 * range-op.h (le_op1_op2_relation): Delete.
10274 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
10276 * range-op-float.cc (foperator_lt): Remove. Move prototypes
10277 to range-op-mixed.h
10278 (operator_lt::fold_range): Rename from foperator_lt.
10279 (operator_lt::op1_range): Ditto.
10280 (float_table::float_table): Remove LT_EXPR.
10281 * range-op-mixed.h (class operator_lt): Combined from integer
10283 * range-op.cc (op_lt): New object.
10284 (unified_table::unified_table): Add LT_EXPR.
10285 (class operator_lt): Move to range-op-mixed.h.
10286 (lt_op1_op2_relation): Fold into
10287 operator_lt::op1_op2_relation.
10288 (integral_table::integral_table): Remove LT_EXPR.
10289 (pointer_table::pointer_table): Remove LT_EXPR.
10290 * range-op.h (lt_op1_op2_relation): Delete.
10292 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
10294 * range-op-float.cc (foperator_not_equal): Remove. Move prototypes
10295 to range-op-mixed.h
10296 (operator_equal::fold_range): Rename from foperator_not_equal.
10297 (operator_equal::op1_range): Ditto.
10298 (float_table::float_table): Remove NE_EXPR.
10299 * range-op-mixed.h (class operator_not_equal): Combined from integer
10301 * range-op.cc (op_equal): New object.
10302 (unified_table::unified_table): Add NE_EXPR.
10303 (class operator_not_equal): Move to range-op-mixed.h.
10304 (not_equal_op1_op2_relation): Fold into
10305 operator_not_equal::op1_op2_relation.
10306 (integral_table::integral_table): Remove NE_EXPR.
10307 (pointer_table::pointer_table): Remove NE_EXPR.
10308 * range-op.h (not_equal_op1_op2_relation): Delete.
10310 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
10312 * range-op-float.cc (foperator_equal): Remove. Move prototypes
10313 to range-op-mixed.h
10314 (operator_equal::fold_range): Rename from foperator_equal.
10315 (operator_equal::op1_range): Ditto.
10316 (float_table::float_table): Remove EQ_EXPR.
10317 * range-op-mixed.h (class operator_equal): Combined from integer
10319 * range-op.cc (op_equal): New object.
10320 (unified_table::unified_table): Add EQ_EXPR.
10321 (class operator_equal): Move to range-op-mixed.h.
10322 (equal_op1_op2_relation): Fold into
10323 operator_equal::op1_op2_relation.
10324 (integral_table::integral_table): Remove EQ_EXPR.
10325 (pointer_table::pointer_table): Remove EQ_EXPR.
10326 * range-op.h (equal_op1_op2_relation): Delete.
10328 2023-06-10 Andrew MacLeod <amacleod@redhat.com>
10330 * range-op-float.cc (class float_table): Move to header.
10331 (float_table::float_table): Move float only operators to...
10332 (range_op_table::initialize_float_ops): Here.
10333 * range-op-mixed.h: New.
10334 * range-op.cc (integral_tree_table, pointer_tree_table): Moved
10336 (float_tree_table): Moved from range-op-float.cc.
10337 (unified_tree_table): New.
10338 (unified_table::unified_table): New. Call initialize routines.
10339 (get_op_handler): Check unified table first.
10340 (range_op_handler::range_op_handler): Handle no type constructor.
10341 (integral_table::integral_table): Move integral only operators to...
10342 (range_op_table::initialize_integral_ops): Here.
10343 (pointer_table::pointer_table): Move pointer only operators to...
10344 (range_op_table::initialize_pointer_ops): Here.
10345 * range-op.h (enum bool_range_state): Move to range-op-mixed.h.
10346 (get_bool_state): Ditto.
10347 (empty_range_varying): Ditto.
10348 (relop_early_resolve): Ditto.
10349 (class range_op_table): Add new init methods for range types.
10350 (class integral_table): Move declaration to here.
10351 (class pointer_table): Move declaration to here.
10352 (class float_table): Move declaration to here.
10354 2023-06-09 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
10355 Richard Sandiford <richard.sandiford@arm.com>
10356 Richard Biener <rguenther@suse.de>
10358 * doc/md.texi: Add SELECT_VL support.
10359 * internal-fn.def (SELECT_VL): Ditto.
10360 * optabs.def (OPTAB_D): Ditto.
10361 * tree-vect-loop-manip.cc (vect_set_loop_controls_directly): Ditto.
10362 * tree-vect-loop.cc (_loop_vec_info::_loop_vec_info): Ditto.
10363 * tree-vect-stmts.cc (get_select_vl_data_ref_ptr): Ditto.
10364 (vectorizable_store): Ditto.
10365 (vectorizable_load): Ditto.
10366 * tree-vectorizer.h (LOOP_VINFO_USING_SELECT_VL_P): Ditto.
10368 2023-06-09 Andrew MacLeod <amacleod@redhat.com>
10371 * ipa-prop.cc (ipa_compute_jump_functions_for_edge): Check param
10374 2023-06-09 Andrew MacLeod <amacleod@redhat.com>
10376 * range-op.cc (range_cast): Move to...
10377 * range-op.h (range_cast): Here and add generic a version.
10379 2023-06-09 Marek Polacek <polacek@redhat.com>
10383 * doc/invoke.texi: Clarify that -Wmissing-field-initializers doesn't
10384 warn about designated initializers in C only.
10386 2023-06-09 Andrew Pinski <apinski@marvell.com>
10388 PR tree-optimization/97711
10389 PR tree-optimization/110155
10390 * match.pd ((zero_one == 0) ? y : z <op> y): Add plus to the op.
10391 ((zero_one != 0) ? z <op> y : y): Likewise.
10393 2023-06-09 Andrew Pinski <apinski@marvell.com>
10395 * match.pd ((zero_one ==/!= 0) ? y : z <op> y): Use
10396 multiply rather than negation/bit_and.
10398 2023-06-09 Andrew Pinski <apinski@marvell.com>
10400 * match.pd (`X & -Y -> X * Y`): Allow for truncation
10401 and the same type for unsigned types.
10403 2023-06-09 Andrew Pinski <apinski@marvell.com>
10405 PR tree-optimization/110165
10406 PR tree-optimization/110166
10407 * match.pd (zero_one_valued_p): Don't accept
10408 signed 1-bit integers.
10410 2023-06-09 Richard Biener <rguenther@suse.de>
10412 * match.pd (two conversions in a row): Use element_precision
10413 to DTRT for VECTOR_TYPE.
10415 2023-06-09 Pan Li <pan2.li@intel.com>
10417 * config/riscv/riscv.md (enabled): Move to another place, and
10418 add fp_vector_disabled to the cond.
10419 (fp_vector_disabled): New attr defined for disabling fp.
10420 * config/riscv/vector-iterators.md: Fix V_WHOLE and V_FRACT.
10422 2023-06-09 Pan Li <pan2.li@intel.com>
10424 * config/riscv/riscv-protos.h (enum frm_field_enum): Adjust
10427 2023-06-09 liuhongt <hongtao.liu@intel.com>
10430 * config/i386/i386.cc (ix86_gimple_fold_builtin): Explicitly
10431 view_convert_expr mask to signed type when folding pblendvb
10434 2023-06-09 liuhongt <hongtao.liu@intel.com>
10437 * config/i386/i386.cc (ix86_gimple_fold_builtin): Fold
10438 _mm{,256,512}_abs_{epi8,epi16,epi32,epi64} into gimple
10439 ABSU_EXPR + VCE, don't fold _mm_abs_{pi8,pi16,pi32} w/o
10441 * config/i386/i386-builtin.def: Replace CODE_FOR_nothing with
10442 real codename for __builtin_ia32_pabs{b,w,d}.
10444 2023-06-08 Andrew MacLeod <amacleod@redhat.com>
10446 * gimple-range-op.cc
10447 (gimple_range_op_handler::gimple_range_op_handler): Adjust.
10448 (gimple_range_op_handler::maybe_builtin_call): Adjust.
10449 * gimple-range-op.h (operand1, operand2): Use m_operator.
10450 * range-op.cc (integral_table, pointer_table): Relocate.
10451 (get_op_handler): Rename from get_handler and handle all types.
10452 (range_op_handler::range_op_handler): Relocate.
10453 (range_op_handler::set_op_handler): Relocate and adjust.
10454 (range_op_handler::range_op_handler): Relocate.
10455 (dispatch_trio): New.
10456 (RO_III, RO_IFI, RO_IFF, RO_FFF, RO_FIF, RO_FII): New consts.
10457 (range_op_handler::dispatch_kind): New.
10458 (range_op_handler::fold_range): Relocate and Use new dispatch value.
10459 (range_op_handler::op1_range): Ditto.
10460 (range_op_handler::op2_range): Ditto.
10461 (range_op_handler::lhs_op1_relation): Ditto.
10462 (range_op_handler::lhs_op2_relation): Ditto.
10463 (range_op_handler::op1_op2_relation): Ditto.
10464 (range_op_handler::set_op_handler): Use m_operator member.
10465 * range-op.h (range_op_handler::operator bool): Use m_operator.
10466 (range_op_handler::dispatch_kind): New.
10467 (range_op_handler::m_valid): Delete.
10468 (range_op_handler::m_int): Delete
10469 (range_op_handler::m_float): Delete
10470 (range_op_handler::m_operator): New.
10471 (range_op_table::operator[]): Relocate from .cc file.
10472 (range_op_table::set): Ditto.
10473 * value-range.h (class vrange): Make range_op_handler a friend.
10475 2023-06-08 Andrew MacLeod <amacleod@redhat.com>
10477 * gimple-range-op.cc (cfn_constant_float_p): Change base class.
10478 (cfn_pass_through_arg1): Adjust using statemenmt.
10479 (cfn_signbit): Change base class, adjust using statement.
10480 (cfn_copysign): Ditto.
10482 (cfn_sincos): Ditto.
10483 * range-op-float.cc (fold_range): Change class to range_operator.
10487 (lhs_op1_relation): Ditto.
10488 (lhs_op2_relation): Ditto.
10489 (op1_op2_relation): Ditto.
10490 (foperator_*): Ditto.
10491 (class float_table): New. Inherit from range_op_table.
10492 (floating_tree_table) Change to range_op_table pointer.
10493 (class floating_op_table): Delete.
10494 * range-op.cc (operator_equal): Adjust using statement.
10495 (operator_not_equal): Ditto.
10496 (operator_lt, operator_le, operator_gt, operator_ge): Ditto.
10497 (operator_minus, operator_cast): Ditto.
10498 (operator_bitwise_and, pointer_plus_operator): Ditto.
10499 (get_float_handle): Change return type.
10500 * range-op.h (range_operator_float): Delete. Relocate all methods
10501 into class range_operator.
10502 (range_op_handler::m_float): Change type to range_operator.
10503 (floating_op_table): Delete.
10504 (floating_tree_table): Change type.
10506 2023-06-08 Andrew MacLeod <amacleod@redhat.com>
10508 * range-op.cc (range_operator::fold_range): Call virtual routine.
10509 (range_operator::update_bitmask): New.
10510 (operator_equal::update_bitmask): New.
10511 (operator_not_equal::update_bitmask): New.
10512 (operator_lt::update_bitmask): New.
10513 (operator_le::update_bitmask): New.
10514 (operator_gt::update_bitmask): New.
10515 (operator_ge::update_bitmask): New.
10516 (operator_ge::update_bitmask): New.
10517 (operator_plus::update_bitmask): New.
10518 (operator_minus::update_bitmask): New.
10519 (operator_pointer_diff::update_bitmask): New.
10520 (operator_min::update_bitmask): New.
10521 (operator_max::update_bitmask): New.
10522 (operator_mult::update_bitmask): New.
10523 (operator_div:operator_div):New.
10524 (operator_div::update_bitmask): New.
10525 (operator_div::m_code): New member.
10526 (operator_exact_divide::operator_exact_divide): New constructor.
10527 (operator_lshift::update_bitmask): New.
10528 (operator_rshift::update_bitmask): New.
10529 (operator_bitwise_and::update_bitmask): New.
10530 (operator_bitwise_or::update_bitmask): New.
10531 (operator_bitwise_xor::update_bitmask): New.
10532 (operator_trunc_mod::update_bitmask): New.
10533 (op_ident, op_unknown, op_ptr_min_max): New.
10534 (op_nop, op_convert): Delete.
10535 (op_ssa, op_paren, op_obj_type): Delete.
10536 (op_realpart, op_imagpart): Delete.
10537 (op_ptr_min, op_ptr_max): Delete.
10538 (pointer_plus_operator:update_bitmask): New.
10539 (range_op_table::set): Do not use m_code.
10540 (integral_table::integral_table): Adjust to single instances.
10541 * range-op.h (range_operator::range_operator): Delete.
10542 (range_operator::m_code): Delete.
10543 (range_operator::update_bitmask): New.
10545 2023-06-08 Andrew MacLeod <amacleod@redhat.com>
10547 * range-op-float.cc (range_operator_float::fold_range): Return
10548 NAN of the result type.
10550 2023-06-08 Jakub Jelinek <jakub@redhat.com>
10552 * optabs.cc (expand_ffs): Add forward declaration.
10553 (expand_doubleword_clz): Rename to ...
10554 (expand_doubleword_clz_ctz_ffs): ... this. Add UNOPTAB argument,
10555 handle also doubleword CTZ and FFS in addition to CLZ.
10556 (expand_unop): Adjust caller. Also call it for doubleword
10557 ctz_optab and ffs_optab.
10559 2023-06-08 Jakub Jelinek <jakub@redhat.com>
10562 * config/i386/i386-expand.cc (ix86_expand_vector_init_general): For
10563 n_words == 2 recurse with mmx_ok as first argument rather than false.
10565 2023-06-07 Roger Sayle <roger@nextmovesoftware.com>
10567 * wide-int.cc (wi::bitreverse_large): Use HOST_WIDE_INT_1U to
10568 avoid sign extension/undefined behaviour when setting each bit.
10570 2023-06-07 Roger Sayle <roger@nextmovesoftware.com>
10571 Uros Bizjak <ubizjak@gmail.com>
10573 * config/i386/i386-expand.cc (ix86_expand_builtin) <handlecarry>:
10574 Use new x86_stc instruction when the carry flag must be set.
10575 * config/i386/i386.cc (ix86_cc_mode): Use CCCmode for *x86_cmc.
10576 (ix86_rtx_costs): Provide accurate rtx_costs for *x86_cmc.
10577 * config/i386/i386.h (TARGET_SLOW_STC): New define.
10578 * config/i386/i386.md (UNSPEC_STC): New UNSPEC for stc.
10579 (x86_stc): New define_insn.
10580 (define_peephole2): Convert x86_stc into alternate implementation
10581 on pentium4 without -Os when a QImode register is available.
10582 (*x86_cmc): New define_insn.
10583 (define_peephole2): Convert *x86_cmc into alternate implementation
10584 on pentium4 without -Os when a QImode register is available.
10585 (*setccc): New define_insn_and_split for a no-op CCCmode move.
10586 (*setcc_qi_negqi_ccc_1_<mode>): New define_insn_and_split to
10587 recognize (and eliminate) the carry flag being copied to itself.
10588 (*setcc_qi_negqi_ccc_2_<mode>): Likewise.
10589 * config/i386/x86-tune.def (X86_TUNE_SLOW_STC): New tuning flag.
10591 2023-06-07 Andrew Pinski <apinski@marvell.com>
10593 * match.pd: Fix comment for the
10594 `(zero_one ==/!= 0) ? y : z <op> y` patterns.
10596 2023-06-07 Jeff Law <jlaw@ventanamicro.com>
10597 Jeff Law <jlaw@ventanamicro.com>
10599 * config/riscv/bitmanip.md (rotrdi3, rotrsi3, rotlsi3): New expanders.
10600 (rotrsi3_sext): Expose generator.
10601 (rotlsi3 pattern): Hide generator.
10602 * config/riscv/riscv-protos.h (riscv_emit_binary): New function
10604 * config/riscv/riscv.cc (riscv_emit_binary): Removed static
10605 * config/riscv/riscv.md (addsi3, subsi3, negsi2): Hide generator.
10606 (mulsi3, <optab>si3): Likewise.
10607 (addsi3, subsi3, negsi2, mulsi3, <optab>si3): New expanders.
10608 (addv<mode>4, subv<mode>4, mulv<mode>4): Use riscv_emit_binary.
10609 (<u>mulsidi3): Likewise.
10610 (addsi3_extended, subsi3_extended, negsi2_extended): Expose generator.
10611 (mulsi3_extended, <optab>si3_extended): Likewise.
10612 (splitter for shadd feeding divison): Update RTL pattern to account
10613 for changes in how 32 bit ops are expanded for TARGET_64BIT.
10614 * loop-iv.cc (get_biv_step_1): Process src of extension when it PLUS.
10616 2023-06-07 Dimitar Dimitrov <dimitar@dinux.eu>
10619 * config/riscv/riscv.cc (riscv_print_operand): Calculate
10620 memmodel only when it is valid.
10622 2023-06-07 Dimitar Dimitrov <dimitar@dinux.eu>
10624 * config/riscv/riscv.cc (riscv_const_insns): Recursively call
10625 for constant element of a vector.
10627 2023-06-07 Jakub Jelinek <jakub@redhat.com>
10629 * match.pd (zero_one_valued_p): Don't handle integer_zerop specially,
10630 instead compare tree_nonzero_bits <= 1U rather than just == 1.
10632 2023-06-07 Alex Coplan <alex.coplan@arm.com>
10635 * config/aarch64/aarch64-builtins.cc (aarch64_general_simulate_builtin):
10637 (aarch64_init_ls64_builtins): ... here. Switch to declaring public ACLE
10638 names for builtins.
10639 (aarch64_general_init_builtins): Ensure we invoke the arm_acle.h
10640 setup if in_lto_p, just like we do for SVE.
10641 * config/aarch64/arm_acle.h: (__arm_ld64b): Delete.
10642 (__arm_st64b): Delete.
10643 (__arm_st64bv): Delete.
10644 (__arm_st64bv0): Delete.
10646 2023-06-07 Alex Coplan <alex.coplan@arm.com>
10649 * config/aarch64/aarch64-builtins.cc (aarch64_expand_builtin_ls64):
10650 Use input operand for the destination address.
10651 * config/aarch64/aarch64.md (st64b): Fix constraint on address
10654 2023-06-07 Alex Coplan <alex.coplan@arm.com>
10657 * config/aarch64/aarch64-builtins.cc (aarch64_init_ls64_builtins_types):
10658 Replace eight consecutive spaces with tabs.
10659 (aarch64_init_ls64_builtins): Likewise.
10660 (aarch64_expand_builtin_ls64): Likewise.
10661 * config/aarch64/aarch64.md (ld64b): Likewise.
10664 (st64bv0): Likewise.
10666 2023-06-07 Vladimir N. Makarov <vmakarov@redhat.com>
10668 * ira-costs.cc: (find_costs_and_classes): Constrain classes of pic
10669 offset table pseudo to a general reg subset.
10671 2023-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10673 * config/aarch64/aarch64-simd.md (aarch64_sqmovun<mode><vczle><vczbe>):
10675 (*aarch64_sqmovun<mode>_insn<vczle><vczbe>): ... This. Reimplement
10677 (aarch64_sqmovun<mode> [SD_HSDI]): Reimplement with RTL codes.
10678 (aarch64_sqxtun2<mode>_le): Likewise.
10679 (aarch64_sqxtun2<mode>_be): Likewise.
10680 (aarch64_sqxtun2<mode>): Adjust for the above.
10681 (aarch64_sqmovun<mode>): New define_expand.
10682 * config/aarch64/iterators.md (UNSPEC_SQXTUN): Delete.
10683 (half_mask): New mode attribute.
10684 * config/aarch64/predicates.md (aarch64_simd_umax_half_mode):
10687 2023-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10689 * config/aarch64/aarch64-simd.md (aarch64_addp<mode><vczle><vczbe>):
10691 (aarch64_addp<mode>_insn): ... This...
10692 (aarch64_addp<mode><vczle><vczbe>_insn): ... And this.
10693 (aarch64_addp<mode>): New define_expand.
10695 2023-06-07 Juzhe-Zhong <juzhe.zhong@rivai.ai>
10697 * config/riscv/riscv-protos.h (expand_vec_perm_const): New function.
10698 * config/riscv/riscv-v.cc
10699 (rvv_builder::can_duplicate_repeating_sequence_p): Support POLY
10701 (rvv_builder::single_step_npatterns_p): New function.
10702 (rvv_builder::npatterns_all_equal_p): Ditto.
10703 (const_vec_all_in_range_p): Support POLY handling.
10704 (gen_const_vector_dup): Ditto.
10705 (emit_vlmax_gather_insn): Add vrgatherei16.
10706 (emit_vlmax_masked_gather_mu_insn): Ditto.
10707 (expand_const_vector): Add VLA SLP const vector support.
10708 (expand_vec_perm): Support POLY.
10709 (struct expand_vec_perm_d): New struct.
10710 (shuffle_generic_patterns): New function.
10711 (expand_vec_perm_const_1): Ditto.
10712 (expand_vec_perm_const): Ditto.
10713 * config/riscv/riscv.cc (riscv_vectorize_vec_perm_const): Ditto.
10714 (TARGET_VECTORIZE_VEC_PERM_CONST): New targethook.
10716 2023-06-07 Andrew Pinski <apinski@marvell.com>
10718 PR middle-end/110117
10719 * expr.cc (expand_single_bit_test): Handle
10720 const_int from expand_expr.
10722 2023-06-07 Andrew Pinski <apinski@marvell.com>
10724 * expr.cc (do_store_flag): Rearrange the
10725 TER code so that it overrides the nonzero bits
10726 info if we had `a & POW2`.
10728 2023-06-07 Andrew Pinski <apinski@marvell.com>
10730 PR tree-optimization/110134
10731 * match.pd (-A CMP -B -> B CMP A): Allow EQ/NE for all integer
10733 (-A CMP CST -> B CMP (-CST)): Likewise.
10735 2023-06-07 Andrew Pinski <apinski@marvell.com>
10737 PR tree-optimization/89263
10738 PR tree-optimization/99069
10739 PR tree-optimization/20083
10740 PR tree-optimization/94898
10741 * match.pd: Add patterns to optimize `a ? onezero : onezero` with
10742 one of the operands are constant.
10744 2023-06-07 Andrew Pinski <apinski@marvell.com>
10746 * match.pd (zero_one_valued_p): Match 0 integer constant
10749 2023-06-07 Pan Li <pan2.li@intel.com>
10751 * config/riscv/riscv-vector-builtins-types.def
10752 (vfloat32mf2_t): Take RVV_REQUIRE_ELEN_FP_16 as requirement.
10753 (vfloat32m1_t): Ditto.
10754 (vfloat32m2_t): Ditto.
10755 (vfloat32m4_t): Ditto.
10756 (vfloat32m8_t): Ditto.
10757 (vint16mf4_t): Ditto.
10758 (vint16mf2_t): Ditto.
10759 (vint16m1_t): Ditto.
10760 (vint16m2_t): Ditto.
10761 (vint16m4_t): Ditto.
10762 (vint16m8_t): Ditto.
10763 (vuint16mf4_t): Ditto.
10764 (vuint16mf2_t): Ditto.
10765 (vuint16m1_t): Ditto.
10766 (vuint16m2_t): Ditto.
10767 (vuint16m4_t): Ditto.
10768 (vuint16m8_t): Ditto.
10769 (vint32mf2_t): Ditto.
10770 (vint32m1_t): Ditto.
10771 (vint32m2_t): Ditto.
10772 (vint32m4_t): Ditto.
10773 (vint32m8_t): Ditto.
10774 (vuint32mf2_t): Ditto.
10775 (vuint32m1_t): Ditto.
10776 (vuint32m2_t): Ditto.
10777 (vuint32m4_t): Ditto.
10778 (vuint32m8_t): Ditto.
10780 2023-06-07 Jason Merrill <jason@redhat.com>
10783 * doc/invoke.texi: Document it.
10785 2023-06-06 Roger Sayle <roger@nextmovesoftware.com>
10787 * doc/rtl.texi (bitreverse, copysign): Document new RTX codes.
10788 * rtl.def (BITREVERSE, COPYSIGN): Define new RTX codes.
10789 * simplify-rtx.cc (simplify_unary_operation_1): Optimize
10790 NOT (BITREVERSE x) as BITREVERSE (NOT x).
10791 Optimize POPCOUNT (BITREVERSE x) as POPCOUNT x.
10792 Optimize PARITY (BITREVERSE x) as PARITY x.
10793 Optimize BITREVERSE (BITREVERSE x) as x.
10794 (simplify_const_unary_operation) <case BITREVERSE>: Evaluate
10795 BITREVERSE of a constant integer at compile-time.
10796 (simplify_binary_operation_1) <case COPYSIGN>: Optimize
10797 COPY_SIGN (x, x) as x. Optimize COPYSIGN (x, C) as ABS x
10798 or NEG (ABS x) for constant C. Optimize COPYSIGN (ABS x, y)
10799 and COPYSIGN (NEG x, y) as COPYSIGN (x, y).
10800 Optimize COPYSIGN (x, ABS y) as ABS x.
10801 Optimize COPYSIGN (COPYSIGN (x, y), z) as COPYSIGN (x, z).
10802 Optimize COPYSIGN (x, COPYSIGN (y, z)) as COPYSIGN (x, z).
10803 (simplify_const_binary_operation): Evaluate COPYSIGN of constant
10804 arguments at compile-time.
10806 2023-06-06 Uros Bizjak <ubizjak@gmail.com>
10808 * rtl.h (function_invariant_p): Change return type from int to bool.
10809 * reload1.cc (function_invariant_p): Change return type from
10810 int to bool and adjust function body accordingly.
10812 2023-06-06 Juzhe-Zhong <juzhe.zhong@rivai.ai>
10814 * config/riscv/autovec-opt.md (*<optab>_fma<mode>): New pattern.
10815 (*single_<optab>mult_plus<mode>): Ditto.
10816 (*double_<optab>mult_plus<mode>): Ditto.
10817 (*sign_zero_extend_fma): Ditto.
10818 (*zero_sign_extend_fma): Ditto.
10819 * config/riscv/riscv-protos.h (enum insn_type): New enum.
10821 2023-06-06 Kwok Cheung Yeung <kcy@codesourcery.com>
10822 Tobias Burnus <tobias@codesourcery.com>
10824 * gimplify.cc (omp_notice_variable): Apply GOVD_MAP_ALLOC_ONLY flag
10825 and defaultmap flags if the defaultmap has GOVD_MAP_FORCE_PRESENT flag
10827 (omp_get_attachment): Handle map clauses with 'present' modifier.
10828 (omp_group_base): Likewise.
10829 (gimplify_scan_omp_clauses): Reorder present maps to come first.
10830 Set GOVD flags for present defaultmaps.
10831 (gimplify_adjust_omp_clauses_1): Set map kind for present defaultmaps.
10832 * omp-low.cc (scan_sharing_clauses): Handle 'always, present' map
10834 (lower_omp_target): Handle map clauses with 'present' modifier.
10835 Handle 'to' and 'from' clauses with 'present'.
10836 * tree-core.h (enum omp_clause_defaultmap_kind): Add
10837 OMP_CLAUSE_DEFAULTMAP_PRESENT defaultmap kind.
10838 * tree-pretty-print.cc (dump_omp_clause): Handle 'map', 'to' and
10839 'from' clauses with 'present' modifier. Handle present defaultmap.
10840 * tree.h (OMP_CLAUSE_MOTION_PRESENT): New #define.
10842 2023-06-06 Segher Boessenkool <segher@kernel.crashing.org>
10844 * config/rs6000/genfusion.pl: Delete some dead code.
10846 2023-06-06 Segher Boessenkool <segher@kernel.crashing.org>
10848 * config/rs6000/genfusion.pl (gen_ld_cmpi_p10_one): New, rewritten and
10850 (gen_ld_cmpi_p10): ... this.
10852 2023-06-06 Jeevitha Palanisamy <jeevitha@linux.ibm.com>
10855 * config/rs6000/rs6000.cc (vec_const_128bit_to_bytes): Remove
10856 duplicate expression.
10858 2023-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10860 * config/aarch64/aarch64-builtins.cc (aarch64_general_gimple_fold_builtin):
10861 Handle unsigned reduc_plus_scal_ builtins.
10862 * config/aarch64/aarch64-simd-builtins.def (addp): Delete DImode instances.
10863 * config/aarch64/aarch64-simd.md (aarch64_addpdi): Delete.
10864 * config/aarch64/arm_neon.h (vpaddd_s64): Reimplement with
10865 __builtin_aarch64_reduc_plus_scal_v2di.
10866 (vpaddd_u64): Reimplement with __builtin_aarch64_reduc_plus_scal_v2di_uu.
10868 2023-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10870 * config/aarch64/aarch64-simd.md (aarch64_<sur>shr_n<mode>): Delete.
10871 (aarch64_<sra_op>rshr_n<mode><vczle><vczbe>_insn): New define_insn.
10872 (aarch64_<sra_op>rshr_n<mode>): New define_expand.
10874 2023-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10876 * config/aarch64/aarch64-simd.md (aarch64_shrn<mode>_insn_le): Delete.
10877 (aarch64_shrn<mode>_insn_be): Delete.
10878 (*aarch64_<srn_op>shrn<mode>_vect): Rename to...
10879 (*aarch64_<srn_op>shrn<mode><vczle><vczbe>): ... This.
10880 (aarch64_shrn<mode>): Remove reference to the above deleted patterns.
10881 (aarch64_rshrn<mode>_insn_le): Delete.
10882 (aarch64_rshrn<mode>_insn_be): Delete.
10883 (aarch64_rshrn<mode><vczle><vczbe>_insn): New define_insn.
10884 (aarch64_rshrn<mode>): Remove references to the above deleted patterns.
10886 2023-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10888 * config/aarch64/aarch64-protos.h (aarch64_parallel_select_half_p):
10890 (aarch64_pars_overlap_p): Likewise.
10891 * config/aarch64/aarch64-simd.md (aarch64_<su>addlv<mode>):
10892 Express in terms of UNSPEC_ADDV.
10893 (*aarch64_<su>addlv<VDQV_L:mode>_ze<GPI:mode>): Likewise.
10894 (*aarch64_<su>addlv<mode>_reduction): Define.
10895 (*aarch64_uaddlv<mode>_reduction_2): Likewise.
10896 * config/aarch64/aarch64.cc (aarch64_parallel_select_half_p): Define.
10897 (aarch64_pars_overlap_p): Likewise.
10898 * config/aarch64/iterators.md (UNSPEC_SADDLV, UNSPEC_UADDLV): Delete.
10899 (VQUADW): New mode attribute.
10900 (VWIDE2X_S): Likewise.
10902 (su): Delete handling of UNSPEC_SADDLV, UNSPEC_UADDLV.
10903 * config/aarch64/predicates.md (vect_par_cnst_select_half): Define.
10905 2023-06-06 Richard Biener <rguenther@suse.de>
10907 PR middle-end/110055
10908 * gimplify.cc (gimplify_target_expr): Do not emit
10909 CLOBBERs for variables which have static storage duration
10910 after gimplifying their initializers.
10912 2023-06-06 Richard Biener <rguenther@suse.de>
10914 PR tree-optimization/109143
10915 * tree-ssa-structalias.cc (solution_set_expand): Avoid
10916 one bitmap iteration and optimize bit range setting.
10918 2023-06-06 Hans-Peter Nilsson <hp@axis.com>
10920 PR bootstrap/110120
10921 * postreload.cc (reload_cse_move2add, move2add_use_add2_insn): Use
10922 XVECEXP, not XEXP, to access first item of a PARALLEL.
10924 2023-06-06 Pan Li <pan2.li@intel.com>
10926 * config/riscv/riscv-vector-builtins-types.def
10927 (vfloat16mf4_t): Add vfloat16mf4_t to WF operations.
10928 (vfloat16mf2_t): Likewise.
10929 (vfloat16m1_t): Likewise.
10930 (vfloat16m2_t): Likewise.
10931 (vfloat16m4_t): Likewise.
10932 (vfloat16m8_t): Likewise.
10933 * config/riscv/vector-iterators.md: Add FP=16 to VWF, VWF_ZVE64,
10934 VWLMUL1, VWLMUL1_ZVE64, vwlmul1 and vwlmul1_zve64.
10936 2023-06-06 Fei Gao <gaofei@eswincomputing.com>
10938 * config/riscv/riscv.cc (riscv_adjust_libcall_cfi_prologue): Use Pmode
10939 for cfi reg/mem machmode
10940 (riscv_adjust_libcall_cfi_epilogue): Use Pmode for cfi reg machmode
10942 2023-06-06 Li Xu <xuli1@eswincomputing.com>
10944 * config/riscv/vector-iterators.md:
10945 Fix 'REQUIREMENT' for machine_mode 'MODE'.
10946 * config/riscv/vector.md (@pred_indexed_<order>store<VNX16_QHS:mode>
10947 <VNX16_QHSI:mode>): change VNX16_QHSI to VNX16_QHSDI.
10948 (@pred_indexed_<order>store<VNX16_QHS:mode><VNX16_QHSDI:mode>): Ditto.
10950 2023-06-06 Pan Li <pan2.li@intel.com>
10952 * config/riscv/vector-iterators.md: Fix typo in mode attr.
10954 2023-06-05 Andre Vieira <andre.simoesdiasvieira@arm.com>
10955 Joel Hutton <joel.hutton@arm.com>
10957 * doc/generic.texi: Remove old tree codes.
10958 * expr.cc (expand_expr_real_2): Remove old tree code cases.
10959 * gimple-pretty-print.cc (dump_binary_rhs): Likewise.
10960 * optabs-tree.cc (optab_for_tree_code): Likewise.
10961 (supportable_half_widening_operation): Likewise.
10962 * tree-cfg.cc (verify_gimple_assign_binary): Likewise.
10963 * tree-inline.cc (estimate_operator_cost): Likewise.
10964 (op_symbol_code): Likewise.
10965 * tree-vect-data-refs.cc (vect_get_smallest_scalar_type): Likewise.
10966 (vect_analyze_data_ref_accesses): Likewise.
10967 * tree-vect-generic.cc (expand_vector_operations_1): Likewise.
10968 * cfgexpand.cc (expand_debug_expr): Likewise.
10969 * tree-vect-stmts.cc (vectorizable_conversion): Likewise.
10970 (supportable_widening_operation): Likewise.
10971 * gimple-range-op.cc (gimple_range_op_handler::maybe_non_standard):
10973 * optabs.def (vec_widen_ssubl_hi_optab, vec_widen_ssubl_lo_optab,
10974 vec_widen_saddl_hi_optab, vec_widen_saddl_lo_optab,
10975 vec_widen_usubl_hi_optab, vec_widen_usubl_lo_optab,
10976 vec_widen_uaddl_hi_optab, vec_widen_uaddl_lo_optab): Remove optabs.
10977 * tree-pretty-print.cc (dump_generic_node): Remove tree code definition.
10978 * tree.def (WIDEN_PLUS_EXPR, WIDEN_MINUS_EXPR, VEC_WIDEN_PLUS_HI_EXPR,
10979 VEC_WIDEN_PLUS_LO_EXPR, VEC_WIDEN_MINUS_HI_EXPR,
10980 VEC_WIDEN_MINUS_LO_EXPR): Likewise.
10982 2023-06-05 Andre Vieira <andre.simoesdiasvieira@arm.com>
10983 Joel Hutton <joel.hutton@arm.com>
10984 Tamar Christina <tamar.christina@arm.com>
10986 * config/aarch64/aarch64-simd.md (vec_widen_<su>addl_lo_<mode>): Rename
10988 (vec_widen_<su>add_lo_<mode>): ... to this.
10989 (vec_widen_<su>addl_hi_<mode>): Rename this ...
10990 (vec_widen_<su>add_hi_<mode>): ... to this.
10991 (vec_widen_<su>subl_lo_<mode>): Rename this ...
10992 (vec_widen_<su>sub_lo_<mode>): ... to this.
10993 (vec_widen_<su>subl_hi_<mode>): Rename this ...
10994 (vec_widen_<su>sub_hi_<mode>): ...to this.
10995 * doc/generic.texi: Document new IFN codes.
10996 * internal-fn.cc (lookup_hilo_internal_fn): Add lookup function.
10997 (commutative_binary_fn_p): Add widen_plus fn's.
10998 (widening_fn_p): New function.
10999 (narrowing_fn_p): New function.
11000 (direct_internal_fn_optab): Change visibility.
11001 * internal-fn.def (DEF_INTERNAL_WIDENING_OPTAB_FN): Macro to define an
11002 internal_fn that expands into multiple internal_fns for widening.
11003 (IFN_VEC_WIDEN_PLUS, IFN_VEC_WIDEN_PLUS_HI, IFN_VEC_WIDEN_PLUS_LO,
11004 IFN_VEC_WIDEN_PLUS_EVEN, IFN_VEC_WIDEN_PLUS_ODD,
11005 IFN_VEC_WIDEN_MINUS, IFN_VEC_WIDEN_MINUS_HI,
11006 IFN_VEC_WIDEN_MINUS_LO, IFN_VEC_WIDEN_MINUS_ODD,
11007 IFN_VEC_WIDEN_MINUS_EVEN): Define widening plus,minus functions.
11008 * internal-fn.h (direct_internal_fn_optab): Declare new prototype.
11009 (lookup_hilo_internal_fn): Likewise.
11010 (widening_fn_p): Likewise.
11011 (Narrowing_fn_p): Likewise.
11012 * optabs.cc (commutative_optab_p): Add widening plus optabs.
11013 * optabs.def (OPTAB_D): Define widen add, sub optabs.
11014 * tree-vect-patterns.cc (vect_recog_widen_op_pattern): Support
11015 patterns with a hi/lo or even/odd split.
11016 (vect_recog_sad_pattern): Refactor to use new IFN codes.
11017 (vect_recog_widen_plus_pattern): Likewise.
11018 (vect_recog_widen_minus_pattern): Likewise.
11019 (vect_recog_average_pattern): Likewise.
11020 * tree-vect-stmts.cc (vectorizable_conversion): Add support for
11022 (supportable_widening_operation): Likewise.
11023 * tree.def (WIDEN_SUM_EXPR): Update example to use new IFNs.
11025 2023-06-05 Andre Vieira <andre.simoesdiasvieira@arm.com>
11026 Joel Hutton <joel.hutton@arm.com>
11028 * tree-vect-patterns.cc: Add include for gimple-iterator.
11029 (vect_recog_widen_op_pattern): Refactor to use code_helper.
11030 (vect_gimple_build): New function.
11031 * tree-vect-stmts.cc (simple_integer_narrowing): Refactor to use
11033 (vectorizable_call): Likewise.
11034 (vect_gen_widened_results_half): Likewise.
11035 (vect_create_vectorized_demotion_stmts): Likewise.
11036 (vect_create_vectorized_promotion_stmts): Likewise.
11037 (vect_create_half_widening_stmts): Likewise.
11038 (vectorizable_conversion): Likewise.
11039 (supportable_widening_operation): Likewise.
11040 (supportable_narrowing_operation): Likewise.
11041 * tree-vectorizer.h (supportable_widening_operation): Change
11042 prototype to use code_helper.
11043 (supportable_narrowing_operation): Likewise.
11044 (vect_gimple_build): New function prototype.
11045 * tree.h (code_helper::safe_as_tree_code): New function.
11046 (code_helper::safe_as_fn_code): New function.
11048 2023-06-05 Roger Sayle <roger@nextmovesoftware.com>
11050 * wide-int.cc (wi::bitreverse_large): New function implementing
11051 bit reversal of an integer.
11052 * wide-int.h (wi::bitreverse): New (template) function prototype.
11053 (bitreverse_large): Prototype helper function/implementation.
11054 (wi::bitreverse): New template wrapper around bitreverse_large.
11056 2023-06-05 Uros Bizjak <ubizjak@gmail.com>
11058 * rtl.h (print_rtl_single): Change return type from int to void.
11059 (print_rtl_single_with_indent): Ditto.
11060 * print-rtl.h (class rtx_writer): Ditto. Change m_sawclose to bool.
11061 * print-rtl.cc (rtx_writer::rtx_writer): Update for m_sawclose change.
11062 (rtx_writer::print_rtx_operand_code_0): Ditto.
11063 (rtx_writer::print_rtx_operand_codes_E_and_V): Ditto.
11064 (rtx_writer::print_rtx_operand_code_i): Ditto.
11065 (rtx_writer::print_rtx_operand_code_u): Ditto.
11066 (rtx_writer::print_rtx_operand): Ditto.
11067 (rtx_writer::print_rtx): Ditto.
11068 (rtx_writer::finish_directive): Ditto.
11069 (print_rtl_single): Change return type from int to void
11070 and adjust function body accordingly.
11071 (rtx_writer::print_rtl_single_with_indent): Ditto.
11073 2023-06-05 Uros Bizjak <ubizjak@gmail.com>
11075 * rtl.h (reg_classes_intersect_p): Change return type from int to bool.
11076 (reg_class_subset_p): Ditto.
11077 * reginfo.cc (reg_classes_intersect_p): Ditto.
11078 (reg_class_subset_p): Ditto.
11080 2023-06-05 Pan Li <pan2.li@intel.com>
11082 * config/riscv/riscv-vector-builtins-types.def
11083 (vfloat32mf2_t): New type for DEF_RVV_WEXTF_OPS.
11084 (vfloat32m1_t): Ditto.
11085 (vfloat32m2_t): Ditto.
11086 (vfloat32m4_t): Ditto.
11087 (vfloat32m8_t): Ditto.
11088 (vint16mf4_t): New type for DEF_RVV_CONVERT_I_OPS.
11089 (vint16mf2_t): Ditto.
11090 (vint16m1_t): Ditto.
11091 (vint16m2_t): Ditto.
11092 (vint16m4_t): Ditto.
11093 (vint16m8_t): Ditto.
11094 (vuint16mf4_t): New type for DEF_RVV_CONVERT_U_OPS.
11095 (vuint16mf2_t): Ditto.
11096 (vuint16m1_t): Ditto.
11097 (vuint16m2_t): Ditto.
11098 (vuint16m4_t): Ditto.
11099 (vuint16m8_t): Ditto.
11100 (vint32mf2_t): New type for DEF_RVV_WCONVERT_I_OPS.
11101 (vint32m1_t): Ditto.
11102 (vint32m2_t): Ditto.
11103 (vint32m4_t): Ditto.
11104 (vint32m8_t): Ditto.
11105 (vuint32mf2_t): New type for DEF_RVV_WCONVERT_U_OPS.
11106 (vuint32m1_t): Ditto.
11107 (vuint32m2_t): Ditto.
11108 (vuint32m4_t): Ditto.
11109 (vuint32m8_t): Ditto.
11110 * config/riscv/vector-iterators.md: Add FP=16 support for V,
11111 VWCONVERTI, VCONVERT, VNCONVERT, VMUL1 and vlmul1.
11113 2023-06-05 Andrew Pinski <apinski@marvell.com>
11115 PR bootstrap/110085
11116 * Makefile.in (clean): Remove the removing of
11117 MULTILIB_DIR/MULTILIB_OPTIONS directories.
11119 2023-06-05 YunQiang Su <yunqiang.su@cipunited.com>
11121 * config/mips/mips-protos.h (mips_emit_speculation_barrier): New
11123 * config/mips/mips.cc (speculation_barrier_libfunc): New static
11125 (mips_init_libfuncs): Initialize it.
11126 (mips_emit_speculation_barrier): New function.
11127 * config/mips/mips.md (speculation_barrier): Call
11128 mips_emit_speculation_barrier.
11130 2023-06-05 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11132 * config/riscv/riscv-v.cc (class rvv_builder): Reorganize functions.
11133 (rvv_builder::can_duplicate_repeating_sequence_p): Ditto.
11134 (rvv_builder::repeating_sequence_use_merge_profitable_p): Ditto.
11135 (rvv_builder::get_merged_repeating_sequence): Ditto.
11136 (rvv_builder::get_merge_scalar_mask): Ditto.
11137 (emit_scalar_move_insn): Ditto.
11138 (emit_vlmax_integer_move_insn): Ditto.
11139 (emit_nonvlmax_integer_move_insn): Ditto.
11140 (emit_vlmax_gather_insn): Ditto.
11141 (emit_vlmax_masked_gather_mu_insn): Ditto.
11142 (get_repeating_sequence_dup_machine_mode): Ditto.
11144 2023-06-05 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11146 * config/riscv/autovec.md: Split arguments.
11147 * config/riscv/riscv-protos.h (expand_vec_perm): Ditto.
11148 * config/riscv/riscv-v.cc (expand_vec_perm): Ditto.
11150 2023-06-04 Andrew Pinski <apinski@marvell.com>
11152 * expr.cc (do_store_flag): Improve for single bit testing
11153 not against zero but against that single bit.
11155 2023-06-04 Andrew Pinski <apinski@marvell.com>
11157 * expr.cc (do_store_flag): Extend the one bit checking case
11158 to handle the case where we don't have an and but rather still
11159 one bit is known to be non-zero.
11161 2023-06-04 Jeff Law <jlaw@ventanamicro.com>
11163 * config/h8300/constraints.md (Zz): Make this a normal
11165 * config/h8300/h8300.cc (TARGET_LRA_P): Remove.
11166 * config/h8300/logical.md (H8/SX bit patterns): Remove.
11168 2023-06-04 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
11170 * config/xtensa/xtensa.md (*btrue_INT_MIN, *eqne_INT_MIN):
11171 New insn_and_split patterns.
11173 2023-06-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11176 * config/riscv/riscv-vector-builtins-bases.cc: Change expand approach.
11177 * config/riscv/vector.md (@vlmul_extx2<mode>): Remove it.
11178 (@vlmul_extx4<mode>): Ditto.
11179 (@vlmul_extx8<mode>): Ditto.
11180 (@vlmul_extx16<mode>): Ditto.
11181 (@vlmul_extx32<mode>): Ditto.
11182 (@vlmul_extx64<mode>): Ditto.
11183 (*vlmul_extx2<mode>): Ditto.
11184 (*vlmul_extx4<mode>): Ditto.
11185 (*vlmul_extx8<mode>): Ditto.
11186 (*vlmul_extx16<mode>): Ditto.
11187 (*vlmul_extx32<mode>): Ditto.
11188 (*vlmul_extx64<mode>): Ditto.
11190 2023-06-04 Pan Li <pan2.li@intel.com>
11192 * config/riscv/riscv-vector-builtins-types.def
11193 (vfloat32mf2_t): Add vfloat32mf2_t type to vfncvt.f.f.w operations.
11194 (vfloat32m1_t): Likewise.
11195 (vfloat32m2_t): Likewise.
11196 (vfloat32m4_t): Likewise.
11197 (vfloat32m8_t): Likewise.
11198 * config/riscv/riscv-vector-builtins.def: Fix typo in comments.
11199 * config/riscv/vector-iterators.md: Add single to half machine
11202 2023-06-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11204 * config/riscv/autovec-opt.md (*<optab>not<mode>): Move to autovec-opt.md.
11205 (*n<optab><mode>): Ditto.
11206 * config/riscv/autovec.md (*<optab>not<mode>): Ditto.
11207 (*n<optab><mode>): Ditto.
11208 * config/riscv/vector.md: Ditto.
11210 2023-06-04 Roger Sayle <roger@nextmovesoftware.com>
11213 * config/i386/i386-features.cc (scalar_chain::convert_compare):
11214 Update or delete REG_EQUAL notes, converting CONST_INT and
11215 CONST_WIDE_INT immediate operands to a suitable CONST_VECTOR.
11217 2023-06-04 Jason Merrill <jason@redhat.com>
11220 * tree-eh.cc (lower_resx): Pass the exception pointer to the
11222 * except.h: Tweak comment.
11224 2023-06-04 Hans-Peter Nilsson <hp@axis.com>
11226 * postreload.cc (move2add_use_add2_insn): Handle
11227 trivial single_sets. Rename variable PAT to SET.
11228 (move2add_use_add3_insn, reload_cse_move2add): Similar.
11230 2023-06-04 Pan Li <pan2.li@intel.com>
11232 * config/riscv/riscv-vector-builtins-types.def
11233 (vfloat16mf4_t): Add the float16 type to DEF_RVV_F_OPS.
11234 (vfloat16mf2_t): Likewise.
11235 (vfloat16m1_t): Likewise.
11236 (vfloat16m2_t): Likewise.
11237 (vfloat16m4_t): Likewise.
11238 (vfloat16m8_t): Likewise.
11239 * config/riscv/riscv.md: Add vfloat16*_t to attr mode.
11240 * config/riscv/vector-iterators.md: Add vfloat16*_t machine mode
11241 to V, V_WHOLE, V_FRACT, VINDEX, VM, VEL and sew.
11242 * config/riscv/vector.md: Add vfloat16*_t machine mode to sew,
11245 2023-06-03 Fei Gao <gaofei@eswincomputing.com>
11247 * config/riscv/riscv.cc (riscv_expand_epilogue): fix cfi issue with
11250 2023-06-03 Die Li <lidie@eswincomputing.com>
11252 * config/riscv/thead.md (*th_cond_gpr_mov<GPR:mode><GPR2:mode>): Delete.
11254 2023-06-03 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11256 * config/riscv/predicates.md: Change INTVAL into UINTVAL.
11258 2023-06-03 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11260 * config/riscv/vector.md: Add vector-opt.md.
11261 * config/riscv/autovec-opt.md: New file.
11263 2023-06-03 liuhongt <hongtao.liu@intel.com>
11265 PR tree-optimization/110067
11266 * gimple-ssa-store-merging.cc (find_bswap_or_nop): Don't try
11267 bswap + rotate when TYPE_PRECISION(n->type) > n->range.
11269 2023-06-03 liuhongt <hongtao.liu@intel.com>
11272 * config/i386/mmx.md (truncv2hiv2qi2): New define_insn.
11273 (truncv2si<mode>2): Ditto.
11275 2023-06-02 Andrew Pinski <apinski@marvell.com>
11277 PR rtl-optimization/102733
11278 * dse.cc (store_info): Add addrspace field.
11279 (record_store): Record the address space
11280 and check to make sure they are the same.
11282 2023-06-02 Andrew Pinski <apinski@marvell.com>
11284 PR rtl-optimization/110042
11285 * ifcvt.cc (bbs_ok_for_cmove_arith): Allow paradoxical subregs.
11286 (bb_valid_for_noce_process_p): Strip the subreg for the SET_DEST.
11288 2023-06-02 Iain Sandoe <iain@sandoe.co.uk>
11291 * config/rs6000/rs6000.cc (darwin_rs6000_special_round_type_align):
11292 Make sure that we do not have a cap on field alignment before altering
11293 the struct layout based on the type alignment of the first entry.
11295 2023-06-02 David Faust <david.faust@oracle.com>
11298 * btfout.cc (btf_absolute_func_id): New function.
11299 (btf_asm_func_type): Call it here. Change index parameter from
11300 size_t to ctf_id_t. Use PRIu64 formatter.
11302 2023-06-02 Alex Coplan <alex.coplan@arm.com>
11304 * btfout.cc (btf_asm_type): Use PRIu64 instead of %lu for uint64_t.
11305 (btf_asm_datasec_type): Likewise.
11307 2023-06-02 Carl Love <cel@us.ibm.com>
11309 * config/rs6000/rs6000-builtins.def (__builtin_altivec_tr_stxvrhx,
11310 __builtin_altivec_tr_stxvrwx): Fix type of third argument.
11312 2023-06-02 Jason Merrill <jason@redhat.com>
11316 * tree.h (DECL_MERGEABLE): New.
11317 * tree-core.h (struct tree_decl_common): Mention it.
11318 * gimplify.cc (gimplify_init_constructor): Check it.
11319 * cgraph.cc (symtab_node::address_can_be_compared_p): Likewise.
11320 * varasm.cc (categorize_decl_for_section): Likewise.
11322 2023-06-02 Uros Bizjak <ubizjak@gmail.com>
11324 * rtl.h (stack_regs_mentioned): Change return type from int to bool.
11325 * reg-stack.cc (struct_block_info_def): Change "done" to bool.
11326 (stack_regs_mentioned_p): Change return type from int to bool
11327 and adjust function body accordingly.
11328 (stack_regs_mentioned): Ditto.
11329 (check_asm_stack_operands): Ditto. Change "malformed_asm"
11331 (move_for_stack_reg): Recode handling of control_flow_insn_deleted.
11332 (swap_rtx_condition_1): Change return type from int to bool
11333 and adjust function body accordingly. Change "r" variable to bool.
11334 (swap_rtx_condition): Change return type from int to bool
11335 and adjust function body accordingly.
11336 (subst_stack_regs_pat): Recode handling of control_flow_insn_deleted.
11337 (subst_stack_regs): Ditto.
11338 (convert_regs_entry): Change return type from int to bool and adjust
11339 function body accordingly. Change "inserted" variable to bool.
11340 (convert_regs_1): Recode handling of control_flow_insn_deleted.
11341 (convert_regs_2): Recode handling of cfg_altered.
11342 (convert_regs): Ditto. Change "inserted" variable to bool.
11344 2023-06-02 Jason Merrill <jason@redhat.com>
11347 * varasm.cc (output_constant) [REAL_TYPE]: Check that sizes match.
11348 (initializer_constant_valid_p_1): Compare float precision.
11350 2023-06-02 Alexander Monakov <amonakov@ispras.ru>
11352 * doc/extend.texi (Vector Extensions): Clarify bitwise shift
11355 2023-06-02 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
11357 * tree-vect-loop-manip.cc (vect_set_loop_controls_directly): Change decrement IV flow.
11358 (vect_set_loop_condition_partial_vectors): Ditto.
11360 2023-06-02 Georg-Johann Lay <avr@gjlay.de>
11363 * config/avr/avr.md: Add an RTL peephole to optimize operations on
11364 non-LD_REGS after a move from LD_REGS.
11365 (piaop): New code iterator.
11367 2023-06-02 Thomas Schwinge <thomas@codesourcery.com>
11370 * doc/install.texi: Document (optional) Perl usage for parallel
11371 testing of libgomp.
11373 2023-06-02 Thomas Schwinge <thomas@codesourcery.com>
11376 * doc/install.texi (Perl): Back to requiring "Perl version 5.6.1 (or
11379 2023-06-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11380 KuanLin Chen <best124612@gmail.com>
11382 * config/riscv/riscv-vector-builtins-bases.cc: Add _mu overloaded intrinsics.
11383 * config/riscv/riscv-vector-builtins-shapes.cc (struct fault_load_def): Ditto.
11385 2023-06-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11387 * config/riscv/riscv-v.cc (expand_vec_series): Optimize reverse series index vector.
11389 2023-06-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11391 * config/riscv/predicates.md: Change INTVAL into UINTVAL.
11393 2023-06-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11395 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_VXRM_ENUM): Add
11397 (DEF_RVV_FRM_ENUM): Ditto.
11399 2023-06-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11401 * config/riscv/riscv-vector-builtins-bases.cc: Change vwadd.wv/vwsub.wv
11402 intrinsic API expander
11403 * config/riscv/vector.md
11404 (@pred_single_widen_<plus_minus:optab><any_extend:su><mode>): Remove it.
11405 (@pred_single_widen_sub<any_extend:su><mode>): New pattern.
11406 (@pred_single_widen_add<any_extend:su><mode>): New pattern.
11408 2023-06-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11410 * config/riscv/autovec.md (vec_perm<mode>): New pattern.
11411 * config/riscv/predicates.md (vector_perm_operand): New predicate.
11412 * config/riscv/riscv-protos.h (enum insn_type): New enum.
11413 (expand_vec_perm): New function.
11414 * config/riscv/riscv-v.cc (const_vec_all_in_range_p): Ditto.
11415 (gen_const_vector_dup): Ditto.
11416 (emit_vlmax_gather_insn): Ditto.
11417 (emit_vlmax_masked_gather_mu_insn): Ditto.
11418 (expand_vec_perm): Ditto.
11420 2023-06-01 Jason Merrill <jason@redhat.com>
11422 * doc/invoke.texi (-Wpedantic): Improve clarity.
11424 2023-06-01 Uros Bizjak <ubizjak@gmail.com>
11426 * rtl.h (exp_equiv_p): Change return type from int to bool.
11427 * cse.cc (mention_regs): Change return type from int to bool
11428 and adjust function body accordingly.
11429 (exp_equiv_p): Ditto.
11430 (insert_regs): Ditto. Change "modified" function argument to bool
11431 and update usage accordingly.
11432 (record_jump_cond): Remove always zero "reversed_nonequality"
11433 function argument and update usage accordingly.
11434 (fold_rtx): Change "changed" variable to bool.
11435 (record_jump_equiv): Remove unneeded "reversed_nonequality" variable.
11436 (is_dead_reg): Change return type from int to bool.
11438 2023-06-01 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
11440 * config/xtensa/xtensa.md (adddi3, subdi3):
11441 New RTL generation patterns implemented according to the instruc-
11442 tion idioms described in the Xtensa ISA reference manual (p. 600).
11444 2023-06-01 Roger Sayle <roger@nextmovesoftware.com>
11445 Uros Bizjak <ubizjak@gmail.com>
11448 * config/i386/i386-builtin.def (__builtin_ia32_ptestz128): Use new
11449 CODE_for_sse4_1_ptestzv2di.
11450 (__builtin_ia32_ptestc128): Use new CODE_for_sse4_1_ptestcv2di.
11451 (__builtin_ia32_ptestz256): Use new CODE_for_avx_ptestzv4di.
11452 (__builtin_ia32_ptestc256): Use new CODE_for_avx_ptestcv4di.
11453 * config/i386/i386-expand.cc (ix86_expand_branch): Use CCZmode
11454 when expanding UNSPEC_PTEST to compare against zero.
11455 * config/i386/i386-features.cc (scalar_chain::convert_compare):
11456 Likewise generate CCZmode UNSPEC_PTESTs when converting comparisons.
11457 (general_scalar_chain::convert_insn): Use CCZmode for COMPARE result.
11458 (timode_scalar_chain::convert_insn): Use CCZmode for COMPARE result.
11459 * config/i386/i386-protos.h (ix86_match_ptest_ccmode): Prototype.
11460 * config/i386/i386.cc (ix86_match_ptest_ccmode): New predicate to
11461 check for suitable matching modes for the UNSPEC_PTEST pattern.
11462 * config/i386/sse.md (define_split): When splitting UNSPEC_MOVMSK
11463 to UNSPEC_PTEST, preserve the FLAG_REG mode as CCZ.
11464 (*<sse4_1>_ptest<mode>): Add asterisk to hide define_insn. Remove
11465 ":CC" mode of FLAGS_REG, instead use ix86_match_ptest_ccmode.
11466 (<sse4_1>_ptestz<mode>): New define_expand to specify CCZ.
11467 (<sse4_1>_ptestc<mode>): New define_expand to specify CCC.
11468 (<sse4_1>_ptest<mode>): A define_expand using CC to preserve the
11470 (*ptest<mode>_and): Specify CCZ to only perform this optimization
11471 when only the Z flag is required.
11473 2023-06-01 Jonathan Wakely <jwakely@redhat.com>
11476 * doc/invoke.texi (x86 Options): Fix description of -m32 option.
11478 2023-06-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11480 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<VDMOV:mode>):
11481 Add =r,m and =r,m alternatives.
11482 (load_pair<DREG:mode><DREG2:mode>): Likewise.
11483 (vec_store_pair<DREG:mode><DREG2:mode>): Likewise.
11485 2023-06-01 Pan Li <pan2.li@intel.com>
11487 * common/config/riscv/riscv-common.cc: Add FP_16 mask to zvfhmin
11489 * config/riscv/genrvv-type-indexer.cc (valid_type): Allow FP16.
11490 (main): Disable FP16 tuple.
11491 * config/riscv/riscv-opts.h (MASK_VECTOR_ELEN_FP_16): New macro.
11492 (TARGET_VECTOR_ELEN_FP_16): Ditto.
11493 * config/riscv/riscv-vector-builtins.cc (check_required_extensions):
11495 * config/riscv/riscv-vector-builtins.def (vfloat16mf4_t): New type.
11496 (vfloat16mf2_t): Ditto.
11497 (vfloat16m1_t): Ditto.
11498 (vfloat16m2_t): Ditto.
11499 (vfloat16m4_t): Ditto.
11500 (vfloat16m8_t): Ditto.
11501 * config/riscv/riscv-vector-builtins.h (RVV_REQUIRE_ELEN_FP_16):
11503 * config/riscv/riscv-vector-switch.def (ENTRY): Allow FP16
11504 machine mode based on TARGET_VECTOR_ELEN_FP_16.
11506 2023-06-01 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11508 * config/riscv/riscv-vector-builtins.cc (register_frm): New function.
11509 (DEF_RVV_FRM_ENUM): New macro.
11510 (handle_pragma_vector): Add FRM enum
11511 * config/riscv/riscv-vector-builtins.def (DEF_RVV_FRM_ENUM): New macro.
11518 2023-05-31 Roger Sayle <roger@nextmovesoftware.com>
11519 Richard Sandiford <richard.sandiford@arm.com>
11521 * fold-const-call.cc (fold_const_call_ss) <CFN_BUILT_IN_BSWAP*>:
11522 Update call to wi::bswap.
11523 * simplify-rtx.cc (simplify_const_unary_operation) <case BSWAP>:
11524 Update call to wi::bswap.
11525 * tree-ssa-ccp.cc (evaluate_stmt) <case BUILT_IN_BSWAP*>:
11526 Update calls to wi::bswap.
11527 * wide-int.cc (wide_int_storage::bswap): Remove/rename to...
11528 (wi::bswap_large): New function, with revised API.
11529 * wide-int.h (wi::bswap): New (template) function prototype.
11530 (wide_int_storage::bswap): Remove method.
11531 (sext_large, zext_large): Consistent indentation/line wrapping.
11532 (bswap_large): Prototype helper function containing implementation.
11533 (wi::bswap): New template wrapper around bswap_large.
11535 2023-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11538 * config/aarch64/aarch64-simd.md (<sur>dot_prod<vsi2qi>): Rename to...
11539 (<sur>dot_prod<vsi2qi><vczle><vczbe>): ... This.
11540 (usdot_prod<vsi2qi>): Rename to...
11541 (usdot_prod<vsi2qi><vczle><vczbe>): ... This.
11542 (aarch64_<sur>dot_lane<vsi2qi>): Rename to...
11543 (aarch64_<sur>dot_lane<vsi2qi><vczle><vczbe>): ... This.
11544 (aarch64_<sur>dot_laneq<vsi2qi>): Rename to...
11545 (aarch64_<sur>dot_laneq<vsi2qi><vczle><vczbe>): ... This.
11546 (aarch64_<DOTPROD_I8MM:sur>dot_lane<VB:isquadop><VS:vsi2qi>): Rename to...
11547 (aarch64_<DOTPROD_I8MM:sur>dot_lane<VB:isquadop><VS:vsi2qi><vczle><vczbe>):
11550 2023-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11553 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh<mode>): Rename to...
11554 (aarch64_sq<r>dmulh<mode><vczle><vczbe>): ... This.
11555 (aarch64_sq<r>dmulh_n<mode>): Rename to...
11556 (aarch64_sq<r>dmulh_n<mode><vczle><vczbe>): ... This.
11557 (aarch64_sq<r>dmulh_lane<mode>): Rename to...
11558 (aarch64_sq<r>dmulh_lane<mode><vczle><vczbe>): ... This.
11559 (aarch64_sq<r>dmulh_laneq<mode>): Rename to...
11560 (aarch64_sq<r>dmulh_laneq<mode><vczle><vczbe>): ... This.
11561 (aarch64_sqrdml<SQRDMLH_AS:rdma_as>h<mode>): Rename to...
11562 (aarch64_sqrdml<SQRDMLH_AS:rdma_as>h<mode><vczle><vczbe>): ... This.
11563 (aarch64_sqrdml<SQRDMLH_AS:rdma_as>h_lane<mode>): Rename to...
11564 (aarch64_sqrdml<SQRDMLH_AS:rdma_as>h_lane<mode><vczle><vczbe>): ... This.
11565 (aarch64_sqrdml<SQRDMLH_AS:rdma_as>h_laneq<mode>): Rename to...
11566 (aarch64_sqrdml<SQRDMLH_AS:rdma_as>h_laneq<mode><vczle><vczbe>): ... This.
11568 2023-05-31 David Faust <david.faust@oracle.com>
11570 * btfout.cc (btf_kind_names): New.
11571 (btf_kind_name): New.
11572 (btf_absolute_var_id): New utility function.
11573 (btf_relative_var_id): Likewise.
11574 (btf_relative_func_id): Likewise.
11575 (btf_absolute_datasec_id): Likewise.
11576 (btf_asm_type_ref): New.
11577 (btf_asm_type): Update asm comments and use btf_asm_type_ref ().
11578 (btf_asm_array): Likewise. Accept ctf_container_ref parameter.
11579 (btf_asm_varent): Likewise.
11580 (btf_asm_func_arg): Likewise.
11581 (btf_asm_datasec_entry): Likewise.
11582 (btf_asm_datasec_type): Likewise.
11583 (btf_asm_func_type): Likewise. Add index parameter.
11584 (btf_asm_enum_const): Likewise.
11585 (btf_asm_sou_member): Likewise.
11586 (output_btf_vars): Update btf_asm_* call accordingly.
11587 (output_asm_btf_sou_fields): Likewise.
11588 (output_asm_btf_enum_list): Likewise.
11589 (output_asm_btf_func_args_list): Likewise.
11590 (output_asm_btf_vlen_bytes): Likewise.
11591 (output_btf_func_types): Add ctf_container_ref parameter.
11592 Pass it to btf_asm_func_type.
11593 (output_btf_datasec_types): Update btf_asm_datsec_type call similarly.
11594 (btf_output): Update output_btf_func_types call similarly.
11596 2023-05-31 David Faust <david.faust@oracle.com>
11598 * btfout.cc (btf_asm_type): Add dedicated cases for BTF_KIND_ARRAY
11599 and BTF_KIND_FWD which do not use the size/type field at all.
11601 2023-05-31 Uros Bizjak <ubizjak@gmail.com>
11603 * rtl.h (subreg_lowpart_p): Change return type from int to bool.
11604 (active_insn_p): Ditto.
11605 (in_sequence_p): Ditto.
11606 (unshare_all_rtl): Change return type from int to void.
11607 * emit-rtl.h (mem_expr_equal_p): Change return type from int to bool.
11608 * emit-rtl.cc (subreg_lowpart_p): Change return type from int to bool
11609 and adjust function body accordingly.
11610 (mem_expr_equal_p): Ditto.
11611 (unshare_all_rtl): Change return type from int to void
11612 and adjust function body accordingly.
11613 (verify_rtx_sharing): Remove unneeded return.
11614 (active_insn_p): Change return type from int to bool
11615 and adjust function body accordingly.
11616 (in_sequence_p): Ditto.
11618 2023-05-31 Uros Bizjak <ubizjak@gmail.com>
11620 * rtl.h (true_dependence): Change return type from int to bool.
11621 (canon_true_dependence): Ditto.
11622 (read_dependence): Ditto.
11623 (anti_dependence): Ditto.
11624 (canon_anti_dependence): Ditto.
11625 (output_dependence): Ditto.
11626 (canon_output_dependence): Ditto.
11627 (may_alias_p): Ditto.
11628 * alias.h (alias_sets_conflict_p): Ditto.
11629 (alias_sets_must_conflict_p): Ditto.
11630 (objects_must_conflict_p): Ditto.
11631 (nonoverlapping_memrefs_p): Ditto.
11632 * alias.cc (rtx_equal_for_memref_p): Remove forward declaration.
11633 (record_set): Ditto.
11634 (base_alias_check): Ditto.
11635 (find_base_value): Ditto.
11636 (mems_in_disjoint_alias_sets_p): Ditto.
11637 (get_alias_set_entry): Ditto.
11638 (decl_for_component_ref): Ditto.
11639 (write_dependence_p): Ditto.
11640 (memory_modified_1): Ditto.
11641 (mems_in_disjoint_alias_set_p): Change return type from int to bool
11642 and adjust function body accordingly.
11643 (alias_sets_conflict_p): Ditto.
11644 (alias_sets_must_conflict_p): Ditto.
11645 (objects_must_conflict_p): Ditto.
11646 (rtx_equal_for_memref_p): Ditto.
11647 (base_alias_check): Ditto.
11648 (read_dependence): Ditto.
11649 (nonoverlapping_memrefs_p): Ditto.
11650 (true_dependence_1): Ditto.
11651 (true_dependence): Ditto.
11652 (canon_true_dependence): Ditto.
11653 (write_dependence_p): Ditto.
11654 (anti_dependence): Ditto.
11655 (canon_anti_dependence): Ditto.
11656 (output_dependence): Ditto.
11657 (canon_output_dependence): Ditto.
11658 (may_alias_p): Ditto.
11659 (init_alias_analysis): Change "changed" variable to bool.
11661 2023-05-31 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11663 * config/riscv/autovec.md (<optab><v_double_trunc><mode>2): Change
11664 expand into define_insn_and_split.
11666 2023-05-31 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11668 * config/riscv/vector.md: Remove FRM.
11670 2023-05-31 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11672 * config/riscv/vector.md: Remove FRM.
11674 2023-05-31 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11676 * config/riscv/vector.md: Remove FRM.
11678 2023-05-31 Christophe Lyon <christophe.lyon@linaro.org>
11681 * config/aarch64/aarch64.md (aarch64_rev16si2_alt3): New
11684 2023-05-31 Richard Biener <rguenther@suse.de>
11687 PR tree-optimization/109143
11688 * tree-ssa-structalias.cc (struct topo_info): Remove.
11689 (init_topo_info): Likewise.
11690 (free_topo_info): Likewise.
11691 (compute_topo_order): Simplify API, put the component
11692 with ESCAPED last so it's processed first.
11693 (topo_visit): Adjust.
11694 (solve_graph): Likewise.
11696 2023-05-31 Richard Biener <rguenther@suse.de>
11698 * tree-ssa-structalias.cc (constraint_stats::num_avoided_edges):
11700 (add_graph_edge): Count redundant edges we avoid to create.
11701 (dump_sa_stats): Dump them.
11702 (ipa_pta_execute): Do not dump generating constraints when
11703 we are not dumping them.
11705 2023-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11707 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<VDMOV:mode>): Rewrite
11708 output template to avoid explicit switch on which_alternative.
11709 (*aarch64_simd_mov<VQMOV:mode>): Likewise.
11710 (and<mode>3): Likewise.
11711 (ior<mode>3): Likewise.
11712 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Likewise.
11714 2023-05-31 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
11716 * config/xtensa/predicates.md (xtensa_bit_join_operator):
11718 * config/xtensa/xtensa.md (ior_op): Remove.
11719 (*shlrd_reg): Rename from "*shlrd_reg_<code>", and add the
11720 insn_and_split pattern of the same name to express and capture
11721 the bit-combining operation with both sides swapped.
11722 In addition, replace use of code iterator with new operator
11724 (*shlrd_const, *shlrd_per_byte):
11725 Likewise regarding the code iterator.
11727 2023-05-31 Cui, Lili <lili.cui@intel.com>
11729 PR tree-optimization/110038
11730 * params.opt: Add a limit on tree-reassoc-width.
11731 * tree-ssa-reassoc.cc
11732 (rewrite_expr_tree_parallel): Add width limit.
11734 2023-05-31 Pan Li <pan2.li@intel.com>
11736 * common/config/riscv/riscv-common.cc:
11737 (riscv_implied_info): Add zvfh item.
11738 (riscv_ext_version_table): Ditto.
11739 (riscv_ext_flag_table): Ditto.
11740 * config/riscv/riscv-opts.h (MASK_ZVFH): New macro.
11741 (TARGET_ZVFH): Ditto.
11743 2023-05-30 liuhongt <hongtao.liu@intel.com>
11745 PR tree-optimization/108804
11746 * tree-vect-patterns.cc (vect_get_range_info): Remove static.
11747 * tree-vect-stmts.cc (vect_create_vectorized_demotion_stmts):
11748 Add new parameter narrow_src_p.
11749 (vectorizable_conversion): Enhance NARROW FLOAT_EXPR
11750 vectorization by truncating to lower precision.
11751 * tree-vectorizer.h (vect_get_range_info): New declare.
11753 2023-05-30 Vladimir N. Makarov <vmakarov@redhat.com>
11755 * lra-int.h (lra_update_sp_offset): Add the prototype.
11756 * lra.cc (setup_sp_offset): Change the return type. Use
11757 lra_update_sp_offset.
11758 * lra-eliminations.cc (lra_update_sp_offset): New function.
11759 (lra_process_new_insns): Push the current insn to reprocess if the
11760 input reload changes sp offset.
11762 2023-05-30 Uros Bizjak <ubizjak@gmail.com>
11765 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi2):
11766 Fix misleading identation.
11768 2023-05-30 Uros Bizjak <ubizjak@gmail.com>
11770 * rtl.h (comparison_dominates_p): Change return type from int to bool.
11771 (condjump_p): Ditto.
11772 (any_condjump_p): Ditto.
11773 (any_uncondjump_p): Ditto.
11774 (simplejump_p): Ditto.
11775 (returnjump_p): Ditto.
11776 (eh_returnjump_p): Ditto.
11777 (onlyjump_p): Ditto.
11778 (invert_jump_1): Ditto.
11779 (invert_jump): Ditto.
11780 (rtx_renumbered_equal_p): Ditto.
11781 (redirect_jump_1): Ditto.
11782 (redirect_jump): Ditto.
11783 (condjump_in_parallel_p): Ditto.
11784 * jump.cc (invert_exp_1): Adjust forward declaration.
11785 (comparison_dominates_p): Change return type from int to bool
11786 and adjust function body accordingly.
11787 (simplejump_p): Ditto.
11788 (condjump_p): Ditto.
11789 (condjump_in_parallel_p): Ditto.
11790 (any_uncondjump_p): Ditto.
11791 (any_condjump_p): Ditto.
11792 (returnjump_p): Ditto.
11793 (eh_returnjump_p): Ditto.
11794 (onlyjump_p): Ditto.
11795 (redirect_jump_1): Ditto.
11796 (redirect_jump): Ditto.
11797 (invert_exp_1): Ditto.
11798 (invert_jump_1): Ditto.
11799 (invert_jump): Ditto.
11800 (rtx_renumbered_equal_p): Ditto.
11802 2023-05-30 Andrew Pinski <apinski@marvell.com>
11804 * fold-const.cc (minmax_from_comparison): Add support for NE_EXPR.
11805 * match.pd ((cond (cmp (convert1? x) c1) (convert2? x) c2) pattern):
11806 Add ne as a possible cmp.
11807 ((a CMP b) ? minmax<a, c> : minmax<b, c> pattern): Likewise.
11809 2023-05-30 Andrew Pinski <apinski@marvell.com>
11811 * match.pd (`(a CMP CST1) ? max<a,CST2> : a`): New
11814 2023-05-30 Roger Sayle <roger@nextmovesoftware.com>
11816 * simplify-rtx.cc (simplify_binary_operation_1) <AND>: Use wide-int
11817 instead of HWI_COMPUTABLE_MODE_P and UINTVAL in transformation of
11818 (and (extend X) C) as (zero_extend (and X C)), to also optimize
11819 modes wider than HOST_WIDE_INT.
11821 2023-05-30 Roger Sayle <roger@nextmovesoftware.com>
11824 * simplify-rtx.cc (simplify_const_relational_operation): Return
11825 early if we have a MODE_CC comparison that isn't a COMPARE against
11828 2023-05-30 Robin Dapp <rdapp@ventanamicro.com>
11830 * config/riscv/riscv.cc (riscv_const_insns): Allow
11831 const_vec_duplicates.
11833 2023-05-30 liuhongt <hongtao.liu@intel.com>
11835 PR middle-end/108938
11836 * gimple-ssa-store-merging.cc (is_bswap_or_nop_p): New
11837 function, cut from original find_bswap_or_nop function.
11838 (find_bswap_or_nop): Add a new parameter, detect bswap +
11839 rotate and save rotate result in the new parameter.
11840 (bswap_replace): Add a new parameter to indicate rotate and
11841 generate rotate stmt if needed.
11842 (maybe_optimize_vector_constructor): Adjust for new rotate
11843 parameter in the upper 2 functions.
11844 (pass_optimize_bswap::execute): Ditto.
11845 (imm_store_chain_info::output_merged_store): Ditto.
11847 2023-05-30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11849 * config/aarch64/aarch64-simd.md (aarch64_<sur>adalp<mode>): Delete.
11850 (aarch64_<su>adalp<mode>): New define_expand.
11851 (*aarch64_<su>adalp<mode><vczle><vczbe>_insn): New define_insn.
11852 (aarch64_<su>addlp<mode>): Convert to define_expand.
11853 (*aarch64_<su>addlp<mode><vczle><vczbe>_insn): New define_insn.
11854 * config/aarch64/iterators.md (UNSPEC_SADDLP, UNSPEC_UADDLP): Delete.
11856 (USADDLP): Likewise.
11857 * config/aarch64/predicates.md (vect_par_cnst_even_or_odd_half): Define.
11859 2023-05-30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11861 * config/aarch64/aarch64-builtins.cc (VAR1): Move to after inclusion of
11862 aarch64-builtin-iterators.h. Add definition to remap shadd, uhadd,
11863 srhadd, urhadd builtin codes for standard optab ones.
11864 * config/aarch64/aarch64-simd.md (<u>avg<mode>3_floor): Rename to...
11865 (<su_optab>avg<mode>3_floor): ... This. Expand to RTL codes rather than
11867 (<u>avg<mode>3_ceil): Rename to...
11868 (<su_optab>avg<mode>3_ceil): ... This. Expand to RTL codes rather than
11870 (aarch64_<su>hsub<mode>): New define_expand.
11871 (aarch64_<sur>h<addsub><mode><vczle><vczbe>): Split into...
11872 (*aarch64_<su>h<ADDSUB:optab><mode><vczle><vczbe>_insn): ... This...
11873 (*aarch64_<su>rhadd<mode><vczle><vczbe>_insn): ... And this.
11875 2023-05-30 Andreas Schwab <schwab@suse.de>
11878 * config/riscv/riscv.cc (riscv_asan_shadow_offset): Update to
11879 match libsanitizer.
11881 2023-05-30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11883 * config/aarch64/aarch64-modes.def (V16HI, V8SI, V4DI, V2TI): New modes.
11884 * config/aarch64/aarch64-protos.h (aarch64_const_vec_rnd_cst_p):
11886 (aarch64_const_vec_rsra_rnd_imm_p): Likewise.
11887 * config/aarch64/aarch64-simd.md (*aarch64_simd_sra<mode>): Rename to...
11888 (aarch64_<sra_op>sra_n<mode>_insn): ... This.
11889 (aarch64_<sra_op>rsra_n<mode>_insn): New define_insn.
11890 (aarch64_<sra_op>sra_n<mode>): New define_expand.
11891 (aarch64_<sra_op>rsra_n<mode>): Likewise.
11892 (aarch64_<sur>sra_n<mode>): Rename to...
11893 (aarch64_<sur>sra_ndi): ... This.
11894 * config/aarch64/aarch64.cc (aarch64_classify_vector_mode): Add
11895 any_target_p argument.
11896 (aarch64_extract_vec_duplicate_wide_int): Define.
11897 (aarch64_const_vec_rsra_rnd_imm_p): Likewise.
11898 (aarch64_const_vec_rnd_cst_p): Likewise.
11899 (aarch64_vector_mode_supported_any_target_p): Likewise.
11900 (TARGET_VECTOR_MODE_SUPPORTED_ANY_TARGET_P): Likewise.
11901 * config/aarch64/iterators.md (UNSPEC_SRSRA, UNSPEC_URSRA): Delete.
11902 (VSRA): Adjust for the above.
11904 (V2XWIDE): New mode_attr.
11905 (vec_or_offset): Likewise.
11906 (SHIFTEXTEND): Likewise.
11907 * config/aarch64/predicates.md (aarch64_simd_rsra_rnd_imm_vec): New
11909 * doc/tm.texi (TARGET_VECTOR_MODE_SUPPORTED_P): Adjust description to
11910 clarify that it applies to current target options.
11911 (TARGET_VECTOR_MODE_SUPPORTED_ANY_TARGET_P): Document.
11912 * doc/tm.texi.in: Regenerate.
11913 * stor-layout.cc (mode_for_vector): Check
11914 vector_mode_supported_any_target_p when iterating through vector modes.
11915 * target.def (TARGET_VECTOR_MODE_SUPPORTED_P): Adjust description to
11916 clarify that it applies to current target options.
11917 (TARGET_VECTOR_MODE_SUPPORTED_ANY_TARGET_P): Define.
11919 2023-05-30 Lili Cui <lili.cui@intel.com>
11921 PR tree-optimization/98350
11922 * tree-ssa-reassoc.cc
11923 (rewrite_expr_tree_parallel): Rewrite this function.
11924 (rank_ops_for_fma): New.
11925 (reassociate_bb): Handle new function.
11927 2023-05-30 Uros Bizjak <ubizjak@gmail.com>
11929 * rtl.h (rtx_addr_can_trap_p): Change return type from int to bool.
11930 (rtx_unstable_p): Ditto.
11931 (reg_mentioned_p): Ditto.
11932 (reg_referenced_p): Ditto.
11933 (reg_used_between_p): Ditto.
11934 (reg_set_between_p): Ditto.
11935 (modified_between_p): Ditto.
11936 (no_labels_between_p): Ditto.
11937 (modified_in_p): Ditto.
11938 (reg_set_p): Ditto.
11939 (multiple_sets): Ditto.
11940 (set_noop_p): Ditto.
11941 (noop_move_p): Ditto.
11942 (reg_overlap_mentioned_p): Ditto.
11943 (dead_or_set_p): Ditto.
11944 (dead_or_set_regno_p): Ditto.
11945 (find_reg_fusage): Ditto.
11946 (find_regno_fusage): Ditto.
11947 (side_effects_p): Ditto.
11948 (volatile_refs_p): Ditto.
11949 (volatile_insn_p): Ditto.
11950 (may_trap_p_1): Ditto.
11951 (may_trap_p): Ditto.
11952 (may_trap_or_fault_p): Ditto.
11953 (computed_jump_p): Ditto.
11954 (auto_inc_p): Ditto.
11955 (loc_mentioned_in_p): Ditto.
11956 * rtlanal.cc (computed_jump_p_1): Adjust forward declaration.
11957 (rtx_unstable_p): Change return type from int to bool
11958 and adjust function body accordingly.
11959 (rtx_addr_can_trap_p): Ditto.
11960 (reg_mentioned_p): Ditto.
11961 (no_labels_between_p): Ditto.
11962 (reg_used_between_p): Ditto.
11963 (reg_referenced_p): Ditto.
11964 (reg_set_between_p): Ditto.
11965 (reg_set_p): Ditto.
11966 (modified_between_p): Ditto.
11967 (modified_in_p): Ditto.
11968 (multiple_sets): Ditto.
11969 (set_noop_p): Ditto.
11970 (noop_move_p): Ditto.
11971 (reg_overlap_mentioned_p): Ditto.
11972 (dead_or_set_p): Ditto.
11973 (dead_or_set_regno_p): Ditto.
11974 (find_reg_fusage): Ditto.
11975 (find_regno_fusage): Ditto.
11976 (remove_node_from_insn_list): Ditto.
11977 (volatile_insn_p): Ditto.
11978 (volatile_refs_p): Ditto.
11979 (side_effects_p): Ditto.
11980 (may_trap_p_1): Ditto.
11981 (may_trap_p): Ditto.
11982 (may_trap_or_fault_p): Ditto.
11983 (computed_jump_p): Ditto.
11984 (auto_inc_p): Ditto.
11985 (loc_mentioned_in_p): Ditto.
11986 * combine.cc (can_combine_p): Update indirect function.
11988 2023-05-30 Juzhe-Zhong <juzhe.zhong@rivai.ai>
11990 * config/riscv/autovec.md (<optab><mode><vconvert>2): New pattern.
11991 * config/riscv/iterators.md: New attribute.
11992 * config/riscv/vector-iterators.md: New attribute.
11994 2023-05-30 From: Juzhe-Zhong <juzhe.zhong@rivai.ai>
11996 * config/riscv/riscv.md: Fix signed and unsigned comparison
11999 2023-05-30 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12001 * config/riscv/autovec.md (fnma<mode>4): New pattern.
12002 (*fnma<mode>): Ditto.
12004 2023-05-29 Die Li <lidie@eswincomputing.com>
12006 * config/riscv/riscv.cc (riscv_expand_conditional_move_onesided):
12008 (riscv_expand_conditional_move): Reuse the TARGET_SFB_ALU expand
12009 process for TARGET_XTHEADCONDMOV
12011 2023-05-29 Uros Bizjak <ubizjak@gmail.com>
12014 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi2): Also require
12015 TARGET_AVX512BW to generate truncv16hiv16qi2.
12017 2023-05-29 Jivan Hakobyan <jivanhakobyan9@gmail.com>
12019 * config/riscv/riscv.md (and<mode>3): New expander.
12020 (*and<mode>3) New pattern.
12021 * config/riscv/predicates.md (arith_operand_or_mode_mask): New
12024 2023-05-29 Pan Li <pan2.li@intel.com>
12026 * config/riscv/riscv-v.cc (emit_vlmax_insn): Remove unnecessary
12027 comments and rename local variables.
12028 (emit_nonvlmax_insn): Diito.
12029 (emit_vlmax_merge_insn): Ditto.
12030 (emit_vlmax_cmp_insn): Ditto.
12031 (emit_vlmax_cmp_mu_insn): Ditto.
12032 (emit_scalar_move_insn): Ditto.
12034 2023-05-29 Pan Li <pan2.li@intel.com>
12036 * config/riscv/riscv-v.cc (emit_vlmax_insn): Eliminate the
12038 (emit_nonvlmax_insn): Ditto.
12039 (emit_vlmax_merge_insn): Ditto.
12040 (emit_vlmax_cmp_insn): Ditto.
12041 (emit_vlmax_cmp_mu_insn): Ditto.
12042 (expand_vec_series): Ditto.
12044 2023-05-29 Pan Li <pan2.li@intel.com>
12046 * config/riscv/riscv-protos.h (enum insn_type): New type.
12047 * config/riscv/riscv-v.cc (RVV_INSN_OPERANDS_MAX): New macro.
12048 (rvv_builder::can_duplicate_repeating_sequence_p): Align the referenced
12050 (rvv_builder::get_merged_repeating_sequence): Ditto.
12051 (rvv_builder::repeating_sequence_use_merge_profitable_p): New function
12052 to evaluate the optimization cost.
12053 (rvv_builder::get_merge_scalar_mask): New function to get the merge
12055 (emit_scalar_move_insn): New function to emit vmv.s.x.
12056 (emit_vlmax_integer_move_insn): New function to emit vlmax vmv.v.x.
12057 (emit_nonvlmax_integer_move_insn): New function to emit nonvlmax
12059 (get_repeating_sequence_dup_machine_mode): New function to get the dup
12061 (expand_vector_init_merge_repeating_sequence): New function to perform
12063 (expand_vec_init): Add this vector init optimization.
12064 * config/riscv/riscv.h (BITS_PER_WORD): New macro.
12066 2023-05-29 Eric Botcazou <ebotcazou@adacore.com>
12068 * tree-ssa-loop-manip.cc (create_iv): Try harder to find a SLOC to
12069 put onto the increment when it is inserted after the position.
12071 2023-05-29 Eric Botcazou <ebotcazou@adacore.com>
12073 * match.pd ((T)P - (T)(P + A) -> -(T) A): Avoid artificial overflow
12076 2023-05-29 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12078 * config/riscv/riscv-vsetvl.cc (source_equal_p): Fix ICE.
12080 2023-05-29 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12082 * config/riscv/autovec.md (fma<mode>4): New pattern.
12083 (*fma<mode>): Ditto.
12084 * config/riscv/riscv-protos.h (enum insn_type): New enum.
12085 (emit_vlmax_ternary_insn): New function.
12086 * config/riscv/riscv-v.cc (emit_vlmax_ternary_insn): Ditto.
12088 2023-05-29 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12090 * config/riscv/vector.md: Fix vimuladd instruction bug.
12092 2023-05-29 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12094 * config/riscv/riscv.cc (global_state_unknown_p): New function.
12095 (riscv_mode_after): Fix incorrect VXM.
12097 2023-05-29 Pan Li <pan2.li@intel.com>
12099 * common/config/riscv/riscv-common.cc:
12100 (riscv_implied_info): Add zvfhmin item.
12101 (riscv_ext_version_table): Ditto.
12102 (riscv_ext_flag_table): Ditto.
12103 * config/riscv/riscv-opts.h (MASK_ZVFHMIN): New macro.
12104 (TARGET_ZFHMIN): Align indent.
12105 (TARGET_ZFH): Ditto.
12106 (TARGET_ZVFHMIN): New macro.
12108 2023-05-27 liuhongt <hongtao.liu@intel.com>
12111 * config/i386/sse.md (*andnot<mode>3): Extend below splitter
12112 to VI_AVX2 to cover more modes.
12114 2023-05-27 liuhongt <hongtao.liu@intel.com>
12116 * config/i386/x86-tune.def (X86_TUNE_AVOID_FALSE_DEP_FOR_BMI):
12117 Remove ATOM and ICELAKE(and later) core processors.
12119 2023-05-26 Robin Dapp <rdapp@ventanamicro.com>
12121 * config/riscv/autovec.md (<optab><mode>2): Add vneg/vnot.
12123 * config/riscv/riscv-protos.h (emit_vlmax_masked_mu_insn):
12125 * config/riscv/riscv-v.cc (emit_vlmax_masked_mu_insn): New
12128 2023-05-26 Robin Dapp <rdapp@ventanamicro.com>
12129 Juzhe Zhong <juzhe.zhong@rivai.ai>
12131 * config/riscv/autovec.md (<optab><v_double_trunc><mode>2): New
12133 (<optab><v_quad_trunc><mode>2): Dito.
12134 (<optab><v_oct_trunc><mode>2): Dito.
12135 (trunc<mode><v_double_trunc>2): Dito.
12136 (trunc<mode><v_quad_trunc>2): Dito.
12137 (trunc<mode><v_oct_trunc>2): Dito.
12138 * config/riscv/riscv-protos.h (vectorize_related_mode): Define.
12139 (autovectorize_vector_modes): Define.
12140 * config/riscv/riscv-v.cc (vectorize_related_mode): Implement
12142 (autovectorize_vector_modes): Implement hook.
12143 * config/riscv/riscv.cc (riscv_autovectorize_vector_modes):
12144 Implement target hook.
12145 (riscv_vectorize_related_mode): Implement target hook.
12146 (TARGET_VECTORIZE_AUTOVECTORIZE_VECTOR_MODES): Define.
12147 (TARGET_VECTORIZE_RELATED_MODE): Define.
12148 * config/riscv/vector-iterators.md: Add lowercase versions of
12149 mode_attr iterators.
12151 2023-05-26 Andrew Stubbs <ams@codesourcery.com>
12152 Tobias Burnus <tobias@codesourcery.com>
12154 * config/gcn/gcn-hsa.h (XNACKOPT): New macro.
12155 (ASM_SPEC): Use XNACKOPT.
12156 * config/gcn/gcn-opts.h (enum sram_ecc_type): Rename to ...
12157 (enum hsaco_attr_type): ... this, and generalize the names.
12158 (TARGET_XNACK): New macro.
12159 * config/gcn/gcn.cc (gcn_option_override): Update to sorry for all
12161 (output_file_start): Update xnack handling.
12162 (gcn_hsa_declare_function_name): Use TARGET_XNACK.
12163 * config/gcn/gcn.opt (-mxnack): Add the "on/off/any" syntax.
12164 (sram_ecc_type): Rename to ...
12165 (hsaco_attr_type: ... this.)
12166 * config/gcn/mkoffload.cc (SET_XNACK_ANY): New macro.
12167 (TEST_XNACK): Delete.
12168 (TEST_XNACK_ANY): New macro.
12169 (TEST_XNACK_ON): New macro.
12170 (main): Support the new -mxnack=on/off/any syntax.
12171 * doc/invoke.texi (-mxnack): Update for new syntax.
12173 2023-05-26 Andrew Pinski <apinski@marvell.com>
12175 * genmatch.cc (emit_debug_printf): New function.
12176 (dt_simplify::gen_1): Emit printf into the code
12177 before the `return true` or returning the folded result
12178 instead of emitting it always.
12180 2023-05-26 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12182 * config/xtensa/xtensa-protos.h
12183 (xtensa_expand_block_set_unrolled_loop,
12184 xtensa_expand_block_set_small_loop): Remove.
12185 (xtensa_expand_block_set): New prototype.
12186 * config/xtensa/xtensa.cc
12187 (xtensa_expand_block_set_libcall): New subfunction.
12188 (xtensa_expand_block_set_unrolled_loop,
12189 xtensa_expand_block_set_small_loop): Rewrite as subfunctions.
12190 (xtensa_expand_block_set): New function that calls the above
12192 * config/xtensa/xtensa.md (memsetsi): Change to invoke only
12193 xtensa_expand_block_set().
12195 2023-05-26 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12197 * config/xtensa/xtensa-protos.h (xtensa_m1_or_1_thru_15):
12199 * config/xtensa/xtensa.cc (xtensa_m1_or_1_thru_15):
12201 * config/xtensa/constraints.md (O):
12202 Change to use the above function.
12203 * config/xtensa/xtensa.md (*subsi3_from_const):
12204 New insn_and_split pattern.
12206 2023-05-26 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12208 * config/xtensa/xtensa.md (*extzvsi-1bit_ashlsi3):
12209 Retract excessive line folding, and correct the value of
12210 the "length" insn attribute related to TARGET_DENSITY.
12211 (*extzvsi-1bit_addsubx): Ditto.
12213 2023-05-26 Uros Bizjak <ubizjak@gmail.com>
12215 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi):
12216 Do not disable call to ix86_expand_vecop_qihi2.
12218 2023-05-26 liuhongt <hongtao.liu@intel.com>
12222 * ira-costs.cc (scan_one_insn): Only use NO_REGS in cost
12223 calculation when !hard_regno_mode_ok for GENERAL_REGS and
12224 mode, otherwise still use GENERAL_REGS.
12226 2023-05-26 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12228 * config/riscv/riscv.cc (vector_zero_call_used_regs): Add
12229 explict VL and drop VL in ops.
12231 2023-05-25 Jin Ma <jinma@linux.alibaba.com>
12233 * sched-deps.cc (sched_macro_fuse_insns): Insns should not be fusion
12234 in different BB blocks.
12236 2023-05-25 Uros Bizjak <ubizjak@gmail.com>
12238 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi2):
12239 Rewrite to expand to 2x-wider (e.g. V16QI -> V16HImode)
12240 instructions when available. Emulate truncation via
12241 ix86_expand_vec_perm_const_1 when native truncate insn
12243 (ix86_expand_vecop_qihi_partial) <case MULT>: Use pmovzx
12244 when available. Trivially rename some variables.
12245 (ix86_expand_vecop_qihi): Unconditionally call ix86_expand_vecop_qihi2.
12246 * config/i386/i386.cc (ix86_multiplication_cost): Rewrite cost
12247 calculation of V*QImode emulations to account for generation of
12248 2x-wider mode instructions.
12249 (ix86_shift_rotate_cost): Update cost calculation of V*QImode
12250 emulations to account for generation of 2x-wider mode instructions.
12252 2023-05-25 Georg-Johann Lay <avr@gjlay.de>
12255 * config/avr/avr.cc (avr_can_inline_p): New static function.
12256 (TARGET_CAN_INLINE_P): Define to that function.
12258 2023-05-25 Georg-Johann Lay <avr@gjlay.de>
12261 * config/avr/avr.md (*movbitqi.0): Rename to *movbit<mode>.0-6.
12262 Handle any bit position and use mode QISI.
12263 * config/avr/avr.cc (avr_rtx_costs_1) [IOR]: Return a cost
12264 of 2 insns for bit-transfer of respective style.
12266 2023-05-25 Christophe Lyon <christophe.lyon@linaro.org>
12268 * config/arm/iterators.md (MVE_6): Remove.
12269 * config/arm/mve.md: Replace MVE_6 with MVE_5.
12271 2023-05-25 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
12272 Richard Sandiford <richard.sandiford@arm.com>
12274 * tree-vect-loop-manip.cc (vect_adjust_loop_lens_control): New
12276 (vect_set_loop_controls_directly): Add decrement IV support.
12277 (vect_set_loop_condition_partial_vectors): Ditto.
12278 * tree-vect-loop.cc (_loop_vec_info::_loop_vec_info): New
12280 * tree-vectorizer.h (LOOP_VINFO_USING_DECREMENTING_IV_P): New
12283 2023-05-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12286 * config/aarch64/aarch64-simd.md (aarch64_fcadd<rot><mode>): Rename to...
12287 (aarch64_fcadd<rot><mode><vczle><vczbe>): ... This.
12288 Fix canonicalization of PLUS operands.
12289 (aarch64_fcmla<rot><mode>): Rename to...
12290 (aarch64_fcmla<rot><mode><vczle><vczbe>): ... This.
12291 Fix canonicalization of PLUS operands.
12292 (aarch64_fcmla_lane<rot><mode>): Rename to...
12293 (aarch64_fcmla_lane<rot><mode><vczle><vczbe>): ... This.
12294 Fix canonicalization of PLUS operands.
12295 (aarch64_fcmla_laneq<rot>v4hf): Rename to...
12296 (aarch64_fcmla_laneq<rot>v4hf<vczle><vczbe>): ... This.
12297 Fix canonicalization of PLUS operands.
12298 (aarch64_fcmlaq_lane<rot><mode>): Fix canonicalization of PLUS operands.
12300 2023-05-25 Chris Sidebottom <chris.sidebottom@arm.com>
12302 * config/arm/arm.md (rbitsi2): Rename to...
12303 (arm_rbit): ... This.
12304 (ctzsi2): Adjust for the above.
12305 (arm_rev16si2): Convert to define_expand.
12306 (arm_rev16si2_alt1): New pattern.
12307 (arm_rev16si2_alt): Rename to...
12308 (*arm_rev16si2_alt2): ... This.
12309 * config/arm/arm_acle.h (__ror, __rorl, __rorll, __clz, __clzl, __clzll,
12310 __cls, __clsl, __clsll, __revsh, __rev, __revl, __revll, __rev16,
12311 __rev16l, __rev16ll, __rbit, __rbitl, __rbitll): Define intrinsics.
12312 * config/arm/arm_acle_builtins.def (rbit, rev16si2): Define builtins.
12314 2023-05-25 Alex Coplan <alex.coplan@arm.com>
12317 * config/arm/arm.md (movdf): Generate temporary pseudo in DImode
12319 * config/arm/vfp.md (no_literal_pool_df_immediate): Rather than punning an
12320 lvalue DFmode pseudo into DImode, use a DImode pseudo and pun it into
12321 DFmode as an rvalue.
12323 2023-05-25 Richard Biener <rguenther@suse.de>
12326 * tree-vect-stmts.cc (vectorizable_condition): For
12327 embedded comparisons also handle the case when the target
12328 only provides vec_cmp and vcond_mask.
12330 2023-05-25 Claudiu Zissulescu <claziss@gmail.com>
12332 * config/arc/arc.cc (arc_call_tls_get_addr): Simplify access using
12335 2023-05-25 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
12337 * config/aarch64/aarch64.cc (scalar_move_insn_p): New function.
12338 (seq_cost_ignoring_scalar_moves): Likewise.
12339 (aarch64_expand_vector_init): Call seq_cost_ignoring_scalar_moves.
12341 2023-05-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12343 * config/aarch64/arm_neon.h (vcage_f64): Reimplement with builtins.
12344 (vcage_f32): Likewise.
12345 (vcages_f32): Likewise.
12346 (vcageq_f32): Likewise.
12347 (vcaged_f64): Likewise.
12348 (vcageq_f64): Likewise.
12349 (vcagts_f32): Likewise.
12350 (vcagt_f32): Likewise.
12351 (vcagt_f64): Likewise.
12352 (vcagtq_f32): Likewise.
12353 (vcagtd_f64): Likewise.
12354 (vcagtq_f64): Likewise.
12355 (vcale_f32): Likewise.
12356 (vcale_f64): Likewise.
12357 (vcaled_f64): Likewise.
12358 (vcales_f32): Likewise.
12359 (vcaleq_f32): Likewise.
12360 (vcaleq_f64): Likewise.
12361 (vcalt_f32): Likewise.
12362 (vcalt_f64): Likewise.
12363 (vcaltd_f64): Likewise.
12364 (vcaltq_f32): Likewise.
12365 (vcaltq_f64): Likewise.
12366 (vcalts_f32): Likewise.
12368 2023-05-25 Hu, Lin1 <lin1.hu@intel.com>
12372 * config/i386/avx512bwintrin.h (_mm512_srli_epi16): Change type from
12373 int to const int or const int to const unsigned int.
12374 (_mm512_mask_srli_epi16): Ditto.
12375 (_mm512_slli_epi16): Ditto.
12376 (_mm512_mask_slli_epi16): Ditto.
12377 (_mm512_maskz_slli_epi16): Ditto.
12378 (_mm512_srai_epi16): Ditto.
12379 (_mm512_mask_srai_epi16): Ditto.
12380 (_mm512_maskz_srai_epi16): Ditto.
12381 * config/i386/avx512fintrin.h (_mm512_slli_epi64): Ditto.
12382 (_mm512_mask_slli_epi64): Ditto.
12383 (_mm512_maskz_slli_epi64): Ditto.
12384 (_mm512_srli_epi64): Ditto.
12385 (_mm512_mask_srli_epi64): Ditto.
12386 (_mm512_maskz_srli_epi64): Ditto.
12387 (_mm512_srai_epi64): Ditto.
12388 (_mm512_mask_srai_epi64): Ditto.
12389 (_mm512_maskz_srai_epi64): Ditto.
12390 (_mm512_slli_epi32): Ditto.
12391 (_mm512_mask_slli_epi32): Ditto.
12392 (_mm512_maskz_slli_epi32): Ditto.
12393 (_mm512_srli_epi32): Ditto.
12394 (_mm512_mask_srli_epi32): Ditto.
12395 (_mm512_maskz_srli_epi32): Ditto.
12396 (_mm512_srai_epi32): Ditto.
12397 (_mm512_mask_srai_epi32): Ditto.
12398 (_mm512_maskz_srai_epi32): Ditto.
12399 * config/i386/avx512vlbwintrin.h (_mm256_mask_srai_epi16): Ditto.
12400 (_mm256_maskz_srai_epi16): Ditto.
12401 (_mm_mask_srai_epi16): Ditto.
12402 (_mm_maskz_srai_epi16): Ditto.
12403 (_mm256_mask_slli_epi16): Ditto.
12404 (_mm256_maskz_slli_epi16): Ditto.
12405 (_mm_mask_slli_epi16): Ditto.
12406 (_mm_maskz_slli_epi16): Ditto.
12407 (_mm_maskz_srli_epi16): Ditto.
12408 * config/i386/avx512vlintrin.h (_mm256_mask_srli_epi32): Ditto.
12409 (_mm256_maskz_srli_epi32): Ditto.
12410 (_mm_mask_srli_epi32): Ditto.
12411 (_mm_maskz_srli_epi32): Ditto.
12412 (_mm256_mask_srli_epi64): Ditto.
12413 (_mm256_maskz_srli_epi64): Ditto.
12414 (_mm_mask_srli_epi64): Ditto.
12415 (_mm_maskz_srli_epi64): Ditto.
12416 (_mm256_mask_srai_epi32): Ditto.
12417 (_mm256_maskz_srai_epi32): Ditto.
12418 (_mm_mask_srai_epi32): Ditto.
12419 (_mm_maskz_srai_epi32): Ditto.
12420 (_mm256_srai_epi64): Ditto.
12421 (_mm256_mask_srai_epi64): Ditto.
12422 (_mm256_maskz_srai_epi64): Ditto.
12423 (_mm_srai_epi64): Ditto.
12424 (_mm_mask_srai_epi64): Ditto.
12425 (_mm_maskz_srai_epi64): Ditto.
12426 (_mm_mask_slli_epi32): Ditto.
12427 (_mm_maskz_slli_epi32): Ditto.
12428 (_mm_mask_slli_epi64): Ditto.
12429 (_mm_maskz_slli_epi64): Ditto.
12430 (_mm256_mask_slli_epi32): Ditto.
12431 (_mm256_maskz_slli_epi32): Ditto.
12432 (_mm256_mask_slli_epi64): Ditto.
12433 (_mm256_maskz_slli_epi64): Ditto.
12435 2023-05-25 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12437 * config/riscv/vector.md: Remove FRM_REGNUM dependency in rtz
12440 2023-05-25 Aldy Hernandez <aldyh@redhat.com>
12442 * data-streamer-in.cc (streamer_read_value_range): Handle NANs.
12443 * data-streamer-out.cc (streamer_write_vrange): Same.
12444 * value-range.h (class vrange): Make streamer_write_vrange a friend.
12446 2023-05-25 Aldy Hernandez <aldyh@redhat.com>
12448 * value-query.cc (range_query::get_tree_range): Set NAN directly
12450 * value-range.cc (frange::set): Assert that bounds are not NAN.
12452 2023-05-25 Aldy Hernandez <aldyh@redhat.com>
12454 * value-range.cc (add_vrange): Handle known NANs.
12456 2023-05-25 Aldy Hernandez <aldyh@redhat.com>
12458 * value-range.h (frange::set_nan): New.
12460 2023-05-25 Alexandre Oliva <oliva@adacore.com>
12463 * emit-rtl.cc (validate_subreg): Reject a SUBREG of a MEM that
12464 requires stricter alignment than MEM's.
12466 2023-05-24 Andrew MacLeod <amacleod@redhat.com>
12468 PR tree-optimization/107822
12469 PR tree-optimization/107986
12470 * Makefile.in (OBJS): Add gimple-range-phi.o.
12471 * gimple-range-cache.h (ranger_cache::m_estimate): New
12472 phi_analyzer pointer member.
12473 * gimple-range-fold.cc (fold_using_range::range_of_phi): Use
12474 phi_analyzer if no loop info is available.
12475 * gimple-range-phi.cc: New file.
12476 * gimple-range-phi.h: New file.
12477 * tree-vrp.cc (execute_ranger_vrp): Utililze a phi_analyzer.
12479 2023-05-24 Andrew MacLeod <amacleod@redhat.com>
12481 * gimple-range-fold.cc (fur_list::fur_list): Add range_query param
12483 (fold_range): Add range_query parameter.
12484 (fur_relation::fur_relation): New.
12485 (fur_relation::trio): New.
12486 (fur_relation::register_relation): New.
12487 (fold_relations): New.
12488 * gimple-range-fold.h (fold_range): Adjust prototypes.
12489 (fold_relations): New.
12491 2023-05-24 Andrew MacLeod <amacleod@redhat.com>
12493 * gimple-range-cache.cc (ssa_cache::range_of_expr): New.
12494 * gimple-range-cache.h (class ssa_cache): Inherit from range_query.
12495 (ranger_cache::const_query): New.
12496 * gimple-range.cc (gimple_ranger::const_query): New.
12497 * gimple-range.h (gimple_ranger::const_query): New prototype.
12499 2023-05-24 Andrew MacLeod <amacleod@redhat.com>
12501 * gimple-range-cache.cc (ssa_cache::dump): Use get_range.
12502 (ssa_cache::dump_range_query): Delete.
12503 (ssa_lazy_cache::dump_range_query): Delete.
12504 (ssa_lazy_cache::get_range): Move from header file.
12505 (ssa_lazy_cache::clear_range): ditto.
12506 (ssa_lazy_cache::clear): Ditto.
12507 * gimple-range-cache.h (class ssa_cache): Virtualize.
12508 (class ssa_lazy_cache): Inherit and virtualize.
12510 2023-05-24 Aldy Hernandez <aldyh@redhat.com>
12512 * value-range.h (vrange::kind): Remove.
12514 2023-05-24 Roger Sayle <roger@nextmovesoftware.com>
12516 PR middle-end/109840
12517 * match.pd <popcount optimizations>: Preserve zero-extension when
12518 optimizing popcount((T)bswap(x)) and popcount((T)rotate(x,y)) as
12519 popcount((T)x), so the popcount's argument keeps the same type.
12520 <parity optimizations>: Likewise preserve extensions when
12521 simplifying parity((T)bswap(x)) and parity((T)rotate(x,y)) as
12522 parity((T)x), so that the parity's argument type is the same.
12524 2023-05-24 Aldy Hernandez <aldyh@redhat.com>
12526 * ipa-cp.cc (ipa_value_range_from_jfunc): Use new ipa_vr API.
12527 (ipcp_store_vr_results): Same.
12528 * ipa-prop.cc (ipa_vr::ipa_vr): New.
12529 (ipa_vr::get_vrange): New.
12530 (ipa_vr::set_unknown): New.
12531 (ipa_vr::streamer_read): New.
12532 (ipa_vr::streamer_write): New.
12533 (write_ipcp_transformation_info): Use new ipa_vr API.
12534 (read_ipcp_transformation_info): Same.
12535 (ipa_vr::nonzero_p): Delete.
12536 (ipcp_update_vr): Use new ipa_vr API.
12537 * ipa-prop.h (class ipa_vr): Provide an API and hide internals.
12538 * ipa-sra.cc (zap_useless_ipcp_results): Use new ipa_vr API.
12540 2023-05-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12542 * config/mcore/mcore.cc (output_inline_const) Make buffer smaller to
12543 silence overflow warnings later on.
12545 2023-05-24 Uros Bizjak <ubizjak@gmail.com>
12547 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi2):
12548 Remove handling of V8QImode.
12549 * config/i386/mmx.md (v<insn>v8qi3): Move from sse.md.
12550 Call ix86_expand_vecop_qihi_partial. Enable for TARGET_MMX_WITH_SSE.
12551 (v<insn>v4qi3): Ditto.
12552 * config/i386/sse.md (v<insn>v8qi3): Remove.
12554 2023-05-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12557 * config/aarch64/aarch64-simd.md (aarch64_simd_lshr<mode>): Rename to...
12558 (aarch64_simd_lshr<mode><vczle><vczbe>): ... This.
12559 (aarch64_simd_ashr<mode>): Rename to...
12560 (aarch64_simd_ashr<mode><vczle><vczbe>): ... This.
12561 (aarch64_simd_imm_shl<mode>): Rename to...
12562 (aarch64_simd_imm_shl<mode><vczle><vczbe>): ... This.
12563 (aarch64_simd_reg_sshl<mode>): Rename to...
12564 (aarch64_simd_reg_sshl<mode><vczle><vczbe>): ... This.
12565 (aarch64_simd_reg_shl<mode>_unsigned): Rename to...
12566 (aarch64_simd_reg_shl<mode>_unsigned<vczle><vczbe>): ... This.
12567 (aarch64_simd_reg_shl<mode>_signed): Rename to...
12568 (aarch64_simd_reg_shl<mode>_signed<vczle><vczbe>): ... This.
12569 (vec_shr_<mode>): Rename to...
12570 (vec_shr_<mode><vczle><vczbe>): ... This.
12571 (aarch64_<sur>shl<mode>): Rename to...
12572 (aarch64_<sur>shl<mode><vczle><vczbe>): ... This.
12573 (aarch64_<sur>q<r>shl<mode>): Rename to...
12574 (aarch64_<sur>q<r>shl<mode><vczle><vczbe>): ... This.
12576 2023-05-24 Richard Biener <rguenther@suse.de>
12579 * config/i386/i386-expand.cc (ix86_expand_vector_init_general):
12580 Perform final vector composition using
12581 ix86_expand_vector_init_general instead of setting
12582 the highpart and lowpart which causes spilling.
12584 2023-05-24 Andrew MacLeod <amacleod@redhat.com>
12586 PR tree-optimization/109695
12587 * gimple-range-cache.cc (ranger_cache::get_global_range): Add
12589 * gimple-range-cache.h (ranger_cache::get_global_range): Ditto.
12590 * gimple-range.cc (gimple_ranger::range_of_stmt): Pass changed
12591 flag to set_global_range.
12592 (gimple_ranger::prefill_stmt_dependencies): Ditto.
12594 2023-05-24 Andrew MacLeod <amacleod@redhat.com>
12596 PR tree-optimization/109695
12597 * gimple-range-cache.cc (temporal_cache::temporal_value): Return
12599 (temporal_cache::current_p): Check always_current method.
12600 (temporal_cache::set_always_current): Add param and set value
12602 (temporal_cache::always_current_p): New.
12603 (ranger_cache::get_global_range): Adjust.
12604 (ranger_cache::set_global_range): set always current first.
12606 2023-05-24 Andrew MacLeod <amacleod@redhat.com>
12608 PR tree-optimization/109695
12609 * gimple-range-cache.cc (ranger_cache::get_global_range): Call
12610 fold_range with global query to choose an initial value.
12612 2023-05-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12614 * config/riscv/riscv-protos.h (enum frm_field_enum): Add FRM_
12617 2023-05-24 Richard Biener <rguenther@suse.de>
12619 PR tree-optimization/109849
12620 * tree-ssa-pre.cc (do_hoist_insertion): Do not intersect
12621 expressions but take the first sets.
12623 2023-05-24 Gaius Mulley <gaiusmod2@gmail.com>
12626 * doc/gm2.texi (High procedure function): New node.
12627 (Using): New menu entry for High procedure function.
12629 2023-05-24 Richard Sandiford <richard.sandiford@arm.com>
12631 PR rtl-optimization/109940
12632 * early-remat.cc (postorder_index): Rename to...
12633 (rpo_index): ...this.
12634 (compare_candidates): Sort by decreasing rpo_index rather than
12635 increasing postorder_index.
12636 (early_remat::sort_candidates): Calculate the forward RPO from
12638 (early_remat::local_phase): Follow forward RPO using DF_FORWARD,
12639 rather than DF_BACKWARD in reverse.
12641 2023-05-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12644 * config/arm/arm-builtins.cc (SAT_BINOP_UNSIGNED_IMM_QUALIFIERS): Use
12645 qualifier_none for the return operand.
12647 2023-05-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12649 * config/riscv/autovec.md (<optab><mode>3): New pattern.
12650 (one_cmpl<mode>2): Ditto.
12651 (*<optab>not<mode>): Ditto.
12652 (*n<optab><mode>): Ditto.
12653 * config/riscv/riscv-v.cc (expand_vec_cmp_float): Change to
12656 2023-05-24 Kewen Lin <linkw@linux.ibm.com>
12658 * tree-vect-slp.cc (vect_transform_slp_perm_load_1): Adjust the
12659 calculation on n_perms by considering nvectors_per_build.
12661 2023-05-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12662 Richard Sandiford <richard.sandiford@arm.com>
12664 * config/riscv/autovec.md (@vcond_mask_<mode><vm>): New pattern.
12665 (vec_cmp<mode><vm>): New pattern.
12666 (vec_cmpu<mode><vm>): New pattern.
12667 (vcond<V:mode><VI:mode>): New pattern.
12668 (vcondu<V:mode><VI:mode>): New pattern.
12669 * config/riscv/riscv-protos.h (enum insn_type): Add new enum.
12670 (emit_vlmax_merge_insn): New function.
12671 (emit_vlmax_cmp_insn): Ditto.
12672 (emit_vlmax_cmp_mu_insn): Ditto.
12673 (expand_vec_cmp): Ditto.
12674 (expand_vec_cmp_float): Ditto.
12675 (expand_vcond): Ditto.
12676 * config/riscv/riscv-v.cc (emit_vlmax_merge_insn): Ditto.
12677 (emit_vlmax_cmp_insn): Ditto.
12678 (emit_vlmax_cmp_mu_insn): Ditto.
12679 (get_cmp_insn_code): Ditto.
12680 (expand_vec_cmp): Ditto.
12681 (expand_vec_cmp_float): Ditto.
12682 (expand_vcond): Ditto.
12684 2023-05-24 Pan Li <pan2.li@intel.com>
12686 * config/riscv/genrvv-type-indexer.cc (main): Add
12687 unsigned_eew*_lmul1_interpret for indexer.
12688 * config/riscv/riscv-vector-builtins-functions.def (vreinterpret):
12689 Register vuint*m1_t interpret function.
12690 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_UNSIGNED_EEW8_LMUL1_INTERPRET_OPS):
12691 New macro for vuint8m1_t.
12692 (DEF_RVV_UNSIGNED_EEW16_LMUL1_INTERPRET_OPS): Likewise.
12693 (DEF_RVV_UNSIGNED_EEW32_LMUL1_INTERPRET_OPS): Likewise.
12694 (DEF_RVV_UNSIGNED_EEW64_LMUL1_INTERPRET_OPS): Likewise.
12695 (vbool1_t): Add to unsigned_eew*_interpret_ops.
12696 (vbool2_t): Likewise.
12697 (vbool4_t): Likewise.
12698 (vbool8_t): Likewise.
12699 (vbool16_t): Likewise.
12700 (vbool32_t): Likewise.
12701 (vbool64_t): Likewise.
12702 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_UNSIGNED_EEW8_LMUL1_INTERPRET_OPS):
12703 New macro for vuint*m1_t.
12704 (DEF_RVV_UNSIGNED_EEW16_LMUL1_INTERPRET_OPS): Likewise.
12705 (DEF_RVV_UNSIGNED_EEW32_LMUL1_INTERPRET_OPS): Likewise.
12706 (DEF_RVV_UNSIGNED_EEW64_LMUL1_INTERPRET_OPS): Likewise.
12707 (required_extensions_p): Add vuint*m1_t interpret case.
12708 * config/riscv/riscv-vector-builtins.def (unsigned_eew8_lmul1_interpret):
12709 Add vuint*m1_t interpret to base type.
12710 (unsigned_eew16_lmul1_interpret): Likewise.
12711 (unsigned_eew32_lmul1_interpret): Likewise.
12712 (unsigned_eew64_lmul1_interpret): Likewise.
12714 2023-05-24 Pan Li <pan2.li@intel.com>
12716 * config/riscv/genrvv-type-indexer.cc (EEW_SIZE_LIST): New macro
12717 for the eew size list.
12718 (LMUL1_LOG2): New macro for the log2 value of lmul=1.
12719 (main): Add signed_eew*_lmul1_interpret for indexer.
12720 * config/riscv/riscv-vector-builtins-functions.def (vreinterpret):
12721 Register vint*m1_t interpret function.
12722 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_SIGNED_EEW8_LMUL1_INTERPRET_OPS):
12723 New macro for vint8m1_t.
12724 (DEF_RVV_SIGNED_EEW16_LMUL1_INTERPRET_OPS): Likewise.
12725 (DEF_RVV_SIGNED_EEW32_LMUL1_INTERPRET_OPS): Likewise.
12726 (DEF_RVV_SIGNED_EEW64_LMUL1_INTERPRET_OPS): Likewise.
12727 (vbool1_t): Add to signed_eew*_interpret_ops.
12728 (vbool2_t): Likewise.
12729 (vbool4_t): Likewise.
12730 (vbool8_t): Likewise.
12731 (vbool16_t): Likewise.
12732 (vbool32_t): Likewise.
12733 (vbool64_t): Likewise.
12734 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_SIGNED_EEW8_LMUL1_INTERPRET_OPS):
12735 New macro for vint*m1_t.
12736 (DEF_RVV_SIGNED_EEW16_LMUL1_INTERPRET_OPS): Likewise.
12737 (DEF_RVV_SIGNED_EEW32_LMUL1_INTERPRET_OPS): Likewise.
12738 (DEF_RVV_SIGNED_EEW64_LMUL1_INTERPRET_OPS): Likewise.
12739 (required_extensions_p): Add vint8m1_t interpret case.
12740 * config/riscv/riscv-vector-builtins.def (signed_eew8_lmul1_interpret):
12741 Add vint*m1_t interpret to base type.
12742 (signed_eew16_lmul1_interpret): Likewise.
12743 (signed_eew32_lmul1_interpret): Likewise.
12744 (signed_eew64_lmul1_interpret): Likewise.
12746 2023-05-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12748 * config/riscv/autovec.md: Adjust for new interface.
12749 * config/riscv/riscv-protos.h (emit_vlmax_insn): Add VL operand.
12750 (emit_nonvlmax_insn): Add AVL operand.
12751 * config/riscv/riscv-v.cc (emit_vlmax_insn): Add VL operand.
12752 (emit_nonvlmax_insn): Add AVL operand.
12753 (sew64_scalar_helper): Adjust for new interface.
12754 (expand_tuple_move): Ditto.
12755 * config/riscv/vector.md: Ditto.
12757 2023-05-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12759 * config/riscv/riscv-v.cc (expand_vec_series): Remove magic number.
12760 (expand_const_vector): Ditto.
12761 (legitimize_move): Ditto.
12762 (sew64_scalar_helper): Ditto.
12763 (expand_tuple_move): Ditto.
12764 (expand_vector_init_insert_elems): Ditto.
12765 * config/riscv/riscv.cc (vector_zero_call_used_regs): Ditto.
12767 2023-05-24 liuhongt <hongtao.liu@intel.com>
12770 * config/i386/i386.cc (ix86_gimple_fold_builtin): Fold
12771 _mm{,256,512}_abs_{epi8,epi16,epi32,epi64} and
12772 _mm_abs_{pi8,pi16,pi32} into gimple ABS_EXPR.
12773 (ix86_masked_all_ones): Handle 64-bit mask.
12774 * config/i386/i386-builtin.def: Replace icode of related
12775 non-mask simd abs builtins with CODE_FOR_nothing.
12777 2023-05-23 Martin Uecker <uecker@tugraz.at>
12780 * function.cc (gimplify_parm_type): Remove function.
12781 (gimplify_parameters): Call gimplify_type_sizes.
12783 2023-05-23 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12785 * config/xtensa/xtensa.md (*addsubx): Rename from '*addx',
12786 and change to also accept '*subx' pattern.
12789 2023-05-23 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12791 * config/xtensa/predicates.md (addsub_operator): New.
12792 * config/xtensa/xtensa.md (*extzvsi-1bit_ashlsi3,
12793 *extzvsi-1bit_addsubx): New insn_and_split patterns.
12794 * config/xtensa/xtensa.cc (xtensa_rtx_costs):
12795 Add a special case about ifcvt 'noce_try_cmove()' to handle
12796 constant loads that do not fit into signed 12 bits in the
12797 patterns added above.
12799 2023-05-23 Richard Biener <rguenther@suse.de>
12801 PR tree-optimization/109747
12802 * tree-vect-slp.cc (vect_prologue_cost_for_slp): Pass down
12803 the SLP node only once to the cost hook.
12805 2023-05-23 Georg-Johann Lay <avr@gjlay.de>
12807 * config/avr/avr.cc (avr_insn_cost): New static function.
12808 (TARGET_INSN_COST): Define to that function.
12810 2023-05-23 Richard Biener <rguenther@suse.de>
12813 * config/i386/i386.cc (ix86_vector_costs::add_stmt_cost):
12814 For vector construction or splats apply GPR->XMM move
12815 costing. QImode memory can be handled directly only
12816 with SSE4.1 pinsrb.
12818 2023-05-23 Richard Biener <rguenther@suse.de>
12820 PR tree-optimization/108752
12821 * tree-vect-stmts.cc (vectorizable_operation): For bit
12822 operations with generic word_mode vectors do not cost
12823 an extra stmt. For plus, minus and negate also cost the
12824 constant materialization.
12826 2023-05-23 Uros Bizjak <ubizjak@gmail.com>
12828 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi_partial):
12829 Call ix86_expand_vec_shift_qihi_constant for shifts
12830 with constant count operand.
12831 * config/i386/i386.cc (ix86_shift_rotate_cost):
12832 Handle V4QImode and V8QImode.
12833 * config/i386/mmx.md (<insn>v8qi3): New insn pattern.
12834 (<insn>v4qi3): Ditto.
12836 2023-05-23 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12838 * config/riscv/vector.md: Add mode.
12840 2023-05-23 Aldy Hernandez <aldyh@redhat.com>
12842 PR tree-optimization/109934
12843 * value-range.cc (irange::invert): Remove buggy special case.
12845 2023-05-23 Richard Biener <rguenther@suse.de>
12847 * tree-ssa-pre.cc (compute_antic_aux): Dump the correct
12850 2023-05-23 Richard Sandiford <richard.sandiford@arm.com>
12853 * config/aarch64/aarch64.cc (aarch64_modes_tieable_p): Allow
12854 subregs between any scalars that are 64 bits or smaller.
12855 * config/aarch64/iterators.md (SUBDI_BITS): New int iterator.
12856 (bits_etype): New int attribute.
12857 * config/aarch64/aarch64.md (*insv_reg<mode>_<SUBDI_BITS>)
12858 (*aarch64_bfi<GPI:mode><ALLX:mode>_<SUBDI_BITS>): New patterns.
12859 (*aarch64_bfidi<ALLX:mode>_subreg_<SUBDI_BITS>): Likewise.
12861 2023-05-23 Richard Sandiford <richard.sandiford@arm.com>
12863 * doc/md.texi: Document that <FOO> can be used to refer to the
12864 numerical value of an int iterator FOO. Tweak other parts of
12865 the int iterator documentation.
12866 * read-rtl.cc (iterator_group::has_self_attr): New field.
12867 (map_attr_string): When has_self_attr is true, make <FOO>
12868 expand to the current value of iterator FOO.
12869 (initialize_iterators): Set has_self_attr for int iterators.
12871 2023-05-23 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12873 * config/riscv/autovec.md: Refactor the framework of RVV auto-vectorization.
12874 * config/riscv/riscv-protos.h (RVV_MISC_OP_NUM): Ditto.
12875 (RVV_UNOP_NUM): New macro.
12876 (RVV_BINOP_NUM): Ditto.
12877 (legitimize_move): Refactor the framework of RVV auto-vectorization.
12878 (emit_vlmax_op): Ditto.
12879 (emit_vlmax_reg_op): Ditto.
12880 (emit_len_op): Ditto.
12881 (emit_len_binop): Ditto.
12882 (emit_vlmax_tany_many): Ditto.
12883 (emit_nonvlmax_tany_many): Ditto.
12884 (sew64_scalar_helper): Ditto.
12885 (expand_tuple_move): Ditto.
12886 * config/riscv/riscv-v.cc (emit_pred_op): Ditto.
12887 (emit_pred_binop): Ditto.
12888 (emit_vlmax_op): Ditto.
12889 (emit_vlmax_tany_many): New function.
12890 (emit_len_op): Remove.
12891 (emit_nonvlmax_tany_many): New function.
12892 (emit_vlmax_reg_op): Remove.
12893 (emit_len_binop): Ditto.
12894 (emit_index_op): Ditto.
12895 (expand_vec_series): Refactor the framework of RVV auto-vectorization.
12896 (expand_const_vector): Ditto.
12897 (legitimize_move): Ditto.
12898 (sew64_scalar_helper): Ditto.
12899 (expand_tuple_move): Ditto.
12900 (expand_vector_init_insert_elems): Ditto.
12901 * config/riscv/riscv.cc (vector_zero_call_used_regs): Ditto.
12902 * config/riscv/vector.md: Ditto.
12904 2023-05-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12907 * config/aarch64/aarch64-simd.md (add_vec_concat_subst_le): Add predicate
12908 and constraint for operand 0.
12909 (add_vec_concat_subst_be): Likewise.
12911 2023-05-23 Richard Biener <rguenther@suse.de>
12913 PR tree-optimization/109849
12914 * tree-ssa-pre.cc (do_hoist_insertion): Compute ANTIC_OUT
12915 and use that to determine what to hoist.
12917 2023-05-23 Eric Botcazou <ebotcazou@adacore.com>
12919 * fold-const.cc (native_encode_initializer) <CONSTRUCTOR>: Apply the
12920 specific treatment for bit-fields only if they have an integral type
12921 and filter out non-integral bit-fields that do not start and end on
12924 2023-05-23 Aldy Hernandez <aldyh@redhat.com>
12926 PR tree-optimization/109920
12927 * value-range.h (RESIZABLE>::~int_range): Use delete[].
12929 2023-05-22 Uros Bizjak <ubizjak@gmail.com>
12931 * config/i386/i386.cc (ix86_shift_rotate_cost): Correct
12932 calcuation of integer vector mode costs to reflect generated
12933 instruction sequences of different integer vector modes and
12934 different target ABIs. Remove "speed" function argument.
12935 (ix86_rtx_costs): Update call for removed function argument.
12936 (ix86_vector_costs::add_stmt_cost): Ditto.
12938 2023-05-22 Aldy Hernandez <aldyh@redhat.com>
12940 * value-range.h (class Value_Range): Implement set_zero,
12941 set_nonzero, and nonzero_p.
12943 2023-05-22 Uros Bizjak <ubizjak@gmail.com>
12945 * config/i386/i386.cc (ix86_multiplication_cost): Add
12946 the cost of a memory read to the cost of V?QImode sequences.
12948 2023-05-22 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12950 * config/riscv/riscv-v.cc: Add "m_" prefix.
12952 2023-05-22 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
12954 * tree-vect-loop.cc (vect_get_loop_len): Fix issue for
12955 multiple-rgroup of length.
12956 * tree-vect-stmts.cc (vectorizable_store): Ditto.
12957 (vectorizable_load): Ditto.
12958 * tree-vectorizer.h (vect_get_loop_len): Ditto.
12960 2023-05-22 Juzhe-Zhong <juzhe.zhong@rivai.ai>
12962 * config/riscv/riscv.cc (riscv_const_insns): Reorganize the
12965 2023-05-22 Kewen Lin <linkw@linux.ibm.com>
12967 * tree-vect-slp.cc (vect_transform_slp_perm_load_1): Refactor the
12968 handling for the case index == count.
12970 2023-05-21 Georg-Johann Lay <avr@gjlay.de>
12973 * config/avr/avr.cc (avr_fold_builtin) [AVR_BUILTIN_INSERT_BITS]:
12974 Don't fold to XOR / AND / XOR if just one bit is copied to the
12977 2023-05-21 Roger Sayle <roger@nextmovesoftware.com>
12979 * config/nvptx/nvptx.cc (nvptx_expand_brev): Expand target
12980 builtin for bit reversal using brev instruction.
12981 (enum nvptx_builtins): Add NVPTX_BUILTIN_BREV and
12982 NVPTX_BUILTIN_BREVLL.
12983 (nvptx_init_builtins): Define "brev" and "brevll".
12984 (nvptx_expand_builtin): Expand NVPTX_BUILTIN_BREV and
12985 NVPTX_BUILTIN_BREVLL via nvptx_expand_brev function.
12986 * doc/extend.texi (Nvidia PTX Builtin-in Functions): New
12987 section, document __builtin_nvptx_brev{,ll}.
12989 2023-05-21 Jakub Jelinek <jakub@redhat.com>
12991 PR tree-optimization/109505
12992 * match.pd ((x | CST1) & CST2 -> (x & CST2) | (CST1 & CST2),
12993 Combine successive equal operations with constants,
12994 (A +- CST1) +- CST2 -> A + CST3, (CST1 - A) +- CST2 -> CST3 - A,
12995 CST1 - (CST2 - A) -> CST3 + A): Use ! on ops with 2 CONSTANT_CLASS_P
12998 2023-05-21 Andrew Pinski <apinski@marvell.com>
13000 * expr.cc (expand_single_bit_test): Correct bitpos for big-endian.
13002 2023-05-21 Pan Li <pan2.li@intel.com>
13004 * config/riscv/genrvv-type-indexer.cc (BOOL_SIZE_LIST): Add the
13005 rest bool size, aka 2, 4, 8, 16, 32, 64.
13006 * config/riscv/riscv-vector-builtins-functions.def (vreinterpret):
13007 Register vbool[2|4|8|16|32|64] interpret function.
13008 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_BOOL2_INTERPRET_OPS):
13009 New macro for vbool2_t.
13010 (DEF_RVV_BOOL4_INTERPRET_OPS): Likewise.
13011 (DEF_RVV_BOOL8_INTERPRET_OPS): Likewise.
13012 (DEF_RVV_BOOL16_INTERPRET_OPS): Likewise.
13013 (DEF_RVV_BOOL32_INTERPRET_OPS): Likewise.
13014 (DEF_RVV_BOOL64_INTERPRET_OPS): Likewise.
13015 (vint8m1_t): Add the type to bool[2|4|8|16|32|64]_interpret_ops.
13016 (vint16m1_t): Likewise.
13017 (vint32m1_t): Likewise.
13018 (vint64m1_t): Likewise.
13019 (vuint8m1_t): Likewise.
13020 (vuint16m1_t): Likewise.
13021 (vuint32m1_t): Likewise.
13022 (vuint64m1_t): Likewise.
13023 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_BOOL2_INTERPRET_OPS):
13024 New macro for vbool2_t.
13025 (DEF_RVV_BOOL4_INTERPRET_OPS): Likewise.
13026 (DEF_RVV_BOOL8_INTERPRET_OPS): Likewise.
13027 (DEF_RVV_BOOL16_INTERPRET_OPS): Likewise.
13028 (DEF_RVV_BOOL32_INTERPRET_OPS): Likewise.
13029 (DEF_RVV_BOOL64_INTERPRET_OPS): Likewise.
13030 (required_extensions_p): Add vbool[2|4|8|16|32|64] interpret case.
13031 * config/riscv/riscv-vector-builtins.def (bool2_interpret): Add
13032 vbool2_t interprect to base type.
13033 (bool4_interpret): Likewise.
13034 (bool8_interpret): Likewise.
13035 (bool16_interpret): Likewise.
13036 (bool32_interpret): Likewise.
13037 (bool64_interpret): Likewise.
13039 2023-05-21 Andrew Pinski <apinski@marvell.com>
13041 PR middle-end/109919
13042 * expr.cc (expand_single_bit_test): Don't use the
13043 target for expand_expr.
13045 2023-05-20 Gerald Pfeifer <gerald@pfeifer.com>
13047 * doc/install.texi (Specific): Remove de facto empty alpha*-*-*
13050 2023-05-20 Pan Li <pan2.li@intel.com>
13052 * mode-switching.cc (entity_map): Initialize the array to zero.
13055 2023-05-20 Triffid Hunter <triffid.hunter@gmail.com>
13058 * config/avr/avr.md (divmodpsi, udivmodpsi, divmodsi, udivmodsi):
13059 Remove superfluous "parallel" in insn pattern.
13060 ([u]divmod<mode>4): Tidy code. Use gcc_unreachable() instead of
13061 printing error text to assembly.
13063 2023-05-20 Andrew Pinski <apinski@marvell.com>
13065 * expr.cc (fold_single_bit_test): Rename to ...
13066 (expand_single_bit_test): This and expand directly.
13067 (do_store_flag): Update for the rename function.
13069 2023-05-20 Andrew Pinski <apinski@marvell.com>
13071 * expr.cc (fold_single_bit_test): Use BIT_FIELD_REF
13072 instead of shift/and.
13074 2023-05-20 Andrew Pinski <apinski@marvell.com>
13076 * expr.cc (fold_single_bit_test): Add an assert
13077 and simplify based on code being NE_EXPR or EQ_EXPR.
13079 2023-05-20 Andrew Pinski <apinski@marvell.com>
13081 * expr.cc (fold_single_bit_test): Take inner and bitnum
13082 instead of arg0 and arg1. Update the code.
13083 (do_store_flag): Don't create a tree when calling
13084 fold_single_bit_test instead just call it with the bitnum
13085 and the inner tree.
13087 2023-05-20 Andrew Pinski <apinski@marvell.com>
13089 * expr.cc (fold_single_bit_test): Use get_def_for_expr
13090 instead of checking the inner's code.
13092 2023-05-20 Andrew Pinski <apinski@marvell.com>
13094 * expr.cc (fold_single_bit_test_into_sign_test): Inline into ...
13095 (fold_single_bit_test): This and simplify.
13097 2023-05-20 Andrew Pinski <apinski@marvell.com>
13099 * fold-const.cc (fold_single_bit_test_into_sign_test): Move to
13101 (fold_single_bit_test): Likewise.
13102 * expr.cc (fold_single_bit_test_into_sign_test): Move from fold-const.cc
13103 (fold_single_bit_test): Likewise and make static.
13104 * fold-const.h (fold_single_bit_test): Remove declaration.
13106 2023-05-20 Die Li <lidie@eswincomputing.com>
13108 * config/riscv/riscv.cc (riscv_expand_conditional_move): Fix mode
13111 2023-05-20 Raphael Moreira Zinsly <rzinsly@ventanamicro.com>
13113 * config/riscv/bitmanip.md (branch<X:mode>_bext): New split pattern.
13115 2023-05-20 Raphael Moreira Zinsly <rzinsly@ventanamicro.com>
13118 * config/riscv/bitmanip.md
13119 (<bitmanip_optab>disi2): Match with any_extend.
13120 (<bitmanip_optab>disi2_sext): New pattern to match
13121 with sign extend using an ANDI instruction.
13123 2023-05-19 Nathan Sidwell <nathan@acm.org>
13126 * opts.h (handle_deferred_dump_options): Declare.
13127 * opts-global.cc (handle_common_deferred_options): Do not handle
13129 (handle_deferred_dump_options): New.
13130 * toplev.cc (toplev::main): Call it after plugin init.
13132 2023-05-19 Joern Rennecke <joern.rennecke@embecosm.com>
13134 * config/riscv/constraints.md (DsS, DsD): Restore agreement
13135 with shiftm1 mode attribute.
13137 2023-05-19 Andrew Pinski <apinski@marvell.com>
13140 * gcc.cc (default_compilers["@c-header"]): Add %w
13141 after the --output-pch.
13143 2023-05-19 Vineet Gupta <vineetg@rivosinc.com>
13145 * config/riscv/riscv.cc (riscv_split_integer): if loval is equal
13146 to hival, ASHIFT the corresponding regs.
13148 2023-05-19 Robin Dapp <rdapp@ventanamicro.com>
13150 * config/riscv/riscv.cc (riscv_const_insns): Remove else.
13152 2023-05-19 Jakub Jelinek <jakub@redhat.com>
13154 PR tree-optimization/105776
13155 * tree-ssa-math-opts.cc (arith_overflow_check_p): If cast_stmt is
13156 non-NULL, allow division statement to have a cast as single imm use
13157 rather than comparison/condition.
13158 (match_arith_overflow): In that case remove the cast stmt in addition
13159 to the division statement.
13161 2023-05-19 Jakub Jelinek <jakub@redhat.com>
13163 PR tree-optimization/101856
13164 * tree-ssa-math-opts.cc (match_arith_overflow): Pattern detect
13165 unsigned __builtin_mul_overflow_p even when umulv4_optab doesn't
13166 support it but umul_highpart_optab does.
13168 2023-05-19 Eric Botcazou <ebotcazou@adacore.com>
13170 * varasm.cc (output_constructor_bitfield): Call tree_to_uhwi instead
13171 of tree_to_shwi on array indices. Minor tweaks.
13173 2023-05-18 Bernhard Reutner-Fischer <aldot@gcc.gnu.org>
13175 * alias.cc (ref_all_alias_ptr_type_p): Use _P() defines from tree.h.
13176 * attribs.cc (diag_attr_exclusions): Ditto.
13177 (decl_attributes): Ditto.
13178 (build_type_attribute_qual_variant): Ditto.
13179 * builtins.cc (fold_builtin_carg): Ditto.
13180 (fold_builtin_next_arg): Ditto.
13181 (do_mpc_arg2): Ditto.
13182 * cfgexpand.cc (expand_return): Ditto.
13183 * cgraph.h (decl_in_symtab_p): Ditto.
13184 (symtab_node::get_create): Ditto.
13185 * dwarf2out.cc (base_type_die): Ditto.
13186 (implicit_ptr_descriptor): Ditto.
13187 (gen_array_type_die): Ditto.
13188 (gen_type_die_with_usage): Ditto.
13189 (optimize_location_into_implicit_ptr): Ditto.
13190 * expr.cc (do_store_flag): Ditto.
13191 * fold-const.cc (negate_expr_p): Ditto.
13192 (fold_negate_expr_1): Ditto.
13193 (fold_convert_const): Ditto.
13194 (fold_convert_loc): Ditto.
13195 (constant_boolean_node): Ditto.
13196 (fold_binary_op_with_conditional_arg): Ditto.
13197 (build_fold_addr_expr_with_type_loc): Ditto.
13198 (fold_comparison): Ditto.
13199 (fold_checksum_tree): Ditto.
13200 (tree_unary_nonnegative_warnv_p): Ditto.
13201 (integer_valued_real_unary_p): Ditto.
13202 (fold_read_from_constant_string): Ditto.
13203 * gcc-rich-location.cc (maybe_range_label_for_tree_type_mismatch::get_text): Ditto.
13204 * gimple-expr.cc (useless_type_conversion_p): Ditto.
13205 (is_gimple_reg): Ditto.
13206 (is_gimple_asm_val): Ditto.
13207 (mark_addressable): Ditto.
13208 * gimple-expr.h (is_gimple_variable): Ditto.
13209 (virtual_operand_p): Ditto.
13210 * gimple-ssa-warn-access.cc (pass_waccess::check_dangling_stores): Ditto.
13211 * gimplify.cc (gimplify_bind_expr): Ditto.
13212 (gimplify_return_expr): Ditto.
13213 (gimple_add_padding_init_for_auto_var): Ditto.
13214 (gimplify_addr_expr): Ditto.
13215 (omp_add_variable): Ditto.
13216 (omp_notice_variable): Ditto.
13217 (omp_get_base_pointer): Ditto.
13218 (omp_strip_components_and_deref): Ditto.
13219 (omp_strip_indirections): Ditto.
13220 (omp_accumulate_sibling_list): Ditto.
13221 (omp_build_struct_sibling_lists): Ditto.
13222 (gimplify_adjust_omp_clauses_1): Ditto.
13223 (gimplify_adjust_omp_clauses): Ditto.
13224 (gimplify_omp_for): Ditto.
13225 (goa_lhs_expr_p): Ditto.
13226 (gimplify_one_sizepos): Ditto.
13227 * graphite-scop-detection.cc (scop_detection::graphite_can_represent_scev): Ditto.
13228 * ipa-devirt.cc (odr_types_equivalent_p): Ditto.
13229 * ipa-prop.cc (ipa_set_jf_constant): Ditto.
13230 (propagate_controlled_uses): Ditto.
13231 * ipa-sra.cc (type_prevails_p): Ditto.
13232 (scan_expr_access): Ditto.
13233 * optabs-tree.cc (optab_for_tree_code): Ditto.
13234 * toplev.cc (wrapup_global_declaration_1): Ditto.
13235 * trans-mem.cc (transaction_invariant_address_p): Ditto.
13236 * tree-cfg.cc (verify_types_in_gimple_reference): Ditto.
13237 (verify_gimple_comparison): Ditto.
13238 (verify_gimple_assign_binary): Ditto.
13239 (verify_gimple_assign_single): Ditto.
13240 * tree-complex.cc (get_component_ssa_name): Ditto.
13241 * tree-emutls.cc (lower_emutls_2): Ditto.
13242 * tree-inline.cc (copy_tree_body_r): Ditto.
13243 (estimate_move_cost): Ditto.
13244 (copy_decl_for_dup_finish): Ditto.
13245 * tree-nested.cc (convert_nonlocal_omp_clauses): Ditto.
13246 (note_nonlocal_vla_type): Ditto.
13247 (convert_local_omp_clauses): Ditto.
13248 (remap_vla_decls): Ditto.
13249 (fixup_vla_decls): Ditto.
13250 * tree-parloops.cc (loop_has_vector_phi_nodes): Ditto.
13251 * tree-pretty-print.cc (print_declaration): Ditto.
13252 (print_call_name): Ditto.
13253 * tree-sra.cc (compare_access_positions): Ditto.
13254 * tree-ssa-alias.cc (compare_type_sizes): Ditto.
13255 * tree-ssa-ccp.cc (get_default_value): Ditto.
13256 * tree-ssa-coalesce.cc (populate_coalesce_list_for_outofssa): Ditto.
13257 * tree-ssa-dom.cc (reduce_vector_comparison_to_scalar_comparison): Ditto.
13258 * tree-ssa-forwprop.cc (can_propagate_from): Ditto.
13259 * tree-ssa-propagate.cc (may_propagate_copy): Ditto.
13260 * tree-ssa-sccvn.cc (fully_constant_vn_reference_p): Ditto.
13261 * tree-ssa-sink.cc (statement_sink_location): Ditto.
13262 * tree-ssa-structalias.cc (type_must_have_pointers): Ditto.
13263 * tree-ssa-ter.cc (find_replaceable_in_bb): Ditto.
13264 * tree-ssa-uninit.cc (warn_uninit): Ditto.
13265 * tree-ssa.cc (maybe_rewrite_mem_ref_base): Ditto.
13266 (non_rewritable_mem_ref_base): Ditto.
13267 * tree-streamer-in.cc (lto_input_ts_type_non_common_tree_pointers): Ditto.
13268 * tree-streamer-out.cc (write_ts_type_non_common_tree_pointers): Ditto.
13269 * tree-vect-generic.cc (do_binop): Ditto.
13271 * tree-vect-stmts.cc (vect_init_vector): Ditto.
13272 * tree-vector-builder.h (tree_vector_builder::note_representative): Ditto.
13273 * tree.cc (sign_mask_for): Ditto.
13274 (verify_type_variant): Ditto.
13275 (gimple_canonical_types_compatible_p): Ditto.
13276 (verify_type): Ditto.
13277 * ubsan.cc (get_ubsan_type_info_for_type): Ditto.
13278 * var-tracking.cc (prepare_call_arguments): Ditto.
13279 (vt_add_function_parameters): Ditto.
13280 * varasm.cc (decode_addr_const): Ditto.
13282 2023-05-18 Bernhard Reutner-Fischer <aldot@gcc.gnu.org>
13284 * omp-low.cc (scan_sharing_clauses): Use _P() defines from tree.h.
13285 (lower_reduction_clauses): Ditto.
13286 (lower_send_clauses): Ditto.
13287 (lower_omp_task_reductions): Ditto.
13288 * omp-oacc-neuter-broadcast.cc (install_var_field): Ditto.
13289 (worker_single_copy): Ditto.
13290 * omp-offload.cc (oacc_rewrite_var_decl): Ditto.
13291 * omp-simd-clone.cc (plausible_type_for_simd_clone): Ditto.
13293 2023-05-18 Bernhard Reutner-Fischer <aldot@gcc.gnu.org>
13295 * lto-streamer-in.cc (lto_input_var_decl_ref): Use _P defines from
13297 (lto_read_body_or_constructor): Ditto.
13298 * lto-streamer-out.cc (tree_is_indexable): Ditto.
13299 (lto_output_var_decl_ref): Ditto.
13300 (DFS::DFS_write_tree_body): Ditto.
13301 (wrap_refs): Ditto.
13302 (write_symbol_extension_info): Ditto.
13304 2023-05-18 Bernhard Reutner-Fischer <aldot@gcc.gnu.org>
13306 * config/aarch64/aarch64.cc (aarch64_short_vector_p): Use _P
13307 defines from tree.h.
13308 (aarch64_mangle_type): Ditto.
13309 * config/alpha/alpha.cc (alpha_in_small_data_p): Ditto.
13310 (alpha_gimplify_va_arg_1): Ditto.
13311 * config/arc/arc.cc (arc_encode_section_info): Ditto.
13312 (arc_is_aux_reg_p): Ditto.
13313 (arc_is_uncached_mem_p): Ditto.
13314 (arc_handle_aux_attribute): Ditto.
13315 * config/arm/arm.cc (arm_handle_isr_attribute): Ditto.
13316 (arm_handle_cmse_nonsecure_call): Ditto.
13317 (arm_set_default_type_attributes): Ditto.
13318 (arm_is_segment_info_known): Ditto.
13319 (arm_mangle_type): Ditto.
13320 * config/arm/unknown-elf.h (IN_NAMED_SECTION_P): Ditto.
13321 * config/avr/avr.cc (avr_lookup_function_attribute1): Ditto.
13322 (avr_decl_absdata_p): Ditto.
13323 (avr_insert_attributes): Ditto.
13324 (avr_section_type_flags): Ditto.
13325 (avr_encode_section_info): Ditto.
13326 * config/bfin/bfin.cc (bfin_handle_l2_attribute): Ditto.
13327 * config/bpf/bpf.cc (bpf_core_compute): Ditto.
13328 * config/c6x/c6x.cc (c6x_in_small_data_p): Ditto.
13329 * config/csky/csky.cc (csky_handle_isr_attribute): Ditto.
13330 (csky_mangle_type): Ditto.
13331 * config/darwin-c.cc (darwin_pragma_unused): Ditto.
13332 * config/darwin.cc (is_objc_metadata): Ditto.
13333 * config/epiphany/epiphany.cc (epiphany_function_ok_for_sibcall): Ditto.
13334 * config/epiphany/epiphany.h (ROUND_TYPE_ALIGN): Ditto.
13335 * config/frv/frv.cc (frv_emit_movsi): Ditto.
13336 * config/gcn/gcn-tree.cc (gcn_lockless_update): Ditto.
13337 * config/gcn/gcn.cc (gcn_asm_output_symbol_ref): Ditto.
13338 * config/h8300/h8300.cc (h8300_encode_section_info): Ditto.
13339 * config/i386/i386-expand.cc: Ditto.
13340 * config/i386/i386.cc (type_natural_mode): Ditto.
13341 (ix86_function_arg): Ditto.
13342 (ix86_data_alignment): Ditto.
13343 (ix86_local_alignment): Ditto.
13344 (ix86_simd_clone_compute_vecsize_and_simdlen): Ditto.
13345 * config/i386/winnt-cxx.cc (i386_pe_type_dllimport_p): Ditto.
13346 (i386_pe_type_dllexport_p): Ditto.
13347 (i386_pe_adjust_class_at_definition): Ditto.
13348 * config/i386/winnt.cc (i386_pe_determine_dllimport_p): Ditto.
13349 (i386_pe_binds_local_p): Ditto.
13350 (i386_pe_section_type_flags): Ditto.
13351 * config/ia64/ia64.cc (ia64_encode_section_info): Ditto.
13352 (ia64_gimplify_va_arg): Ditto.
13353 (ia64_in_small_data_p): Ditto.
13354 * config/iq2000/iq2000.cc (iq2000_function_arg): Ditto.
13355 * config/lm32/lm32.cc (lm32_in_small_data_p): Ditto.
13356 * config/loongarch/loongarch.cc (loongarch_handle_model_attribute): Ditto.
13357 * config/m32c/m32c.cc (m32c_insert_attributes): Ditto.
13358 * config/mcore/mcore.cc (mcore_mark_dllimport): Ditto.
13359 (mcore_encode_section_info): Ditto.
13360 * config/microblaze/microblaze.cc (microblaze_elf_in_small_data_p): Ditto.
13361 * config/mips/mips.cc (mips_output_aligned_decl_common): Ditto.
13362 * config/mmix/mmix.cc (mmix_encode_section_info): Ditto.
13363 * config/nvptx/nvptx.cc (nvptx_encode_section_info): Ditto.
13364 (pass_in_memory): Ditto.
13365 (nvptx_generate_vector_shuffle): Ditto.
13366 (nvptx_lockless_update): Ditto.
13367 * config/pa/pa.cc (pa_function_arg_padding): Ditto.
13368 (pa_function_value): Ditto.
13369 (pa_function_arg): Ditto.
13370 * config/pa/pa.h (IN_NAMED_SECTION_P): Ditto.
13371 (TEXT_SPACE_P): Ditto.
13372 * config/pa/som.h (MAKE_DECL_ONE_ONLY): Ditto.
13373 * config/pdp11/pdp11.cc (pdp11_return_in_memory): Ditto.
13374 * config/riscv/riscv.cc (riscv_in_small_data_p): Ditto.
13375 (riscv_mangle_type): Ditto.
13376 * config/rl78/rl78.cc (rl78_insert_attributes): Ditto.
13377 (rl78_addsi3_internal): Ditto.
13378 * config/rs6000/aix.h (ROUND_TYPE_ALIGN): Ditto.
13379 * config/rs6000/darwin.h (ROUND_TYPE_ALIGN): Ditto.
13380 * config/rs6000/freebsd64.h (ROUND_TYPE_ALIGN): Ditto.
13381 * config/rs6000/linux64.h (ROUND_TYPE_ALIGN): Ditto.
13382 * config/rs6000/rs6000-call.cc (rs6000_function_arg_boundary): Ditto.
13383 (rs6000_function_arg_advance_1): Ditto.
13384 (rs6000_function_arg): Ditto.
13385 (rs6000_pass_by_reference): Ditto.
13386 * config/rs6000/rs6000-logue.cc (rs6000_function_ok_for_sibcall): Ditto.
13387 * config/rs6000/rs6000.cc (rs6000_data_alignment): Ditto.
13388 (rs6000_set_default_type_attributes): Ditto.
13389 (rs6000_elf_in_small_data_p): Ditto.
13390 (IN_NAMED_SECTION): Ditto.
13391 (rs6000_xcoff_encode_section_info): Ditto.
13392 (rs6000_function_value): Ditto.
13393 (invalid_arg_for_unprototyped_fn): Ditto.
13394 * config/s390/s390-c.cc (s390_fn_types_compatible): Ditto.
13395 (s390_vec_n_elem): Ditto.
13396 * config/s390/s390.cc (s390_check_type_for_vector_abi): Ditto.
13397 (s390_function_arg_integer): Ditto.
13398 (s390_return_in_memory): Ditto.
13399 (s390_encode_section_info): Ditto.
13400 * config/sh/sh.cc (sh_gimplify_va_arg_expr): Ditto.
13401 (sh_function_value): Ditto.
13402 * config/sol2.cc (solaris_insert_attributes): Ditto.
13403 * config/sparc/sparc.cc (function_arg_slotno): Ditto.
13404 * config/sparc/sparc.h (ROUND_TYPE_ALIGN): Ditto.
13405 * config/stormy16/stormy16.cc (xstormy16_encode_section_info): Ditto.
13406 (xstormy16_handle_below100_attribute): Ditto.
13407 * config/v850/v850.cc (v850_encode_section_info): Ditto.
13408 (v850_insert_attributes): Ditto.
13409 * config/visium/visium.cc (visium_pass_by_reference): Ditto.
13410 (visium_return_in_memory): Ditto.
13411 * config/xtensa/xtensa.cc (xtensa_multibss_section_type_flags): Ditto.
13413 2023-05-18 Uros Bizjak <ubizjak@gmail.com>
13415 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi_partial): New.
13416 (ix86_expand_vecop_qihi): Add op2vec bool variable.
13417 Do not set REG_EQUAL note.
13418 * config/i386/i386-protos.h (ix86_expand_vecop_qihi_partial):
13420 * config/i386/i386.cc (ix86_multiplication_cost): Handle
13421 V4QImode and V8QImode.
13422 * config/i386/mmx.md (mulv8qi3): New expander.
13424 * config/i386/sse.md (mulv8qi3): Remove.
13426 2023-05-18 Georg-Johann Lay <avr@gjlay.de>
13428 * config/avr/gen-avr-mmcu-specs.cc: Remove stale */ after // comment.
13430 2023-05-18 Jonathan Wakely <jwakely@redhat.com>
13432 PR bootstrap/105831
13433 * config.gcc: Use = operator instead of ==.
13435 2023-05-18 Michael Bäuerle <micha@NetBSD.org>
13437 PR bootstrap/105831
13438 * config/nvptx/gen-opt.sh: Use = operator instead of ==.
13439 * configure.ac: Likewise.
13440 * configure: Regenerate.
13442 2023-05-18 Stam Markianos-Wright <stam.markianos-wright@arm.com>
13444 * config/arm/arm_mve.h: (__ARM_mve_typeid): Add more pointer types.
13445 (__ARM_mve_coerce1): Remove.
13446 (__ARM_mve_coerce2): Remove.
13447 (__ARM_mve_coerce3): Remove.
13448 (__ARM_mve_coerce_i_scalar): New.
13449 (__ARM_mve_coerce_s8_ptr): New.
13450 (__ARM_mve_coerce_u8_ptr): New.
13451 (__ARM_mve_coerce_s16_ptr): New.
13452 (__ARM_mve_coerce_u16_ptr): New.
13453 (__ARM_mve_coerce_s32_ptr): New.
13454 (__ARM_mve_coerce_u32_ptr): New.
13455 (__ARM_mve_coerce_s64_ptr): New.
13456 (__ARM_mve_coerce_u64_ptr): New.
13457 (__ARM_mve_coerce_f_scalar): New.
13458 (__ARM_mve_coerce_f16_ptr): New.
13459 (__ARM_mve_coerce_f32_ptr): New.
13460 (__arm_vst4q): Change _coerce_ overloads.
13461 (__arm_vbicq): Change _coerce_ overloads.
13462 (__arm_vld1q): Change _coerce_ overloads.
13463 (__arm_vld1q_z): Change _coerce_ overloads.
13464 (__arm_vld2q): Change _coerce_ overloads.
13465 (__arm_vld4q): Change _coerce_ overloads.
13466 (__arm_vldrhq_gather_offset): Change _coerce_ overloads.
13467 (__arm_vldrhq_gather_offset_z): Change _coerce_ overloads.
13468 (__arm_vldrhq_gather_shifted_offset): Change _coerce_ overloads.
13469 (__arm_vldrhq_gather_shifted_offset_z): Change _coerce_ overloads.
13470 (__arm_vldrwq_gather_offset): Change _coerce_ overloads.
13471 (__arm_vldrwq_gather_offset_z): Change _coerce_ overloads.
13472 (__arm_vldrwq_gather_shifted_offset): Change _coerce_ overloads.
13473 (__arm_vldrwq_gather_shifted_offset_z): Change _coerce_ overloads.
13474 (__arm_vst1q_p): Change _coerce_ overloads.
13475 (__arm_vst2q): Change _coerce_ overloads.
13476 (__arm_vst1q): Change _coerce_ overloads.
13477 (__arm_vstrhq): Change _coerce_ overloads.
13478 (__arm_vstrhq_p): Change _coerce_ overloads.
13479 (__arm_vstrhq_scatter_offset_p): Change _coerce_ overloads.
13480 (__arm_vstrhq_scatter_offset): Change _coerce_ overloads.
13481 (__arm_vstrhq_scatter_shifted_offset_p): Change _coerce_ overloads.
13482 (__arm_vstrhq_scatter_shifted_offset): Change _coerce_ overloads.
13483 (__arm_vstrwq_p): Change _coerce_ overloads.
13484 (__arm_vstrwq): Change _coerce_ overloads.
13485 (__arm_vstrwq_scatter_offset): Change _coerce_ overloads.
13486 (__arm_vstrwq_scatter_offset_p): Change _coerce_ overloads.
13487 (__arm_vstrwq_scatter_shifted_offset): Change _coerce_ overloads.
13488 (__arm_vstrwq_scatter_shifted_offset_p): Change _coerce_ overloads.
13489 (__arm_vsetq_lane): Change _coerce_ overloads.
13490 (__arm_vldrbq_gather_offset): Change _coerce_ overloads.
13491 (__arm_vdwdupq_x_u8): Change _coerce_ overloads.
13492 (__arm_vdwdupq_x_u16): Change _coerce_ overloads.
13493 (__arm_vdwdupq_x_u32): Change _coerce_ overloads.
13494 (__arm_viwdupq_x_u8): Change _coerce_ overloads.
13495 (__arm_viwdupq_x_u16): Change _coerce_ overloads.
13496 (__arm_viwdupq_x_u32): Change _coerce_ overloads.
13497 (__arm_vidupq_x_u8): Change _coerce_ overloads.
13498 (__arm_vddupq_x_u8): Change _coerce_ overloads.
13499 (__arm_vidupq_x_u16): Change _coerce_ overloads.
13500 (__arm_vddupq_x_u16): Change _coerce_ overloads.
13501 (__arm_vidupq_x_u32): Change _coerce_ overloads.
13502 (__arm_vddupq_x_u32): Change _coerce_ overloads.
13503 (__arm_vldrdq_gather_offset): Change _coerce_ overloads.
13504 (__arm_vldrdq_gather_offset_z): Change _coerce_ overloads.
13505 (__arm_vldrdq_gather_shifted_offset): Change _coerce_ overloads.
13506 (__arm_vldrdq_gather_shifted_offset_z): Change _coerce_ overloads.
13507 (__arm_vldrbq_gather_offset_z): Change _coerce_ overloads.
13508 (__arm_vidupq_u16): Change _coerce_ overloads.
13509 (__arm_vidupq_u32): Change _coerce_ overloads.
13510 (__arm_vidupq_u8): Change _coerce_ overloads.
13511 (__arm_vddupq_u16): Change _coerce_ overloads.
13512 (__arm_vddupq_u32): Change _coerce_ overloads.
13513 (__arm_vddupq_u8): Change _coerce_ overloads.
13514 (__arm_viwdupq_m): Change _coerce_ overloads.
13515 (__arm_viwdupq_u16): Change _coerce_ overloads.
13516 (__arm_viwdupq_u32): Change _coerce_ overloads.
13517 (__arm_viwdupq_u8): Change _coerce_ overloads.
13518 (__arm_vdwdupq_m): Change _coerce_ overloads.
13519 (__arm_vdwdupq_u16): Change _coerce_ overloads.
13520 (__arm_vdwdupq_u32): Change _coerce_ overloads.
13521 (__arm_vdwdupq_u8): Change _coerce_ overloads.
13522 (__arm_vstrbq): Change _coerce_ overloads.
13523 (__arm_vstrbq_p): Change _coerce_ overloads.
13524 (__arm_vstrbq_scatter_offset_p): Change _coerce_ overloads.
13525 (__arm_vstrdq_scatter_offset_p): Change _coerce_ overloads.
13526 (__arm_vstrdq_scatter_offset): Change _coerce_ overloads.
13527 (__arm_vstrdq_scatter_shifted_offset_p): Change _coerce_ overloads.
13528 (__arm_vstrdq_scatter_shifted_offset): Change _coerce_ overloads.
13530 2023-05-18 Stam Markianos-Wright <stam.markianos-wright@arm.com>
13532 * config/arm/arm_mve.h (__arm_vbicq): Change coerce on
13535 2023-05-18 Stam Markianos-Wright <stam.markianos-wright@arm.com>
13537 * config/arm/arm_mve.h (__arm_vadcq_s32): Fix arithmetic.
13538 (__arm_vadcq_u32): Likewise.
13539 (__arm_vadcq_m_s32): Likewise.
13540 (__arm_vadcq_m_u32): Likewise.
13541 (__arm_vsbcq_s32): Likewise.
13542 (__arm_vsbcq_u32): Likewise.
13543 (__arm_vsbcq_m_s32): Likewise.
13544 (__arm_vsbcq_m_u32): Likewise.
13545 * config/arm/mve.md (get_fpscr_nzcvqc): Make unspec_volatile.
13547 2023-05-18 Andrea Corallo <andrea.corallo@arm.com>
13549 * config/arm/mve.md (mve_vrndq_m_f<mode>, mve_vrev64q_f<mode>)
13550 (mve_vrev32q_fv8hf, mve_vcvttq_f32_f16v4sf)
13551 (mve_vcvtbq_f32_f16v4sf, mve_vcvtq_to_f_<supf><mode>)
13552 (mve_vrev64q_<supf><mode>, mve_vcvtq_from_f_<supf><mode>)
13553 (mve_vmovltq_<supf><mode>, mve_vmovlbq_<supf><mode>)
13554 (mve_vcvtpq_<supf><mode>, mve_vcvtnq_<supf><mode>)
13555 (mve_vcvtmq_<supf><mode>, mve_vcvtaq_<supf><mode>)
13556 (mve_vmvnq_n_<supf><mode>, mve_vrev16q_<supf>v16qi)
13557 (mve_vctp<MVE_vctp>q<MVE_vpred>, mve_vbrsrq_n_f<mode>)
13558 (mve_vbrsrq_n_<supf><mode>, mve_vandq_f<mode>, mve_vbicq_f<mode>)
13559 (mve_vctp<MVE_vctp>q_m<MVE_vpred>, mve_vcvtbq_f16_f32v8hf)
13560 (mve_vcvttq_f16_f32v8hf, mve_veorq_f<mode>)
13561 (mve_vmlaldavxq_s<mode>, mve_vmlsldavq_s<mode>)
13562 (mve_vmlsldavxq_s<mode>, mve_vornq_f<mode>, mve_vorrq_f<mode>)
13563 (mve_vrmlaldavhxq_sv4si, mve_vcvtq_m_to_f_<supf><mode>)
13564 (mve_vshlcq_<supf><mode>, mve_vmvnq_m_<supf><mode>)
13565 (mve_vpselq_<supf><mode>, mve_vcvtbq_m_f16_f32v8hf)
13566 (mve_vcvtbq_m_f32_f16v4sf, mve_vcvttq_m_f16_f32v8hf)
13567 (mve_vcvttq_m_f32_f16v4sf, mve_vmlaldavq_p_<supf><mode>)
13568 (mve_vmlsldavaq_s<mode>, mve_vmlsldavaxq_s<mode>)
13569 (mve_vmlsldavq_p_s<mode>, mve_vmlsldavxq_p_s<mode>)
13570 (mve_vmvnq_m_n_<supf><mode>, mve_vorrq_m_n_<supf><mode>)
13571 (mve_vpselq_f<mode>, mve_vrev32q_m_fv8hf)
13572 (mve_vrev32q_m_<supf><mode>, mve_vrev64q_m_f<mode>)
13573 (mve_vrmlaldavhaxq_sv4si, mve_vrmlaldavhxq_p_sv4si)
13574 (mve_vrmlsldavhaxq_sv4si, mve_vrmlsldavhq_p_sv4si)
13575 (mve_vrmlsldavhxq_p_sv4si, mve_vrev16q_m_<supf>v16qi)
13576 (mve_vrmlaldavhq_p_<supf>v4si, mve_vrmlsldavhaq_sv4si)
13577 (mve_vandq_m_<supf><mode>, mve_vbicq_m_<supf><mode>)
13578 (mve_veorq_m_<supf><mode>, mve_vornq_m_<supf><mode>)
13579 (mve_vorrq_m_<supf><mode>, mve_vandq_m_f<mode>)
13580 (mve_vbicq_m_f<mode>, mve_veorq_m_f<mode>, mve_vornq_m_f<mode>)
13581 (mve_vorrq_m_f<mode>)
13582 (mve_vstrdq_scatter_shifted_offset_p_<supf>v2di_insn)
13583 (mve_vstrdq_scatter_shifted_offset_<supf>v2di_insn)
13584 (mve_vstrdq_scatter_base_wb_p_<supf>v2di) : Fix spacing and
13585 capitalization in the emitted asm.
13587 2023-05-18 Andrea Corallo <andrea.corallo@arm.com>
13589 * config/arm/constraints.md (mve_vldrd_immediate): Move it to
13591 (Ri): Move constraint definition from predicates.md.
13592 (Rl): Define new constraint.
13593 * config/arm/mve.md (mve_vstrwq_scatter_base_wb_p_<supf>v4si): Add
13594 missing constraint.
13595 (mve_vstrwq_scatter_base_wb_p_fv4sf): Add missing Up constraint
13596 for op 1, use mve_vstrw_immediate predicate and Rl constraint for
13597 op 2. Fix asm output spacing.
13598 (mve_vstrdq_scatter_base_wb_p_<supf>v2di): Add missing constraint.
13599 * config/arm/predicates.md (Ri) Move constraint to constraints.md
13600 (mve_vldrd_immediate): Move it from
13602 (mve_vstrw_immediate): New predicate.
13604 2023-05-18 Pan Li <pan2.li@intel.com>
13605 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
13606 Kito Cheng <kito.cheng@sifive.com>
13607 Richard Biener <rguenther@suse.de>
13608 Richard Sandiford <richard.sandiford@arm.com>
13610 * combine.cc (struct reg_stat_type): Extend machine_mode to 16 bits.
13611 * cse.cc (struct qty_table_elem): Extend machine_mode to 16 bits
13612 (struct table_elt): Extend machine_mode to 16 bits.
13613 (struct set): Ditto.
13614 * genmodes.cc (emit_mode_wider): Extend type from char to short.
13615 (emit_mode_complex): Ditto.
13616 (emit_mode_inner): Ditto.
13617 (emit_class_narrowest_mode): Ditto.
13618 * genopinit.cc (main): Extend the machine_mode limit.
13619 * ira-int.h (struct ira_allocno): Extend machine_mode to 16 bits and
13620 re-ordered the struct fields for padding.
13621 * machmode.h (MACHINE_MODE_BITSIZE): New macro.
13622 (GET_MODE_2XWIDER_MODE): Extend type from char to short.
13623 (get_mode_alignment): Extend type from char to short.
13624 * ree.cc (struct ext_modified): Extend machine_mode to 16 bits and
13625 removed the ATTRIBUTE_PACKED.
13626 * rtl-ssa/accesses.h: Extend machine_mode to 16 bits, narrow
13627 * rtl-ssa/internals.inl (rtl_ssa::access_info): Adjust the assignment.
13628 m_kind to 2 bits and remove m_spare.
13629 * rtl.h (RTX_CODE_BITSIZE): New macro.
13630 (struct rtx_def): Swap both the bit size and location between the
13631 rtx_code and the machine_mode.
13632 (subreg_shape::unique_id): Extend the machine_mode limit.
13633 * rtlanal.h: Extend machine_mode to 16 bits.
13634 * tree-core.h (struct tree_type_common): Extend machine_mode to 16
13635 bits and re-ordered the struct fields for padding.
13636 (struct tree_decl_common): Extend machine_mode to 16 bits.
13638 2023-05-17 Jin Ma <jinma@linux.alibaba.com>
13640 * genrecog.cc (print_nonbool_test): Fix type error of
13641 switch (SUBREG_BYTE (op))'.
13643 2023-05-17 Jin Ma <jinma@linux.alibaba.com>
13645 * common/config/riscv/riscv-common.cc: Remove
13646 trailing spaces on lines.
13647 * config/riscv/riscv.cc (riscv_legitimize_move): Likewise.
13648 * config/riscv/riscv.h (enum reg_class): Likewise.
13649 * config/riscv/riscv.md: Likewise.
13651 2023-05-17 John David Anglin <danglin@gcc.gnu.org>
13653 * config/pa/pa.md (clear_cache): New.
13655 2023-05-17 Arsen Arsenović <arsen@aarsen.me>
13657 * doc/extend.texi (C++ Concepts) <forall>: Remove extraneous
13658 parenthesis. Fix misnamed index entry.
13659 <concept>: Fix misnamed index entry.
13661 2023-05-17 Jivan Hakobyan <jivanhakobyan9@gmail.com>
13663 * config/riscv/riscv.md (*<optab><GPR:mode>3_mask): New pattern,
13665 (*<optab>si3_mask, *<optab>di3_mask): Here.
13666 (*<optab>si3_mask_1, *<optab>di3_mask_1): And here.
13667 * config/riscv/bitmanip.md (*<bitmanip_optab><GPR:mode>3_mask): New
13669 (*<bitmanip_optab>si3_sext_mask): Likewise.
13670 * config/riscv/iterators.md (shiftm1): Use const_si_mask_operand
13671 and const_di_mask_operand.
13672 (bitmanip_rotate): New iterator.
13673 (bitmanip_optab): Add rotates.
13674 * config/riscv/predicates.md (const_si_mask_operand): Renamed
13675 from const31_operand. Generalize to handle more mask constants.
13676 (const_di_mask_operand): Similarly.
13678 2023-05-17 Jakub Jelinek <jakub@redhat.com>
13681 * config/i386/i386-builtin-types.def (FLOAT128): Use
13682 float128t_type_node rather than float128_type_node.
13684 2023-05-17 Alexander Monakov <amonakov@ispras.ru>
13686 * tree-ssa-math-opts.cc (convert_mult_to_fma): Enable only for
13687 FP_CONTRACT_FAST (no functional change).
13689 2023-05-17 Uros Bizjak <ubizjak@gmail.com>
13691 * config/i386/i386.cc (ix86_multiplication_cost): Correct
13692 calcuation of integer vector mode costs to reflect generated
13693 instruction sequences of different integer vector modes and
13694 different target ABIs.
13696 2023-05-17 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13698 * config/riscv/riscv-opts.h (enum riscv_entity): New enum.
13699 * config/riscv/riscv.cc (riscv_emit_mode_set): New function.
13700 (riscv_mode_needed): Ditto.
13701 (riscv_mode_after): Ditto.
13702 (riscv_mode_entry): Ditto.
13703 (riscv_mode_exit): Ditto.
13704 (riscv_mode_priority): Ditto.
13705 (TARGET_MODE_EMIT): New target hook.
13706 (TARGET_MODE_NEEDED): Ditto.
13707 (TARGET_MODE_AFTER): Ditto.
13708 (TARGET_MODE_ENTRY): Ditto.
13709 (TARGET_MODE_EXIT): Ditto.
13710 (TARGET_MODE_PRIORITY): Ditto.
13711 * config/riscv/riscv.h (OPTIMIZE_MODE_SWITCHING): Ditto.
13712 (NUM_MODES_FOR_MODE_SWITCHING): Ditto.
13713 * config/riscv/riscv.md: Add csrwvxrm.
13714 * config/riscv/vector.md (rnu,rne,rdn,rod,none): New attribute.
13715 (vxrmsi): New pattern.
13717 2023-05-17 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13719 * config/riscv/riscv-vector-builtins-bases.cc: Introduce rounding mode.
13720 * config/riscv/riscv-vector-builtins-shapes.cc (struct alu_def): Ditto.
13721 (struct narrow_alu_def): Ditto.
13722 * config/riscv/riscv-vector-builtins.cc (function_builder::apply_predication): Ditto.
13723 (function_expander::use_exact_insn): Ditto.
13724 * config/riscv/riscv-vector-builtins.h (function_checker::arg_num): New function.
13725 (function_base::has_rounding_mode_operand_p): New function.
13727 2023-05-17 Andrew Pinski <apinski@marvell.com>
13729 * tree-ssa-forwprop.cc (simplify_builtin_call): Check
13730 against 0 instead of calling integer_zerop.
13732 2023-05-17 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13734 * config/riscv/riscv-vector-builtins.cc (register_vxrm): New function.
13735 (DEF_RVV_VXRM_ENUM): New macro.
13736 (handle_pragma_vector): Add vxrm enum register.
13737 * config/riscv/riscv-vector-builtins.def (DEF_RVV_VXRM_ENUM): New macro.
13743 2023-05-17 Aldy Hernandez <aldyh@redhat.com>
13745 * value-range.h (Value_Range::operator=): New.
13747 2023-05-17 Aldy Hernandez <aldyh@redhat.com>
13749 * value-range.cc (vrange::operator=): Add a stub to copy
13750 unsupported ranges.
13751 * value-range.h (is_a <unsupported_range>): New.
13752 (Value_Range::operator=): Support copying unsupported ranges.
13754 2023-05-17 Aldy Hernandez <aldyh@redhat.com>
13756 * data-streamer-in.cc (streamer_read_real_value): New.
13757 (streamer_read_value_range): New.
13758 * data-streamer-out.cc (streamer_write_real_value): New.
13759 (streamer_write_vrange): New.
13760 * data-streamer.h (streamer_write_vrange): New.
13761 (streamer_read_value_range): New.
13763 2023-05-17 Jonathan Wakely <jwakely@redhat.com>
13766 * doc/invoke.texi (Code Gen Options): Note that -fshort-enums
13767 is ignored for a fixed underlying type.
13768 (C++ Dialect Options): Likewise for -fstrict-enums.
13770 2023-05-17 Tobias Burnus <tobias@codesourcery.com>
13772 * gimplify.cc (gimplify_scan_omp_clauses): Remove Fortran
13775 2023-05-17 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
13777 * config/s390/s390.cc (TARGET_ATOMIC_ALIGN_FOR_MODE):
13779 (s390_atomic_align_for_mode): New.
13781 2023-05-17 Jakub Jelinek <jakub@redhat.com>
13783 * wide-int.cc (wi::from_array): Add missing closing paren in function
13786 2023-05-17 Kewen Lin <linkw@linux.ibm.com>
13788 * tree-vect-loop.cc (vect_analyze_loop_1): Don't retry analysis with
13789 suggested unroll factor once the previous analysis fails.
13791 2023-05-17 Pan Li <pan2.li@intel.com>
13793 * config/riscv/genrvv-type-indexer.cc (BOOL_SIZE_LIST): New
13795 (main): Add bool1 to the type indexer.
13796 * config/riscv/riscv-vector-builtins-functions.def
13797 (vreinterpret): Register vbool1 interpret function.
13798 * config/riscv/riscv-vector-builtins-types.def
13799 (DEF_RVV_BOOL1_INTERPRET_OPS): New macro.
13800 (vint8m1_t): Add the type to bool1_interpret_ops.
13801 (vint16m1_t): Ditto.
13802 (vint32m1_t): Ditto.
13803 (vint64m1_t): Ditto.
13804 (vuint8m1_t): Ditto.
13805 (vuint16m1_t): Ditto.
13806 (vuint32m1_t): Ditto.
13807 (vuint64m1_t): Ditto.
13808 * config/riscv/riscv-vector-builtins.cc
13809 (DEF_RVV_BOOL1_INTERPRET_OPS): New macro.
13810 (required_extensions_p): Add bool1 interpret case.
13811 * config/riscv/riscv-vector-builtins.def
13812 (bool1_interpret): Add bool1 interpret to base type.
13813 * config/riscv/vector.md (@vreinterpret<mode>): Add new expand
13814 with VB dest for vreinterpret.
13816 2023-05-17 Jiufu Guo <guojiufu@linux.ibm.com>
13819 * config/rs6000/rs6000.cc (rs6000_emit_set_long_const): Support building
13820 constants through "lis; xoris".
13822 2023-05-16 Ajit Kumar Agarwal <aagarwa1@linux.ibm.com>
13824 * common/config/rs6000/rs6000-common.cc: Add REE pass as a
13825 default rs6000 target pass for O2 and above.
13826 * doc/invoke.texi: Document -free
13828 2023-05-16 Kito Cheng <kito.cheng@sifive.com>
13830 * common/config/riscv/riscv-common.cc (riscv_compute_multilib):
13831 Fix wrong select_kind...
13833 2023-05-16 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
13835 * config/s390/s390-protos.h (s390_expand_setmem): Change
13836 function signature.
13837 * config/s390/s390.cc (s390_expand_setmem): For memset's less
13838 than or equal to 256 byte do not perform a libc call.
13839 * config/s390/s390.md: Change expander into a version which
13842 2023-05-16 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
13844 * config/s390/s390-protos.h (s390_expand_movmem): New.
13845 * config/s390/s390.cc (s390_expand_movmem): New.
13846 * config/s390/s390.md (movmem<mode>): New.
13850 2023-05-16 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
13852 * config/s390/s390-protos.h (s390_expand_cpymem): Change
13853 function signature.
13854 * config/s390/s390.cc (s390_expand_cpymem): For memcpy's less
13855 than or equal to 256 byte do not perform a libc call.
13856 (s390_expand_insv): Adapt new function signature of
13857 s390_expand_cpymem.
13858 * config/s390/s390.md: Change expander into a version which
13861 2023-05-16 Andrew Pinski <apinski@marvell.com>
13863 PR tree-optimization/109424
13864 * match.pd: Add patterns for min/max of zero_one_valued
13867 2023-05-16 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13869 * config/riscv/riscv-protos.h (enum frm_field_enum): New enum.
13870 * config/riscv/riscv-vector-builtins.cc
13871 (function_expander::use_ternop_insn): Add default rounding mode.
13872 (function_expander::use_widen_ternop_insn): Ditto.
13873 * config/riscv/riscv.cc (riscv_hard_regno_nregs): Add FRM REGNUM.
13874 (riscv_hard_regno_mode_ok): Ditto.
13875 (riscv_conditional_register_usage): Ditto.
13876 * config/riscv/riscv.h (DWARF_FRAME_REGNUM): Ditto.
13877 (FRM_REG_P): Ditto.
13878 (RISCV_DWARF_FRM): Ditto.
13879 * config/riscv/riscv.md: Ditto.
13880 * config/riscv/vector-iterators.md: split no frm and has frm operations.
13881 * config/riscv/vector.md (@pred_<optab><mode>_scalar): New pattern.
13882 (@pred_<optab><mode>): Ditto.
13884 2023-05-15 Aldy Hernandez <aldyh@redhat.com>
13886 PR tree-optimization/109695
13887 * value-range.cc (irange::operator=): Resize range.
13888 (irange::union_): Same.
13889 (irange::intersect): Same.
13890 (irange::invert): Same.
13891 (int_range_max): Default to 3 sub-ranges and resize as needed.
13892 * value-range.h (irange::maybe_resize): New.
13894 (int_range::int_range): Adjust for resizing.
13895 (int_range::operator=): Same.
13897 2023-05-15 Aldy Hernandez <aldyh@redhat.com>
13899 * ipa-cp.cc (ipcp_vr_lattice::meet_with_1): Avoid unnecessary
13901 * value-range.cc (irange::union_nonzero_bits): Return TRUE only
13902 when range changed.
13904 2023-05-15 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13906 * config/riscv/riscv-protos.h (enum vxrm_field_enum): New enum.
13907 * config/riscv/riscv-vector-builtins.cc
13908 (function_expander::use_exact_insn): Add default rounding mode operand.
13909 * config/riscv/riscv.cc (riscv_hard_regno_nregs): Add VXRM_REGNUM.
13910 (riscv_hard_regno_mode_ok): Ditto.
13911 (riscv_conditional_register_usage): Ditto.
13912 * config/riscv/riscv.h (DWARF_FRAME_REGNUM): Ditto.
13913 (VXRM_REG_P): Ditto.
13914 (RISCV_DWARF_VXRM): Ditto.
13915 * config/riscv/riscv.md: Ditto.
13916 * config/riscv/vector.md: Ditto
13918 2023-05-15 Pan Li <pan2.li@intel.com>
13920 * optabs.cc (maybe_gen_insn): Add case to generate instruction
13921 that has 11 operands.
13923 2023-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13925 * config/aarch64/aarch64.cc (aarch64_rtx_costs, NEG case): Add costing
13926 logic for vector modes.
13928 2023-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13931 * config/aarch64/aarch64-simd.md (aarch64_cm<optab><mode>): Rename to...
13932 (aarch64_cm<optab><mode><vczle><vczbe>): ... This.
13933 (aarch64_cmtst<mode>): Rename to...
13934 (aarch64_cmtst<mode><vczle><vczbe>): ... This.
13935 (*aarch64_cmtst_same_<mode>): Rename to...
13936 (*aarch64_cmtst_same_<mode><vczle><vczbe>): ... This.
13937 (*aarch64_cmtstdi): Rename to...
13938 (*aarch64_cmtstdi<vczle><vczbe>): ... This.
13939 (aarch64_fac<optab><mode>): Rename to...
13940 (aarch64_fac<optab><mode><vczle><vczbe>): ... This.
13942 2023-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13945 * config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>): Rename to...
13946 (aarch64_s<optab><mode><vczle><vczbe>): ... This.
13948 2023-05-15 Pan Li <pan2.li@intel.com>
13949 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13950 kito-cheng <kito.cheng@sifive.com>
13952 * config/riscv/riscv-v.cc (const_vlmax_p): New function for
13953 deciding the mode is constant or not.
13954 (set_len_and_policy): Optimize VLS-VLMAX code gen to vsetivli.
13956 2023-05-15 Richard Biener <rguenther@suse.de>
13958 PR tree-optimization/109848
13959 * tree-ssa-forwprop.cc (pass_forwprop::execute): Put the
13960 TARGET_MEM_REF address preparation before the store, not
13963 2023-05-15 Juzhe-Zhong <juzhe.zhong@rivai.ai>
13965 * config/riscv/riscv.cc
13966 (riscv_vectorize_preferred_vector_alignment): New function.
13967 (TARGET_VECTORIZE_PREFERRED_VECTOR_ALIGNMENT): New target hook.
13969 2023-05-14 Andrew Pinski <apinski@marvell.com>
13971 PR tree-optimization/109829
13972 * match.pd: Add pattern for `signbit(x) !=/== 0 ? x : -x`.
13974 2023-05-14 Uros Bizjak <ubizjak@gmail.com>
13977 * config/i386/i386.cc: Revert the 2023-05-11 change.
13978 (ix86_widen_mult_cost): Return high value instead of
13979 ICEing for unsupported modes.
13981 2023-05-14 Ard Biesheuvel <ardb@kernel.org>
13983 * config/i386/i386.cc (x86_function_profiler): Take
13984 ix86_direct_extern_access into account when generating calls
13987 2023-05-14 Pan Li <pan2.li@intel.com>
13989 * config/riscv/riscv-vector-builtins.cc (required_extensions_p):
13990 Refactor the or pattern to switch cases.
13992 2023-05-13 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
13994 * config/aarch64/aarch64.cc (aarch64_expand_vector_init_fallback): Rename
13995 aarch64_expand_vector_init to this, and remove interleaving case.
13996 Recursively call aarch64_expand_vector_init_fallback, instead of
13997 aarch64_expand_vector_init.
13998 (aarch64_unzip_vector_init): New function.
13999 (aarch64_expand_vector_init): Likewise.
14001 2023-05-13 Kito Cheng <kito.cheng@sifive.com>
14003 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::cleanup_insns):
14004 Pull out function call from the gcc_assert.
14006 2023-05-13 Kito Cheng <kito.cheng@sifive.com>
14008 * config/riscv/riscv-vsetvl.cc (vlmul_to_str): New.
14009 (policy_to_str): New.
14010 (vector_insn_info::dump): Use vlmul_to_str and policy_to_str.
14012 2023-05-13 Andrew Pinski <apinski@marvell.com>
14014 PR tree-optimization/109834
14015 * match.pd (popcount(bswap(x))->popcount(x)): Fix up unsigned type checking.
14016 (popcount(rotate(x,y))->popcount(x)): Likewise.
14018 2023-05-12 Uros Bizjak <ubizjak@gmail.com>
14020 * config/i386/i386-expand.cc (ix86_expand_vecop_qihi2): Also
14021 reject ymm instructions for TARGET_PREFER_AVX128. Use generic
14022 gen_extend_insn to generate zero/sign extension instructions.
14024 (ix86_expand_vecop_qihi): Initialize interleave functions
14025 for MULT code only. Fix comments.
14027 2023-05-12 Uros Bizjak <ubizjak@gmail.com>
14030 * config/i386/mmx.md (mulv2si3): Remove expander.
14031 (mulv2si3): Rename insn pattern from *mulv2si.
14033 2023-05-12 Tobias Burnus <tobias@codesourcery.com>
14035 PR libstdc++/109816
14036 * lto-cgraph.cc (output_symtab): Guard lto_output_toplevel_asms by
14037 '!lto_stream_offload_p'.
14039 2023-05-12 Kito Cheng <kito.cheng@sifive.com>
14040 Juzhe-Zhong <juzhe.zhong@rivai.ai>
14043 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::get_vsetvl_at_end): New.
14044 (local_avl_compatible_p): New.
14045 (pass_vsetvl::local_eliminate_vsetvl_insn): Enhance local optimizations
14046 for LCM, rewrite as a backward algorithm.
14047 (pass_vsetvl::cleanup_insns): Use new local_eliminate_vsetvl_insn
14048 interface, handle a BB at once.
14050 2023-05-12 Richard Biener <rguenther@suse.de>
14052 PR tree-optimization/64731
14053 * tree-ssa-forwprop.cc (pass_forwprop::execute): Also
14054 handle TARGET_MEM_REF destinations of stores from vector
14057 2023-05-12 Richard Biener <rguenther@suse.de>
14059 PR tree-optimization/109791
14060 * match.pd (minus (convert ADDR_EXPR@0) (convert (pointer_plus @1 @2))):
14062 (minus (convert (pointer_plus @1 @2)) (convert ADDR_EXPR@0)):
14065 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14067 * config/arm/arm-mve-builtins-base.cc (vsriq): New.
14068 * config/arm/arm-mve-builtins-base.def (vsriq): New.
14069 * config/arm/arm-mve-builtins-base.h (vsriq): New.
14070 * config/arm/arm-mve-builtins.cc
14071 (function_instance::has_inactive_argument): Handle vsriq.
14072 * config/arm/arm_mve.h (vsriq): Remove.
14074 (vsriq_n_u8): Remove.
14075 (vsriq_n_s8): Remove.
14076 (vsriq_n_u16): Remove.
14077 (vsriq_n_s16): Remove.
14078 (vsriq_n_u32): Remove.
14079 (vsriq_n_s32): Remove.
14080 (vsriq_m_n_s8): Remove.
14081 (vsriq_m_n_u8): Remove.
14082 (vsriq_m_n_s16): Remove.
14083 (vsriq_m_n_u16): Remove.
14084 (vsriq_m_n_s32): Remove.
14085 (vsriq_m_n_u32): Remove.
14086 (__arm_vsriq_n_u8): Remove.
14087 (__arm_vsriq_n_s8): Remove.
14088 (__arm_vsriq_n_u16): Remove.
14089 (__arm_vsriq_n_s16): Remove.
14090 (__arm_vsriq_n_u32): Remove.
14091 (__arm_vsriq_n_s32): Remove.
14092 (__arm_vsriq_m_n_s8): Remove.
14093 (__arm_vsriq_m_n_u8): Remove.
14094 (__arm_vsriq_m_n_s16): Remove.
14095 (__arm_vsriq_m_n_u16): Remove.
14096 (__arm_vsriq_m_n_s32): Remove.
14097 (__arm_vsriq_m_n_u32): Remove.
14098 (__arm_vsriq): Remove.
14099 (__arm_vsriq_m): Remove.
14101 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14103 * config/arm/iterators.md (mve_insn): Add vsri.
14104 * config/arm/mve.md (mve_vsriq_n_<supf><mode>): Rename into ...
14105 (@mve_<mve_insn>q_n_<supf><mode>): .,. this.
14106 (mve_vsriq_m_n_<supf><mode>): Rename into ...
14107 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
14109 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14111 * config/arm/arm-mve-builtins-shapes.cc (ternary_rshift): New.
14112 * config/arm/arm-mve-builtins-shapes.h (ternary_rshift): New.
14114 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14116 * config/arm/arm-mve-builtins-base.cc (vsliq): New.
14117 * config/arm/arm-mve-builtins-base.def (vsliq): New.
14118 * config/arm/arm-mve-builtins-base.h (vsliq): New.
14119 * config/arm/arm-mve-builtins.cc
14120 (function_instance::has_inactive_argument): Handle vsliq.
14121 * config/arm/arm_mve.h (vsliq): Remove.
14123 (vsliq_n_u8): Remove.
14124 (vsliq_n_s8): Remove.
14125 (vsliq_n_u16): Remove.
14126 (vsliq_n_s16): Remove.
14127 (vsliq_n_u32): Remove.
14128 (vsliq_n_s32): Remove.
14129 (vsliq_m_n_s8): Remove.
14130 (vsliq_m_n_s32): Remove.
14131 (vsliq_m_n_s16): Remove.
14132 (vsliq_m_n_u8): Remove.
14133 (vsliq_m_n_u32): Remove.
14134 (vsliq_m_n_u16): Remove.
14135 (__arm_vsliq_n_u8): Remove.
14136 (__arm_vsliq_n_s8): Remove.
14137 (__arm_vsliq_n_u16): Remove.
14138 (__arm_vsliq_n_s16): Remove.
14139 (__arm_vsliq_n_u32): Remove.
14140 (__arm_vsliq_n_s32): Remove.
14141 (__arm_vsliq_m_n_s8): Remove.
14142 (__arm_vsliq_m_n_s32): Remove.
14143 (__arm_vsliq_m_n_s16): Remove.
14144 (__arm_vsliq_m_n_u8): Remove.
14145 (__arm_vsliq_m_n_u32): Remove.
14146 (__arm_vsliq_m_n_u16): Remove.
14147 (__arm_vsliq): Remove.
14148 (__arm_vsliq_m): Remove.
14150 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14152 * config/arm/iterators.md (mve_insn>): Add vsli.
14153 * config/arm/mve.md (mve_vsliq_n_<supf><mode>): Rename into ...
14154 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
14155 (mve_vsliq_m_n_<supf><mode>): Rename into ...
14156 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
14158 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14160 * config/arm/arm-mve-builtins-shapes.cc (ternary_lshift): New.
14161 * config/arm/arm-mve-builtins-shapes.h (ternary_lshift): New.
14163 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14165 * config/arm/arm-mve-builtins-base.cc (vpselq): New.
14166 * config/arm/arm-mve-builtins-base.def (vpselq): New.
14167 * config/arm/arm-mve-builtins-base.h (vpselq): New.
14168 * config/arm/arm_mve.h (vpselq): Remove.
14169 (vpselq_u8): Remove.
14170 (vpselq_s8): Remove.
14171 (vpselq_u16): Remove.
14172 (vpselq_s16): Remove.
14173 (vpselq_u32): Remove.
14174 (vpselq_s32): Remove.
14175 (vpselq_u64): Remove.
14176 (vpselq_s64): Remove.
14177 (vpselq_f16): Remove.
14178 (vpselq_f32): Remove.
14179 (__arm_vpselq_u8): Remove.
14180 (__arm_vpselq_s8): Remove.
14181 (__arm_vpselq_u16): Remove.
14182 (__arm_vpselq_s16): Remove.
14183 (__arm_vpselq_u32): Remove.
14184 (__arm_vpselq_s32): Remove.
14185 (__arm_vpselq_u64): Remove.
14186 (__arm_vpselq_s64): Remove.
14187 (__arm_vpselq_f16): Remove.
14188 (__arm_vpselq_f32): Remove.
14189 (__arm_vpselq): Remove.
14191 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14193 * config/arm/arm-mve-builtins-shapes.cc (vpsel): New.
14194 * config/arm/arm-mve-builtins-shapes.h (vpsel): New.
14196 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14198 * config/arm/arm.cc (arm_expand_vcond): Use gen_mve_q instead of
14200 * config/arm/iterators.md (MVE_VPSELQ_F): New.
14201 (mve_insn): Add vpsel.
14202 * config/arm/mve.md (@mve_vpselq_<supf><mode>): Rename into ...
14203 (@mve_<mve_insn>q_<supf><mode>): ... this.
14204 (@mve_vpselq_f<mode>): Rename into ...
14205 (@mve_<mve_insn>q_f<mode>): ... this.
14207 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14209 * config/arm/arm-mve-builtins-base.cc (vfmaq, vfmasq, vfmsq): New.
14210 * config/arm/arm-mve-builtins-base.def (vfmaq, vfmasq, vfmsq): New.
14211 * config/arm/arm-mve-builtins-base.h (vfmaq, vfmasq, vfmsq): New.
14212 * config/arm/arm-mve-builtins.cc
14213 (function_instance::has_inactive_argument): Handle vfmaq, vfmasq,
14215 * config/arm/arm_mve.h (vfmaq): Remove.
14219 (vfmasq_m): Remove.
14221 (vfmaq_f16): Remove.
14222 (vfmaq_n_f16): Remove.
14223 (vfmasq_n_f16): Remove.
14224 (vfmsq_f16): Remove.
14225 (vfmaq_f32): Remove.
14226 (vfmaq_n_f32): Remove.
14227 (vfmasq_n_f32): Remove.
14228 (vfmsq_f32): Remove.
14229 (vfmaq_m_f32): Remove.
14230 (vfmaq_m_f16): Remove.
14231 (vfmaq_m_n_f32): Remove.
14232 (vfmaq_m_n_f16): Remove.
14233 (vfmasq_m_n_f32): Remove.
14234 (vfmasq_m_n_f16): Remove.
14235 (vfmsq_m_f32): Remove.
14236 (vfmsq_m_f16): Remove.
14237 (__arm_vfmaq_f16): Remove.
14238 (__arm_vfmaq_n_f16): Remove.
14239 (__arm_vfmasq_n_f16): Remove.
14240 (__arm_vfmsq_f16): Remove.
14241 (__arm_vfmaq_f32): Remove.
14242 (__arm_vfmaq_n_f32): Remove.
14243 (__arm_vfmasq_n_f32): Remove.
14244 (__arm_vfmsq_f32): Remove.
14245 (__arm_vfmaq_m_f32): Remove.
14246 (__arm_vfmaq_m_f16): Remove.
14247 (__arm_vfmaq_m_n_f32): Remove.
14248 (__arm_vfmaq_m_n_f16): Remove.
14249 (__arm_vfmasq_m_n_f32): Remove.
14250 (__arm_vfmasq_m_n_f16): Remove.
14251 (__arm_vfmsq_m_f32): Remove.
14252 (__arm_vfmsq_m_f16): Remove.
14253 (__arm_vfmaq): Remove.
14254 (__arm_vfmasq): Remove.
14255 (__arm_vfmsq): Remove.
14256 (__arm_vfmaq_m): Remove.
14257 (__arm_vfmasq_m): Remove.
14258 (__arm_vfmsq_m): Remove.
14260 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14262 * config/arm/iterators.md (MVE_FP_M_BINARY): Add VFMAQ_M_F,
14264 (MVE_FP_M_N_BINARY): Add VFMAQ_M_N_F, VFMASQ_M_N_F.
14265 (MVE_VFMxQ_F, MVE_VFMAxQ_N_F): New.
14266 (mve_insn): Add vfma, vfmas, vfms.
14267 * config/arm/mve.md (mve_vfmaq_f<mode>, mve_vfmsq_f<mode>): Merge
14269 (@mve_<mve_insn>q_f<mode>): ... this.
14270 (mve_vfmaq_n_f<mode>, mve_vfmasq_n_f<mode>): Merge into ...
14271 (@mve_<mve_insn>q_n_f<mode>): ... this.
14272 (mve_vfmaq_m_f<mode>, mve_vfmsq_m_f<mode>): Merge into
14273 @mve_<mve_insn>q_m_f<mode>.
14274 (mve_vfmaq_m_n_f<mode>, mve_vfmasq_m_n_f<mode>): Merge into
14275 @mve_<mve_insn>q_m_n_f<mode>.
14277 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14279 * config/arm/arm-mve-builtins-shapes.cc (ternary_opt_n): New.
14280 * config/arm/arm-mve-builtins-shapes.h (ternary_opt_n): New.
14282 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14284 * config/arm/arm-mve-builtins-base.cc
14285 (FUNCTION_WITH_RTX_M_N_NO_F): New.
14287 * config/arm/arm-mve-builtins-base.def (vmvnq): New.
14288 * config/arm/arm-mve-builtins-base.h (vmvnq): New.
14289 * config/arm/arm_mve.h (vmvnq): Remove.
14292 (vmvnq_s8): Remove.
14293 (vmvnq_s16): Remove.
14294 (vmvnq_s32): Remove.
14295 (vmvnq_n_s16): Remove.
14296 (vmvnq_n_s32): Remove.
14297 (vmvnq_u8): Remove.
14298 (vmvnq_u16): Remove.
14299 (vmvnq_u32): Remove.
14300 (vmvnq_n_u16): Remove.
14301 (vmvnq_n_u32): Remove.
14302 (vmvnq_m_u8): Remove.
14303 (vmvnq_m_s8): Remove.
14304 (vmvnq_m_u16): Remove.
14305 (vmvnq_m_s16): Remove.
14306 (vmvnq_m_u32): Remove.
14307 (vmvnq_m_s32): Remove.
14308 (vmvnq_m_n_s16): Remove.
14309 (vmvnq_m_n_u16): Remove.
14310 (vmvnq_m_n_s32): Remove.
14311 (vmvnq_m_n_u32): Remove.
14312 (vmvnq_x_s8): Remove.
14313 (vmvnq_x_s16): Remove.
14314 (vmvnq_x_s32): Remove.
14315 (vmvnq_x_u8): Remove.
14316 (vmvnq_x_u16): Remove.
14317 (vmvnq_x_u32): Remove.
14318 (vmvnq_x_n_s16): Remove.
14319 (vmvnq_x_n_s32): Remove.
14320 (vmvnq_x_n_u16): Remove.
14321 (vmvnq_x_n_u32): Remove.
14322 (__arm_vmvnq_s8): Remove.
14323 (__arm_vmvnq_s16): Remove.
14324 (__arm_vmvnq_s32): Remove.
14325 (__arm_vmvnq_n_s16): Remove.
14326 (__arm_vmvnq_n_s32): Remove.
14327 (__arm_vmvnq_u8): Remove.
14328 (__arm_vmvnq_u16): Remove.
14329 (__arm_vmvnq_u32): Remove.
14330 (__arm_vmvnq_n_u16): Remove.
14331 (__arm_vmvnq_n_u32): Remove.
14332 (__arm_vmvnq_m_u8): Remove.
14333 (__arm_vmvnq_m_s8): Remove.
14334 (__arm_vmvnq_m_u16): Remove.
14335 (__arm_vmvnq_m_s16): Remove.
14336 (__arm_vmvnq_m_u32): Remove.
14337 (__arm_vmvnq_m_s32): Remove.
14338 (__arm_vmvnq_m_n_s16): Remove.
14339 (__arm_vmvnq_m_n_u16): Remove.
14340 (__arm_vmvnq_m_n_s32): Remove.
14341 (__arm_vmvnq_m_n_u32): Remove.
14342 (__arm_vmvnq_x_s8): Remove.
14343 (__arm_vmvnq_x_s16): Remove.
14344 (__arm_vmvnq_x_s32): Remove.
14345 (__arm_vmvnq_x_u8): Remove.
14346 (__arm_vmvnq_x_u16): Remove.
14347 (__arm_vmvnq_x_u32): Remove.
14348 (__arm_vmvnq_x_n_s16): Remove.
14349 (__arm_vmvnq_x_n_s32): Remove.
14350 (__arm_vmvnq_x_n_u16): Remove.
14351 (__arm_vmvnq_x_n_u32): Remove.
14352 (__arm_vmvnq): Remove.
14353 (__arm_vmvnq_m): Remove.
14354 (__arm_vmvnq_x): Remove.
14356 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14358 * config/arm/iterators.md (mve_insn): Add vmvn.
14359 * config/arm/mve.md (mve_vmvnq_n_<supf><mode>): Rename into ...
14360 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
14361 (mve_vmvnq_m_<supf><mode>): Rename into ...
14362 (@mve_<mve_insn>q_m_<supf><mode>): ... this.
14363 (mve_vmvnq_m_n_<supf><mode>): Rename into ...
14364 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
14366 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14368 * config/arm/arm-mve-builtins-shapes.cc (mvn): New.
14369 * config/arm/arm-mve-builtins-shapes.h (mvn): New.
14371 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14373 * config/arm/arm-mve-builtins-base.cc (vbrsrq): New.
14374 * config/arm/arm-mve-builtins-base.def (vbrsrq): New.
14375 * config/arm/arm-mve-builtins-base.h (vbrsrq): New.
14376 * config/arm/arm_mve.h (vbrsrq): Remove.
14377 (vbrsrq_m): Remove.
14378 (vbrsrq_x): Remove.
14379 (vbrsrq_n_f16): Remove.
14380 (vbrsrq_n_f32): Remove.
14381 (vbrsrq_n_u8): Remove.
14382 (vbrsrq_n_s8): Remove.
14383 (vbrsrq_n_u16): Remove.
14384 (vbrsrq_n_s16): Remove.
14385 (vbrsrq_n_u32): Remove.
14386 (vbrsrq_n_s32): Remove.
14387 (vbrsrq_m_n_s8): Remove.
14388 (vbrsrq_m_n_s32): Remove.
14389 (vbrsrq_m_n_s16): Remove.
14390 (vbrsrq_m_n_u8): Remove.
14391 (vbrsrq_m_n_u32): Remove.
14392 (vbrsrq_m_n_u16): Remove.
14393 (vbrsrq_m_n_f32): Remove.
14394 (vbrsrq_m_n_f16): Remove.
14395 (vbrsrq_x_n_s8): Remove.
14396 (vbrsrq_x_n_s16): Remove.
14397 (vbrsrq_x_n_s32): Remove.
14398 (vbrsrq_x_n_u8): Remove.
14399 (vbrsrq_x_n_u16): Remove.
14400 (vbrsrq_x_n_u32): Remove.
14401 (vbrsrq_x_n_f16): Remove.
14402 (vbrsrq_x_n_f32): Remove.
14403 (__arm_vbrsrq_n_u8): Remove.
14404 (__arm_vbrsrq_n_s8): Remove.
14405 (__arm_vbrsrq_n_u16): Remove.
14406 (__arm_vbrsrq_n_s16): Remove.
14407 (__arm_vbrsrq_n_u32): Remove.
14408 (__arm_vbrsrq_n_s32): Remove.
14409 (__arm_vbrsrq_m_n_s8): Remove.
14410 (__arm_vbrsrq_m_n_s32): Remove.
14411 (__arm_vbrsrq_m_n_s16): Remove.
14412 (__arm_vbrsrq_m_n_u8): Remove.
14413 (__arm_vbrsrq_m_n_u32): Remove.
14414 (__arm_vbrsrq_m_n_u16): Remove.
14415 (__arm_vbrsrq_x_n_s8): Remove.
14416 (__arm_vbrsrq_x_n_s16): Remove.
14417 (__arm_vbrsrq_x_n_s32): Remove.
14418 (__arm_vbrsrq_x_n_u8): Remove.
14419 (__arm_vbrsrq_x_n_u16): Remove.
14420 (__arm_vbrsrq_x_n_u32): Remove.
14421 (__arm_vbrsrq_n_f16): Remove.
14422 (__arm_vbrsrq_n_f32): Remove.
14423 (__arm_vbrsrq_m_n_f32): Remove.
14424 (__arm_vbrsrq_m_n_f16): Remove.
14425 (__arm_vbrsrq_x_n_f16): Remove.
14426 (__arm_vbrsrq_x_n_f32): Remove.
14427 (__arm_vbrsrq): Remove.
14428 (__arm_vbrsrq_m): Remove.
14429 (__arm_vbrsrq_x): Remove.
14431 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14433 * config/arm/iterators.md (MVE_VBRSR_M_N_FP, MVE_VBRSR_N_FP): New.
14434 (mve_insn): Add vbrsr.
14435 * config/arm/mve.md (mve_vbrsrq_n_f<mode>): Rename into ...
14436 (@mve_<mve_insn>q_n_f<mode>): ... this.
14437 (mve_vbrsrq_n_<supf><mode>): Rename into ...
14438 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
14439 (mve_vbrsrq_m_n_<supf><mode>): Rename into ...
14440 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
14441 (mve_vbrsrq_m_n_f<mode>): Rename into ...
14442 (@mve_<mve_insn>q_m_n_f<mode>): ... this.
14444 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14446 * config/arm/arm-mve-builtins-shapes.cc (binary_imm32): New.
14447 * config/arm/arm-mve-builtins-shapes.h (binary_imm32): New.
14449 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14451 * config/arm/arm-mve-builtins-base.cc (vqshluq): New.
14452 * config/arm/arm-mve-builtins-base.def (vqshluq): New.
14453 * config/arm/arm-mve-builtins-base.h (vqshluq): New.
14454 * config/arm/arm_mve.h (vqshluq): Remove.
14455 (vqshluq_m): Remove.
14456 (vqshluq_n_s8): Remove.
14457 (vqshluq_n_s16): Remove.
14458 (vqshluq_n_s32): Remove.
14459 (vqshluq_m_n_s8): Remove.
14460 (vqshluq_m_n_s16): Remove.
14461 (vqshluq_m_n_s32): Remove.
14462 (__arm_vqshluq_n_s8): Remove.
14463 (__arm_vqshluq_n_s16): Remove.
14464 (__arm_vqshluq_n_s32): Remove.
14465 (__arm_vqshluq_m_n_s8): Remove.
14466 (__arm_vqshluq_m_n_s16): Remove.
14467 (__arm_vqshluq_m_n_s32): Remove.
14468 (__arm_vqshluq): Remove.
14469 (__arm_vqshluq_m): Remove.
14471 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14473 * config/arm/iterators.md (mve_insn): Add vqshlu.
14474 (supf): Add VQSHLUQ_M_N_S, VQSHLUQ_N_S.
14475 (VQSHLUQ_M_N, VQSHLUQ_N): New.
14476 * config/arm/mve.md (mve_vqshluq_n_s<mode>): Change name into ...
14477 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
14478 (mve_vqshluq_m_n_s<mode>): Change name into ...
14479 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
14481 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14483 * config/arm/arm-mve-builtins-shapes.cc
14484 (binary_lshift_unsigned): New.
14485 * config/arm/arm-mve-builtins-shapes.h
14486 (binary_lshift_unsigned): New.
14488 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14490 * config/arm/arm-mve-builtins-base.cc (vrmlaldavhaq)
14491 (vrmlaldavhaxq, vrmlsldavhaq, vrmlsldavhaxq): New.
14492 * config/arm/arm-mve-builtins-base.def (vrmlaldavhaq)
14493 (vrmlaldavhaxq, vrmlsldavhaq, vrmlsldavhaxq): New.
14494 * config/arm/arm-mve-builtins-base.h (vrmlaldavhaq)
14495 (vrmlaldavhaxq, vrmlsldavhaq, vrmlsldavhaxq): New.
14496 * config/arm/arm-mve-builtins-functions.h: Handle vrmlaldavhaq,
14497 vrmlaldavhaxq, vrmlsldavhaq, vrmlsldavhaxq.
14498 * config/arm/arm_mve.h (vrmlaldavhaq): Remove.
14499 (vrmlaldavhaxq): Remove.
14500 (vrmlsldavhaq): Remove.
14501 (vrmlsldavhaxq): Remove.
14502 (vrmlaldavhaq_p): Remove.
14503 (vrmlaldavhaxq_p): Remove.
14504 (vrmlsldavhaq_p): Remove.
14505 (vrmlsldavhaxq_p): Remove.
14506 (vrmlaldavhaq_s32): Remove.
14507 (vrmlaldavhaq_u32): Remove.
14508 (vrmlaldavhaxq_s32): Remove.
14509 (vrmlsldavhaq_s32): Remove.
14510 (vrmlsldavhaxq_s32): Remove.
14511 (vrmlaldavhaq_p_s32): Remove.
14512 (vrmlaldavhaq_p_u32): Remove.
14513 (vrmlaldavhaxq_p_s32): Remove.
14514 (vrmlsldavhaq_p_s32): Remove.
14515 (vrmlsldavhaxq_p_s32): Remove.
14516 (__arm_vrmlaldavhaq_s32): Remove.
14517 (__arm_vrmlaldavhaq_u32): Remove.
14518 (__arm_vrmlaldavhaxq_s32): Remove.
14519 (__arm_vrmlsldavhaq_s32): Remove.
14520 (__arm_vrmlsldavhaxq_s32): Remove.
14521 (__arm_vrmlaldavhaq_p_s32): Remove.
14522 (__arm_vrmlaldavhaq_p_u32): Remove.
14523 (__arm_vrmlaldavhaxq_p_s32): Remove.
14524 (__arm_vrmlsldavhaq_p_s32): Remove.
14525 (__arm_vrmlsldavhaxq_p_s32): Remove.
14526 (__arm_vrmlaldavhaq): Remove.
14527 (__arm_vrmlaldavhaxq): Remove.
14528 (__arm_vrmlsldavhaq): Remove.
14529 (__arm_vrmlsldavhaxq): Remove.
14530 (__arm_vrmlaldavhaq_p): Remove.
14531 (__arm_vrmlaldavhaxq_p): Remove.
14532 (__arm_vrmlsldavhaq_p): Remove.
14533 (__arm_vrmlsldavhaxq_p): Remove.
14535 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14537 * config/arm/iterators.md (MVE_VRMLxLDAVHAxQ)
14538 (MVE_VRMLxLDAVHAxQ_P): New.
14539 (mve_insn): Add vrmlaldavha, vrmlaldavhax, vrmlsldavha,
14541 (supf): Add VRMLALDAVHAXQ_P_S, VRMLALDAVHAXQ_S, VRMLSLDAVHAQ_P_S,
14542 VRMLSLDAVHAQ_S, VRMLSLDAVHAXQ_P_S, VRMLSLDAVHAXQ_S,
14544 * config/arm/mve.md (mve_vrmlaldavhaq_<supf>v4si)
14545 (mve_vrmlaldavhaxq_sv4si, mve_vrmlsldavhaxq_sv4si)
14546 (mve_vrmlsldavhaq_sv4si): Merge into ...
14547 (@mve_<mve_insn>q_<supf>v4si): ... this.
14548 (mve_vrmlaldavhaq_p_sv4si, mve_vrmlaldavhaq_p_uv4si)
14549 (mve_vrmlaldavhaxq_p_sv4si, mve_vrmlsldavhaq_p_sv4si)
14550 (mve_vrmlsldavhaxq_p_sv4si): Merge into ...
14551 (@mve_<mve_insn>q_p_<supf>v4si): ... this.
14553 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14555 * config/arm/arm-mve-builtins-base.cc (vqdmullbq, vqdmulltq): New.
14556 * config/arm/arm-mve-builtins-base.def (vqdmullbq, vqdmulltq):
14558 * config/arm/arm-mve-builtins-base.h (vqdmullbq, vqdmulltq): New.
14559 * config/arm/arm_mve.h (vqdmulltq): Remove.
14560 (vqdmullbq): Remove.
14561 (vqdmullbq_m): Remove.
14562 (vqdmulltq_m): Remove.
14563 (vqdmulltq_s16): Remove.
14564 (vqdmulltq_n_s16): Remove.
14565 (vqdmullbq_s16): Remove.
14566 (vqdmullbq_n_s16): Remove.
14567 (vqdmulltq_s32): Remove.
14568 (vqdmulltq_n_s32): Remove.
14569 (vqdmullbq_s32): Remove.
14570 (vqdmullbq_n_s32): Remove.
14571 (vqdmullbq_m_n_s32): Remove.
14572 (vqdmullbq_m_n_s16): Remove.
14573 (vqdmullbq_m_s32): Remove.
14574 (vqdmullbq_m_s16): Remove.
14575 (vqdmulltq_m_n_s32): Remove.
14576 (vqdmulltq_m_n_s16): Remove.
14577 (vqdmulltq_m_s32): Remove.
14578 (vqdmulltq_m_s16): Remove.
14579 (__arm_vqdmulltq_s16): Remove.
14580 (__arm_vqdmulltq_n_s16): Remove.
14581 (__arm_vqdmullbq_s16): Remove.
14582 (__arm_vqdmullbq_n_s16): Remove.
14583 (__arm_vqdmulltq_s32): Remove.
14584 (__arm_vqdmulltq_n_s32): Remove.
14585 (__arm_vqdmullbq_s32): Remove.
14586 (__arm_vqdmullbq_n_s32): Remove.
14587 (__arm_vqdmullbq_m_n_s32): Remove.
14588 (__arm_vqdmullbq_m_n_s16): Remove.
14589 (__arm_vqdmullbq_m_s32): Remove.
14590 (__arm_vqdmullbq_m_s16): Remove.
14591 (__arm_vqdmulltq_m_n_s32): Remove.
14592 (__arm_vqdmulltq_m_n_s16): Remove.
14593 (__arm_vqdmulltq_m_s32): Remove.
14594 (__arm_vqdmulltq_m_s16): Remove.
14595 (__arm_vqdmulltq): Remove.
14596 (__arm_vqdmullbq): Remove.
14597 (__arm_vqdmullbq_m): Remove.
14598 (__arm_vqdmulltq_m): Remove.
14600 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14602 * config/arm/iterators.md (MVE_VQDMULLxQ, MVE_VQDMULLxQ_M)
14603 (MVE_VQDMULLxQ_M_N, MVE_VQDMULLxQ_N): New.
14604 (mve_insn): Add vqdmullb, vqdmullt.
14605 (supf): Add VQDMULLBQ_S, VQDMULLBQ_M_S, VQDMULLBQ_M_N_S,
14606 VQDMULLBQ_N_S, VQDMULLTQ_S, VQDMULLTQ_M_S, VQDMULLTQ_M_N_S,
14608 * config/arm/mve.md (mve_vqdmullbq_n_s<mode>)
14609 (mve_vqdmulltq_n_s<mode>): Merge into ...
14610 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
14611 (mve_vqdmullbq_s<mode>, mve_vqdmulltq_s<mode>): Merge into ...
14612 (@mve_<mve_insn>q_<supf><mode>): ... this.
14613 (mve_vqdmullbq_m_n_s<mode>, mve_vqdmulltq_m_n_s<mode>): Merge into
14615 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
14616 (mve_vqdmullbq_m_s<mode>, mve_vqdmulltq_m_s<mode>): Merge into ...
14617 (@mve_<mve_insn>q_m_<supf><mode>): ... this.
14619 2023-05-12 Christophe Lyon <christophe.lyon@arm.com>
14621 * config/arm/arm-mve-builtins-shapes.cc (binary_widen_opt_n): New.
14622 * config/arm/arm-mve-builtins-shapes.h (binary_widen_opt_n): New.
14624 2023-05-12 Kito Cheng <kito.cheng@sifive.com>
14626 * common/config/riscv/riscv-common.cc (riscv_select_multilib_by_abi):
14627 Drop unused parameter.
14628 (riscv_select_multilib): Ditto.
14629 (riscv_compute_multilib): Update call site of
14630 riscv_select_multilib_by_abi and riscv_select_multilib_by_abi.
14632 2023-05-12 Juzhe Zhong <juzhe.zhong@rivai.ai>
14634 * config/riscv/autovec.md (vec_init<mode><vel>): New pattern.
14635 * config/riscv/riscv-protos.h (expand_vec_init): New function.
14636 * config/riscv/riscv-v.cc (class rvv_builder): New class.
14637 (rvv_builder::can_duplicate_repeating_sequence_p): New function.
14638 (rvv_builder::get_merged_repeating_sequence): Ditto.
14639 (expand_vector_init_insert_elems): Ditto.
14640 (expand_vec_init): Ditto.
14641 * config/riscv/vector-iterators.md: New attribute.
14643 2023-05-12 Haochen Gui <guihaoc@gcc.gnu.org>
14645 * config/rs6000/rs6000-builtins.def
14646 (__builtin_vsx_scalar_insert_exp): Replace bif-pattern from xsiexpdp
14648 (__builtin_vsx_scalar_insert_exp_dp): Replace bif-pattern from
14649 xsiexpdpf to xsiexpdpf_di.
14650 * config/rs6000/vsx.md (xsiexpdp): Rename to...
14651 (xsiexpdp_<mode>): ..., set the mode of second operand to GPR and
14652 replace TARGET_64BIT with TARGET_POWERPC64.
14653 (xsiexpdpf): Rename to...
14654 (xsiexpdpf_<mode>): ..., set the mode of second operand to GPR and
14655 replace TARGET_64BIT with TARGET_POWERPC64.
14657 2023-05-12 Haochen Gui <guihaoc@gcc.gnu.org>
14659 * config/rs6000/rs6000-builtins.def
14660 (__builtin_vsx_scalar_extract_sig): Set return type to const signed
14662 * config/rs6000/vsx.md (xsxsigdp): Replace TARGET_64BIT with
14665 2023-05-12 Haochen Gui <guihaoc@gcc.gnu.org>
14667 * config/rs6000/rs6000-builtins.def
14668 (__builtin_vsx_scalar_extract_exp): Set return type to const signed
14669 int and set its bif-pattern to xsxexpdp_si, move it from power9-64
14671 * config/rs6000/vsx.md (xsxexpdp): Rename to ...
14672 (xsxexpdp_<mode>): ..., set mode of operand 0 to GPR and remove
14673 TARGET_64BIT check.
14674 * doc/extend.texi (scalar_extract_exp): Remove 64-bit environment
14675 requirement when it has a 64-bit argument.
14677 2023-05-12 Pan Li <pan2.li@intel.com>
14678 Richard Sandiford <richard.sandiford@arm.com>
14679 Richard Biener <rguenther@suse.de>
14680 Jakub Jelinek <jakub@redhat.com>
14682 * mux-utils.h: Add overload operator == and != for pointer_mux.
14683 * var-tracking.cc: Included mux-utils.h for pointer_tmux.
14684 (decl_or_value): Changed from void * to pointer_mux<tree_node, rtx_def>.
14685 (dv_is_decl_p): Reconciled to the new type, aka pointer_mux.
14686 (dv_as_decl): Ditto.
14687 (dv_as_opaque): Removed due to unnecessary.
14688 (struct variable_hasher): Take decl_or_value as compare_type.
14689 (variable_hasher::equal): Diito.
14690 (dv_from_decl): Reconciled to the new type, aka pointer_mux.
14691 (dv_from_value): Ditto.
14692 (attrs_list_member): Ditto.
14693 (vars_copy): Ditto.
14694 (var_reg_decl_set): Ditto.
14695 (var_reg_delete_and_set): Ditto.
14696 (find_loc_in_1pdv): Ditto.
14697 (canonicalize_values_star): Ditto.
14698 (variable_post_merge_new_vals): Ditto.
14699 (dump_onepart_variable_differences): Ditto.
14700 (variable_different_p): Ditto.
14701 (set_slot_part): Ditto.
14702 (clobber_slot_part): Ditto.
14703 (clobber_variable_part): Ditto.
14705 2023-05-11 mtsamis <manolis.tsamis@vrull.eu>
14707 * match.pd: simplify vector shift + bit_and + multiply.
14709 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
14711 * config/arm/arm-mve-builtins-base.cc (vmlaq, vmlasq, vqdmlahq)
14712 (vqdmlashq, vqrdmlahq, vqrdmlashq): New.
14713 * config/arm/arm-mve-builtins-base.def (vmlaq, vmlasq, vqdmlahq)
14714 (vqdmlashq, vqrdmlahq, vqrdmlashq): New.
14715 * config/arm/arm-mve-builtins-base.h (vmlaq, vmlasq, vqdmlahq)
14716 (vqdmlashq, vqrdmlahq, vqrdmlashq): New.
14717 * config/arm/arm-mve-builtins.cc
14718 (function_instance::has_inactive_argument): Handle vmlaq, vmlasq,
14719 vqdmlahq, vqdmlashq, vqrdmlahq, vqrdmlashq.
14720 * config/arm/arm_mve.h (vqrdmlashq): Remove.
14721 (vqrdmlahq): Remove.
14722 (vqdmlashq): Remove.
14723 (vqdmlahq): Remove.
14727 (vmlasq_m): Remove.
14728 (vqdmlashq_m): Remove.
14729 (vqdmlahq_m): Remove.
14730 (vqrdmlahq_m): Remove.
14731 (vqrdmlashq_m): Remove.
14732 (vmlasq_n_u8): Remove.
14733 (vmlaq_n_u8): Remove.
14734 (vqrdmlashq_n_s8): Remove.
14735 (vqrdmlahq_n_s8): Remove.
14736 (vqdmlahq_n_s8): Remove.
14737 (vqdmlashq_n_s8): Remove.
14738 (vmlasq_n_s8): Remove.
14739 (vmlaq_n_s8): Remove.
14740 (vmlasq_n_u16): Remove.
14741 (vmlaq_n_u16): Remove.
14742 (vqrdmlashq_n_s16): Remove.
14743 (vqrdmlahq_n_s16): Remove.
14744 (vqdmlashq_n_s16): Remove.
14745 (vqdmlahq_n_s16): Remove.
14746 (vmlasq_n_s16): Remove.
14747 (vmlaq_n_s16): Remove.
14748 (vmlasq_n_u32): Remove.
14749 (vmlaq_n_u32): Remove.
14750 (vqrdmlashq_n_s32): Remove.
14751 (vqrdmlahq_n_s32): Remove.
14752 (vqdmlashq_n_s32): Remove.
14753 (vqdmlahq_n_s32): Remove.
14754 (vmlasq_n_s32): Remove.
14755 (vmlaq_n_s32): Remove.
14756 (vmlaq_m_n_s8): Remove.
14757 (vmlaq_m_n_s32): Remove.
14758 (vmlaq_m_n_s16): Remove.
14759 (vmlaq_m_n_u8): Remove.
14760 (vmlaq_m_n_u32): Remove.
14761 (vmlaq_m_n_u16): Remove.
14762 (vmlasq_m_n_s8): Remove.
14763 (vmlasq_m_n_s32): Remove.
14764 (vmlasq_m_n_s16): Remove.
14765 (vmlasq_m_n_u8): Remove.
14766 (vmlasq_m_n_u32): Remove.
14767 (vmlasq_m_n_u16): Remove.
14768 (vqdmlashq_m_n_s8): Remove.
14769 (vqdmlashq_m_n_s32): Remove.
14770 (vqdmlashq_m_n_s16): Remove.
14771 (vqdmlahq_m_n_s8): Remove.
14772 (vqdmlahq_m_n_s32): Remove.
14773 (vqdmlahq_m_n_s16): Remove.
14774 (vqrdmlahq_m_n_s8): Remove.
14775 (vqrdmlahq_m_n_s32): Remove.
14776 (vqrdmlahq_m_n_s16): Remove.
14777 (vqrdmlashq_m_n_s8): Remove.
14778 (vqrdmlashq_m_n_s32): Remove.
14779 (vqrdmlashq_m_n_s16): Remove.
14780 (__arm_vmlasq_n_u8): Remove.
14781 (__arm_vmlaq_n_u8): Remove.
14782 (__arm_vqrdmlashq_n_s8): Remove.
14783 (__arm_vqdmlashq_n_s8): Remove.
14784 (__arm_vqrdmlahq_n_s8): Remove.
14785 (__arm_vqdmlahq_n_s8): Remove.
14786 (__arm_vmlasq_n_s8): Remove.
14787 (__arm_vmlaq_n_s8): Remove.
14788 (__arm_vmlasq_n_u16): Remove.
14789 (__arm_vmlaq_n_u16): Remove.
14790 (__arm_vqrdmlashq_n_s16): Remove.
14791 (__arm_vqdmlashq_n_s16): Remove.
14792 (__arm_vqrdmlahq_n_s16): Remove.
14793 (__arm_vqdmlahq_n_s16): Remove.
14794 (__arm_vmlasq_n_s16): Remove.
14795 (__arm_vmlaq_n_s16): Remove.
14796 (__arm_vmlasq_n_u32): Remove.
14797 (__arm_vmlaq_n_u32): Remove.
14798 (__arm_vqrdmlashq_n_s32): Remove.
14799 (__arm_vqdmlashq_n_s32): Remove.
14800 (__arm_vqrdmlahq_n_s32): Remove.
14801 (__arm_vqdmlahq_n_s32): Remove.
14802 (__arm_vmlasq_n_s32): Remove.
14803 (__arm_vmlaq_n_s32): Remove.
14804 (__arm_vmlaq_m_n_s8): Remove.
14805 (__arm_vmlaq_m_n_s32): Remove.
14806 (__arm_vmlaq_m_n_s16): Remove.
14807 (__arm_vmlaq_m_n_u8): Remove.
14808 (__arm_vmlaq_m_n_u32): Remove.
14809 (__arm_vmlaq_m_n_u16): Remove.
14810 (__arm_vmlasq_m_n_s8): Remove.
14811 (__arm_vmlasq_m_n_s32): Remove.
14812 (__arm_vmlasq_m_n_s16): Remove.
14813 (__arm_vmlasq_m_n_u8): Remove.
14814 (__arm_vmlasq_m_n_u32): Remove.
14815 (__arm_vmlasq_m_n_u16): Remove.
14816 (__arm_vqdmlahq_m_n_s8): Remove.
14817 (__arm_vqdmlahq_m_n_s32): Remove.
14818 (__arm_vqdmlahq_m_n_s16): Remove.
14819 (__arm_vqrdmlahq_m_n_s8): Remove.
14820 (__arm_vqrdmlahq_m_n_s32): Remove.
14821 (__arm_vqrdmlahq_m_n_s16): Remove.
14822 (__arm_vqrdmlashq_m_n_s8): Remove.
14823 (__arm_vqrdmlashq_m_n_s32): Remove.
14824 (__arm_vqrdmlashq_m_n_s16): Remove.
14825 (__arm_vqdmlashq_m_n_s8): Remove.
14826 (__arm_vqdmlashq_m_n_s16): Remove.
14827 (__arm_vqdmlashq_m_n_s32): Remove.
14828 (__arm_vmlasq): Remove.
14829 (__arm_vmlaq): Remove.
14830 (__arm_vqrdmlashq): Remove.
14831 (__arm_vqdmlashq): Remove.
14832 (__arm_vqrdmlahq): Remove.
14833 (__arm_vqdmlahq): Remove.
14834 (__arm_vmlaq_m): Remove.
14835 (__arm_vmlasq_m): Remove.
14836 (__arm_vqdmlahq_m): Remove.
14837 (__arm_vqrdmlahq_m): Remove.
14838 (__arm_vqrdmlashq_m): Remove.
14839 (__arm_vqdmlashq_m): Remove.
14841 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
14843 * config/arm/iterators.md (MVE_VMLxQ_N): New.
14844 (mve_insn): Add vmla, vmlas, vqdmlah, vqdmlash, vqrdmlah,
14846 (supf): Add VQDMLAHQ_N_S, VQDMLASHQ_N_S, VQRDMLAHQ_N_S,
14848 * config/arm/mve.md (mve_vmlaq_n_<supf><mode>)
14849 (mve_vmlasq_n_<supf><mode>, mve_vqdmlahq_n_<supf><mode>)
14850 (mve_vqdmlashq_n_<supf><mode>, mve_vqrdmlahq_n_<supf><mode>)
14851 (mve_vqrdmlashq_n_<supf><mode>): Merge into ...
14852 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
14854 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
14856 * config/arm/arm-mve-builtins-shapes.cc (ternary_n): New.
14857 * config/arm/arm-mve-builtins-shapes.h (ternary_n): New.
14859 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
14861 * config/arm/arm-mve-builtins-base.cc (vqdmladhq, vqdmladhxq)
14862 (vqdmlsdhq, vqdmlsdhxq, vqrdmladhq, vqrdmladhxq, vqrdmlsdhq)
14863 (vqrdmlsdhxq): New.
14864 * config/arm/arm-mve-builtins-base.def (vqdmladhq, vqdmladhxq)
14865 (vqdmlsdhq, vqdmlsdhxq, vqrdmladhq, vqrdmladhxq, vqrdmlsdhq)
14866 (vqrdmlsdhxq): New.
14867 * config/arm/arm-mve-builtins-base.h (vqdmladhq, vqdmladhxq)
14868 (vqdmlsdhq, vqdmlsdhxq, vqrdmladhq, vqrdmladhxq, vqrdmlsdhq)
14869 (vqrdmlsdhxq): New.
14870 * config/arm/arm-mve-builtins.cc
14871 (function_instance::has_inactive_argument): Handle vqrdmladhq,
14872 vqrdmladhxq, vqrdmlsdhq, vqrdmlsdhxq vqdmladhq, vqdmladhxq,
14873 vqdmlsdhq, vqdmlsdhxq.
14874 * config/arm/arm_mve.h (vqrdmlsdhxq): Remove.
14875 (vqrdmlsdhq): Remove.
14876 (vqrdmladhxq): Remove.
14877 (vqrdmladhq): Remove.
14878 (vqdmlsdhxq): Remove.
14879 (vqdmlsdhq): Remove.
14880 (vqdmladhxq): Remove.
14881 (vqdmladhq): Remove.
14882 (vqdmladhq_m): Remove.
14883 (vqdmladhxq_m): Remove.
14884 (vqdmlsdhq_m): Remove.
14885 (vqdmlsdhxq_m): Remove.
14886 (vqrdmladhq_m): Remove.
14887 (vqrdmladhxq_m): Remove.
14888 (vqrdmlsdhq_m): Remove.
14889 (vqrdmlsdhxq_m): Remove.
14890 (vqrdmlsdhxq_s8): Remove.
14891 (vqrdmlsdhq_s8): Remove.
14892 (vqrdmladhxq_s8): Remove.
14893 (vqrdmladhq_s8): Remove.
14894 (vqdmlsdhxq_s8): Remove.
14895 (vqdmlsdhq_s8): Remove.
14896 (vqdmladhxq_s8): Remove.
14897 (vqdmladhq_s8): Remove.
14898 (vqrdmlsdhxq_s16): Remove.
14899 (vqrdmlsdhq_s16): Remove.
14900 (vqrdmladhxq_s16): Remove.
14901 (vqrdmladhq_s16): Remove.
14902 (vqdmlsdhxq_s16): Remove.
14903 (vqdmlsdhq_s16): Remove.
14904 (vqdmladhxq_s16): Remove.
14905 (vqdmladhq_s16): Remove.
14906 (vqrdmlsdhxq_s32): Remove.
14907 (vqrdmlsdhq_s32): Remove.
14908 (vqrdmladhxq_s32): Remove.
14909 (vqrdmladhq_s32): Remove.
14910 (vqdmlsdhxq_s32): Remove.
14911 (vqdmlsdhq_s32): Remove.
14912 (vqdmladhxq_s32): Remove.
14913 (vqdmladhq_s32): Remove.
14914 (vqdmladhq_m_s8): Remove.
14915 (vqdmladhq_m_s32): Remove.
14916 (vqdmladhq_m_s16): Remove.
14917 (vqdmladhxq_m_s8): Remove.
14918 (vqdmladhxq_m_s32): Remove.
14919 (vqdmladhxq_m_s16): Remove.
14920 (vqdmlsdhq_m_s8): Remove.
14921 (vqdmlsdhq_m_s32): Remove.
14922 (vqdmlsdhq_m_s16): Remove.
14923 (vqdmlsdhxq_m_s8): Remove.
14924 (vqdmlsdhxq_m_s32): Remove.
14925 (vqdmlsdhxq_m_s16): Remove.
14926 (vqrdmladhq_m_s8): Remove.
14927 (vqrdmladhq_m_s32): Remove.
14928 (vqrdmladhq_m_s16): Remove.
14929 (vqrdmladhxq_m_s8): Remove.
14930 (vqrdmladhxq_m_s32): Remove.
14931 (vqrdmladhxq_m_s16): Remove.
14932 (vqrdmlsdhq_m_s8): Remove.
14933 (vqrdmlsdhq_m_s32): Remove.
14934 (vqrdmlsdhq_m_s16): Remove.
14935 (vqrdmlsdhxq_m_s8): Remove.
14936 (vqrdmlsdhxq_m_s32): Remove.
14937 (vqrdmlsdhxq_m_s16): Remove.
14938 (__arm_vqrdmlsdhxq_s8): Remove.
14939 (__arm_vqrdmlsdhq_s8): Remove.
14940 (__arm_vqrdmladhxq_s8): Remove.
14941 (__arm_vqrdmladhq_s8): Remove.
14942 (__arm_vqdmlsdhxq_s8): Remove.
14943 (__arm_vqdmlsdhq_s8): Remove.
14944 (__arm_vqdmladhxq_s8): Remove.
14945 (__arm_vqdmladhq_s8): Remove.
14946 (__arm_vqrdmlsdhxq_s16): Remove.
14947 (__arm_vqrdmlsdhq_s16): Remove.
14948 (__arm_vqrdmladhxq_s16): Remove.
14949 (__arm_vqrdmladhq_s16): Remove.
14950 (__arm_vqdmlsdhxq_s16): Remove.
14951 (__arm_vqdmlsdhq_s16): Remove.
14952 (__arm_vqdmladhxq_s16): Remove.
14953 (__arm_vqdmladhq_s16): Remove.
14954 (__arm_vqrdmlsdhxq_s32): Remove.
14955 (__arm_vqrdmlsdhq_s32): Remove.
14956 (__arm_vqrdmladhxq_s32): Remove.
14957 (__arm_vqrdmladhq_s32): Remove.
14958 (__arm_vqdmlsdhxq_s32): Remove.
14959 (__arm_vqdmlsdhq_s32): Remove.
14960 (__arm_vqdmladhxq_s32): Remove.
14961 (__arm_vqdmladhq_s32): Remove.
14962 (__arm_vqdmladhq_m_s8): Remove.
14963 (__arm_vqdmladhq_m_s32): Remove.
14964 (__arm_vqdmladhq_m_s16): Remove.
14965 (__arm_vqdmladhxq_m_s8): Remove.
14966 (__arm_vqdmladhxq_m_s32): Remove.
14967 (__arm_vqdmladhxq_m_s16): Remove.
14968 (__arm_vqdmlsdhq_m_s8): Remove.
14969 (__arm_vqdmlsdhq_m_s32): Remove.
14970 (__arm_vqdmlsdhq_m_s16): Remove.
14971 (__arm_vqdmlsdhxq_m_s8): Remove.
14972 (__arm_vqdmlsdhxq_m_s32): Remove.
14973 (__arm_vqdmlsdhxq_m_s16): Remove.
14974 (__arm_vqrdmladhq_m_s8): Remove.
14975 (__arm_vqrdmladhq_m_s32): Remove.
14976 (__arm_vqrdmladhq_m_s16): Remove.
14977 (__arm_vqrdmladhxq_m_s8): Remove.
14978 (__arm_vqrdmladhxq_m_s32): Remove.
14979 (__arm_vqrdmladhxq_m_s16): Remove.
14980 (__arm_vqrdmlsdhq_m_s8): Remove.
14981 (__arm_vqrdmlsdhq_m_s32): Remove.
14982 (__arm_vqrdmlsdhq_m_s16): Remove.
14983 (__arm_vqrdmlsdhxq_m_s8): Remove.
14984 (__arm_vqrdmlsdhxq_m_s32): Remove.
14985 (__arm_vqrdmlsdhxq_m_s16): Remove.
14986 (__arm_vqrdmlsdhxq): Remove.
14987 (__arm_vqrdmlsdhq): Remove.
14988 (__arm_vqrdmladhxq): Remove.
14989 (__arm_vqrdmladhq): Remove.
14990 (__arm_vqdmlsdhxq): Remove.
14991 (__arm_vqdmlsdhq): Remove.
14992 (__arm_vqdmladhxq): Remove.
14993 (__arm_vqdmladhq): Remove.
14994 (__arm_vqdmladhq_m): Remove.
14995 (__arm_vqdmladhxq_m): Remove.
14996 (__arm_vqdmlsdhq_m): Remove.
14997 (__arm_vqdmlsdhxq_m): Remove.
14998 (__arm_vqrdmladhq_m): Remove.
14999 (__arm_vqrdmladhxq_m): Remove.
15000 (__arm_vqrdmlsdhq_m): Remove.
15001 (__arm_vqrdmlsdhxq_m): Remove.
15003 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15005 * config/arm/iterators.md (MVE_VQxDMLxDHxQ_S): New.
15006 (mve_insn): Add vqdmladh, vqdmladhx, vqdmlsdh, vqdmlsdhx,
15007 vqrdmladh, vqrdmladhx, vqrdmlsdh, vqrdmlsdhx.
15008 (supf): Add VQDMLADHQ_S, VQDMLADHXQ_S, VQDMLSDHQ_S, VQDMLSDHXQ_S,
15009 VQRDMLADHQ_S,VQRDMLADHXQ_S, VQRDMLSDHQ_S, VQRDMLSDHXQ_S.
15010 * config/arm/mve.md (mve_vqrdmladhq_s<mode>)
15011 (mve_vqrdmladhxq_s<mode>, mve_vqrdmlsdhq_s<mode>)
15012 (mve_vqrdmlsdhxq_s<mode>, mve_vqdmlsdhxq_s<mode>)
15013 (mve_vqdmlsdhq_s<mode>, mve_vqdmladhxq_s<mode>)
15014 (mve_vqdmladhq_s<mode>): Merge into ...
15015 (@mve_<mve_insn>q_<supf><mode>): ... this.
15017 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15019 * config/arm/arm-mve-builtins-shapes.cc (ternary): New.
15020 * config/arm/arm-mve-builtins-shapes.h (ternary): New.
15022 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15024 * config/arm/arm-mve-builtins-base.cc (vmlaldavaq, vmlaldavaxq)
15025 (vmlsldavaq, vmlsldavaxq): New.
15026 * config/arm/arm-mve-builtins-base.def (vmlaldavaq, vmlaldavaxq)
15027 (vmlsldavaq, vmlsldavaxq): New.
15028 * config/arm/arm-mve-builtins-base.h (vmlaldavaq, vmlaldavaxq)
15029 (vmlsldavaq, vmlsldavaxq): New.
15030 * config/arm/arm_mve.h (vmlaldavaq): Remove.
15031 (vmlaldavaxq): Remove.
15032 (vmlsldavaq): Remove.
15033 (vmlsldavaxq): Remove.
15034 (vmlaldavaq_p): Remove.
15035 (vmlaldavaxq_p): Remove.
15036 (vmlsldavaq_p): Remove.
15037 (vmlsldavaxq_p): Remove.
15038 (vmlaldavaq_s16): Remove.
15039 (vmlaldavaxq_s16): Remove.
15040 (vmlsldavaq_s16): Remove.
15041 (vmlsldavaxq_s16): Remove.
15042 (vmlaldavaq_u16): Remove.
15043 (vmlaldavaq_s32): Remove.
15044 (vmlaldavaxq_s32): Remove.
15045 (vmlsldavaq_s32): Remove.
15046 (vmlsldavaxq_s32): Remove.
15047 (vmlaldavaq_u32): Remove.
15048 (vmlaldavaq_p_s32): Remove.
15049 (vmlaldavaq_p_s16): Remove.
15050 (vmlaldavaq_p_u32): Remove.
15051 (vmlaldavaq_p_u16): Remove.
15052 (vmlaldavaxq_p_s32): Remove.
15053 (vmlaldavaxq_p_s16): Remove.
15054 (vmlsldavaq_p_s32): Remove.
15055 (vmlsldavaq_p_s16): Remove.
15056 (vmlsldavaxq_p_s32): Remove.
15057 (vmlsldavaxq_p_s16): Remove.
15058 (__arm_vmlaldavaq_s16): Remove.
15059 (__arm_vmlaldavaxq_s16): Remove.
15060 (__arm_vmlsldavaq_s16): Remove.
15061 (__arm_vmlsldavaxq_s16): Remove.
15062 (__arm_vmlaldavaq_u16): Remove.
15063 (__arm_vmlaldavaq_s32): Remove.
15064 (__arm_vmlaldavaxq_s32): Remove.
15065 (__arm_vmlsldavaq_s32): Remove.
15066 (__arm_vmlsldavaxq_s32): Remove.
15067 (__arm_vmlaldavaq_u32): Remove.
15068 (__arm_vmlaldavaq_p_s32): Remove.
15069 (__arm_vmlaldavaq_p_s16): Remove.
15070 (__arm_vmlaldavaq_p_u32): Remove.
15071 (__arm_vmlaldavaq_p_u16): Remove.
15072 (__arm_vmlaldavaxq_p_s32): Remove.
15073 (__arm_vmlaldavaxq_p_s16): Remove.
15074 (__arm_vmlsldavaq_p_s32): Remove.
15075 (__arm_vmlsldavaq_p_s16): Remove.
15076 (__arm_vmlsldavaxq_p_s32): Remove.
15077 (__arm_vmlsldavaxq_p_s16): Remove.
15078 (__arm_vmlaldavaq): Remove.
15079 (__arm_vmlaldavaxq): Remove.
15080 (__arm_vmlsldavaq): Remove.
15081 (__arm_vmlsldavaxq): Remove.
15082 (__arm_vmlaldavaq_p): Remove.
15083 (__arm_vmlaldavaxq_p): Remove.
15084 (__arm_vmlsldavaq_p): Remove.
15085 (__arm_vmlsldavaxq_p): Remove.
15087 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15089 * config/arm/iterators.md (MVE_VMLxLDAVAxQ, MVE_VMLxLDAVAxQ_P):
15091 (mve_insn): Add vmlaldava, vmlaldavax, vmlsldava, vmlsldavax.
15092 (supf): Add VMLALDAVAXQ_P_S, VMLALDAVAXQ_S, VMLSLDAVAQ_P_S,
15093 VMLSLDAVAQ_S, VMLSLDAVAXQ_P_S, VMLSLDAVAXQ_S.
15094 * config/arm/mve.md (mve_vmlaldavaq_<supf><mode>)
15095 (mve_vmlsldavaq_s<mode>, mve_vmlsldavaxq_s<mode>)
15096 (mve_vmlaldavaxq_s<mode>): Merge into ...
15097 (@mve_<mve_insn>q_<supf><mode>): ... this.
15098 (mve_vmlaldavaq_p_<supf><mode>, mve_vmlaldavaxq_p_<supf><mode>)
15099 (mve_vmlsldavaq_p_s<mode>, mve_vmlsldavaxq_p_s<mode>): Merge into
15101 (@mve_<mve_insn>q_p_<supf><mode>): ... this.
15103 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15105 * config/arm/arm-mve-builtins-shapes.cc (binary_acca_int64): New.
15106 * config/arm/arm-mve-builtins-shapes.h (binary_acca_int64): New.
15108 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15110 * config/arm/arm-mve-builtins-base.cc (vrmlaldavhq, vrmlaldavhxq)
15111 (vrmlsldavhq, vrmlsldavhxq): New.
15112 * config/arm/arm-mve-builtins-base.def (vrmlaldavhq, vrmlaldavhxq)
15113 (vrmlsldavhq, vrmlsldavhxq): New.
15114 * config/arm/arm-mve-builtins-base.h (vrmlaldavhq, vrmlaldavhxq)
15115 (vrmlsldavhq, vrmlsldavhxq): New.
15116 * config/arm/arm-mve-builtins-functions.h
15117 (unspec_mve_function_exact_insn_pred_p): Handle vrmlaldavhq,
15118 vrmlaldavhxq, vrmlsldavhq, vrmlsldavhxq.
15119 * config/arm/arm_mve.h (vrmlaldavhq): Remove.
15120 (vrmlsldavhxq): Remove.
15121 (vrmlsldavhq): Remove.
15122 (vrmlaldavhxq): Remove.
15123 (vrmlaldavhq_p): Remove.
15124 (vrmlaldavhxq_p): Remove.
15125 (vrmlsldavhq_p): Remove.
15126 (vrmlsldavhxq_p): Remove.
15127 (vrmlaldavhq_u32): Remove.
15128 (vrmlsldavhxq_s32): Remove.
15129 (vrmlsldavhq_s32): Remove.
15130 (vrmlaldavhxq_s32): Remove.
15131 (vrmlaldavhq_s32): Remove.
15132 (vrmlaldavhq_p_s32): Remove.
15133 (vrmlaldavhxq_p_s32): Remove.
15134 (vrmlsldavhq_p_s32): Remove.
15135 (vrmlsldavhxq_p_s32): Remove.
15136 (vrmlaldavhq_p_u32): Remove.
15137 (__arm_vrmlaldavhq_u32): Remove.
15138 (__arm_vrmlsldavhxq_s32): Remove.
15139 (__arm_vrmlsldavhq_s32): Remove.
15140 (__arm_vrmlaldavhxq_s32): Remove.
15141 (__arm_vrmlaldavhq_s32): Remove.
15142 (__arm_vrmlaldavhq_p_s32): Remove.
15143 (__arm_vrmlaldavhxq_p_s32): Remove.
15144 (__arm_vrmlsldavhq_p_s32): Remove.
15145 (__arm_vrmlsldavhxq_p_s32): Remove.
15146 (__arm_vrmlaldavhq_p_u32): Remove.
15147 (__arm_vrmlaldavhq): Remove.
15148 (__arm_vrmlsldavhxq): Remove.
15149 (__arm_vrmlsldavhq): Remove.
15150 (__arm_vrmlaldavhxq): Remove.
15151 (__arm_vrmlaldavhq_p): Remove.
15152 (__arm_vrmlaldavhxq_p): Remove.
15153 (__arm_vrmlsldavhq_p): Remove.
15154 (__arm_vrmlsldavhxq_p): Remove.
15156 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15158 * config/arm/iterators.md (MVE_VRMLxLDAVxQ, MVE_VRMLxLDAVHxQ_P):
15160 (mve_insn): Add vrmlaldavh, vrmlaldavhx, vrmlsldavh, vrmlsldavhx.
15161 (supf): Add VRMLALDAVHXQ_P_S, VRMLALDAVHXQ_S, VRMLSLDAVHQ_P_S,
15162 VRMLSLDAVHQ_S, VRMLSLDAVHXQ_P_S, VRMLSLDAVHXQ_S.
15163 * config/arm/mve.md (mve_vrmlaldavhxq_sv4si)
15164 (mve_vrmlsldavhq_sv4si, mve_vrmlsldavhxq_sv4si)
15165 (mve_vrmlaldavhq_<supf>v4si): Merge into ...
15166 (@mve_<mve_insn>q_<supf>v4si): ... this.
15167 (mve_vrmlaldavhxq_p_sv4si, mve_vrmlsldavhq_p_sv4si)
15168 (mve_vrmlsldavhxq_p_sv4si, mve_vrmlaldavhq_p_<supf>v4si): Merge
15170 (@mve_<mve_insn>q_p_<supf>v4si): ... this.
15172 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15174 * config/arm/arm-mve-builtins-base.cc (vmlaldavq, vmlaldavxq)
15175 (vmlsldavq, vmlsldavxq): New.
15176 * config/arm/arm-mve-builtins-base.def (vmlaldavq, vmlaldavxq)
15177 (vmlsldavq, vmlsldavxq): New.
15178 * config/arm/arm-mve-builtins-base.h (vmlaldavq, vmlaldavxq)
15179 (vmlsldavq, vmlsldavxq): New.
15180 * config/arm/arm_mve.h (vmlaldavq): Remove.
15181 (vmlsldavxq): Remove.
15182 (vmlsldavq): Remove.
15183 (vmlaldavxq): Remove.
15184 (vmlaldavq_p): Remove.
15185 (vmlaldavxq_p): Remove.
15186 (vmlsldavq_p): Remove.
15187 (vmlsldavxq_p): Remove.
15188 (vmlaldavq_u16): Remove.
15189 (vmlsldavxq_s16): Remove.
15190 (vmlsldavq_s16): Remove.
15191 (vmlaldavxq_s16): Remove.
15192 (vmlaldavq_s16): Remove.
15193 (vmlaldavq_u32): Remove.
15194 (vmlsldavxq_s32): Remove.
15195 (vmlsldavq_s32): Remove.
15196 (vmlaldavxq_s32): Remove.
15197 (vmlaldavq_s32): Remove.
15198 (vmlaldavq_p_s16): Remove.
15199 (vmlaldavxq_p_s16): Remove.
15200 (vmlsldavq_p_s16): Remove.
15201 (vmlsldavxq_p_s16): Remove.
15202 (vmlaldavq_p_u16): Remove.
15203 (vmlaldavq_p_s32): Remove.
15204 (vmlaldavxq_p_s32): Remove.
15205 (vmlsldavq_p_s32): Remove.
15206 (vmlsldavxq_p_s32): Remove.
15207 (vmlaldavq_p_u32): Remove.
15208 (__arm_vmlaldavq_u16): Remove.
15209 (__arm_vmlsldavxq_s16): Remove.
15210 (__arm_vmlsldavq_s16): Remove.
15211 (__arm_vmlaldavxq_s16): Remove.
15212 (__arm_vmlaldavq_s16): Remove.
15213 (__arm_vmlaldavq_u32): Remove.
15214 (__arm_vmlsldavxq_s32): Remove.
15215 (__arm_vmlsldavq_s32): Remove.
15216 (__arm_vmlaldavxq_s32): Remove.
15217 (__arm_vmlaldavq_s32): Remove.
15218 (__arm_vmlaldavq_p_s16): Remove.
15219 (__arm_vmlaldavxq_p_s16): Remove.
15220 (__arm_vmlsldavq_p_s16): Remove.
15221 (__arm_vmlsldavxq_p_s16): Remove.
15222 (__arm_vmlaldavq_p_u16): Remove.
15223 (__arm_vmlaldavq_p_s32): Remove.
15224 (__arm_vmlaldavxq_p_s32): Remove.
15225 (__arm_vmlsldavq_p_s32): Remove.
15226 (__arm_vmlsldavxq_p_s32): Remove.
15227 (__arm_vmlaldavq_p_u32): Remove.
15228 (__arm_vmlaldavq): Remove.
15229 (__arm_vmlsldavxq): Remove.
15230 (__arm_vmlsldavq): Remove.
15231 (__arm_vmlaldavxq): Remove.
15232 (__arm_vmlaldavq_p): Remove.
15233 (__arm_vmlaldavxq_p): Remove.
15234 (__arm_vmlsldavq_p): Remove.
15235 (__arm_vmlsldavxq_p): Remove.
15237 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15239 * config/arm/iterators.md (MVE_VMLxLDAVxQ, MVE_VMLxLDAVxQ_P): New.
15240 (mve_insn): Add vmlaldav, vmlaldavx, vmlsldav, vmlsldavx.
15241 (supf): Add VMLALDAVXQ_S, VMLSLDAVQ_S, VMLSLDAVXQ_S,
15242 VMLALDAVXQ_P_S, VMLSLDAVQ_P_S, VMLSLDAVXQ_P_S.
15243 * config/arm/mve.md (mve_vmlaldavq_<supf><mode>)
15244 (mve_vmlaldavxq_s<mode>, mve_vmlsldavq_s<mode>)
15245 (mve_vmlsldavxq_s<mode>): Merge into ...
15246 (@mve_<mve_insn>q_<supf><mode>): ... this.
15247 (mve_vmlaldavq_p_<supf><mode>, mve_vmlaldavxq_p_s<mode>)
15248 (mve_vmlsldavq_p_s<mode>, mve_vmlsldavxq_p_s<mode>): Merge into
15250 (@mve_<mve_insn>q_p_<supf><mode>): ... this.
15252 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15254 * config/arm/arm-mve-builtins-shapes.cc (binary_acc_int64): New.
15255 * config/arm/arm-mve-builtins-shapes.h (binary_acc_int64): New.
15257 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15259 * config/arm/arm-mve-builtins-base.cc (vabavq): New.
15260 * config/arm/arm-mve-builtins-base.def (vabavq): New.
15261 * config/arm/arm-mve-builtins-base.h (vabavq): New.
15262 * config/arm/arm_mve.h (vabavq): Remove.
15263 (vabavq_p): Remove.
15264 (vabavq_s8): Remove.
15265 (vabavq_s16): Remove.
15266 (vabavq_s32): Remove.
15267 (vabavq_u8): Remove.
15268 (vabavq_u16): Remove.
15269 (vabavq_u32): Remove.
15270 (vabavq_p_s8): Remove.
15271 (vabavq_p_u8): Remove.
15272 (vabavq_p_s16): Remove.
15273 (vabavq_p_u16): Remove.
15274 (vabavq_p_s32): Remove.
15275 (vabavq_p_u32): Remove.
15276 (__arm_vabavq_s8): Remove.
15277 (__arm_vabavq_s16): Remove.
15278 (__arm_vabavq_s32): Remove.
15279 (__arm_vabavq_u8): Remove.
15280 (__arm_vabavq_u16): Remove.
15281 (__arm_vabavq_u32): Remove.
15282 (__arm_vabavq_p_s8): Remove.
15283 (__arm_vabavq_p_u8): Remove.
15284 (__arm_vabavq_p_s16): Remove.
15285 (__arm_vabavq_p_u16): Remove.
15286 (__arm_vabavq_p_s32): Remove.
15287 (__arm_vabavq_p_u32): Remove.
15288 (__arm_vabavq): Remove.
15289 (__arm_vabavq_p): Remove.
15291 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15293 * config/arm/iterators.md (mve_insn): Add vabav.
15294 * config/arm/mve.md (mve_vabavq_<supf><mode>): Rename into ...
15295 (@mve_<mve_insn>q_<supf><mode>): ... this,.
15296 (mve_vabavq_p_<supf><mode>): Rename into ...
15297 (@mve_<mve_insn>q_p_<supf><mode>): ... this,.
15299 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15301 * config/arm/arm-mve-builtins-base.cc (vmladavaxq, vmladavaq)
15302 (vmlsdavaq, vmlsdavaxq): New.
15303 * config/arm/arm-mve-builtins-base.def (vmladavaxq, vmladavaq)
15304 (vmlsdavaq, vmlsdavaxq): New.
15305 * config/arm/arm-mve-builtins-base.h (vmladavaxq, vmladavaq)
15306 (vmlsdavaq, vmlsdavaxq): New.
15307 * config/arm/arm_mve.h (vmladavaq): Remove.
15308 (vmlsdavaxq): Remove.
15309 (vmlsdavaq): Remove.
15310 (vmladavaxq): Remove.
15311 (vmladavaq_p): Remove.
15312 (vmladavaxq_p): Remove.
15313 (vmlsdavaq_p): Remove.
15314 (vmlsdavaxq_p): Remove.
15315 (vmladavaq_u8): Remove.
15316 (vmlsdavaxq_s8): Remove.
15317 (vmlsdavaq_s8): Remove.
15318 (vmladavaxq_s8): Remove.
15319 (vmladavaq_s8): Remove.
15320 (vmladavaq_u16): Remove.
15321 (vmlsdavaxq_s16): Remove.
15322 (vmlsdavaq_s16): Remove.
15323 (vmladavaxq_s16): Remove.
15324 (vmladavaq_s16): Remove.
15325 (vmladavaq_u32): Remove.
15326 (vmlsdavaxq_s32): Remove.
15327 (vmlsdavaq_s32): Remove.
15328 (vmladavaxq_s32): Remove.
15329 (vmladavaq_s32): Remove.
15330 (vmladavaq_p_s8): Remove.
15331 (vmladavaq_p_s32): Remove.
15332 (vmladavaq_p_s16): Remove.
15333 (vmladavaq_p_u8): Remove.
15334 (vmladavaq_p_u32): Remove.
15335 (vmladavaq_p_u16): Remove.
15336 (vmladavaxq_p_s8): Remove.
15337 (vmladavaxq_p_s32): Remove.
15338 (vmladavaxq_p_s16): Remove.
15339 (vmlsdavaq_p_s8): Remove.
15340 (vmlsdavaq_p_s32): Remove.
15341 (vmlsdavaq_p_s16): Remove.
15342 (vmlsdavaxq_p_s8): Remove.
15343 (vmlsdavaxq_p_s32): Remove.
15344 (vmlsdavaxq_p_s16): Remove.
15345 (__arm_vmladavaq_u8): Remove.
15346 (__arm_vmlsdavaxq_s8): Remove.
15347 (__arm_vmlsdavaq_s8): Remove.
15348 (__arm_vmladavaxq_s8): Remove.
15349 (__arm_vmladavaq_s8): Remove.
15350 (__arm_vmladavaq_u16): Remove.
15351 (__arm_vmlsdavaxq_s16): Remove.
15352 (__arm_vmlsdavaq_s16): Remove.
15353 (__arm_vmladavaxq_s16): Remove.
15354 (__arm_vmladavaq_s16): Remove.
15355 (__arm_vmladavaq_u32): Remove.
15356 (__arm_vmlsdavaxq_s32): Remove.
15357 (__arm_vmlsdavaq_s32): Remove.
15358 (__arm_vmladavaxq_s32): Remove.
15359 (__arm_vmladavaq_s32): Remove.
15360 (__arm_vmladavaq_p_s8): Remove.
15361 (__arm_vmladavaq_p_s32): Remove.
15362 (__arm_vmladavaq_p_s16): Remove.
15363 (__arm_vmladavaq_p_u8): Remove.
15364 (__arm_vmladavaq_p_u32): Remove.
15365 (__arm_vmladavaq_p_u16): Remove.
15366 (__arm_vmladavaxq_p_s8): Remove.
15367 (__arm_vmladavaxq_p_s32): Remove.
15368 (__arm_vmladavaxq_p_s16): Remove.
15369 (__arm_vmlsdavaq_p_s8): Remove.
15370 (__arm_vmlsdavaq_p_s32): Remove.
15371 (__arm_vmlsdavaq_p_s16): Remove.
15372 (__arm_vmlsdavaxq_p_s8): Remove.
15373 (__arm_vmlsdavaxq_p_s32): Remove.
15374 (__arm_vmlsdavaxq_p_s16): Remove.
15375 (__arm_vmladavaq): Remove.
15376 (__arm_vmlsdavaxq): Remove.
15377 (__arm_vmlsdavaq): Remove.
15378 (__arm_vmladavaxq): Remove.
15379 (__arm_vmladavaq_p): Remove.
15380 (__arm_vmladavaxq_p): Remove.
15381 (__arm_vmlsdavaq_p): Remove.
15382 (__arm_vmlsdavaxq_p): Remove.
15384 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15386 * config/arm/arm-mve-builtins-shapes.cc (binary_acca_int32): New.
15387 * config/arm/arm-mve-builtins-shapes.h (binary_acca_int32): New.
15389 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15391 * config/arm/arm-mve-builtins-base.cc (vmladavq, vmladavxq)
15392 (vmlsdavq, vmlsdavxq): New.
15393 * config/arm/arm-mve-builtins-base.def (vmladavq, vmladavxq)
15394 (vmlsdavq, vmlsdavxq): New.
15395 * config/arm/arm-mve-builtins-base.h (vmladavq, vmladavxq)
15396 (vmlsdavq, vmlsdavxq): New.
15397 * config/arm/arm_mve.h (vmladavq): Remove.
15398 (vmlsdavxq): Remove.
15399 (vmlsdavq): Remove.
15400 (vmladavxq): Remove.
15401 (vmladavq_p): Remove.
15402 (vmlsdavxq_p): Remove.
15403 (vmlsdavq_p): Remove.
15404 (vmladavxq_p): Remove.
15405 (vmladavq_u8): Remove.
15406 (vmlsdavxq_s8): Remove.
15407 (vmlsdavq_s8): Remove.
15408 (vmladavxq_s8): Remove.
15409 (vmladavq_s8): Remove.
15410 (vmladavq_u16): Remove.
15411 (vmlsdavxq_s16): Remove.
15412 (vmlsdavq_s16): Remove.
15413 (vmladavxq_s16): Remove.
15414 (vmladavq_s16): Remove.
15415 (vmladavq_u32): Remove.
15416 (vmlsdavxq_s32): Remove.
15417 (vmlsdavq_s32): Remove.
15418 (vmladavxq_s32): Remove.
15419 (vmladavq_s32): Remove.
15420 (vmladavq_p_u8): Remove.
15421 (vmlsdavxq_p_s8): Remove.
15422 (vmlsdavq_p_s8): Remove.
15423 (vmladavxq_p_s8): Remove.
15424 (vmladavq_p_s8): Remove.
15425 (vmladavq_p_u16): Remove.
15426 (vmlsdavxq_p_s16): Remove.
15427 (vmlsdavq_p_s16): Remove.
15428 (vmladavxq_p_s16): Remove.
15429 (vmladavq_p_s16): Remove.
15430 (vmladavq_p_u32): Remove.
15431 (vmlsdavxq_p_s32): Remove.
15432 (vmlsdavq_p_s32): Remove.
15433 (vmladavxq_p_s32): Remove.
15434 (vmladavq_p_s32): Remove.
15435 (__arm_vmladavq_u8): Remove.
15436 (__arm_vmlsdavxq_s8): Remove.
15437 (__arm_vmlsdavq_s8): Remove.
15438 (__arm_vmladavxq_s8): Remove.
15439 (__arm_vmladavq_s8): Remove.
15440 (__arm_vmladavq_u16): Remove.
15441 (__arm_vmlsdavxq_s16): Remove.
15442 (__arm_vmlsdavq_s16): Remove.
15443 (__arm_vmladavxq_s16): Remove.
15444 (__arm_vmladavq_s16): Remove.
15445 (__arm_vmladavq_u32): Remove.
15446 (__arm_vmlsdavxq_s32): Remove.
15447 (__arm_vmlsdavq_s32): Remove.
15448 (__arm_vmladavxq_s32): Remove.
15449 (__arm_vmladavq_s32): Remove.
15450 (__arm_vmladavq_p_u8): Remove.
15451 (__arm_vmlsdavxq_p_s8): Remove.
15452 (__arm_vmlsdavq_p_s8): Remove.
15453 (__arm_vmladavxq_p_s8): Remove.
15454 (__arm_vmladavq_p_s8): Remove.
15455 (__arm_vmladavq_p_u16): Remove.
15456 (__arm_vmlsdavxq_p_s16): Remove.
15457 (__arm_vmlsdavq_p_s16): Remove.
15458 (__arm_vmladavxq_p_s16): Remove.
15459 (__arm_vmladavq_p_s16): Remove.
15460 (__arm_vmladavq_p_u32): Remove.
15461 (__arm_vmlsdavxq_p_s32): Remove.
15462 (__arm_vmlsdavq_p_s32): Remove.
15463 (__arm_vmladavxq_p_s32): Remove.
15464 (__arm_vmladavq_p_s32): Remove.
15465 (__arm_vmladavq): Remove.
15466 (__arm_vmlsdavxq): Remove.
15467 (__arm_vmlsdavq): Remove.
15468 (__arm_vmladavxq): Remove.
15469 (__arm_vmladavq_p): Remove.
15470 (__arm_vmlsdavxq_p): Remove.
15471 (__arm_vmlsdavq_p): Remove.
15472 (__arm_vmladavxq_p): Remove.
15474 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15476 * config/arm/iterators.md (MVE_VMLxDAVQ, MVE_VMLxDAVQ_P)
15477 (MVE_VMLxDAVAQ, MVE_VMLxDAVAQ_P): New.
15478 (mve_insn): Add vmladava, vmladavax, vmladav, vmladavx, vmlsdava,
15479 vmlsdavax, vmlsdav, vmlsdavx.
15480 (supf): Add VMLADAVAXQ_P_S, VMLADAVAXQ_S, VMLADAVXQ_P_S,
15481 VMLADAVXQ_S, VMLSDAVAQ_P_S, VMLSDAVAQ_S, VMLSDAVAXQ_P_S,
15482 VMLSDAVAXQ_S, VMLSDAVQ_P_S, VMLSDAVQ_S, VMLSDAVXQ_P_S,
15484 * config/arm/mve.md (mve_vmladavq_<supf><mode>)
15485 (mve_vmladavxq_s<mode>, mve_vmlsdavq_s<mode>)
15486 (mve_vmlsdavxq_s<mode>): Merge into ...
15487 (@mve_<mve_insn>q_<supf><mode>): ... this.
15488 (mve_vmlsdavaq_s<mode>, mve_vmladavaxq_s<mode>)
15489 (mve_vmlsdavaxq_s<mode>, mve_vmladavaq_<supf><mode>): Merge into
15491 (@mve_<mve_insn>q_<supf><mode>): ... this.
15492 (mve_vmladavq_p_<supf><mode>, mve_vmladavxq_p_s<mode>)
15493 (mve_vmlsdavq_p_s<mode>, mve_vmlsdavxq_p_s<mode>): Merge into ...
15494 (@mve_<mve_insn>q_p_<supf><mode>): ... this.
15495 (mve_vmladavaq_p_<supf><mode>, mve_vmladavaxq_p_s<mode>)
15496 (mve_vmlsdavaq_p_s<mode>, mve_vmlsdavaxq_p_s<mode>): Merge into
15498 (@mve_<mve_insn>q_p_<supf><mode>): ... this.
15500 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15502 * config/arm/arm-mve-builtins-shapes.cc (binary_acc_int32): New.
15503 * config/arm/arm-mve-builtins-shapes.h (binary_acc_int32): New.
15505 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15507 * config/arm/arm-mve-builtins-base.cc (vaddlvaq): New.
15508 * config/arm/arm-mve-builtins-base.def (vaddlvaq): New.
15509 * config/arm/arm-mve-builtins-base.h (vaddlvaq): New.
15510 * config/arm/arm_mve.h (vaddlvaq): Remove.
15511 (vaddlvaq_p): Remove.
15512 (vaddlvaq_u32): Remove.
15513 (vaddlvaq_s32): Remove.
15514 (vaddlvaq_p_s32): Remove.
15515 (vaddlvaq_p_u32): Remove.
15516 (__arm_vaddlvaq_u32): Remove.
15517 (__arm_vaddlvaq_s32): Remove.
15518 (__arm_vaddlvaq_p_s32): Remove.
15519 (__arm_vaddlvaq_p_u32): Remove.
15520 (__arm_vaddlvaq): Remove.
15521 (__arm_vaddlvaq_p): Remove.
15523 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15525 * config/arm/arm-mve-builtins-shapes.cc (unary_widen_acc): New.
15526 * config/arm/arm-mve-builtins-shapes.h (unary_widen_acc): New.
15528 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15530 * config/arm/iterators.md (mve_insn): Add vaddlva.
15531 * config/arm/mve.md (mve_vaddlvaq_<supf>v4si): Rename into ...
15532 (@mve_<mve_insn>q_<supf>v4si): ... this.
15533 (mve_vaddlvaq_p_<supf>v4si): Rename into ...
15534 (@mve_<mve_insn>q_p_<supf>v4si): ... this.
15536 2023-05-11 Uros Bizjak <ubizjak@gmail.com>
15539 * config/i386/i386.cc (ix86_widen_mult_cost):
15540 Handle V4HImode and V2SImode.
15542 2023-05-11 Andrew Pinski <apinski@marvell.com>
15544 * tree-ssa-dce.cc (simple_dce_from_worklist): For ssa names
15545 defined by a phi node with more than one uses, allow for the
15546 only uses are in that same defining statement.
15548 2023-05-11 Robin Dapp <rdapp@ventanamicro.com>
15550 * config/riscv/riscv.cc (riscv_const_insns): Add permissible
15553 2023-05-11 Pan Li <pan2.li@intel.com>
15555 * config/riscv/vector.md: Add comments for simplifying to vmset.
15557 2023-05-11 Robin Dapp <rdapp@ventanamicro.com>
15559 * config/riscv/autovec.md (<optab><mode>3): Add scalar shift
15561 (v<optab><mode>3): Add vector shift pattern.
15562 * config/riscv/vector-iterators.md: New iterator.
15564 2023-05-11 Robin Dapp <rdapp@ventanamicro.com>
15566 * config/riscv/autovec.md: Use renamed functions.
15567 * config/riscv/riscv-protos.h (emit_vlmax_op): Rename.
15568 (emit_vlmax_reg_op): To this.
15569 (emit_nonvlmax_op): Rename.
15570 (emit_len_op): To this.
15571 (emit_nonvlmax_binop): Rename.
15572 (emit_len_binop): To this.
15573 * config/riscv/riscv-v.cc (emit_pred_op): Add default parameter.
15574 (emit_pred_binop): Remove vlmax_p.
15575 (emit_vlmax_op): Rename.
15576 (emit_vlmax_reg_op): To this.
15577 (emit_nonvlmax_op): Rename.
15578 (emit_len_op): To this.
15579 (emit_nonvlmax_binop): Rename.
15580 (emit_len_binop): To this.
15581 (sew64_scalar_helper): Use renamed functions.
15582 (expand_tuple_move): Use renamed functions.
15583 * config/riscv/riscv.cc (vector_zero_call_used_regs): Use
15585 * config/riscv/vector.md: Use renamed functions.
15587 2023-05-11 Robin Dapp <rdapp@ventanamicro.com>
15588 Michael Collison <collison@rivosinc.com>
15590 * config/riscv/autovec.md (<optab><mode>3): Add integer binops.
15591 * config/riscv/riscv-protos.h (emit_nonvlmax_binop): Declare.
15592 * config/riscv/riscv-v.cc (emit_pred_op): New function.
15593 (set_expander_dest_and_mask): New function.
15594 (emit_pred_binop): New function.
15595 (emit_nonvlmax_binop): New function.
15597 2023-05-11 Pan Li <pan2.li@intel.com>
15599 * cfgloopmanip.cc (create_empty_loop_on_edge): Add PLUS_EXPR.
15600 * gimple-loop-interchange.cc
15601 (tree_loop_interchange::map_inductions_to_loop): Ditto.
15602 * tree-ssa-loop-ivcanon.cc (create_canonical_iv): Ditto.
15603 * tree-ssa-loop-ivopts.cc (create_new_iv): Ditto.
15604 * tree-ssa-loop-manip.cc (create_iv): Ditto.
15605 (tree_transform_and_unroll_loop): Ditto.
15606 (canonicalize_loop_ivs): Ditto.
15607 * tree-ssa-loop-manip.h (create_iv): Ditto.
15608 * tree-vect-data-refs.cc (vect_create_data_ref_ptr): Ditto.
15609 * tree-vect-loop-manip.cc (vect_set_loop_controls_directly):
15611 (vect_set_loop_condition_normal): Ditto.
15612 * tree-vect-loop.cc (vect_create_epilog_for_reduction): Ditto.
15613 * tree-vect-stmts.cc (vectorizable_store): Ditto.
15614 (vectorizable_load): Ditto.
15616 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15618 * config/arm/arm-mve-builtins-base.cc (vmovlbq, vmovltq): New.
15619 * config/arm/arm-mve-builtins-base.def (vmovlbq, vmovltq): New.
15620 * config/arm/arm-mve-builtins-base.h (vmovlbq, vmovltq): New.
15621 * config/arm/arm_mve.h (vmovlbq): Remove.
15623 (vmovlbq_m): Remove.
15624 (vmovltq_m): Remove.
15625 (vmovlbq_x): Remove.
15626 (vmovltq_x): Remove.
15627 (vmovlbq_s8): Remove.
15628 (vmovlbq_s16): Remove.
15629 (vmovltq_s8): Remove.
15630 (vmovltq_s16): Remove.
15631 (vmovltq_u8): Remove.
15632 (vmovltq_u16): Remove.
15633 (vmovlbq_u8): Remove.
15634 (vmovlbq_u16): Remove.
15635 (vmovlbq_m_s8): Remove.
15636 (vmovltq_m_s8): Remove.
15637 (vmovlbq_m_u8): Remove.
15638 (vmovltq_m_u8): Remove.
15639 (vmovlbq_m_s16): Remove.
15640 (vmovltq_m_s16): Remove.
15641 (vmovlbq_m_u16): Remove.
15642 (vmovltq_m_u16): Remove.
15643 (vmovlbq_x_s8): Remove.
15644 (vmovlbq_x_s16): Remove.
15645 (vmovlbq_x_u8): Remove.
15646 (vmovlbq_x_u16): Remove.
15647 (vmovltq_x_s8): Remove.
15648 (vmovltq_x_s16): Remove.
15649 (vmovltq_x_u8): Remove.
15650 (vmovltq_x_u16): Remove.
15651 (__arm_vmovlbq_s8): Remove.
15652 (__arm_vmovlbq_s16): Remove.
15653 (__arm_vmovltq_s8): Remove.
15654 (__arm_vmovltq_s16): Remove.
15655 (__arm_vmovltq_u8): Remove.
15656 (__arm_vmovltq_u16): Remove.
15657 (__arm_vmovlbq_u8): Remove.
15658 (__arm_vmovlbq_u16): Remove.
15659 (__arm_vmovlbq_m_s8): Remove.
15660 (__arm_vmovltq_m_s8): Remove.
15661 (__arm_vmovlbq_m_u8): Remove.
15662 (__arm_vmovltq_m_u8): Remove.
15663 (__arm_vmovlbq_m_s16): Remove.
15664 (__arm_vmovltq_m_s16): Remove.
15665 (__arm_vmovlbq_m_u16): Remove.
15666 (__arm_vmovltq_m_u16): Remove.
15667 (__arm_vmovlbq_x_s8): Remove.
15668 (__arm_vmovlbq_x_s16): Remove.
15669 (__arm_vmovlbq_x_u8): Remove.
15670 (__arm_vmovlbq_x_u16): Remove.
15671 (__arm_vmovltq_x_s8): Remove.
15672 (__arm_vmovltq_x_s16): Remove.
15673 (__arm_vmovltq_x_u8): Remove.
15674 (__arm_vmovltq_x_u16): Remove.
15675 (__arm_vmovlbq): Remove.
15676 (__arm_vmovltq): Remove.
15677 (__arm_vmovlbq_m): Remove.
15678 (__arm_vmovltq_m): Remove.
15679 (__arm_vmovlbq_x): Remove.
15680 (__arm_vmovltq_x): Remove.
15682 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15684 * config/arm/arm-mve-builtins-shapes.cc (unary_widen): New.
15685 * config/arm/arm-mve-builtins-shapes.h (unary_widen): New.
15687 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15689 * config/arm/iterators.md (mve_insn): Add vmovlb, vmovlt.
15690 (VMOVLBQ, VMOVLTQ): Merge into ...
15691 (VMOVLxQ): ... this.
15692 (VMOVLTQ_M, VMOVLBQ_M): Merge into ...
15693 (VMOVLxQ_M): ... this.
15694 * config/arm/mve.md (mve_vmovltq_<supf><mode>)
15695 (mve_vmovlbq_<supf><mode>): Merge into ...
15696 (@mve_<mve_insn>q_<supf><mode>): ... this.
15697 (mve_vmovlbq_m_<supf><mode>, mve_vmovltq_m_<supf><mode>): Merge
15699 (@mve_<mve_insn>q_m_<supf><mode>): ... this.
15701 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15703 * config/arm/arm-mve-builtins-base.cc (vaddlvq): New.
15704 * config/arm/arm-mve-builtins-base.def (vaddlvq): New.
15705 * config/arm/arm-mve-builtins-base.h (vaddlvq): New.
15706 * config/arm/arm-mve-builtins-functions.h
15707 (unspec_mve_function_exact_insn_pred_p): Handle vaddlvq.
15708 * config/arm/arm_mve.h (vaddlvq): Remove.
15709 (vaddlvq_p): Remove.
15710 (vaddlvq_s32): Remove.
15711 (vaddlvq_u32): Remove.
15712 (vaddlvq_p_s32): Remove.
15713 (vaddlvq_p_u32): Remove.
15714 (__arm_vaddlvq_s32): Remove.
15715 (__arm_vaddlvq_u32): Remove.
15716 (__arm_vaddlvq_p_s32): Remove.
15717 (__arm_vaddlvq_p_u32): Remove.
15718 (__arm_vaddlvq): Remove.
15719 (__arm_vaddlvq_p): Remove.
15721 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15723 * config/arm/iterators.md (mve_insn): Add vaddlv.
15724 * config/arm/mve.md (mve_vaddlvq_<supf>v4si): Rename into ...
15725 (@mve_<mve_insn>q_<supf>v4si): ... this.
15726 (mve_vaddlvq_p_<supf>v4si): Rename into ...
15727 (@mve_<mve_insn>q_p_<supf>v4si): ... this.
15729 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15731 * config/arm/arm-mve-builtins-shapes.cc (unary_acc): New.
15732 * config/arm/arm-mve-builtins-shapes.h (unary_acc): New.
15734 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15736 * config/arm/arm-mve-builtins-base.cc (vaddvaq): New.
15737 * config/arm/arm-mve-builtins-base.def (vaddvaq): New.
15738 * config/arm/arm-mve-builtins-base.h (vaddvaq): New.
15739 * config/arm/arm_mve.h (vaddvaq): Remove.
15740 (vaddvaq_p): Remove.
15741 (vaddvaq_u8): Remove.
15742 (vaddvaq_s8): Remove.
15743 (vaddvaq_u16): Remove.
15744 (vaddvaq_s16): Remove.
15745 (vaddvaq_u32): Remove.
15746 (vaddvaq_s32): Remove.
15747 (vaddvaq_p_u8): Remove.
15748 (vaddvaq_p_s8): Remove.
15749 (vaddvaq_p_u16): Remove.
15750 (vaddvaq_p_s16): Remove.
15751 (vaddvaq_p_u32): Remove.
15752 (vaddvaq_p_s32): Remove.
15753 (__arm_vaddvaq_u8): Remove.
15754 (__arm_vaddvaq_s8): Remove.
15755 (__arm_vaddvaq_u16): Remove.
15756 (__arm_vaddvaq_s16): Remove.
15757 (__arm_vaddvaq_u32): Remove.
15758 (__arm_vaddvaq_s32): Remove.
15759 (__arm_vaddvaq_p_u8): Remove.
15760 (__arm_vaddvaq_p_s8): Remove.
15761 (__arm_vaddvaq_p_u16): Remove.
15762 (__arm_vaddvaq_p_s16): Remove.
15763 (__arm_vaddvaq_p_u32): Remove.
15764 (__arm_vaddvaq_p_s32): Remove.
15765 (__arm_vaddvaq): Remove.
15766 (__arm_vaddvaq_p): Remove.
15768 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15770 * config/arm/arm-mve-builtins-shapes.cc (unary_int32_acc): New.
15771 * config/arm/arm-mve-builtins-shapes.h (unary_int32_acc): New.
15773 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15775 * config/arm/iterators.md (mve_insn): Add vaddva.
15776 * config/arm/mve.md (mve_vaddvaq_<supf><mode>): Rename into ...
15777 (@mve_<mve_insn>q_<supf><mode>): ... this.
15778 (mve_vaddvaq_p_<supf><mode>): Rename into ...
15779 (@mve_<mve_insn>q_p_<supf><mode>): ... this.
15781 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15783 * config/arm/arm-mve-builtins-base.cc (vaddvq): New.
15784 * config/arm/arm-mve-builtins-base.def (vaddvq): New.
15785 * config/arm/arm-mve-builtins-base.h (vaddvq): New.
15786 * config/arm/arm_mve.h (vaddvq): Remove.
15787 (vaddvq_p): Remove.
15788 (vaddvq_s8): Remove.
15789 (vaddvq_s16): Remove.
15790 (vaddvq_s32): Remove.
15791 (vaddvq_u8): Remove.
15792 (vaddvq_u16): Remove.
15793 (vaddvq_u32): Remove.
15794 (vaddvq_p_u8): Remove.
15795 (vaddvq_p_s8): Remove.
15796 (vaddvq_p_u16): Remove.
15797 (vaddvq_p_s16): Remove.
15798 (vaddvq_p_u32): Remove.
15799 (vaddvq_p_s32): Remove.
15800 (__arm_vaddvq_s8): Remove.
15801 (__arm_vaddvq_s16): Remove.
15802 (__arm_vaddvq_s32): Remove.
15803 (__arm_vaddvq_u8): Remove.
15804 (__arm_vaddvq_u16): Remove.
15805 (__arm_vaddvq_u32): Remove.
15806 (__arm_vaddvq_p_u8): Remove.
15807 (__arm_vaddvq_p_s8): Remove.
15808 (__arm_vaddvq_p_u16): Remove.
15809 (__arm_vaddvq_p_s16): Remove.
15810 (__arm_vaddvq_p_u32): Remove.
15811 (__arm_vaddvq_p_s32): Remove.
15812 (__arm_vaddvq): Remove.
15813 (__arm_vaddvq_p): Remove.
15815 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15817 * config/arm/arm-mve-builtins-shapes.cc (unary_int32): New.
15818 * config/arm/arm-mve-builtins-shapes.h (unary_int32): New.
15820 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15822 * config/arm/iterators.md (mve_insn): Add vaddv.
15823 * config/arm/mve.md (@mve_vaddvq_<supf><mode>): Rename into ...
15824 (@mve_<mve_insn>q_<supf><mode>): ... this.
15825 (mve_vaddvq_p_<supf><mode>): Rename into ...
15826 (@mve_<mve_insn>q_p_<supf><mode>): ... this.
15827 * config/arm/vec-common.md: Use gen_mve_q instead of
15830 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15832 * config/arm/arm-mve-builtins-base.cc (FUNCTION_ONLY_N): New.
15834 * config/arm/arm-mve-builtins-base.def (vdupq): New.
15835 * config/arm/arm-mve-builtins-base.h: (vdupq): New.
15836 * config/arm/arm_mve.h (vdupq_n): Remove.
15838 (vdupq_n_f16): Remove.
15839 (vdupq_n_f32): Remove.
15840 (vdupq_n_s8): Remove.
15841 (vdupq_n_s16): Remove.
15842 (vdupq_n_s32): Remove.
15843 (vdupq_n_u8): Remove.
15844 (vdupq_n_u16): Remove.
15845 (vdupq_n_u32): Remove.
15846 (vdupq_m_n_u8): Remove.
15847 (vdupq_m_n_s8): Remove.
15848 (vdupq_m_n_u16): Remove.
15849 (vdupq_m_n_s16): Remove.
15850 (vdupq_m_n_u32): Remove.
15851 (vdupq_m_n_s32): Remove.
15852 (vdupq_m_n_f16): Remove.
15853 (vdupq_m_n_f32): Remove.
15854 (vdupq_x_n_s8): Remove.
15855 (vdupq_x_n_s16): Remove.
15856 (vdupq_x_n_s32): Remove.
15857 (vdupq_x_n_u8): Remove.
15858 (vdupq_x_n_u16): Remove.
15859 (vdupq_x_n_u32): Remove.
15860 (vdupq_x_n_f16): Remove.
15861 (vdupq_x_n_f32): Remove.
15862 (__arm_vdupq_n_s8): Remove.
15863 (__arm_vdupq_n_s16): Remove.
15864 (__arm_vdupq_n_s32): Remove.
15865 (__arm_vdupq_n_u8): Remove.
15866 (__arm_vdupq_n_u16): Remove.
15867 (__arm_vdupq_n_u32): Remove.
15868 (__arm_vdupq_m_n_u8): Remove.
15869 (__arm_vdupq_m_n_s8): Remove.
15870 (__arm_vdupq_m_n_u16): Remove.
15871 (__arm_vdupq_m_n_s16): Remove.
15872 (__arm_vdupq_m_n_u32): Remove.
15873 (__arm_vdupq_m_n_s32): Remove.
15874 (__arm_vdupq_x_n_s8): Remove.
15875 (__arm_vdupq_x_n_s16): Remove.
15876 (__arm_vdupq_x_n_s32): Remove.
15877 (__arm_vdupq_x_n_u8): Remove.
15878 (__arm_vdupq_x_n_u16): Remove.
15879 (__arm_vdupq_x_n_u32): Remove.
15880 (__arm_vdupq_n_f16): Remove.
15881 (__arm_vdupq_n_f32): Remove.
15882 (__arm_vdupq_m_n_f16): Remove.
15883 (__arm_vdupq_m_n_f32): Remove.
15884 (__arm_vdupq_x_n_f16): Remove.
15885 (__arm_vdupq_x_n_f32): Remove.
15886 (__arm_vdupq_n): Remove.
15887 (__arm_vdupq_m): Remove.
15889 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15891 * config/arm/arm-mve-builtins-shapes.cc (unary_n): New.
15892 * config/arm/arm-mve-builtins-shapes.h (unary_n): New.
15894 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15896 * config/arm/iterators.md (MVE_FP_M_N_VDUPQ_ONLY)
15897 (MVE_FP_N_VDUPQ_ONLY): New.
15898 (mve_insn): Add vdupq.
15899 * config/arm/mve.md (mve_vdupq_n_f<mode>): Rename into ...
15900 (@mve_<mve_insn>q_n_f<mode>): ... this.
15901 (mve_vdupq_n_<supf><mode>): Rename into ...
15902 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
15903 (mve_vdupq_m_n_<supf><mode>): Rename into ...
15904 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
15905 (mve_vdupq_m_n_f<mode>): Rename into ...
15906 (@mve_<mve_insn>q_m_n_f<mode>): ... this.
15908 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
15910 * config/arm/arm-mve-builtins-base.cc (vrev16q, vrev32q, vrev64q):
15912 * config/arm/arm-mve-builtins-base.def (vrev16q, vrev32q)
15914 * config/arm/arm-mve-builtins-base.h (vrev16q, vrev32q)
15916 * config/arm/arm_mve.h (vrev16q): Remove.
15919 (vrev64q_m): Remove.
15920 (vrev16q_m): Remove.
15921 (vrev32q_m): Remove.
15922 (vrev16q_x): Remove.
15923 (vrev32q_x): Remove.
15924 (vrev64q_x): Remove.
15925 (vrev64q_f16): Remove.
15926 (vrev64q_f32): Remove.
15927 (vrev32q_f16): Remove.
15928 (vrev16q_s8): Remove.
15929 (vrev32q_s8): Remove.
15930 (vrev32q_s16): Remove.
15931 (vrev64q_s8): Remove.
15932 (vrev64q_s16): Remove.
15933 (vrev64q_s32): Remove.
15934 (vrev64q_u8): Remove.
15935 (vrev64q_u16): Remove.
15936 (vrev64q_u32): Remove.
15937 (vrev32q_u8): Remove.
15938 (vrev32q_u16): Remove.
15939 (vrev16q_u8): Remove.
15940 (vrev64q_m_u8): Remove.
15941 (vrev64q_m_s8): Remove.
15942 (vrev64q_m_u16): Remove.
15943 (vrev64q_m_s16): Remove.
15944 (vrev64q_m_u32): Remove.
15945 (vrev64q_m_s32): Remove.
15946 (vrev16q_m_s8): Remove.
15947 (vrev32q_m_f16): Remove.
15948 (vrev16q_m_u8): Remove.
15949 (vrev32q_m_s8): Remove.
15950 (vrev64q_m_f16): Remove.
15951 (vrev32q_m_u8): Remove.
15952 (vrev32q_m_s16): Remove.
15953 (vrev64q_m_f32): Remove.
15954 (vrev32q_m_u16): Remove.
15955 (vrev16q_x_s8): Remove.
15956 (vrev16q_x_u8): Remove.
15957 (vrev32q_x_s8): Remove.
15958 (vrev32q_x_s16): Remove.
15959 (vrev32q_x_u8): Remove.
15960 (vrev32q_x_u16): Remove.
15961 (vrev64q_x_s8): Remove.
15962 (vrev64q_x_s16): Remove.
15963 (vrev64q_x_s32): Remove.
15964 (vrev64q_x_u8): Remove.
15965 (vrev64q_x_u16): Remove.
15966 (vrev64q_x_u32): Remove.
15967 (vrev32q_x_f16): Remove.
15968 (vrev64q_x_f16): Remove.
15969 (vrev64q_x_f32): Remove.
15970 (__arm_vrev16q_s8): Remove.
15971 (__arm_vrev32q_s8): Remove.
15972 (__arm_vrev32q_s16): Remove.
15973 (__arm_vrev64q_s8): Remove.
15974 (__arm_vrev64q_s16): Remove.
15975 (__arm_vrev64q_s32): Remove.
15976 (__arm_vrev64q_u8): Remove.
15977 (__arm_vrev64q_u16): Remove.
15978 (__arm_vrev64q_u32): Remove.
15979 (__arm_vrev32q_u8): Remove.
15980 (__arm_vrev32q_u16): Remove.
15981 (__arm_vrev16q_u8): Remove.
15982 (__arm_vrev64q_m_u8): Remove.
15983 (__arm_vrev64q_m_s8): Remove.
15984 (__arm_vrev64q_m_u16): Remove.
15985 (__arm_vrev64q_m_s16): Remove.
15986 (__arm_vrev64q_m_u32): Remove.
15987 (__arm_vrev64q_m_s32): Remove.
15988 (__arm_vrev16q_m_s8): Remove.
15989 (__arm_vrev16q_m_u8): Remove.
15990 (__arm_vrev32q_m_s8): Remove.
15991 (__arm_vrev32q_m_u8): Remove.
15992 (__arm_vrev32q_m_s16): Remove.
15993 (__arm_vrev32q_m_u16): Remove.
15994 (__arm_vrev16q_x_s8): Remove.
15995 (__arm_vrev16q_x_u8): Remove.
15996 (__arm_vrev32q_x_s8): Remove.
15997 (__arm_vrev32q_x_s16): Remove.
15998 (__arm_vrev32q_x_u8): Remove.
15999 (__arm_vrev32q_x_u16): Remove.
16000 (__arm_vrev64q_x_s8): Remove.
16001 (__arm_vrev64q_x_s16): Remove.
16002 (__arm_vrev64q_x_s32): Remove.
16003 (__arm_vrev64q_x_u8): Remove.
16004 (__arm_vrev64q_x_u16): Remove.
16005 (__arm_vrev64q_x_u32): Remove.
16006 (__arm_vrev64q_f16): Remove.
16007 (__arm_vrev64q_f32): Remove.
16008 (__arm_vrev32q_f16): Remove.
16009 (__arm_vrev32q_m_f16): Remove.
16010 (__arm_vrev64q_m_f16): Remove.
16011 (__arm_vrev64q_m_f32): Remove.
16012 (__arm_vrev32q_x_f16): Remove.
16013 (__arm_vrev64q_x_f16): Remove.
16014 (__arm_vrev64q_x_f32): Remove.
16015 (__arm_vrev16q): Remove.
16016 (__arm_vrev32q): Remove.
16017 (__arm_vrev64q): Remove.
16018 (__arm_vrev64q_m): Remove.
16019 (__arm_vrev16q_m): Remove.
16020 (__arm_vrev32q_m): Remove.
16021 (__arm_vrev16q_x): Remove.
16022 (__arm_vrev32q_x): Remove.
16023 (__arm_vrev64q_x): Remove.
16025 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16027 * config/arm/iterators.md (MVE_V8HF, MVE_V16QI)
16028 (MVE_FP_VREV64Q_ONLY, MVE_FP_M_VREV64Q_ONLY, MVE_FP_VREV32Q_ONLY)
16029 (MVE_FP_M_VREV32Q_ONLY): New iterators.
16030 (mve_insn): Add vrev16q, vrev32q, vrev64q.
16031 * config/arm/mve.md (mve_vrev64q_f<mode>): Rename into ...
16032 (@mve_<mve_insn>q_f<mode>): ... this
16033 (mve_vrev32q_fv8hf): Rename into @mve_<mve_insn>q_f<mode>.
16034 (mve_vrev64q_<supf><mode>): Rename into ...
16035 (@mve_<mve_insn>q_<supf><mode>): ... this.
16036 (mve_vrev32q_<supf><mode>): Rename into
16037 @mve_<mve_insn>q_<supf><mode>.
16038 (mve_vrev16q_<supf>v16qi): Rename into
16039 @mve_<mve_insn>q_<supf><mode>.
16040 (mve_vrev64q_m_<supf><mode>): Rename into
16041 @mve_<mve_insn>q_m_<supf><mode>.
16042 (mve_vrev32q_m_fv8hf): Rename into @mve_<mve_insn>q_m_f<mode>.
16043 (mve_vrev32q_m_<supf><mode>): Rename into
16044 @mve_<mve_insn>q_m_<supf><mode>.
16045 (mve_vrev64q_m_f<mode>): Rename into @mve_<mve_insn>q_m_f<mode>.
16046 (mve_vrev16q_m_<supf>v16qi): Rename into
16047 @mve_<mve_insn>q_m_<supf><mode>.
16049 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16051 * config/arm/arm-mve-builtins-base.cc (vcmpeqq, vcmpneq, vcmpgeq)
16052 (vcmpgtq, vcmpleq, vcmpltq, vcmpcsq, vcmphiq): New.
16053 * config/arm/arm-mve-builtins-base.def (vcmpeqq, vcmpneq, vcmpgeq)
16054 (vcmpgtq, vcmpleq, vcmpltq, vcmpcsq, vcmphiq): New.
16055 * config/arm/arm-mve-builtins-base.h (vcmpeqq, vcmpneq, vcmpgeq)
16056 (vcmpgtq, vcmpleq, vcmpltq, vcmpcsq, vcmphiq): New.
16057 * config/arm/arm-mve-builtins-functions.h (class
16058 unspec_based_mve_function_exact_insn_vcmp): New.
16059 * config/arm/arm-mve-builtins.cc
16060 (function_instance::has_inactive_argument): Handle vcmp.
16061 * config/arm/arm_mve.h (vcmpneq): Remove.
16069 (vcmpneq_m): Remove.
16070 (vcmphiq_m): Remove.
16071 (vcmpeqq_m): Remove.
16072 (vcmpcsq_m): Remove.
16073 (vcmpcsq_m_n): Remove.
16074 (vcmpltq_m): Remove.
16075 (vcmpleq_m): Remove.
16076 (vcmpgtq_m): Remove.
16077 (vcmpgeq_m): Remove.
16078 (vcmpneq_s8): Remove.
16079 (vcmpneq_s16): Remove.
16080 (vcmpneq_s32): Remove.
16081 (vcmpneq_u8): Remove.
16082 (vcmpneq_u16): Remove.
16083 (vcmpneq_u32): Remove.
16084 (vcmpneq_n_u8): Remove.
16085 (vcmphiq_u8): Remove.
16086 (vcmphiq_n_u8): Remove.
16087 (vcmpeqq_u8): Remove.
16088 (vcmpeqq_n_u8): Remove.
16089 (vcmpcsq_u8): Remove.
16090 (vcmpcsq_n_u8): Remove.
16091 (vcmpneq_n_s8): Remove.
16092 (vcmpltq_s8): Remove.
16093 (vcmpltq_n_s8): Remove.
16094 (vcmpleq_s8): Remove.
16095 (vcmpleq_n_s8): Remove.
16096 (vcmpgtq_s8): Remove.
16097 (vcmpgtq_n_s8): Remove.
16098 (vcmpgeq_s8): Remove.
16099 (vcmpgeq_n_s8): Remove.
16100 (vcmpeqq_s8): Remove.
16101 (vcmpeqq_n_s8): Remove.
16102 (vcmpneq_n_u16): Remove.
16103 (vcmphiq_u16): Remove.
16104 (vcmphiq_n_u16): Remove.
16105 (vcmpeqq_u16): Remove.
16106 (vcmpeqq_n_u16): Remove.
16107 (vcmpcsq_u16): Remove.
16108 (vcmpcsq_n_u16): Remove.
16109 (vcmpneq_n_s16): Remove.
16110 (vcmpltq_s16): Remove.
16111 (vcmpltq_n_s16): Remove.
16112 (vcmpleq_s16): Remove.
16113 (vcmpleq_n_s16): Remove.
16114 (vcmpgtq_s16): Remove.
16115 (vcmpgtq_n_s16): Remove.
16116 (vcmpgeq_s16): Remove.
16117 (vcmpgeq_n_s16): Remove.
16118 (vcmpeqq_s16): Remove.
16119 (vcmpeqq_n_s16): Remove.
16120 (vcmpneq_n_u32): Remove.
16121 (vcmphiq_u32): Remove.
16122 (vcmphiq_n_u32): Remove.
16123 (vcmpeqq_u32): Remove.
16124 (vcmpeqq_n_u32): Remove.
16125 (vcmpcsq_u32): Remove.
16126 (vcmpcsq_n_u32): Remove.
16127 (vcmpneq_n_s32): Remove.
16128 (vcmpltq_s32): Remove.
16129 (vcmpltq_n_s32): Remove.
16130 (vcmpleq_s32): Remove.
16131 (vcmpleq_n_s32): Remove.
16132 (vcmpgtq_s32): Remove.
16133 (vcmpgtq_n_s32): Remove.
16134 (vcmpgeq_s32): Remove.
16135 (vcmpgeq_n_s32): Remove.
16136 (vcmpeqq_s32): Remove.
16137 (vcmpeqq_n_s32): Remove.
16138 (vcmpneq_n_f16): Remove.
16139 (vcmpneq_f16): Remove.
16140 (vcmpltq_n_f16): Remove.
16141 (vcmpltq_f16): Remove.
16142 (vcmpleq_n_f16): Remove.
16143 (vcmpleq_f16): Remove.
16144 (vcmpgtq_n_f16): Remove.
16145 (vcmpgtq_f16): Remove.
16146 (vcmpgeq_n_f16): Remove.
16147 (vcmpgeq_f16): Remove.
16148 (vcmpeqq_n_f16): Remove.
16149 (vcmpeqq_f16): Remove.
16150 (vcmpneq_n_f32): Remove.
16151 (vcmpneq_f32): Remove.
16152 (vcmpltq_n_f32): Remove.
16153 (vcmpltq_f32): Remove.
16154 (vcmpleq_n_f32): Remove.
16155 (vcmpleq_f32): Remove.
16156 (vcmpgtq_n_f32): Remove.
16157 (vcmpgtq_f32): Remove.
16158 (vcmpgeq_n_f32): Remove.
16159 (vcmpgeq_f32): Remove.
16160 (vcmpeqq_n_f32): Remove.
16161 (vcmpeqq_f32): Remove.
16162 (vcmpeqq_m_f16): Remove.
16163 (vcmpeqq_m_f32): Remove.
16164 (vcmpneq_m_u8): Remove.
16165 (vcmpneq_m_n_u8): Remove.
16166 (vcmphiq_m_u8): Remove.
16167 (vcmphiq_m_n_u8): Remove.
16168 (vcmpeqq_m_u8): Remove.
16169 (vcmpeqq_m_n_u8): Remove.
16170 (vcmpcsq_m_u8): Remove.
16171 (vcmpcsq_m_n_u8): Remove.
16172 (vcmpneq_m_s8): Remove.
16173 (vcmpneq_m_n_s8): Remove.
16174 (vcmpltq_m_s8): Remove.
16175 (vcmpltq_m_n_s8): Remove.
16176 (vcmpleq_m_s8): Remove.
16177 (vcmpleq_m_n_s8): Remove.
16178 (vcmpgtq_m_s8): Remove.
16179 (vcmpgtq_m_n_s8): Remove.
16180 (vcmpgeq_m_s8): Remove.
16181 (vcmpgeq_m_n_s8): Remove.
16182 (vcmpeqq_m_s8): Remove.
16183 (vcmpeqq_m_n_s8): Remove.
16184 (vcmpneq_m_u16): Remove.
16185 (vcmpneq_m_n_u16): Remove.
16186 (vcmphiq_m_u16): Remove.
16187 (vcmphiq_m_n_u16): Remove.
16188 (vcmpeqq_m_u16): Remove.
16189 (vcmpeqq_m_n_u16): Remove.
16190 (vcmpcsq_m_u16): Remove.
16191 (vcmpcsq_m_n_u16): Remove.
16192 (vcmpneq_m_s16): Remove.
16193 (vcmpneq_m_n_s16): Remove.
16194 (vcmpltq_m_s16): Remove.
16195 (vcmpltq_m_n_s16): Remove.
16196 (vcmpleq_m_s16): Remove.
16197 (vcmpleq_m_n_s16): Remove.
16198 (vcmpgtq_m_s16): Remove.
16199 (vcmpgtq_m_n_s16): Remove.
16200 (vcmpgeq_m_s16): Remove.
16201 (vcmpgeq_m_n_s16): Remove.
16202 (vcmpeqq_m_s16): Remove.
16203 (vcmpeqq_m_n_s16): Remove.
16204 (vcmpneq_m_u32): Remove.
16205 (vcmpneq_m_n_u32): Remove.
16206 (vcmphiq_m_u32): Remove.
16207 (vcmphiq_m_n_u32): Remove.
16208 (vcmpeqq_m_u32): Remove.
16209 (vcmpeqq_m_n_u32): Remove.
16210 (vcmpcsq_m_u32): Remove.
16211 (vcmpcsq_m_n_u32): Remove.
16212 (vcmpneq_m_s32): Remove.
16213 (vcmpneq_m_n_s32): Remove.
16214 (vcmpltq_m_s32): Remove.
16215 (vcmpltq_m_n_s32): Remove.
16216 (vcmpleq_m_s32): Remove.
16217 (vcmpleq_m_n_s32): Remove.
16218 (vcmpgtq_m_s32): Remove.
16219 (vcmpgtq_m_n_s32): Remove.
16220 (vcmpgeq_m_s32): Remove.
16221 (vcmpgeq_m_n_s32): Remove.
16222 (vcmpeqq_m_s32): Remove.
16223 (vcmpeqq_m_n_s32): Remove.
16224 (vcmpeqq_m_n_f16): Remove.
16225 (vcmpgeq_m_f16): Remove.
16226 (vcmpgeq_m_n_f16): Remove.
16227 (vcmpgtq_m_f16): Remove.
16228 (vcmpgtq_m_n_f16): Remove.
16229 (vcmpleq_m_f16): Remove.
16230 (vcmpleq_m_n_f16): Remove.
16231 (vcmpltq_m_f16): Remove.
16232 (vcmpltq_m_n_f16): Remove.
16233 (vcmpneq_m_f16): Remove.
16234 (vcmpneq_m_n_f16): Remove.
16235 (vcmpeqq_m_n_f32): Remove.
16236 (vcmpgeq_m_f32): Remove.
16237 (vcmpgeq_m_n_f32): Remove.
16238 (vcmpgtq_m_f32): Remove.
16239 (vcmpgtq_m_n_f32): Remove.
16240 (vcmpleq_m_f32): Remove.
16241 (vcmpleq_m_n_f32): Remove.
16242 (vcmpltq_m_f32): Remove.
16243 (vcmpltq_m_n_f32): Remove.
16244 (vcmpneq_m_f32): Remove.
16245 (vcmpneq_m_n_f32): Remove.
16246 (__arm_vcmpneq_s8): Remove.
16247 (__arm_vcmpneq_s16): Remove.
16248 (__arm_vcmpneq_s32): Remove.
16249 (__arm_vcmpneq_u8): Remove.
16250 (__arm_vcmpneq_u16): Remove.
16251 (__arm_vcmpneq_u32): Remove.
16252 (__arm_vcmpneq_n_u8): Remove.
16253 (__arm_vcmphiq_u8): Remove.
16254 (__arm_vcmphiq_n_u8): Remove.
16255 (__arm_vcmpeqq_u8): Remove.
16256 (__arm_vcmpeqq_n_u8): Remove.
16257 (__arm_vcmpcsq_u8): Remove.
16258 (__arm_vcmpcsq_n_u8): Remove.
16259 (__arm_vcmpneq_n_s8): Remove.
16260 (__arm_vcmpltq_s8): Remove.
16261 (__arm_vcmpltq_n_s8): Remove.
16262 (__arm_vcmpleq_s8): Remove.
16263 (__arm_vcmpleq_n_s8): Remove.
16264 (__arm_vcmpgtq_s8): Remove.
16265 (__arm_vcmpgtq_n_s8): Remove.
16266 (__arm_vcmpgeq_s8): Remove.
16267 (__arm_vcmpgeq_n_s8): Remove.
16268 (__arm_vcmpeqq_s8): Remove.
16269 (__arm_vcmpeqq_n_s8): Remove.
16270 (__arm_vcmpneq_n_u16): Remove.
16271 (__arm_vcmphiq_u16): Remove.
16272 (__arm_vcmphiq_n_u16): Remove.
16273 (__arm_vcmpeqq_u16): Remove.
16274 (__arm_vcmpeqq_n_u16): Remove.
16275 (__arm_vcmpcsq_u16): Remove.
16276 (__arm_vcmpcsq_n_u16): Remove.
16277 (__arm_vcmpneq_n_s16): Remove.
16278 (__arm_vcmpltq_s16): Remove.
16279 (__arm_vcmpltq_n_s16): Remove.
16280 (__arm_vcmpleq_s16): Remove.
16281 (__arm_vcmpleq_n_s16): Remove.
16282 (__arm_vcmpgtq_s16): Remove.
16283 (__arm_vcmpgtq_n_s16): Remove.
16284 (__arm_vcmpgeq_s16): Remove.
16285 (__arm_vcmpgeq_n_s16): Remove.
16286 (__arm_vcmpeqq_s16): Remove.
16287 (__arm_vcmpeqq_n_s16): Remove.
16288 (__arm_vcmpneq_n_u32): Remove.
16289 (__arm_vcmphiq_u32): Remove.
16290 (__arm_vcmphiq_n_u32): Remove.
16291 (__arm_vcmpeqq_u32): Remove.
16292 (__arm_vcmpeqq_n_u32): Remove.
16293 (__arm_vcmpcsq_u32): Remove.
16294 (__arm_vcmpcsq_n_u32): Remove.
16295 (__arm_vcmpneq_n_s32): Remove.
16296 (__arm_vcmpltq_s32): Remove.
16297 (__arm_vcmpltq_n_s32): Remove.
16298 (__arm_vcmpleq_s32): Remove.
16299 (__arm_vcmpleq_n_s32): Remove.
16300 (__arm_vcmpgtq_s32): Remove.
16301 (__arm_vcmpgtq_n_s32): Remove.
16302 (__arm_vcmpgeq_s32): Remove.
16303 (__arm_vcmpgeq_n_s32): Remove.
16304 (__arm_vcmpeqq_s32): Remove.
16305 (__arm_vcmpeqq_n_s32): Remove.
16306 (__arm_vcmpneq_m_u8): Remove.
16307 (__arm_vcmpneq_m_n_u8): Remove.
16308 (__arm_vcmphiq_m_u8): Remove.
16309 (__arm_vcmphiq_m_n_u8): Remove.
16310 (__arm_vcmpeqq_m_u8): Remove.
16311 (__arm_vcmpeqq_m_n_u8): Remove.
16312 (__arm_vcmpcsq_m_u8): Remove.
16313 (__arm_vcmpcsq_m_n_u8): Remove.
16314 (__arm_vcmpneq_m_s8): Remove.
16315 (__arm_vcmpneq_m_n_s8): Remove.
16316 (__arm_vcmpltq_m_s8): Remove.
16317 (__arm_vcmpltq_m_n_s8): Remove.
16318 (__arm_vcmpleq_m_s8): Remove.
16319 (__arm_vcmpleq_m_n_s8): Remove.
16320 (__arm_vcmpgtq_m_s8): Remove.
16321 (__arm_vcmpgtq_m_n_s8): Remove.
16322 (__arm_vcmpgeq_m_s8): Remove.
16323 (__arm_vcmpgeq_m_n_s8): Remove.
16324 (__arm_vcmpeqq_m_s8): Remove.
16325 (__arm_vcmpeqq_m_n_s8): Remove.
16326 (__arm_vcmpneq_m_u16): Remove.
16327 (__arm_vcmpneq_m_n_u16): Remove.
16328 (__arm_vcmphiq_m_u16): Remove.
16329 (__arm_vcmphiq_m_n_u16): Remove.
16330 (__arm_vcmpeqq_m_u16): Remove.
16331 (__arm_vcmpeqq_m_n_u16): Remove.
16332 (__arm_vcmpcsq_m_u16): Remove.
16333 (__arm_vcmpcsq_m_n_u16): Remove.
16334 (__arm_vcmpneq_m_s16): Remove.
16335 (__arm_vcmpneq_m_n_s16): Remove.
16336 (__arm_vcmpltq_m_s16): Remove.
16337 (__arm_vcmpltq_m_n_s16): Remove.
16338 (__arm_vcmpleq_m_s16): Remove.
16339 (__arm_vcmpleq_m_n_s16): Remove.
16340 (__arm_vcmpgtq_m_s16): Remove.
16341 (__arm_vcmpgtq_m_n_s16): Remove.
16342 (__arm_vcmpgeq_m_s16): Remove.
16343 (__arm_vcmpgeq_m_n_s16): Remove.
16344 (__arm_vcmpeqq_m_s16): Remove.
16345 (__arm_vcmpeqq_m_n_s16): Remove.
16346 (__arm_vcmpneq_m_u32): Remove.
16347 (__arm_vcmpneq_m_n_u32): Remove.
16348 (__arm_vcmphiq_m_u32): Remove.
16349 (__arm_vcmphiq_m_n_u32): Remove.
16350 (__arm_vcmpeqq_m_u32): Remove.
16351 (__arm_vcmpeqq_m_n_u32): Remove.
16352 (__arm_vcmpcsq_m_u32): Remove.
16353 (__arm_vcmpcsq_m_n_u32): Remove.
16354 (__arm_vcmpneq_m_s32): Remove.
16355 (__arm_vcmpneq_m_n_s32): Remove.
16356 (__arm_vcmpltq_m_s32): Remove.
16357 (__arm_vcmpltq_m_n_s32): Remove.
16358 (__arm_vcmpleq_m_s32): Remove.
16359 (__arm_vcmpleq_m_n_s32): Remove.
16360 (__arm_vcmpgtq_m_s32): Remove.
16361 (__arm_vcmpgtq_m_n_s32): Remove.
16362 (__arm_vcmpgeq_m_s32): Remove.
16363 (__arm_vcmpgeq_m_n_s32): Remove.
16364 (__arm_vcmpeqq_m_s32): Remove.
16365 (__arm_vcmpeqq_m_n_s32): Remove.
16366 (__arm_vcmpneq_n_f16): Remove.
16367 (__arm_vcmpneq_f16): Remove.
16368 (__arm_vcmpltq_n_f16): Remove.
16369 (__arm_vcmpltq_f16): Remove.
16370 (__arm_vcmpleq_n_f16): Remove.
16371 (__arm_vcmpleq_f16): Remove.
16372 (__arm_vcmpgtq_n_f16): Remove.
16373 (__arm_vcmpgtq_f16): Remove.
16374 (__arm_vcmpgeq_n_f16): Remove.
16375 (__arm_vcmpgeq_f16): Remove.
16376 (__arm_vcmpeqq_n_f16): Remove.
16377 (__arm_vcmpeqq_f16): Remove.
16378 (__arm_vcmpneq_n_f32): Remove.
16379 (__arm_vcmpneq_f32): Remove.
16380 (__arm_vcmpltq_n_f32): Remove.
16381 (__arm_vcmpltq_f32): Remove.
16382 (__arm_vcmpleq_n_f32): Remove.
16383 (__arm_vcmpleq_f32): Remove.
16384 (__arm_vcmpgtq_n_f32): Remove.
16385 (__arm_vcmpgtq_f32): Remove.
16386 (__arm_vcmpgeq_n_f32): Remove.
16387 (__arm_vcmpgeq_f32): Remove.
16388 (__arm_vcmpeqq_n_f32): Remove.
16389 (__arm_vcmpeqq_f32): Remove.
16390 (__arm_vcmpeqq_m_f16): Remove.
16391 (__arm_vcmpeqq_m_f32): Remove.
16392 (__arm_vcmpeqq_m_n_f16): Remove.
16393 (__arm_vcmpgeq_m_f16): Remove.
16394 (__arm_vcmpgeq_m_n_f16): Remove.
16395 (__arm_vcmpgtq_m_f16): Remove.
16396 (__arm_vcmpgtq_m_n_f16): Remove.
16397 (__arm_vcmpleq_m_f16): Remove.
16398 (__arm_vcmpleq_m_n_f16): Remove.
16399 (__arm_vcmpltq_m_f16): Remove.
16400 (__arm_vcmpltq_m_n_f16): Remove.
16401 (__arm_vcmpneq_m_f16): Remove.
16402 (__arm_vcmpneq_m_n_f16): Remove.
16403 (__arm_vcmpeqq_m_n_f32): Remove.
16404 (__arm_vcmpgeq_m_f32): Remove.
16405 (__arm_vcmpgeq_m_n_f32): Remove.
16406 (__arm_vcmpgtq_m_f32): Remove.
16407 (__arm_vcmpgtq_m_n_f32): Remove.
16408 (__arm_vcmpleq_m_f32): Remove.
16409 (__arm_vcmpleq_m_n_f32): Remove.
16410 (__arm_vcmpltq_m_f32): Remove.
16411 (__arm_vcmpltq_m_n_f32): Remove.
16412 (__arm_vcmpneq_m_f32): Remove.
16413 (__arm_vcmpneq_m_n_f32): Remove.
16414 (__arm_vcmpneq): Remove.
16415 (__arm_vcmphiq): Remove.
16416 (__arm_vcmpeqq): Remove.
16417 (__arm_vcmpcsq): Remove.
16418 (__arm_vcmpltq): Remove.
16419 (__arm_vcmpleq): Remove.
16420 (__arm_vcmpgtq): Remove.
16421 (__arm_vcmpgeq): Remove.
16422 (__arm_vcmpneq_m): Remove.
16423 (__arm_vcmphiq_m): Remove.
16424 (__arm_vcmpeqq_m): Remove.
16425 (__arm_vcmpcsq_m): Remove.
16426 (__arm_vcmpltq_m): Remove.
16427 (__arm_vcmpleq_m): Remove.
16428 (__arm_vcmpgtq_m): Remove.
16429 (__arm_vcmpgeq_m): Remove.
16431 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16433 * config/arm/arm-mve-builtins-shapes.cc (cmp): New.
16434 * config/arm/arm-mve-builtins-shapes.h (cmp): New.
16436 2023-05-11 Christophe Lyon <christophe.lyon@arm.com>
16438 * config/arm/iterators.md (MVE_CMP_M, MVE_CMP_M_F, MVE_CMP_M_N)
16439 (MVE_CMP_M_N_F, mve_cmp_op1): New.
16442 * config/arm/mve.md (mve_vcmp<mve_cmp_op>q_n_<mode>): Rename into ...
16443 (@mve_vcmp<mve_cmp_op>q_n_<mode>): ... this.
16444 (mve_vcmpeqq_m_f<mode>, mve_vcmpgeq_m_f<mode>)
16445 (mve_vcmpgtq_m_f<mode>, mve_vcmpleq_m_f<mode>)
16446 (mve_vcmpltq_m_f<mode>, mve_vcmpneq_m_f<mode>): Merge into ...
16447 (@mve_vcmp<mve_cmp_op1>q_m_f<mode>): ... this.
16448 (mve_vcmpcsq_m_u<mode>, mve_vcmpeqq_m_<supf><mode>)
16449 (mve_vcmpgeq_m_s<mode>, mve_vcmpgtq_m_s<mode>)
16450 (mve_vcmphiq_m_u<mode>, mve_vcmpleq_m_s<mode>)
16451 (mve_vcmpltq_m_s<mode>, mve_vcmpneq_m_<supf><mode>): Merge into
16453 (@mve_vcmp<mve_cmp_op1>q_m_<supf><mode>): ... this.
16454 (mve_vcmpcsq_m_n_u<mode>, mve_vcmpeqq_m_n_<supf><mode>)
16455 (mve_vcmpgeq_m_n_s<mode>, mve_vcmpgtq_m_n_s<mode>)
16456 (mve_vcmphiq_m_n_u<mode>, mve_vcmpleq_m_n_s<mode>)
16457 (mve_vcmpltq_m_n_s<mode>, mve_vcmpneq_m_n_<supf><mode>): Merge
16459 (@mve_vcmp<mve_cmp_op1>q_m_n_<supf><mode>): ... this.
16460 (mve_vcmpeqq_m_n_f<mode>, mve_vcmpgeq_m_n_f<mode>)
16461 (mve_vcmpgtq_m_n_f<mode>, mve_vcmpleq_m_n_f<mode>)
16462 (mve_vcmpltq_m_n_f<mode>, mve_vcmpneq_m_n_f<mode>): Merge into ...
16463 (@mve_vcmp<mve_cmp_op1>q_m_n_f<mode>): ... this.
16465 2023-05-11 Roger Sayle <roger@nextmovesoftware.com>
16467 * match.pd <popcount optimizations>: Simplify popcount(X|Y) +
16468 popcount(X&Y) as popcount(X)+popcount(Y). Likewise, simplify
16469 popcount(X)+popcount(Y)-popcount(X&Y) as popcount(X|Y), and
16472 2023-05-11 Roger Sayle <roger@nextmovesoftware.com>
16474 * match.pd <popcount optimizations>: Simplify popcount(bswap(x))
16475 as popcount(x). Simplify popcount(rotate(x,y)) as popcount(x).
16476 <parity optimizations>: Simplify parity(bswap(x)) as parity(x).
16477 Simplify parity(rotate(x,y)) as parity(x).
16479 2023-05-11 Juzhe-Zhong <juzhe.zhong@rivai.ai>
16481 * config/riscv/autovec.md (@vec_series<mode>): New pattern
16482 * config/riscv/riscv-protos.h (expand_vec_series): New function.
16483 * config/riscv/riscv-v.cc (emit_binop): Ditto.
16484 (emit_index_op): Ditto.
16485 (expand_vec_series): Ditto.
16486 (expand_const_vector): Add series vector handling.
16487 * config/riscv/riscv.cc (riscv_const_insns): Enable series vector for testing.
16489 2023-05-10 Roger Sayle <roger@nextmovesoftware.com>
16491 * config/i386/i386.md (*concat<mode><dwi>3_1): Use preferred
16492 [(const_int 0)] idiom, instead of [(clobber (const_int 0))].
16493 (*concat<mode><dwi>3_2): Likewise.
16494 (*concat<mode><dwi>3_3): Likewise.
16495 (*concat<mode><dwi>3_4): Likewise.
16496 (*concat<mode><dwi>3_5): Likewise.
16497 (*concat<mode><dwi>3_6): Likewise.
16498 (*concat<mode><dwi>3_7): Likewise.
16500 2023-05-10 Uros Bizjak <ubizjak@gmail.com>
16503 * config/i386/mmx.md (sse4_1_<code>v2qiv2si2): New insn pattern.
16504 (<insn>v4qiv4hi2): New expander.
16505 (<insn>v2hiv2si2): Ditto.
16506 (<insn>v2qiv2si2): Ditto.
16507 (<insn>v2qiv2hi2): Ditto.
16509 2023-05-10 Jeff Law <jlaw@ventanamicro>
16511 * config/h8300/constraints.md (Q): Make this a special memory
16515 2023-05-10 Jakub Jelinek <jakub@redhat.com>
16518 * ipa-prop.cc (ipa_get_callee_param_type): Don't return TREE_VALUE (t)
16519 if t is void_list_node.
16521 2023-05-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16523 * config/aarch64/aarch64-simd.md (aarch64_sqmovun<mode>_insn_le): Delete.
16524 (aarch64_sqmovun<mode>_insn_be): Delete.
16525 (aarch64_sqmovun<mode><vczle><vczbe>): New define_insn.
16526 (aarch64_sqmovun<mode>): Delete expander.
16528 2023-05-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16531 * config/aarch64/aarch64-simd.md (aarch64_<PERMUTE:perm_insn><mode>):
16533 (aarch64_<PERMUTE:perm_insn><mode><vczle><vczbe>): ... This.
16534 (aarch64_rev<REVERSE:rev_op><mode>): Rename to...
16535 (aarch64_rev<REVERSE:rev_op><mode><vczle><vczbe>): ... This.
16537 2023-05-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16540 * config/aarch64/aarch64-simd.md (aarch64_<su_optab>q<addsub><mode>):
16542 (aarch64_<su_optab>q<addsub><mode><vczle><vczbe>): ... This.
16543 (aarch64_<sur>qadd<mode>): Rename to...
16544 (aarch64_<sur>qadd<mode><vczle><vczbe>): ... This.
16546 2023-05-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16548 * config/aarch64/aarch64-simd.md
16549 (aarch64_<sur>q<r>shr<u>n_n<mode>_insn_le): Delete.
16550 (aarch64_<sur>q<r>shr<u>n_n<mode>_insn_be): Delete.
16551 (aarch64_<sur>q<r>shr<u>n_n<mode>_insn<vczle><vczbe>): New define_insn.
16552 (aarch64_<sur>q<r>shr<u>n_n<mode>): Simplify expander.
16554 2023-05-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16557 * config/aarch64/aarch64-simd.md (aarch64_xtn<mode>_insn_le): Delete.
16558 (aarch64_xtn<mode>_insn_be): Likewise.
16559 (trunc<mode><Vnarrowq>2): Rename to...
16560 (trunc<mode><Vnarrowq>2<vczle><vczbe>): ... This.
16561 (aarch64_xtn<mode>): Move under the above. Just emit the truncate RTL.
16562 (aarch64_<su>qmovn<mode>): Likewise.
16563 (aarch64_<su>qmovn<mode><vczle><vczbe>): New define_insn.
16564 (aarch64_<su>qmovn<mode>_insn_le): Delete.
16565 (aarch64_<su>qmovn<mode>_insn_be): Likewise.
16567 2023-05-10 Li Xu <xuli1@eswincomputing.com>
16569 * config/riscv/riscv-vsetvl.cc (gen_vsetvl_pat): For vfmv.f.s/vmv.x.s
16570 intruction replace null avl with (const_int 0).
16572 2023-05-10 Juzhe-Zhong <juzhe.zhong@rivai.ai>
16574 * config/riscv/riscv.cc (riscv_support_vector_misalignment): Fix
16577 2023-05-10 Juzhe-Zhong <juzhe.zhong@rivai.ai>
16580 * config/riscv/riscv-vsetvl.cc (avl_source_has_vsetvl_p): New function.
16581 (source_equal_p): Fix dead loop in vsetvl avl checking.
16583 2023-05-10 Hans-Peter Nilsson <hp@axis.com>
16585 * config/cris/cris.cc (cris_postdbr_cmpelim): Correct mode
16586 of modeadjusted_dccr.
16588 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16590 * config/arm/arm-mve-builtins-base.cc (vmaxaq, vminaq): New.
16591 * config/arm/arm-mve-builtins-base.def (vmaxaq, vminaq): New.
16592 * config/arm/arm-mve-builtins-base.h (vmaxaq, vminaq): New.
16593 * config/arm/arm-mve-builtins.cc
16594 (function_instance::has_inactive_argument): Handle vmaxaq and
16596 * config/arm/arm_mve.h (vminaq): Remove.
16598 (vminaq_m): Remove.
16599 (vmaxaq_m): Remove.
16600 (vminaq_s8): Remove.
16601 (vmaxaq_s8): Remove.
16602 (vminaq_s16): Remove.
16603 (vmaxaq_s16): Remove.
16604 (vminaq_s32): Remove.
16605 (vmaxaq_s32): Remove.
16606 (vminaq_m_s8): Remove.
16607 (vmaxaq_m_s8): Remove.
16608 (vminaq_m_s16): Remove.
16609 (vmaxaq_m_s16): Remove.
16610 (vminaq_m_s32): Remove.
16611 (vmaxaq_m_s32): Remove.
16612 (__arm_vminaq_s8): Remove.
16613 (__arm_vmaxaq_s8): Remove.
16614 (__arm_vminaq_s16): Remove.
16615 (__arm_vmaxaq_s16): Remove.
16616 (__arm_vminaq_s32): Remove.
16617 (__arm_vmaxaq_s32): Remove.
16618 (__arm_vminaq_m_s8): Remove.
16619 (__arm_vmaxaq_m_s8): Remove.
16620 (__arm_vminaq_m_s16): Remove.
16621 (__arm_vmaxaq_m_s16): Remove.
16622 (__arm_vminaq_m_s32): Remove.
16623 (__arm_vmaxaq_m_s32): Remove.
16624 (__arm_vminaq): Remove.
16625 (__arm_vmaxaq): Remove.
16626 (__arm_vminaq_m): Remove.
16627 (__arm_vmaxaq_m): Remove.
16629 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16631 * config/arm/iterators.md (MVE_VMAXAVMINAQ, MVE_VMAXAVMINAQ_M):
16633 (mve_insn): Add vmaxa, vmina.
16634 (supf): Add VMAXAQ_S, VMAXAQ_M_S, VMINAQ_S, VMINAQ_M_S.
16635 * config/arm/mve.md (mve_vmaxaq_s<mode>, mve_vminaq_s<mode>):
16637 (@mve_<mve_insn>q_<supf><mode>): ... this.
16638 (mve_vmaxaq_m_s<mode>, mve_vminaq_m_s<mode>): Merge into ...
16639 (@mve_<mve_insn>q_m_<supf><mode>): ... this.
16641 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16643 * config/arm/arm-mve-builtins-shapes.cc (binary_maxamina): New.
16644 * config/arm/arm-mve-builtins-shapes.h (binary_maxamina): New.
16646 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16648 * config/arm/arm-mve-builtins-base.cc (vmaxnmaq, vminnmaq): New.
16649 * config/arm/arm-mve-builtins-base.def (vmaxnmaq, vminnmaq): New.
16650 * config/arm/arm-mve-builtins-base.h (vmaxnmaq, vminnmaq): New.
16651 * config/arm/arm-mve-builtins.cc
16652 (function_instance::has_inactive_argument): Handle vmaxnmaq and
16654 * config/arm/arm_mve.h (vminnmaq): Remove.
16655 (vmaxnmaq): Remove.
16656 (vmaxnmaq_m): Remove.
16657 (vminnmaq_m): Remove.
16658 (vminnmaq_f16): Remove.
16659 (vmaxnmaq_f16): Remove.
16660 (vminnmaq_f32): Remove.
16661 (vmaxnmaq_f32): Remove.
16662 (vmaxnmaq_m_f16): Remove.
16663 (vminnmaq_m_f16): Remove.
16664 (vmaxnmaq_m_f32): Remove.
16665 (vminnmaq_m_f32): Remove.
16666 (__arm_vminnmaq_f16): Remove.
16667 (__arm_vmaxnmaq_f16): Remove.
16668 (__arm_vminnmaq_f32): Remove.
16669 (__arm_vmaxnmaq_f32): Remove.
16670 (__arm_vmaxnmaq_m_f16): Remove.
16671 (__arm_vminnmaq_m_f16): Remove.
16672 (__arm_vmaxnmaq_m_f32): Remove.
16673 (__arm_vminnmaq_m_f32): Remove.
16674 (__arm_vminnmaq): Remove.
16675 (__arm_vmaxnmaq): Remove.
16676 (__arm_vmaxnmaq_m): Remove.
16677 (__arm_vminnmaq_m): Remove.
16679 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16681 * config/arm/iterators.md (MVE_VMAXNMA_VMINNMAQ)
16682 (MVE_VMAXNMA_VMINNMAQ_M): New.
16683 (mve_insn): Add vmaxnma, vminnma.
16684 * config/arm/mve.md (mve_vmaxnmaq_f<mode>, mve_vminnmaq_f<mode>):
16686 (@mve_<mve_insn>q_f<mode>): ... this.
16687 (mve_vmaxnmaq_m_f<mode>, mve_vminnmaq_m_f<mode>): Merge into ...
16688 (@mve_<mve_insn>q_m_f<mode>): ... this.
16690 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16692 * config/arm/arm-mve-builtins-base.cc (FUNCTION_PRED_P_F): New.
16693 (vmaxnmavq, vmaxnmvq, vminnmavq, vminnmvq): New.
16694 * config/arm/arm-mve-builtins-base.def (vmaxnmavq, vmaxnmvq)
16695 (vminnmavq, vminnmvq): New.
16696 * config/arm/arm-mve-builtins-base.h (vmaxnmavq, vmaxnmvq)
16697 (vminnmavq, vminnmvq): New.
16698 * config/arm/arm_mve.h (vminnmvq): Remove.
16699 (vminnmavq): Remove.
16700 (vmaxnmvq): Remove.
16701 (vmaxnmavq): Remove.
16702 (vmaxnmavq_p): Remove.
16703 (vmaxnmvq_p): Remove.
16704 (vminnmavq_p): Remove.
16705 (vminnmvq_p): Remove.
16706 (vminnmvq_f16): Remove.
16707 (vminnmavq_f16): Remove.
16708 (vmaxnmvq_f16): Remove.
16709 (vmaxnmavq_f16): Remove.
16710 (vminnmvq_f32): Remove.
16711 (vminnmavq_f32): Remove.
16712 (vmaxnmvq_f32): Remove.
16713 (vmaxnmavq_f32): Remove.
16714 (vmaxnmavq_p_f16): Remove.
16715 (vmaxnmvq_p_f16): Remove.
16716 (vminnmavq_p_f16): Remove.
16717 (vminnmvq_p_f16): Remove.
16718 (vmaxnmavq_p_f32): Remove.
16719 (vmaxnmvq_p_f32): Remove.
16720 (vminnmavq_p_f32): Remove.
16721 (vminnmvq_p_f32): Remove.
16722 (__arm_vminnmvq_f16): Remove.
16723 (__arm_vminnmavq_f16): Remove.
16724 (__arm_vmaxnmvq_f16): Remove.
16725 (__arm_vmaxnmavq_f16): Remove.
16726 (__arm_vminnmvq_f32): Remove.
16727 (__arm_vminnmavq_f32): Remove.
16728 (__arm_vmaxnmvq_f32): Remove.
16729 (__arm_vmaxnmavq_f32): Remove.
16730 (__arm_vmaxnmavq_p_f16): Remove.
16731 (__arm_vmaxnmvq_p_f16): Remove.
16732 (__arm_vminnmavq_p_f16): Remove.
16733 (__arm_vminnmvq_p_f16): Remove.
16734 (__arm_vmaxnmavq_p_f32): Remove.
16735 (__arm_vmaxnmvq_p_f32): Remove.
16736 (__arm_vminnmavq_p_f32): Remove.
16737 (__arm_vminnmvq_p_f32): Remove.
16738 (__arm_vminnmvq): Remove.
16739 (__arm_vminnmavq): Remove.
16740 (__arm_vmaxnmvq): Remove.
16741 (__arm_vmaxnmavq): Remove.
16742 (__arm_vmaxnmavq_p): Remove.
16743 (__arm_vmaxnmvq_p): Remove.
16744 (__arm_vminnmavq_p): Remove.
16745 (__arm_vminnmvq_p): Remove.
16746 (__arm_vmaxnmavq_m): Remove.
16747 (__arm_vmaxnmvq_m): Remove.
16749 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16751 * config/arm/arm-mve-builtins-functions.h
16752 (unspec_mve_function_exact_insn_pred_p): Use code_for_mve_q_p_f.
16754 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16756 * config/arm/iterators.md (MVE_VMAXNMxV_MINNMxVQ)
16757 (MVE_VMAXNMxV_MINNMxVQ_P): New.
16758 (mve_insn): Add vmaxnmav, vmaxnmv, vminnmav, vminnmv.
16759 * config/arm/mve.md (mve_vmaxnmavq_f<mode>, mve_vmaxnmvq_f<mode>)
16760 (mve_vminnmavq_f<mode>, mve_vminnmvq_f<mode>): Merge into ...
16761 (@mve_<mve_insn>q_f<mode>): ... this.
16762 (mve_vmaxnmavq_p_f<mode>, mve_vmaxnmvq_p_f<mode>)
16763 (mve_vminnmavq_p_f<mode>, mve_vminnmvq_p_f<mode>): Merge into ...
16764 (@mve_<mve_insn>q_p_f<mode>): ... this.
16766 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16768 * config/arm/arm-mve-builtins-base.cc (vmaxnmq, vminnmq): New.
16769 * config/arm/arm-mve-builtins-base.def (vmaxnmq, vminnmq): New.
16770 * config/arm/arm-mve-builtins-base.h (vmaxnmq, vminnmq): New.
16771 * config/arm/arm_mve.h (vminnmq): Remove.
16773 (vmaxnmq_m): Remove.
16774 (vminnmq_m): Remove.
16775 (vminnmq_x): Remove.
16776 (vmaxnmq_x): Remove.
16777 (vminnmq_f16): Remove.
16778 (vmaxnmq_f16): Remove.
16779 (vminnmq_f32): Remove.
16780 (vmaxnmq_f32): Remove.
16781 (vmaxnmq_m_f32): Remove.
16782 (vmaxnmq_m_f16): Remove.
16783 (vminnmq_m_f32): Remove.
16784 (vminnmq_m_f16): Remove.
16785 (vminnmq_x_f16): Remove.
16786 (vminnmq_x_f32): Remove.
16787 (vmaxnmq_x_f16): Remove.
16788 (vmaxnmq_x_f32): Remove.
16789 (__arm_vminnmq_f16): Remove.
16790 (__arm_vmaxnmq_f16): Remove.
16791 (__arm_vminnmq_f32): Remove.
16792 (__arm_vmaxnmq_f32): Remove.
16793 (__arm_vmaxnmq_m_f32): Remove.
16794 (__arm_vmaxnmq_m_f16): Remove.
16795 (__arm_vminnmq_m_f32): Remove.
16796 (__arm_vminnmq_m_f16): Remove.
16797 (__arm_vminnmq_x_f16): Remove.
16798 (__arm_vminnmq_x_f32): Remove.
16799 (__arm_vmaxnmq_x_f16): Remove.
16800 (__arm_vmaxnmq_x_f32): Remove.
16801 (__arm_vminnmq): Remove.
16802 (__arm_vmaxnmq): Remove.
16803 (__arm_vmaxnmq_m): Remove.
16804 (__arm_vminnmq_m): Remove.
16805 (__arm_vminnmq_x): Remove.
16806 (__arm_vmaxnmq_x): Remove.
16808 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16810 * config/arm/iterators.md (MAX_MIN_F): New.
16811 (MVE_FP_M_BINARY): Add VMAXNMQ_M_F, VMINNMQ_M_F.
16812 (mve_insn): Add vmaxnm, vminnm.
16813 (max_min_f_str): New.
16814 * config/arm/mve.md (mve_vmaxnmq_f<mode>, mve_vminnmq_f<mode>):
16816 (@mve_<max_min_f_str>q_f<mode>): ... this.
16817 (mve_vmaxnmq_m_f<mode>, mve_vminnmq_m_f<mode>): Merge into ...
16818 (@mve_<mve_insn>q_m_f<mode>): ... this.
16820 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16822 * config/arm/vec-common.md (smin<mode>3): Use VDQWH iterator.
16823 (smax<mode>3): Likewise.
16825 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16827 * config/arm/arm-mve-builtins-base.cc (FUNCTION_PRED_P_S_U)
16828 (FUNCTION_PRED_P_S): New.
16829 (vmaxavq, vminavq, vmaxvq, vminvq): New.
16830 * config/arm/arm-mve-builtins-base.def (vmaxavq, vminavq, vmaxvq)
16832 * config/arm/arm-mve-builtins-base.h (vmaxavq, vminavq, vmaxvq)
16834 * config/arm/arm_mve.h (vminvq): Remove.
16836 (vminvq_p): Remove.
16837 (vmaxvq_p): Remove.
16838 (vminvq_u8): Remove.
16839 (vmaxvq_u8): Remove.
16840 (vminvq_s8): Remove.
16841 (vmaxvq_s8): Remove.
16842 (vminvq_u16): Remove.
16843 (vmaxvq_u16): Remove.
16844 (vminvq_s16): Remove.
16845 (vmaxvq_s16): Remove.
16846 (vminvq_u32): Remove.
16847 (vmaxvq_u32): Remove.
16848 (vminvq_s32): Remove.
16849 (vmaxvq_s32): Remove.
16850 (vminvq_p_u8): Remove.
16851 (vmaxvq_p_u8): Remove.
16852 (vminvq_p_s8): Remove.
16853 (vmaxvq_p_s8): Remove.
16854 (vminvq_p_u16): Remove.
16855 (vmaxvq_p_u16): Remove.
16856 (vminvq_p_s16): Remove.
16857 (vmaxvq_p_s16): Remove.
16858 (vminvq_p_u32): Remove.
16859 (vmaxvq_p_u32): Remove.
16860 (vminvq_p_s32): Remove.
16861 (vmaxvq_p_s32): Remove.
16862 (__arm_vminvq_u8): Remove.
16863 (__arm_vmaxvq_u8): Remove.
16864 (__arm_vminvq_s8): Remove.
16865 (__arm_vmaxvq_s8): Remove.
16866 (__arm_vminvq_u16): Remove.
16867 (__arm_vmaxvq_u16): Remove.
16868 (__arm_vminvq_s16): Remove.
16869 (__arm_vmaxvq_s16): Remove.
16870 (__arm_vminvq_u32): Remove.
16871 (__arm_vmaxvq_u32): Remove.
16872 (__arm_vminvq_s32): Remove.
16873 (__arm_vmaxvq_s32): Remove.
16874 (__arm_vminvq_p_u8): Remove.
16875 (__arm_vmaxvq_p_u8): Remove.
16876 (__arm_vminvq_p_s8): Remove.
16877 (__arm_vmaxvq_p_s8): Remove.
16878 (__arm_vminvq_p_u16): Remove.
16879 (__arm_vmaxvq_p_u16): Remove.
16880 (__arm_vminvq_p_s16): Remove.
16881 (__arm_vmaxvq_p_s16): Remove.
16882 (__arm_vminvq_p_u32): Remove.
16883 (__arm_vmaxvq_p_u32): Remove.
16884 (__arm_vminvq_p_s32): Remove.
16885 (__arm_vmaxvq_p_s32): Remove.
16886 (__arm_vminvq): Remove.
16887 (__arm_vmaxvq): Remove.
16888 (__arm_vminvq_p): Remove.
16889 (__arm_vmaxvq_p): Remove.
16892 (vminavq_p): Remove.
16893 (vmaxavq_p): Remove.
16894 (vminavq_s8): Remove.
16895 (vmaxavq_s8): Remove.
16896 (vminavq_s16): Remove.
16897 (vmaxavq_s16): Remove.
16898 (vminavq_s32): Remove.
16899 (vmaxavq_s32): Remove.
16900 (vminavq_p_s8): Remove.
16901 (vmaxavq_p_s8): Remove.
16902 (vminavq_p_s16): Remove.
16903 (vmaxavq_p_s16): Remove.
16904 (vminavq_p_s32): Remove.
16905 (vmaxavq_p_s32): Remove.
16906 (__arm_vminavq_s8): Remove.
16907 (__arm_vmaxavq_s8): Remove.
16908 (__arm_vminavq_s16): Remove.
16909 (__arm_vmaxavq_s16): Remove.
16910 (__arm_vminavq_s32): Remove.
16911 (__arm_vmaxavq_s32): Remove.
16912 (__arm_vminavq_p_s8): Remove.
16913 (__arm_vmaxavq_p_s8): Remove.
16914 (__arm_vminavq_p_s16): Remove.
16915 (__arm_vmaxavq_p_s16): Remove.
16916 (__arm_vminavq_p_s32): Remove.
16917 (__arm_vmaxavq_p_s32): Remove.
16918 (__arm_vminavq): Remove.
16919 (__arm_vmaxavq): Remove.
16920 (__arm_vminavq_p): Remove.
16921 (__arm_vmaxavq_p): Remove.
16923 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16925 * config/arm/iterators.md (MVE_VMAXVQ_VMINVQ, MVE_VMAXVQ_VMINVQ_P): New.
16926 (mve_insn): Add vmaxav, vmaxv, vminav, vminv.
16927 (supf): Add VMAXAVQ_S, VMAXAVQ_P_S, VMINAVQ_S, VMINAVQ_P_S.
16928 * config/arm/mve.md (mve_vmaxavq_s<mode>, mve_vmaxvq_<supf><mode>)
16929 (mve_vminavq_s<mode>, mve_vminvq_<supf><mode>): Merge into ...
16930 (@mve_<mve_insn>q_<supf><mode>): ... this.
16931 (mve_vmaxavq_p_s<mode>, mve_vmaxvq_p_<supf><mode>)
16932 (mve_vminavq_p_s<mode>, mve_vminvq_p_<supf><mode>): Merge into ...
16933 (@mve_<mve_insn>q_p_<supf><mode>): ... this.
16935 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16937 * config/arm/arm-mve-builtins-functions.h (class
16938 unspec_mve_function_exact_insn_pred_p): New.
16940 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16942 * config/arm/arm-mve-builtins-shapes.cc (binary_maxavminav): New.
16943 * config/arm/arm-mve-builtins-shapes.h (binary_maxavminav): New.
16945 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16947 * config/arm/arm-mve-builtins-shapes.cc (binary_maxvminv): New.
16948 * config/arm/arm-mve-builtins-shapes.h (binary_maxvminv): New.
16950 2023-05-09 Richard Sandiford <richard.sandiford@arm.com>
16952 * config/aarch64/aarch64-protos.h (aarch64_adjust_reg_alloc_order):
16954 * config/aarch64/aarch64.h (REG_ALLOC_ORDER): Define.
16955 (ADJUST_REG_ALLOC_ORDER): Likewise.
16956 * config/aarch64/aarch64.cc (aarch64_adjust_reg_alloc_order): New
16958 * config/aarch64/aarch64-sve.md (*vcond_mask_<mode><vpred>): Use
16959 Upa rather than Upl for unpredicated movprfx alternatives.
16961 2023-05-09 Jeff Law <jlaw@ventanamicro>
16963 * config/h8300/testcompare.md: Add peephole2 which uses a memory
16964 load to set flags, thus eliminating a compare against zero.
16966 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
16968 * config/arm/arm-mve-builtins-base.cc (vshllbq, vshlltq): New.
16969 * config/arm/arm-mve-builtins-base.def (vshllbq, vshlltq): New.
16970 * config/arm/arm-mve-builtins-base.h (vshllbq, vshlltq): New.
16971 * config/arm/arm_mve.h (vshlltq): Remove.
16973 (vshllbq_m): Remove.
16974 (vshlltq_m): Remove.
16975 (vshllbq_x): Remove.
16976 (vshlltq_x): Remove.
16977 (vshlltq_n_u8): Remove.
16978 (vshllbq_n_u8): Remove.
16979 (vshlltq_n_s8): Remove.
16980 (vshllbq_n_s8): Remove.
16981 (vshlltq_n_u16): Remove.
16982 (vshllbq_n_u16): Remove.
16983 (vshlltq_n_s16): Remove.
16984 (vshllbq_n_s16): Remove.
16985 (vshllbq_m_n_s8): Remove.
16986 (vshllbq_m_n_s16): Remove.
16987 (vshllbq_m_n_u8): Remove.
16988 (vshllbq_m_n_u16): Remove.
16989 (vshlltq_m_n_s8): Remove.
16990 (vshlltq_m_n_s16): Remove.
16991 (vshlltq_m_n_u8): Remove.
16992 (vshlltq_m_n_u16): Remove.
16993 (vshllbq_x_n_s8): Remove.
16994 (vshllbq_x_n_s16): Remove.
16995 (vshllbq_x_n_u8): Remove.
16996 (vshllbq_x_n_u16): Remove.
16997 (vshlltq_x_n_s8): Remove.
16998 (vshlltq_x_n_s16): Remove.
16999 (vshlltq_x_n_u8): Remove.
17000 (vshlltq_x_n_u16): Remove.
17001 (__arm_vshlltq_n_u8): Remove.
17002 (__arm_vshllbq_n_u8): Remove.
17003 (__arm_vshlltq_n_s8): Remove.
17004 (__arm_vshllbq_n_s8): Remove.
17005 (__arm_vshlltq_n_u16): Remove.
17006 (__arm_vshllbq_n_u16): Remove.
17007 (__arm_vshlltq_n_s16): Remove.
17008 (__arm_vshllbq_n_s16): Remove.
17009 (__arm_vshllbq_m_n_s8): Remove.
17010 (__arm_vshllbq_m_n_s16): Remove.
17011 (__arm_vshllbq_m_n_u8): Remove.
17012 (__arm_vshllbq_m_n_u16): Remove.
17013 (__arm_vshlltq_m_n_s8): Remove.
17014 (__arm_vshlltq_m_n_s16): Remove.
17015 (__arm_vshlltq_m_n_u8): Remove.
17016 (__arm_vshlltq_m_n_u16): Remove.
17017 (__arm_vshllbq_x_n_s8): Remove.
17018 (__arm_vshllbq_x_n_s16): Remove.
17019 (__arm_vshllbq_x_n_u8): Remove.
17020 (__arm_vshllbq_x_n_u16): Remove.
17021 (__arm_vshlltq_x_n_s8): Remove.
17022 (__arm_vshlltq_x_n_s16): Remove.
17023 (__arm_vshlltq_x_n_u8): Remove.
17024 (__arm_vshlltq_x_n_u16): Remove.
17025 (__arm_vshlltq): Remove.
17026 (__arm_vshllbq): Remove.
17027 (__arm_vshllbq_m): Remove.
17028 (__arm_vshlltq_m): Remove.
17029 (__arm_vshllbq_x): Remove.
17030 (__arm_vshlltq_x): Remove.
17032 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17034 * config/arm/iterators.md (mve_insn): Add vshllb, vshllt.
17035 (VSHLLBQ_N, VSHLLTQ_N): Remove.
17037 (VSHLLBQ_M_N, VSHLLTQ_M_N): Remove.
17038 (VSHLLxQ_M_N): New.
17039 * config/arm/mve.md (mve_vshllbq_n_<supf><mode>)
17040 (mve_vshlltq_n_<supf><mode>): Merge into ...
17041 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
17042 (mve_vshllbq_m_n_<supf><mode>, mve_vshlltq_m_n_<supf><mode>):
17044 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
17046 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17048 * config/arm/arm-mve-builtins-shapes.cc (binary_widen_n): New.
17049 * config/arm/arm-mve-builtins-shapes.h (binary_widen_n): New.
17051 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17053 * config/arm/arm-mve-builtins-base.cc (vmovnbq, vmovntq, vqmovnbq)
17054 (vqmovntq, vqmovunbq, vqmovuntq): New.
17055 * config/arm/arm-mve-builtins-base.def (vmovnbq, vmovntq)
17056 (vqmovnbq, vqmovntq, vqmovunbq, vqmovuntq): New.
17057 * config/arm/arm-mve-builtins-base.h (vmovnbq, vmovntq, vqmovnbq)
17058 (vqmovntq, vqmovunbq, vqmovuntq): New.
17059 * config/arm/arm-mve-builtins.cc
17060 (function_instance::has_inactive_argument): Handle vmovnbq,
17061 vmovntq, vqmovnbq, vqmovntq, vqmovunbq, vqmovuntq.
17062 * config/arm/arm_mve.h (vqmovntq): Remove.
17063 (vqmovnbq): Remove.
17064 (vqmovnbq_m): Remove.
17065 (vqmovntq_m): Remove.
17066 (vqmovntq_u16): Remove.
17067 (vqmovnbq_u16): Remove.
17068 (vqmovntq_s16): Remove.
17069 (vqmovnbq_s16): Remove.
17070 (vqmovntq_u32): Remove.
17071 (vqmovnbq_u32): Remove.
17072 (vqmovntq_s32): Remove.
17073 (vqmovnbq_s32): Remove.
17074 (vqmovnbq_m_s16): Remove.
17075 (vqmovntq_m_s16): Remove.
17076 (vqmovnbq_m_u16): Remove.
17077 (vqmovntq_m_u16): Remove.
17078 (vqmovnbq_m_s32): Remove.
17079 (vqmovntq_m_s32): Remove.
17080 (vqmovnbq_m_u32): Remove.
17081 (vqmovntq_m_u32): Remove.
17082 (__arm_vqmovntq_u16): Remove.
17083 (__arm_vqmovnbq_u16): Remove.
17084 (__arm_vqmovntq_s16): Remove.
17085 (__arm_vqmovnbq_s16): Remove.
17086 (__arm_vqmovntq_u32): Remove.
17087 (__arm_vqmovnbq_u32): Remove.
17088 (__arm_vqmovntq_s32): Remove.
17089 (__arm_vqmovnbq_s32): Remove.
17090 (__arm_vqmovnbq_m_s16): Remove.
17091 (__arm_vqmovntq_m_s16): Remove.
17092 (__arm_vqmovnbq_m_u16): Remove.
17093 (__arm_vqmovntq_m_u16): Remove.
17094 (__arm_vqmovnbq_m_s32): Remove.
17095 (__arm_vqmovntq_m_s32): Remove.
17096 (__arm_vqmovnbq_m_u32): Remove.
17097 (__arm_vqmovntq_m_u32): Remove.
17098 (__arm_vqmovntq): Remove.
17099 (__arm_vqmovnbq): Remove.
17100 (__arm_vqmovnbq_m): Remove.
17101 (__arm_vqmovntq_m): Remove.
17104 (vmovnbq_m): Remove.
17105 (vmovntq_m): Remove.
17106 (vmovntq_u16): Remove.
17107 (vmovnbq_u16): Remove.
17108 (vmovntq_s16): Remove.
17109 (vmovnbq_s16): Remove.
17110 (vmovntq_u32): Remove.
17111 (vmovnbq_u32): Remove.
17112 (vmovntq_s32): Remove.
17113 (vmovnbq_s32): Remove.
17114 (vmovnbq_m_s16): Remove.
17115 (vmovntq_m_s16): Remove.
17116 (vmovnbq_m_u16): Remove.
17117 (vmovntq_m_u16): Remove.
17118 (vmovnbq_m_s32): Remove.
17119 (vmovntq_m_s32): Remove.
17120 (vmovnbq_m_u32): Remove.
17121 (vmovntq_m_u32): Remove.
17122 (__arm_vmovntq_u16): Remove.
17123 (__arm_vmovnbq_u16): Remove.
17124 (__arm_vmovntq_s16): Remove.
17125 (__arm_vmovnbq_s16): Remove.
17126 (__arm_vmovntq_u32): Remove.
17127 (__arm_vmovnbq_u32): Remove.
17128 (__arm_vmovntq_s32): Remove.
17129 (__arm_vmovnbq_s32): Remove.
17130 (__arm_vmovnbq_m_s16): Remove.
17131 (__arm_vmovntq_m_s16): Remove.
17132 (__arm_vmovnbq_m_u16): Remove.
17133 (__arm_vmovntq_m_u16): Remove.
17134 (__arm_vmovnbq_m_s32): Remove.
17135 (__arm_vmovntq_m_s32): Remove.
17136 (__arm_vmovnbq_m_u32): Remove.
17137 (__arm_vmovntq_m_u32): Remove.
17138 (__arm_vmovntq): Remove.
17139 (__arm_vmovnbq): Remove.
17140 (__arm_vmovnbq_m): Remove.
17141 (__arm_vmovntq_m): Remove.
17142 (vqmovuntq): Remove.
17143 (vqmovunbq): Remove.
17144 (vqmovunbq_m): Remove.
17145 (vqmovuntq_m): Remove.
17146 (vqmovuntq_s16): Remove.
17147 (vqmovunbq_s16): Remove.
17148 (vqmovuntq_s32): Remove.
17149 (vqmovunbq_s32): Remove.
17150 (vqmovunbq_m_s16): Remove.
17151 (vqmovuntq_m_s16): Remove.
17152 (vqmovunbq_m_s32): Remove.
17153 (vqmovuntq_m_s32): Remove.
17154 (__arm_vqmovuntq_s16): Remove.
17155 (__arm_vqmovunbq_s16): Remove.
17156 (__arm_vqmovuntq_s32): Remove.
17157 (__arm_vqmovunbq_s32): Remove.
17158 (__arm_vqmovunbq_m_s16): Remove.
17159 (__arm_vqmovuntq_m_s16): Remove.
17160 (__arm_vqmovunbq_m_s32): Remove.
17161 (__arm_vqmovuntq_m_s32): Remove.
17162 (__arm_vqmovuntq): Remove.
17163 (__arm_vqmovunbq): Remove.
17164 (__arm_vqmovunbq_m): Remove.
17165 (__arm_vqmovuntq_m): Remove.
17167 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17169 * config/arm/iterators.md (MVE_MOVN, MVE_MOVN_M): New.
17170 (mve_insn): Add vmovnb, vmovnt, vqmovnb, vqmovnt, vqmovunb,
17173 (supf): Add VQMOVUNBQ_M_S, VQMOVUNBQ_S, VQMOVUNTQ_M_S,
17175 * config/arm/mve.md (mve_vmovnbq_<supf><mode>)
17176 (mve_vmovntq_<supf><mode>, mve_vqmovnbq_<supf><mode>)
17177 (mve_vqmovntq_<supf><mode>, mve_vqmovunbq_s<mode>)
17178 (mve_vqmovuntq_s<mode>): Merge into ...
17179 (@mve_<mve_insn>q_<supf><mode>): ... this.
17180 (mve_vmovnbq_m_<supf><mode>, mve_vmovntq_m_<supf><mode>)
17181 (mve_vqmovnbq_m_<supf><mode>, mve_vqmovntq_m_<supf><mode>)
17182 (mve_vqmovunbq_m_s<mode>, mve_vqmovuntq_m_s<mode>): Merge into ...
17183 (@mve_<mve_insn>q_m_<supf><mode>): ... this.
17185 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17187 * config/arm/arm-mve-builtins-shapes.cc (binary_move_narrow): New.
17188 (binary_move_narrow_unsigned): New.
17189 * config/arm/arm-mve-builtins-shapes.h (binary_move_narrow): New.
17190 (binary_move_narrow_unsigned): New.
17192 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17194 * config/arm/arm-mve-builtins-base.cc (FUNCTION_ONLY_F): New.
17195 (vrndaq, vrndmq, vrndnq, vrndpq, vrndq, vrndxq): New.
17196 * config/arm/arm-mve-builtins-base.def (vrndaq, vrndmq, vrndnq)
17197 (vrndpq, vrndq, vrndxq): New.
17198 * config/arm/arm-mve-builtins-base.h (vrndaq, vrndmq, vrndnq)
17199 (vrndpq, vrndq, vrndxq): New.
17200 * config/arm/arm_mve.h (vrndxq): Remove.
17206 (vrndaq_m): Remove.
17207 (vrndmq_m): Remove.
17208 (vrndnq_m): Remove.
17209 (vrndpq_m): Remove.
17211 (vrndxq_m): Remove.
17213 (vrndnq_x): Remove.
17214 (vrndmq_x): Remove.
17215 (vrndpq_x): Remove.
17216 (vrndaq_x): Remove.
17217 (vrndxq_x): Remove.
17218 (vrndxq_f16): Remove.
17219 (vrndxq_f32): Remove.
17220 (vrndq_f16): Remove.
17221 (vrndq_f32): Remove.
17222 (vrndpq_f16): Remove.
17223 (vrndpq_f32): Remove.
17224 (vrndnq_f16): Remove.
17225 (vrndnq_f32): Remove.
17226 (vrndmq_f16): Remove.
17227 (vrndmq_f32): Remove.
17228 (vrndaq_f16): Remove.
17229 (vrndaq_f32): Remove.
17230 (vrndaq_m_f16): Remove.
17231 (vrndmq_m_f16): Remove.
17232 (vrndnq_m_f16): Remove.
17233 (vrndpq_m_f16): Remove.
17234 (vrndq_m_f16): Remove.
17235 (vrndxq_m_f16): Remove.
17236 (vrndaq_m_f32): Remove.
17237 (vrndmq_m_f32): Remove.
17238 (vrndnq_m_f32): Remove.
17239 (vrndpq_m_f32): Remove.
17240 (vrndq_m_f32): Remove.
17241 (vrndxq_m_f32): Remove.
17242 (vrndq_x_f16): Remove.
17243 (vrndq_x_f32): Remove.
17244 (vrndnq_x_f16): Remove.
17245 (vrndnq_x_f32): Remove.
17246 (vrndmq_x_f16): Remove.
17247 (vrndmq_x_f32): Remove.
17248 (vrndpq_x_f16): Remove.
17249 (vrndpq_x_f32): Remove.
17250 (vrndaq_x_f16): Remove.
17251 (vrndaq_x_f32): Remove.
17252 (vrndxq_x_f16): Remove.
17253 (vrndxq_x_f32): Remove.
17254 (__arm_vrndxq_f16): Remove.
17255 (__arm_vrndxq_f32): Remove.
17256 (__arm_vrndq_f16): Remove.
17257 (__arm_vrndq_f32): Remove.
17258 (__arm_vrndpq_f16): Remove.
17259 (__arm_vrndpq_f32): Remove.
17260 (__arm_vrndnq_f16): Remove.
17261 (__arm_vrndnq_f32): Remove.
17262 (__arm_vrndmq_f16): Remove.
17263 (__arm_vrndmq_f32): Remove.
17264 (__arm_vrndaq_f16): Remove.
17265 (__arm_vrndaq_f32): Remove.
17266 (__arm_vrndaq_m_f16): Remove.
17267 (__arm_vrndmq_m_f16): Remove.
17268 (__arm_vrndnq_m_f16): Remove.
17269 (__arm_vrndpq_m_f16): Remove.
17270 (__arm_vrndq_m_f16): Remove.
17271 (__arm_vrndxq_m_f16): Remove.
17272 (__arm_vrndaq_m_f32): Remove.
17273 (__arm_vrndmq_m_f32): Remove.
17274 (__arm_vrndnq_m_f32): Remove.
17275 (__arm_vrndpq_m_f32): Remove.
17276 (__arm_vrndq_m_f32): Remove.
17277 (__arm_vrndxq_m_f32): Remove.
17278 (__arm_vrndq_x_f16): Remove.
17279 (__arm_vrndq_x_f32): Remove.
17280 (__arm_vrndnq_x_f16): Remove.
17281 (__arm_vrndnq_x_f32): Remove.
17282 (__arm_vrndmq_x_f16): Remove.
17283 (__arm_vrndmq_x_f32): Remove.
17284 (__arm_vrndpq_x_f16): Remove.
17285 (__arm_vrndpq_x_f32): Remove.
17286 (__arm_vrndaq_x_f16): Remove.
17287 (__arm_vrndaq_x_f32): Remove.
17288 (__arm_vrndxq_x_f16): Remove.
17289 (__arm_vrndxq_x_f32): Remove.
17290 (__arm_vrndxq): Remove.
17291 (__arm_vrndq): Remove.
17292 (__arm_vrndpq): Remove.
17293 (__arm_vrndnq): Remove.
17294 (__arm_vrndmq): Remove.
17295 (__arm_vrndaq): Remove.
17296 (__arm_vrndaq_m): Remove.
17297 (__arm_vrndmq_m): Remove.
17298 (__arm_vrndnq_m): Remove.
17299 (__arm_vrndpq_m): Remove.
17300 (__arm_vrndq_m): Remove.
17301 (__arm_vrndxq_m): Remove.
17302 (__arm_vrndq_x): Remove.
17303 (__arm_vrndnq_x): Remove.
17304 (__arm_vrndmq_x): Remove.
17305 (__arm_vrndpq_x): Remove.
17306 (__arm_vrndaq_x): Remove.
17307 (__arm_vrndxq_x): Remove.
17309 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17311 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITHOUT_N_NO_U_F): New.
17312 (vabsq, vnegq, vclsq, vclzq, vqabsq, vqnegq): New.
17313 * config/arm/arm-mve-builtins-base.def (vabsq, vnegq, vclsq)
17314 (vclzq, vqabsq, vqnegq): New.
17315 * config/arm/arm-mve-builtins-base.h (vabsq, vnegq, vclsq, vclzq)
17316 (vqabsq, vqnegq): New.
17317 * config/arm/arm_mve.h (vabsq): Remove.
17320 (vabsq_f16): Remove.
17321 (vabsq_f32): Remove.
17322 (vabsq_s8): Remove.
17323 (vabsq_s16): Remove.
17324 (vabsq_s32): Remove.
17325 (vabsq_m_s8): Remove.
17326 (vabsq_m_s16): Remove.
17327 (vabsq_m_s32): Remove.
17328 (vabsq_m_f16): Remove.
17329 (vabsq_m_f32): Remove.
17330 (vabsq_x_s8): Remove.
17331 (vabsq_x_s16): Remove.
17332 (vabsq_x_s32): Remove.
17333 (vabsq_x_f16): Remove.
17334 (vabsq_x_f32): Remove.
17335 (__arm_vabsq_s8): Remove.
17336 (__arm_vabsq_s16): Remove.
17337 (__arm_vabsq_s32): Remove.
17338 (__arm_vabsq_m_s8): Remove.
17339 (__arm_vabsq_m_s16): Remove.
17340 (__arm_vabsq_m_s32): Remove.
17341 (__arm_vabsq_x_s8): Remove.
17342 (__arm_vabsq_x_s16): Remove.
17343 (__arm_vabsq_x_s32): Remove.
17344 (__arm_vabsq_f16): Remove.
17345 (__arm_vabsq_f32): Remove.
17346 (__arm_vabsq_m_f16): Remove.
17347 (__arm_vabsq_m_f32): Remove.
17348 (__arm_vabsq_x_f16): Remove.
17349 (__arm_vabsq_x_f32): Remove.
17350 (__arm_vabsq): Remove.
17351 (__arm_vabsq_m): Remove.
17352 (__arm_vabsq_x): Remove.
17356 (vnegq_f16): Remove.
17357 (vnegq_f32): Remove.
17358 (vnegq_s8): Remove.
17359 (vnegq_s16): Remove.
17360 (vnegq_s32): Remove.
17361 (vnegq_m_s8): Remove.
17362 (vnegq_m_s16): Remove.
17363 (vnegq_m_s32): Remove.
17364 (vnegq_m_f16): Remove.
17365 (vnegq_m_f32): Remove.
17366 (vnegq_x_s8): Remove.
17367 (vnegq_x_s16): Remove.
17368 (vnegq_x_s32): Remove.
17369 (vnegq_x_f16): Remove.
17370 (vnegq_x_f32): Remove.
17371 (__arm_vnegq_s8): Remove.
17372 (__arm_vnegq_s16): Remove.
17373 (__arm_vnegq_s32): Remove.
17374 (__arm_vnegq_m_s8): Remove.
17375 (__arm_vnegq_m_s16): Remove.
17376 (__arm_vnegq_m_s32): Remove.
17377 (__arm_vnegq_x_s8): Remove.
17378 (__arm_vnegq_x_s16): Remove.
17379 (__arm_vnegq_x_s32): Remove.
17380 (__arm_vnegq_f16): Remove.
17381 (__arm_vnegq_f32): Remove.
17382 (__arm_vnegq_m_f16): Remove.
17383 (__arm_vnegq_m_f32): Remove.
17384 (__arm_vnegq_x_f16): Remove.
17385 (__arm_vnegq_x_f32): Remove.
17386 (__arm_vnegq): Remove.
17387 (__arm_vnegq_m): Remove.
17388 (__arm_vnegq_x): Remove.
17392 (vclsq_s8): Remove.
17393 (vclsq_s16): Remove.
17394 (vclsq_s32): Remove.
17395 (vclsq_m_s8): Remove.
17396 (vclsq_m_s16): Remove.
17397 (vclsq_m_s32): Remove.
17398 (vclsq_x_s8): Remove.
17399 (vclsq_x_s16): Remove.
17400 (vclsq_x_s32): Remove.
17401 (__arm_vclsq_s8): Remove.
17402 (__arm_vclsq_s16): Remove.
17403 (__arm_vclsq_s32): Remove.
17404 (__arm_vclsq_m_s8): Remove.
17405 (__arm_vclsq_m_s16): Remove.
17406 (__arm_vclsq_m_s32): Remove.
17407 (__arm_vclsq_x_s8): Remove.
17408 (__arm_vclsq_x_s16): Remove.
17409 (__arm_vclsq_x_s32): Remove.
17410 (__arm_vclsq): Remove.
17411 (__arm_vclsq_m): Remove.
17412 (__arm_vclsq_x): Remove.
17416 (vclzq_s8): Remove.
17417 (vclzq_s16): Remove.
17418 (vclzq_s32): Remove.
17419 (vclzq_u8): Remove.
17420 (vclzq_u16): Remove.
17421 (vclzq_u32): Remove.
17422 (vclzq_m_u8): Remove.
17423 (vclzq_m_s8): Remove.
17424 (vclzq_m_u16): Remove.
17425 (vclzq_m_s16): Remove.
17426 (vclzq_m_u32): Remove.
17427 (vclzq_m_s32): Remove.
17428 (vclzq_x_s8): Remove.
17429 (vclzq_x_s16): Remove.
17430 (vclzq_x_s32): Remove.
17431 (vclzq_x_u8): Remove.
17432 (vclzq_x_u16): Remove.
17433 (vclzq_x_u32): Remove.
17434 (__arm_vclzq_s8): Remove.
17435 (__arm_vclzq_s16): Remove.
17436 (__arm_vclzq_s32): Remove.
17437 (__arm_vclzq_u8): Remove.
17438 (__arm_vclzq_u16): Remove.
17439 (__arm_vclzq_u32): Remove.
17440 (__arm_vclzq_m_u8): Remove.
17441 (__arm_vclzq_m_s8): Remove.
17442 (__arm_vclzq_m_u16): Remove.
17443 (__arm_vclzq_m_s16): Remove.
17444 (__arm_vclzq_m_u32): Remove.
17445 (__arm_vclzq_m_s32): Remove.
17446 (__arm_vclzq_x_s8): Remove.
17447 (__arm_vclzq_x_s16): Remove.
17448 (__arm_vclzq_x_s32): Remove.
17449 (__arm_vclzq_x_u8): Remove.
17450 (__arm_vclzq_x_u16): Remove.
17451 (__arm_vclzq_x_u32): Remove.
17452 (__arm_vclzq): Remove.
17453 (__arm_vclzq_m): Remove.
17454 (__arm_vclzq_x): Remove.
17457 (vqnegq_m): Remove.
17458 (vqabsq_m): Remove.
17459 (vqabsq_s8): Remove.
17460 (vqabsq_s16): Remove.
17461 (vqabsq_s32): Remove.
17462 (vqnegq_s8): Remove.
17463 (vqnegq_s16): Remove.
17464 (vqnegq_s32): Remove.
17465 (vqnegq_m_s8): Remove.
17466 (vqabsq_m_s8): Remove.
17467 (vqnegq_m_s16): Remove.
17468 (vqabsq_m_s16): Remove.
17469 (vqnegq_m_s32): Remove.
17470 (vqabsq_m_s32): Remove.
17471 (__arm_vqabsq_s8): Remove.
17472 (__arm_vqabsq_s16): Remove.
17473 (__arm_vqabsq_s32): Remove.
17474 (__arm_vqnegq_s8): Remove.
17475 (__arm_vqnegq_s16): Remove.
17476 (__arm_vqnegq_s32): Remove.
17477 (__arm_vqnegq_m_s8): Remove.
17478 (__arm_vqabsq_m_s8): Remove.
17479 (__arm_vqnegq_m_s16): Remove.
17480 (__arm_vqabsq_m_s16): Remove.
17481 (__arm_vqnegq_m_s32): Remove.
17482 (__arm_vqabsq_m_s32): Remove.
17483 (__arm_vqabsq): Remove.
17484 (__arm_vqnegq): Remove.
17485 (__arm_vqnegq_m): Remove.
17486 (__arm_vqabsq_m): Remove.
17488 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17490 * config/arm/iterators.md (MVE_INT_M_UNARY, MVE_INT_UNARY)
17491 (MVE_FP_UNARY, MVE_FP_M_UNARY): New.
17492 (mve_insn): Add vabs, vcls, vclz, vneg, vqabs, vqneg, vrnda,
17493 vrndm, vrndn, vrndp, vrnd, vrndx.
17494 (isu): Add VABSQ_M_S, VCLSQ_M_S, VCLZQ_M_S, VCLZQ_M_U, VNEGQ_M_S,
17495 VQABSQ_M_S, VQNEGQ_M_S.
17497 * config/arm/mve.md (mve_vrndq_m_f<mode>, mve_vrndxq_f<mode>)
17498 (mve_vrndq_f<mode>, mve_vrndpq_f<mode>, mve_vrndnq_f<mode>)
17499 (mve_vrndmq_f<mode>, mve_vrndaq_f<mode>): Merge into ...
17500 (@mve_<mve_insn>q_f<mode>): ... this.
17501 (mve_vnegq_f<mode>, mve_vabsq_f<mode>): Merge into ...
17502 (mve_v<absneg_str>q_f<mode>): ... this.
17503 (mve_vnegq_s<mode>, mve_vabsq_s<mode>): Merge into ...
17504 (mve_v<absneg_str>q_s<mode>): ... this.
17505 (mve_vclsq_s<mode>, mve_vqnegq_s<mode>, mve_vqabsq_s<mode>): Merge into ...
17506 (@mve_<mve_insn>q_<supf><mode>): ... this.
17507 (mve_vabsq_m_s<mode>, mve_vclsq_m_s<mode>)
17508 (mve_vclzq_m_<supf><mode>, mve_vnegq_m_s<mode>)
17509 (mve_vqabsq_m_s<mode>, mve_vqnegq_m_s<mode>): Merge into ...
17510 (@mve_<mve_insn>q_m_<supf><mode>): ... this.
17511 (mve_vabsq_m_f<mode>, mve_vnegq_m_f<mode>, mve_vrndaq_m_f<mode>)
17512 (mve_vrndmq_m_f<mode>, mve_vrndnq_m_f<mode>, mve_vrndpq_m_f<mode>)
17513 (mve_vrndxq_m_f<mode>): Merge into ...
17514 (@mve_<mve_insn>q_m_f<mode>): ... this.
17516 2023-05-09 Christophe Lyon <christophe.lyon@arm.com>
17518 * config/arm/arm-mve-builtins-shapes.cc (unary): New.
17519 * config/arm/arm-mve-builtins-shapes.h (unary): New.
17521 2023-05-09 Jakub Jelinek <jakub@redhat.com>
17523 * mux-utils.h: Fix comment typo, avoides -> avoids.
17525 2023-05-09 Jakub Jelinek <jakub@redhat.com>
17527 PR tree-optimization/109778
17528 * wide-int.h (wi::lrotate, wi::rrotate): Call wi::lrshift on
17529 wi::zext (x, width) rather than x if width != precision, rather
17530 than using wi::zext (right, width) after the shift.
17531 * tree-ssa-ccp.cc (bit_value_binop): Call wi::ext on the results
17532 of wi::lrotate or wi::rrotate.
17534 2023-05-09 Alexander Monakov <amonakov@ispras.ru>
17536 * genmatch.cc (get_out_file): Make static and rename to ...
17537 (choose_output): ... this. Reimplement. Update all uses ...
17538 (decision_tree::gen): ... here and ...
17541 2023-05-09 Alexander Monakov <amonakov@ispras.ru>
17543 * genmatch.cc (showUsage): Reimplement as ...
17544 (usage): ...this. Adjust all uses.
17545 (main): Print usage when no arguments. Add missing 'return 1'.
17547 2023-05-09 Alexander Monakov <amonakov@ispras.ru>
17549 * genmatch.cc (header_file): Make static.
17550 (emit_func): Rename to...
17551 (fp_decl): ... this. Adjust all uses.
17552 (fp_decl_done): New function. Use it...
17553 (decision_tree::gen): ... here and...
17554 (write_predicate): ... here.
17557 2023-05-09 Richard Sandiford <richard.sandiford@arm.com>
17559 * ira-conflicts.cc (can_use_same_reg_p): Skip over non-matching
17562 2023-05-08 Roger Sayle <roger@nextmovesoftware.com>
17563 Uros Bizjak <ubizjak@gmail.com>
17565 * config/i386/i386.md (any_or_plus): Move definition earlier.
17566 (*insvti_highpart_1): New define_insn_and_split to overwrite
17567 (insv) the highpart of a TImode register/memory.
17569 2023-05-08 Eugene Rozenfeld <erozen@microsoft.com>
17571 * auto-profile.cc (auto_profile): Check todo from early_inline
17572 to see if cleanup_tree_vfg needs to be called.
17573 (early_inline): Return todo from early_inliner.
17575 2023-05-08 Kito Cheng <kito.cheng@sifive.com>
17577 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::get_vector_info):
17579 (pass_vsetvl::get_block_info): New.
17580 (pass_vsetvl::update_vector_info): New.
17581 (pass_vsetvl::simple_vsetvl): Use get_vector_info.
17582 (pass_vsetvl::compute_local_backward_infos): Ditto.
17583 (pass_vsetvl::transfer_before): Ditto.
17584 (pass_vsetvl::transfer_after): Ditto.
17585 (pass_vsetvl::emit_local_forward_vsetvls): Ditto.
17586 (pass_vsetvl::local_eliminate_vsetvl_insn): Ditto.
17587 (pass_vsetvl::cleanup_insns): Ditto.
17588 (pass_vsetvl::compute_local_backward_infos): Use
17589 update_vector_info.
17591 2023-05-08 Jeff Law <jlaw@ventanamicro>
17593 * config/stormy16/stormy16.md (zero_extendhisi2): Fix length.
17595 2023-05-08 Richard Biener <rguenther@suse.de>
17596 Michael Meissner <meissner@linux.ibm.com>
17598 PR middle-end/108623
17599 * tree-core.h (tree_type_common): Bump up precision field to 16 bits.
17600 Align bit fields > 1 bit to at least an 8-bit boundary.
17602 2023-05-08 Andrew Pinski <apinski@marvell.com>
17604 PR tree-optimization/109424
17605 PR tree-optimization/59424
17606 * tree-ssa-phiopt.cc (factor_out_conditional_conversion): Rename to ...
17607 (factor_out_conditional_operation): This and add support for all unary
17609 (pass_phiopt::execute): Update call to factor_out_conditional_conversion
17610 to call factor_out_conditional_operation instead.
17612 2023-05-08 Andrew Pinski <apinski@marvell.com>
17614 * tree-ssa-phiopt.cc (pass_phiopt::execute): Loop
17615 over factor_out_conditional_conversion.
17617 2023-05-08 Andrew Pinski <apinski@marvell.com>
17619 PR tree-optimization/49959
17620 PR tree-optimization/103771
17621 * tree-ssa-phiopt.cc (pass_phiopt::execute): Support
17622 Diamond shapped bb form for factor_out_conditional_conversion.
17624 2023-05-08 Juzhe-Zhong <juzhe.zhong@rivai.ai>
17626 * config/riscv/autovec.md (movmisalign<mode>): New pattern.
17627 * config/riscv/riscv-protos.h (riscv_vector_mask_mode_p): Delete.
17628 (riscv_vector_get_mask_mode): Ditto.
17629 (get_mask_policy_no_pred): Ditto.
17630 (get_tail_policy_no_pred): Ditto.
17631 (get_mask_mode): New function.
17632 * config/riscv/riscv-v.cc (get_mask_policy_no_pred): Delete.
17633 (get_tail_policy_no_pred): Ditto.
17634 (riscv_vector_mask_mode_p): Ditto.
17635 (riscv_vector_get_mask_mode): Ditto.
17636 (get_mask_mode): New function.
17637 * config/riscv/riscv-vector-builtins.cc (use_real_merge_p): Remove
17639 (get_tail_policy_for_pred): Ditto.
17640 * config/riscv/riscv-vector-builtins.h (get_tail_policy_for_pred): Ditto.
17641 (get_mask_policy_for_pred): Ditto
17642 * config/riscv/riscv.cc (riscv_get_mask_mode): Refine codes.
17644 2023-05-08 Kito Cheng <kito.cheng@sifive.com>
17646 * common/config/riscv/riscv-common.cc (riscv_select_multilib_by_abi): New.
17647 (riscv_select_multilib): New.
17648 (riscv_compute_multilib): Extract logic to riscv_select_multilib and
17649 also handle select_by_abi.
17650 * config/riscv/elf.h (RISCV_USE_CUSTOMISED_MULTI_LIB): Change it
17651 to select_by_abi_arch_cmodel from 1.
17652 * config/riscv/linux.h (RISCV_USE_CUSTOMISED_MULTI_LIB): Define.
17653 * config/riscv/riscv-opts.h (enum riscv_multilib_select_kind): New.
17655 2023-05-08 Alexander Monakov <amonakov@ispras.ru>
17657 * Makefile.in: (gimple-match-head.o-warn): Remove.
17658 (GIMPLE_MATCH_PD_SEQ_SRC): Do not depend on
17659 gimple-match-exports.cc.
17660 (gimple-match-auto.h): Only depend on s-gimple-match.
17661 (generic-match-auto.h): Likewise.
17663 2023-05-08 Andrew Pinski <apinski@marvell.com>
17665 PR tree-optimization/109691
17666 * tree-ssa-dce.cc (simple_dce_from_worklist): Add need_eh_cleanup
17668 If the removed statement can throw, have need_eh_cleanup
17669 include the bb of that statement.
17670 * tree-ssa-dce.h (simple_dce_from_worklist): Update declaration.
17671 * tree-ssa-propagate.cc (struct prop_stats_d): Remove
17673 (substitute_and_fold_dom_walker::substitute_and_fold_dom_walker):
17674 Initialize dceworklist instead of stmts_to_remove.
17675 (substitute_and_fold_dom_walker::~substitute_and_fold_dom_walker):
17676 Destore dceworklist instead of stmts_to_remove.
17677 (substitute_and_fold_dom_walker::before_dom_children):
17678 Set dceworklist instead of adding to stmts_to_remove.
17679 (substitute_and_fold_engine::substitute_and_fold):
17680 Call simple_dce_from_worklist instead of poping
17682 Don't update the stat on removal statements.
17684 2023-05-07 Andrew Pinski <apinski@marvell.com>
17687 * config/aarch64/aarch64-builtins.cc (aarch64_simd_switcher::aarch64_simd_switcher):
17688 Change argument type to aarch64_feature_flags.
17689 * config/aarch64/aarch64-protos.h (aarch64_simd_switcher): Change
17690 constructor argument type to aarch64_feature_flags.
17691 Change m_old_asm_isa_flags to be aarch64_feature_flags.
17693 2023-05-07 Jiufu Guo <guojiufu@linux.ibm.com>
17695 * config/rs6000/rs6000.cc (rs6000_emit_set_long_const): Generate
17696 more parallel code if can_create_pseudo_p.
17698 2023-05-07 Roger Sayle <roger@nextmovesoftware.com>
17701 * lower-subreg.cc (resolve_simple_move): Don't emit a clobber
17702 immediately before moving a multi-word register by parts.
17704 2023-05-06 Jeff Law <jlaw@ventanamicro>
17706 * config/riscv/riscv-v.cc (riscv_vector_preferred_simd_mode): Delete.
17708 2023-05-06 Michael Collison <collison@rivosinc.com>
17710 * tree-vect-slp.cc (can_duplicate_and_interleave_p):
17711 Check that GET_MODE_NUNITS is a multiple of 2.
17713 2023-05-06 Michael Collison <collison@rivosinc.com>
17715 * config/riscv/riscv.cc
17716 (riscv_estimated_poly_value): Implement
17717 TARGET_ESTIMATED_POLY_VALUE.
17718 (riscv_preferred_simd_mode): Implement
17719 TARGET_VECTORIZE_PREFERRED_SIMD_MODE.
17720 (riscv_get_mask_mode): Implement TARGET_VECTORIZE_GET_MASK_MODE.
17721 (riscv_empty_mask_is_expensive): Implement
17722 TARGET_VECTORIZE_EMPTY_MASK_IS_EXPENSIVE.
17723 (riscv_vectorize_create_costs): Implement
17724 TARGET_VECTORIZE_CREATE_COSTS.
17725 (riscv_support_vector_misalignment): Implement
17726 TARGET_VECTORIZE_SUPPORT_VECTOR_MISALIGNMENT.
17727 (TARGET_ESTIMATED_POLY_VALUE): Register target macro.
17728 (TARGET_VECTORIZE_GET_MASK_MODE): Ditto.
17729 (TARGET_VECTORIZE_EMPTY_MASK_IS_EXPENSIVE): Ditto.
17730 (TARGET_VECTORIZE_SUPPORT_VECTOR_MISALIGNMENT): Ditto.
17732 2023-05-06 Jeff Law <jlaw@ventanamicro>
17734 * config/riscv/riscv-v.cc (autovec_use_vlmax_p): Remove
17735 duplicate definition.
17737 2023-05-06 Michael Collison <collison@rivosinc.com>
17739 * config/riscv/riscv-v.cc (autovec_use_vlmax_p): New function.
17740 (riscv_vector_preferred_simd_mode): Ditto.
17741 (get_mask_policy_no_pred): Ditto.
17742 (get_tail_policy_no_pred): Ditto.
17743 (riscv_vector_mask_mode_p): Ditto.
17744 (riscv_vector_get_mask_mode): Ditto.
17746 2023-05-06 Michael Collison <collison@rivosinc.com>
17748 * config/riscv/riscv-vector-builtins.cc (get_tail_policy_for_pred):
17749 Remove static declaration to to make externally visible.
17750 (get_mask_policy_for_pred): Ditto.
17751 * config/riscv/riscv-vector-builtins.h (get_tail_policy_for_pred):
17752 New external declaration.
17753 (get_mask_policy_for_pred): Ditto.
17755 2023-05-06 Michael Collison <collison@rivosinc.com>
17757 * config/riscv/riscv-protos.h (riscv_vector_mask_mode_p): New.
17758 (riscv_vector_get_mask_mode): Ditto.
17759 (get_mask_policy_no_pred): Ditto.
17760 (get_tail_policy_no_pred): Ditto.
17762 2023-05-06 Xi Ruoyao <xry111@xry111.site>
17764 * config/loongarch/loongarch.h (struct machine_function): Add
17765 reg_is_wrapped_separately array for register wrapping
17767 * config/loongarch/loongarch.cc
17768 (loongarch_get_separate_components): New function.
17769 (loongarch_components_for_bb): Likewise.
17770 (loongarch_disqualify_components): Likewise.
17771 (loongarch_process_components): Likewise.
17772 (loongarch_emit_prologue_components): Likewise.
17773 (loongarch_emit_epilogue_components): Likewise.
17774 (loongarch_set_handled_components): Likewise.
17775 (TARGET_SHRINK_WRAP_GET_SEPARATE_COMPONENTS): Define.
17776 (TARGET_SHRINK_WRAP_COMPONENTS_FOR_BB): Likewise.
17777 (TARGET_SHRINK_WRAP_DISQUALIFY_COMPONENTS): Likewise.
17778 (TARGET_SHRINK_WRAP_EMIT_PROLOGUE_COMPONENTS): Likewise.
17779 (TARGET_SHRINK_WRAP_EMIT_EPILOGUE_COMPONENTS): Likewise.
17780 (TARGET_SHRINK_WRAP_SET_HANDLED_COMPONENTS): Likewise.
17781 (loongarch_for_each_saved_reg): Skip registers that are wrapped
17784 2023-05-06 Xi Ruoyao <xry111@xry111.site>
17787 * Makefile.in (s-macro_list): Pass -nostdinc to
17790 2023-05-06 Juzhe-Zhong <juzhe.zhong@rivai.ai>
17792 * config/riscv/riscv-protos.h (preferred_simd_mode): New function.
17793 * config/riscv/riscv-v.cc (autovec_use_vlmax_p): Ditto.
17794 (preferred_simd_mode): Ditto.
17795 * config/riscv/riscv.cc (riscv_get_arg_info): Handle RVV type in function arg.
17796 (riscv_convert_vector_bits): Adjust for RVV auto-vectorization.
17797 (riscv_preferred_simd_mode): New function.
17798 (TARGET_VECTORIZE_PREFERRED_SIMD_MODE): New target hook support.
17799 * config/riscv/vector.md: Add autovec.md.
17800 * config/riscv/autovec.md: New file.
17802 2023-05-06 Jakub Jelinek <jakub@redhat.com>
17804 * real.h (dconst_pi): Define.
17805 (dconst_e_ptr): Formatting fix.
17806 (dconst_pi_ptr): Declare.
17807 * real.cc (dconst_pi_ptr): New function.
17808 * gimple-range-op.cc (cfn_sincos::fold_range): Intersect the generic
17809 boundaries range with range computed from sin/cos of the particular
17810 bounds if the argument range is shorter than 2*pi.
17811 (cfn_sincos::op1_range): Take bulps into account when determining
17812 which result ranges are always invalid or behave like known NAN.
17814 2023-05-06 Aldy Hernandez <aldyh@redhat.com>
17816 * gimple-range-cache.cc (sbr_sparse_bitmap::set_bb_range): Do not
17817 pass type to vrange_storage::equal_p.
17818 * value-range-storage.cc (vrange_storage::equal_p): Remove type.
17819 (irange_storage::equal_p): Same.
17820 (frange_storage::equal_p): Same.
17821 * value-range-storage.h (class frange_storage): Same.
17823 2023-05-06 Juzhe-Zhong <juzhe.zhong@rivai.ai>
17826 * config/riscv/riscv-vsetvl.cc (local_eliminate_vsetvl_insn): Remove it.
17827 (pass_vsetvl::local_eliminate_vsetvl_insn): New function.
17829 2023-05-06 liuhongt <hongtao.liu@intel.com>
17831 * combine.cc (maybe_swap_commutative_operands): Canonicalize
17832 vec_merge when mask is constant.
17833 * doc/md.texi: Document vec_merge canonicalization.
17835 2023-05-06 Jakub Jelinek <jakub@redhat.com>
17837 * value-range.h (frange_arithmetic): Declare.
17838 * range-op-float.cc (frange_arithmetic): No longer static.
17839 * gimple-range-op.cc (frange_mpfr_arg1): New function.
17840 (cfn_sqrt::fold_range): Intersect the generic boundaries range
17841 with range computed from sqrt of the particular bounds.
17842 (cfn_sqrt::op1_range): Intersect the generic boundaries range
17843 with range computed from squared particular bounds.
17845 2023-05-06 Jakub Jelinek <jakub@redhat.com>
17847 * Makefile.in (check_p_numbers): Rename to one_to_9999, move
17848 earlier with helper variables also renamed.
17849 (MATCH_SPLUT_SEQ): Use $(wordlist 1,$(NUM_MATCH_SPLITS),$(one_to_9999))
17850 instead of $(shell seq 1 $(NUM_MATCH_SPLITS)).
17851 (check_p_subdirs): Use $(one_to_9999) instead of $(check_p_numbers).
17853 2023-05-06 Hans-Peter Nilsson <hp@axis.com>
17855 * config/cris/cris.md (splitop): Add PLUS.
17856 * config/cris/cris.cc (cris_split_constant): Also handle
17857 PLUS when a split into two insns may be useful.
17859 2023-05-05 Hans-Peter Nilsson <hp@axis.com>
17861 * config/cris/cris.md (movandsplit1): New define_peephole2.
17863 2023-05-05 Hans-Peter Nilsson <hp@axis.com>
17865 * config/cris/cris.md (lsrandsplit1): New define_peephole2.
17867 2023-05-05 Hans-Peter Nilsson <hp@axis.com>
17869 * doc/md.texi (define_peephole2): Document order of scanning.
17871 2023-05-05 Pan Li <pan2.li@intel.com>
17872 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
17874 * config/riscv/vector.md: Allow const as the operand of RVV
17875 indexed load/store.
17877 2023-05-05 Pan Li <pan2.li@intel.com>
17879 * config/riscv/riscv.h (VECTOR_STORE_FLAG_VALUE): Add new macro
17880 consumed by simplify_rtx.
17882 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
17884 * config/arm/arm-mve-builtins-base.cc (vrshrq, vshrq): New.
17885 * config/arm/arm-mve-builtins-base.def (vrshrq, vshrq): New.
17886 * config/arm/arm-mve-builtins-base.h (vrshrq, vshrq): New.
17887 * config/arm/arm_mve.h (vshrq): Remove.
17889 (vrshrq_m): Remove.
17891 (vrshrq_x): Remove.
17893 (vshrq_n_s8): Remove.
17894 (vshrq_n_s16): Remove.
17895 (vshrq_n_s32): Remove.
17896 (vshrq_n_u8): Remove.
17897 (vshrq_n_u16): Remove.
17898 (vshrq_n_u32): Remove.
17899 (vrshrq_n_u8): Remove.
17900 (vrshrq_n_s8): Remove.
17901 (vrshrq_n_u16): Remove.
17902 (vrshrq_n_s16): Remove.
17903 (vrshrq_n_u32): Remove.
17904 (vrshrq_n_s32): Remove.
17905 (vrshrq_m_n_s8): Remove.
17906 (vrshrq_m_n_s32): Remove.
17907 (vrshrq_m_n_s16): Remove.
17908 (vrshrq_m_n_u8): Remove.
17909 (vrshrq_m_n_u32): Remove.
17910 (vrshrq_m_n_u16): Remove.
17911 (vshrq_m_n_s8): Remove.
17912 (vshrq_m_n_s32): Remove.
17913 (vshrq_m_n_s16): Remove.
17914 (vshrq_m_n_u8): Remove.
17915 (vshrq_m_n_u32): Remove.
17916 (vshrq_m_n_u16): Remove.
17917 (vrshrq_x_n_s8): Remove.
17918 (vrshrq_x_n_s16): Remove.
17919 (vrshrq_x_n_s32): Remove.
17920 (vrshrq_x_n_u8): Remove.
17921 (vrshrq_x_n_u16): Remove.
17922 (vrshrq_x_n_u32): Remove.
17923 (vshrq_x_n_s8): Remove.
17924 (vshrq_x_n_s16): Remove.
17925 (vshrq_x_n_s32): Remove.
17926 (vshrq_x_n_u8): Remove.
17927 (vshrq_x_n_u16): Remove.
17928 (vshrq_x_n_u32): Remove.
17929 (__arm_vshrq_n_s8): Remove.
17930 (__arm_vshrq_n_s16): Remove.
17931 (__arm_vshrq_n_s32): Remove.
17932 (__arm_vshrq_n_u8): Remove.
17933 (__arm_vshrq_n_u16): Remove.
17934 (__arm_vshrq_n_u32): Remove.
17935 (__arm_vrshrq_n_u8): Remove.
17936 (__arm_vrshrq_n_s8): Remove.
17937 (__arm_vrshrq_n_u16): Remove.
17938 (__arm_vrshrq_n_s16): Remove.
17939 (__arm_vrshrq_n_u32): Remove.
17940 (__arm_vrshrq_n_s32): Remove.
17941 (__arm_vrshrq_m_n_s8): Remove.
17942 (__arm_vrshrq_m_n_s32): Remove.
17943 (__arm_vrshrq_m_n_s16): Remove.
17944 (__arm_vrshrq_m_n_u8): Remove.
17945 (__arm_vrshrq_m_n_u32): Remove.
17946 (__arm_vrshrq_m_n_u16): Remove.
17947 (__arm_vshrq_m_n_s8): Remove.
17948 (__arm_vshrq_m_n_s32): Remove.
17949 (__arm_vshrq_m_n_s16): Remove.
17950 (__arm_vshrq_m_n_u8): Remove.
17951 (__arm_vshrq_m_n_u32): Remove.
17952 (__arm_vshrq_m_n_u16): Remove.
17953 (__arm_vrshrq_x_n_s8): Remove.
17954 (__arm_vrshrq_x_n_s16): Remove.
17955 (__arm_vrshrq_x_n_s32): Remove.
17956 (__arm_vrshrq_x_n_u8): Remove.
17957 (__arm_vrshrq_x_n_u16): Remove.
17958 (__arm_vrshrq_x_n_u32): Remove.
17959 (__arm_vshrq_x_n_s8): Remove.
17960 (__arm_vshrq_x_n_s16): Remove.
17961 (__arm_vshrq_x_n_s32): Remove.
17962 (__arm_vshrq_x_n_u8): Remove.
17963 (__arm_vshrq_x_n_u16): Remove.
17964 (__arm_vshrq_x_n_u32): Remove.
17965 (__arm_vshrq): Remove.
17966 (__arm_vrshrq): Remove.
17967 (__arm_vrshrq_m): Remove.
17968 (__arm_vshrq_m): Remove.
17969 (__arm_vrshrq_x): Remove.
17970 (__arm_vshrq_x): Remove.
17972 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
17974 * config/arm/iterators.md (MVE_VSHRQ_M_N, MVE_VSHRQ_N): New.
17975 (mve_insn): Add vrshr, vshr.
17976 * config/arm/mve.md (mve_vshrq_n_<supf><mode>)
17977 (mve_vrshrq_n_<supf><mode>): Merge into ...
17978 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
17979 (mve_vrshrq_m_n_<supf><mode>, mve_vshrq_m_n_<supf><mode>): Merge
17981 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
17983 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
17985 * config/arm/arm-mve-builtins-shapes.cc (binary_rshift): New.
17986 * config/arm/arm-mve-builtins-shapes.h (binary_rshift): New.
17988 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
17990 * config/arm/arm-mve-builtins-base.cc (FUNCTION_ONLY_N_NO_U_F): New.
17991 (vqshrunbq, vqshruntq, vqrshrunbq, vqrshruntq): New.
17992 * config/arm/arm-mve-builtins-base.def (vqshrunbq, vqshruntq)
17993 (vqrshrunbq, vqrshruntq): New.
17994 * config/arm/arm-mve-builtins-base.h (vqshrunbq, vqshruntq)
17995 (vqrshrunbq, vqrshruntq): New.
17996 * config/arm/arm-mve-builtins.cc
17997 (function_instance::has_inactive_argument): Handle vqshrunbq,
17998 vqshruntq, vqrshrunbq, vqrshruntq.
17999 * config/arm/arm_mve.h (vqrshrunbq): Remove.
18000 (vqrshruntq): Remove.
18001 (vqrshrunbq_m): Remove.
18002 (vqrshruntq_m): Remove.
18003 (vqrshrunbq_n_s16): Remove.
18004 (vqrshrunbq_n_s32): Remove.
18005 (vqrshruntq_n_s16): Remove.
18006 (vqrshruntq_n_s32): Remove.
18007 (vqrshrunbq_m_n_s32): Remove.
18008 (vqrshrunbq_m_n_s16): Remove.
18009 (vqrshruntq_m_n_s32): Remove.
18010 (vqrshruntq_m_n_s16): Remove.
18011 (__arm_vqrshrunbq_n_s16): Remove.
18012 (__arm_vqrshrunbq_n_s32): Remove.
18013 (__arm_vqrshruntq_n_s16): Remove.
18014 (__arm_vqrshruntq_n_s32): Remove.
18015 (__arm_vqrshrunbq_m_n_s32): Remove.
18016 (__arm_vqrshrunbq_m_n_s16): Remove.
18017 (__arm_vqrshruntq_m_n_s32): Remove.
18018 (__arm_vqrshruntq_m_n_s16): Remove.
18019 (__arm_vqrshrunbq): Remove.
18020 (__arm_vqrshruntq): Remove.
18021 (__arm_vqrshrunbq_m): Remove.
18022 (__arm_vqrshruntq_m): Remove.
18023 (vqshrunbq): Remove.
18024 (vqshruntq): Remove.
18025 (vqshrunbq_m): Remove.
18026 (vqshruntq_m): Remove.
18027 (vqshrunbq_n_s16): Remove.
18028 (vqshruntq_n_s16): Remove.
18029 (vqshrunbq_n_s32): Remove.
18030 (vqshruntq_n_s32): Remove.
18031 (vqshrunbq_m_n_s32): Remove.
18032 (vqshrunbq_m_n_s16): Remove.
18033 (vqshruntq_m_n_s32): Remove.
18034 (vqshruntq_m_n_s16): Remove.
18035 (__arm_vqshrunbq_n_s16): Remove.
18036 (__arm_vqshruntq_n_s16): Remove.
18037 (__arm_vqshrunbq_n_s32): Remove.
18038 (__arm_vqshruntq_n_s32): Remove.
18039 (__arm_vqshrunbq_m_n_s32): Remove.
18040 (__arm_vqshrunbq_m_n_s16): Remove.
18041 (__arm_vqshruntq_m_n_s32): Remove.
18042 (__arm_vqshruntq_m_n_s16): Remove.
18043 (__arm_vqshrunbq): Remove.
18044 (__arm_vqshruntq): Remove.
18045 (__arm_vqshrunbq_m): Remove.
18046 (__arm_vqshruntq_m): Remove.
18048 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18050 * config/arm/iterators.md (MVE_SHRN_N): Add VQRSHRUNBQ,
18051 VQRSHRUNTQ, VQSHRUNBQ, VQSHRUNTQ.
18052 (MVE_SHRN_M_N): Likewise.
18053 (mve_insn): Add vqrshrunb, vqrshrunt, vqshrunb, vqshrunt.
18054 (isu): Add VQRSHRUNBQ, VQRSHRUNTQ, VQSHRUNBQ, VQSHRUNTQ.
18056 * config/arm/mve.md (mve_vqrshrunbq_n_s<mode>): Remove.
18057 (mve_vqrshruntq_n_s<mode>): Remove.
18058 (mve_vqshrunbq_n_s<mode>): Remove.
18059 (mve_vqshruntq_n_s<mode>): Remove.
18060 (mve_vqrshrunbq_m_n_s<mode>): Remove.
18061 (mve_vqrshruntq_m_n_s<mode>): Remove.
18062 (mve_vqshrunbq_m_n_s<mode>): Remove.
18063 (mve_vqshruntq_m_n_s<mode>): Remove.
18065 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18067 * config/arm/arm-mve-builtins-shapes.cc
18068 (binary_rshift_narrow_unsigned): New.
18069 * config/arm/arm-mve-builtins-shapes.h
18070 (binary_rshift_narrow_unsigned): New.
18072 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18074 * config/arm/arm-mve-builtins-base.cc (FUNCTION_ONLY_N_NO_F): New.
18075 (vshrnbq, vshrntq, vrshrnbq, vrshrntq, vqshrnbq, vqshrntq)
18076 (vqrshrnbq, vqrshrntq): New.
18077 * config/arm/arm-mve-builtins-base.def (vshrnbq, vshrntq)
18078 (vrshrnbq, vrshrntq, vqshrnbq, vqshrntq, vqrshrnbq, vqrshrntq):
18080 * config/arm/arm-mve-builtins-base.h (vshrnbq, vshrntq, vrshrnbq)
18081 (vrshrntq, vqshrnbq, vqshrntq, vqrshrnbq, vqrshrntq): New.
18082 * config/arm/arm-mve-builtins.cc
18083 (function_instance::has_inactive_argument): Handle vshrnbq,
18084 vshrntq, vrshrnbq, vrshrntq, vqshrnbq, vqshrntq, vqrshrnbq,
18086 * config/arm/arm_mve.h (vshrnbq): Remove.
18088 (vshrnbq_m): Remove.
18089 (vshrntq_m): Remove.
18090 (vshrnbq_n_s16): Remove.
18091 (vshrntq_n_s16): Remove.
18092 (vshrnbq_n_u16): Remove.
18093 (vshrntq_n_u16): Remove.
18094 (vshrnbq_n_s32): Remove.
18095 (vshrntq_n_s32): Remove.
18096 (vshrnbq_n_u32): Remove.
18097 (vshrntq_n_u32): Remove.
18098 (vshrnbq_m_n_s32): Remove.
18099 (vshrnbq_m_n_s16): Remove.
18100 (vshrnbq_m_n_u32): Remove.
18101 (vshrnbq_m_n_u16): Remove.
18102 (vshrntq_m_n_s32): Remove.
18103 (vshrntq_m_n_s16): Remove.
18104 (vshrntq_m_n_u32): Remove.
18105 (vshrntq_m_n_u16): Remove.
18106 (__arm_vshrnbq_n_s16): Remove.
18107 (__arm_vshrntq_n_s16): Remove.
18108 (__arm_vshrnbq_n_u16): Remove.
18109 (__arm_vshrntq_n_u16): Remove.
18110 (__arm_vshrnbq_n_s32): Remove.
18111 (__arm_vshrntq_n_s32): Remove.
18112 (__arm_vshrnbq_n_u32): Remove.
18113 (__arm_vshrntq_n_u32): Remove.
18114 (__arm_vshrnbq_m_n_s32): Remove.
18115 (__arm_vshrnbq_m_n_s16): Remove.
18116 (__arm_vshrnbq_m_n_u32): Remove.
18117 (__arm_vshrnbq_m_n_u16): Remove.
18118 (__arm_vshrntq_m_n_s32): Remove.
18119 (__arm_vshrntq_m_n_s16): Remove.
18120 (__arm_vshrntq_m_n_u32): Remove.
18121 (__arm_vshrntq_m_n_u16): Remove.
18122 (__arm_vshrnbq): Remove.
18123 (__arm_vshrntq): Remove.
18124 (__arm_vshrnbq_m): Remove.
18125 (__arm_vshrntq_m): Remove.
18126 (vrshrnbq): Remove.
18127 (vrshrntq): Remove.
18128 (vrshrnbq_m): Remove.
18129 (vrshrntq_m): Remove.
18130 (vrshrnbq_n_s16): Remove.
18131 (vrshrntq_n_s16): Remove.
18132 (vrshrnbq_n_u16): Remove.
18133 (vrshrntq_n_u16): Remove.
18134 (vrshrnbq_n_s32): Remove.
18135 (vrshrntq_n_s32): Remove.
18136 (vrshrnbq_n_u32): Remove.
18137 (vrshrntq_n_u32): Remove.
18138 (vrshrnbq_m_n_s32): Remove.
18139 (vrshrnbq_m_n_s16): Remove.
18140 (vrshrnbq_m_n_u32): Remove.
18141 (vrshrnbq_m_n_u16): Remove.
18142 (vrshrntq_m_n_s32): Remove.
18143 (vrshrntq_m_n_s16): Remove.
18144 (vrshrntq_m_n_u32): Remove.
18145 (vrshrntq_m_n_u16): Remove.
18146 (__arm_vrshrnbq_n_s16): Remove.
18147 (__arm_vrshrntq_n_s16): Remove.
18148 (__arm_vrshrnbq_n_u16): Remove.
18149 (__arm_vrshrntq_n_u16): Remove.
18150 (__arm_vrshrnbq_n_s32): Remove.
18151 (__arm_vrshrntq_n_s32): Remove.
18152 (__arm_vrshrnbq_n_u32): Remove.
18153 (__arm_vrshrntq_n_u32): Remove.
18154 (__arm_vrshrnbq_m_n_s32): Remove.
18155 (__arm_vrshrnbq_m_n_s16): Remove.
18156 (__arm_vrshrnbq_m_n_u32): Remove.
18157 (__arm_vrshrnbq_m_n_u16): Remove.
18158 (__arm_vrshrntq_m_n_s32): Remove.
18159 (__arm_vrshrntq_m_n_s16): Remove.
18160 (__arm_vrshrntq_m_n_u32): Remove.
18161 (__arm_vrshrntq_m_n_u16): Remove.
18162 (__arm_vrshrnbq): Remove.
18163 (__arm_vrshrntq): Remove.
18164 (__arm_vrshrnbq_m): Remove.
18165 (__arm_vrshrntq_m): Remove.
18166 (vqshrnbq): Remove.
18167 (vqshrntq): Remove.
18168 (vqshrnbq_m): Remove.
18169 (vqshrntq_m): Remove.
18170 (vqshrnbq_n_s16): Remove.
18171 (vqshrntq_n_s16): Remove.
18172 (vqshrnbq_n_u16): Remove.
18173 (vqshrntq_n_u16): Remove.
18174 (vqshrnbq_n_s32): Remove.
18175 (vqshrntq_n_s32): Remove.
18176 (vqshrnbq_n_u32): Remove.
18177 (vqshrntq_n_u32): Remove.
18178 (vqshrnbq_m_n_s32): Remove.
18179 (vqshrnbq_m_n_s16): Remove.
18180 (vqshrnbq_m_n_u32): Remove.
18181 (vqshrnbq_m_n_u16): Remove.
18182 (vqshrntq_m_n_s32): Remove.
18183 (vqshrntq_m_n_s16): Remove.
18184 (vqshrntq_m_n_u32): Remove.
18185 (vqshrntq_m_n_u16): Remove.
18186 (__arm_vqshrnbq_n_s16): Remove.
18187 (__arm_vqshrntq_n_s16): Remove.
18188 (__arm_vqshrnbq_n_u16): Remove.
18189 (__arm_vqshrntq_n_u16): Remove.
18190 (__arm_vqshrnbq_n_s32): Remove.
18191 (__arm_vqshrntq_n_s32): Remove.
18192 (__arm_vqshrnbq_n_u32): Remove.
18193 (__arm_vqshrntq_n_u32): Remove.
18194 (__arm_vqshrnbq_m_n_s32): Remove.
18195 (__arm_vqshrnbq_m_n_s16): Remove.
18196 (__arm_vqshrnbq_m_n_u32): Remove.
18197 (__arm_vqshrnbq_m_n_u16): Remove.
18198 (__arm_vqshrntq_m_n_s32): Remove.
18199 (__arm_vqshrntq_m_n_s16): Remove.
18200 (__arm_vqshrntq_m_n_u32): Remove.
18201 (__arm_vqshrntq_m_n_u16): Remove.
18202 (__arm_vqshrnbq): Remove.
18203 (__arm_vqshrntq): Remove.
18204 (__arm_vqshrnbq_m): Remove.
18205 (__arm_vqshrntq_m): Remove.
18206 (vqrshrnbq): Remove.
18207 (vqrshrntq): Remove.
18208 (vqrshrnbq_m): Remove.
18209 (vqrshrntq_m): Remove.
18210 (vqrshrnbq_n_s16): Remove.
18211 (vqrshrnbq_n_u16): Remove.
18212 (vqrshrnbq_n_s32): Remove.
18213 (vqrshrnbq_n_u32): Remove.
18214 (vqrshrntq_n_s16): Remove.
18215 (vqrshrntq_n_u16): Remove.
18216 (vqrshrntq_n_s32): Remove.
18217 (vqrshrntq_n_u32): Remove.
18218 (vqrshrnbq_m_n_s32): Remove.
18219 (vqrshrnbq_m_n_s16): Remove.
18220 (vqrshrnbq_m_n_u32): Remove.
18221 (vqrshrnbq_m_n_u16): Remove.
18222 (vqrshrntq_m_n_s32): Remove.
18223 (vqrshrntq_m_n_s16): Remove.
18224 (vqrshrntq_m_n_u32): Remove.
18225 (vqrshrntq_m_n_u16): Remove.
18226 (__arm_vqrshrnbq_n_s16): Remove.
18227 (__arm_vqrshrnbq_n_u16): Remove.
18228 (__arm_vqrshrnbq_n_s32): Remove.
18229 (__arm_vqrshrnbq_n_u32): Remove.
18230 (__arm_vqrshrntq_n_s16): Remove.
18231 (__arm_vqrshrntq_n_u16): Remove.
18232 (__arm_vqrshrntq_n_s32): Remove.
18233 (__arm_vqrshrntq_n_u32): Remove.
18234 (__arm_vqrshrnbq_m_n_s32): Remove.
18235 (__arm_vqrshrnbq_m_n_s16): Remove.
18236 (__arm_vqrshrnbq_m_n_u32): Remove.
18237 (__arm_vqrshrnbq_m_n_u16): Remove.
18238 (__arm_vqrshrntq_m_n_s32): Remove.
18239 (__arm_vqrshrntq_m_n_s16): Remove.
18240 (__arm_vqrshrntq_m_n_u32): Remove.
18241 (__arm_vqrshrntq_m_n_u16): Remove.
18242 (__arm_vqrshrnbq): Remove.
18243 (__arm_vqrshrntq): Remove.
18244 (__arm_vqrshrnbq_m): Remove.
18245 (__arm_vqrshrntq_m): Remove.
18247 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18249 * config/arm/iterators.md (MVE_SHRN_N, MVE_SHRN_M_N): New.
18250 (mve_insn): Add vqrshrnb, vqrshrnt, vqshrnb, vqshrnt, vrshrnb,
18251 vrshrnt, vshrnb, vshrnt.
18253 * config/arm/mve.md (mve_vqrshrnbq_n_<supf><mode>)
18254 (mve_vqrshrntq_n_<supf><mode>, mve_vqshrnbq_n_<supf><mode>)
18255 (mve_vqshrntq_n_<supf><mode>, mve_vrshrnbq_n_<supf><mode>)
18256 (mve_vrshrntq_n_<supf><mode>, mve_vshrnbq_n_<supf><mode>)
18257 (mve_vshrntq_n_<supf><mode>): Merge into ...
18258 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
18259 (mve_vqrshrnbq_m_n_<supf><mode>, mve_vqrshrntq_m_n_<supf><mode>)
18260 (mve_vqshrnbq_m_n_<supf><mode>, mve_vqshrntq_m_n_<supf><mode>)
18261 (mve_vrshrnbq_m_n_<supf><mode>, mve_vrshrntq_m_n_<supf><mode>)
18262 (mve_vshrnbq_m_n_<supf><mode>, mve_vshrntq_m_n_<supf><mode>):
18264 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
18266 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18268 * config/arm/arm-mve-builtins-shapes.cc (binary_rshift_narrow):
18270 * config/arm/arm-mve-builtins-shapes.h (binary_rshift_narrow): New.
18272 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18274 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITH_RTX_M_NO_F): New.
18275 (vmaxq, vminq): New.
18276 * config/arm/arm-mve-builtins-base.def (vmaxq, vminq): New.
18277 * config/arm/arm-mve-builtins-base.h (vmaxq, vminq): New.
18278 * config/arm/arm_mve.h (vminq): Remove.
18284 (vminq_u8): Remove.
18285 (vmaxq_u8): Remove.
18286 (vminq_s8): Remove.
18287 (vmaxq_s8): Remove.
18288 (vminq_u16): Remove.
18289 (vmaxq_u16): Remove.
18290 (vminq_s16): Remove.
18291 (vmaxq_s16): Remove.
18292 (vminq_u32): Remove.
18293 (vmaxq_u32): Remove.
18294 (vminq_s32): Remove.
18295 (vmaxq_s32): Remove.
18296 (vmaxq_m_s8): Remove.
18297 (vmaxq_m_s32): Remove.
18298 (vmaxq_m_s16): Remove.
18299 (vmaxq_m_u8): Remove.
18300 (vmaxq_m_u32): Remove.
18301 (vmaxq_m_u16): Remove.
18302 (vminq_m_s8): Remove.
18303 (vminq_m_s32): Remove.
18304 (vminq_m_s16): Remove.
18305 (vminq_m_u8): Remove.
18306 (vminq_m_u32): Remove.
18307 (vminq_m_u16): Remove.
18308 (vminq_x_s8): Remove.
18309 (vminq_x_s16): Remove.
18310 (vminq_x_s32): Remove.
18311 (vminq_x_u8): Remove.
18312 (vminq_x_u16): Remove.
18313 (vminq_x_u32): Remove.
18314 (vmaxq_x_s8): Remove.
18315 (vmaxq_x_s16): Remove.
18316 (vmaxq_x_s32): Remove.
18317 (vmaxq_x_u8): Remove.
18318 (vmaxq_x_u16): Remove.
18319 (vmaxq_x_u32): Remove.
18320 (__arm_vminq_u8): Remove.
18321 (__arm_vmaxq_u8): Remove.
18322 (__arm_vminq_s8): Remove.
18323 (__arm_vmaxq_s8): Remove.
18324 (__arm_vminq_u16): Remove.
18325 (__arm_vmaxq_u16): Remove.
18326 (__arm_vminq_s16): Remove.
18327 (__arm_vmaxq_s16): Remove.
18328 (__arm_vminq_u32): Remove.
18329 (__arm_vmaxq_u32): Remove.
18330 (__arm_vminq_s32): Remove.
18331 (__arm_vmaxq_s32): Remove.
18332 (__arm_vmaxq_m_s8): Remove.
18333 (__arm_vmaxq_m_s32): Remove.
18334 (__arm_vmaxq_m_s16): Remove.
18335 (__arm_vmaxq_m_u8): Remove.
18336 (__arm_vmaxq_m_u32): Remove.
18337 (__arm_vmaxq_m_u16): Remove.
18338 (__arm_vminq_m_s8): Remove.
18339 (__arm_vminq_m_s32): Remove.
18340 (__arm_vminq_m_s16): Remove.
18341 (__arm_vminq_m_u8): Remove.
18342 (__arm_vminq_m_u32): Remove.
18343 (__arm_vminq_m_u16): Remove.
18344 (__arm_vminq_x_s8): Remove.
18345 (__arm_vminq_x_s16): Remove.
18346 (__arm_vminq_x_s32): Remove.
18347 (__arm_vminq_x_u8): Remove.
18348 (__arm_vminq_x_u16): Remove.
18349 (__arm_vminq_x_u32): Remove.
18350 (__arm_vmaxq_x_s8): Remove.
18351 (__arm_vmaxq_x_s16): Remove.
18352 (__arm_vmaxq_x_s32): Remove.
18353 (__arm_vmaxq_x_u8): Remove.
18354 (__arm_vmaxq_x_u16): Remove.
18355 (__arm_vmaxq_x_u32): Remove.
18356 (__arm_vminq): Remove.
18357 (__arm_vmaxq): Remove.
18358 (__arm_vmaxq_m): Remove.
18359 (__arm_vminq_m): Remove.
18360 (__arm_vminq_x): Remove.
18361 (__arm_vmaxq_x): Remove.
18363 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18365 * config/arm/iterators.md (MAX_MIN_SU): New.
18366 (max_min_su_str): New.
18367 (max_min_supf): New.
18368 * config/arm/mve.md (mve_vmaxq_s<mode>, mve_vmaxq_u<mode>)
18369 (mve_vminq_s<mode>, mve_vminq_u<mode>): Merge into ...
18370 (mve_<max_min_su_str>q_<max_min_supf><mode>): ... this.
18372 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18374 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITH_M_N_R): New.
18375 (vqshlq, vshlq): New.
18376 * config/arm/arm-mve-builtins-base.def (vqshlq, vshlq): New.
18377 * config/arm/arm-mve-builtins-base.h (vqshlq, vshlq): New.
18378 * config/arm/arm_mve.h (vshlq): Remove.
18381 (vshlq_m_r): Remove.
18383 (vshlq_m_n): Remove.
18385 (vshlq_x_n): Remove.
18386 (vshlq_s8): Remove.
18387 (vshlq_s16): Remove.
18388 (vshlq_s32): Remove.
18389 (vshlq_u8): Remove.
18390 (vshlq_u16): Remove.
18391 (vshlq_u32): Remove.
18392 (vshlq_r_u8): Remove.
18393 (vshlq_n_u8): Remove.
18394 (vshlq_r_s8): Remove.
18395 (vshlq_n_s8): Remove.
18396 (vshlq_r_u16): Remove.
18397 (vshlq_n_u16): Remove.
18398 (vshlq_r_s16): Remove.
18399 (vshlq_n_s16): Remove.
18400 (vshlq_r_u32): Remove.
18401 (vshlq_n_u32): Remove.
18402 (vshlq_r_s32): Remove.
18403 (vshlq_n_s32): Remove.
18404 (vshlq_m_r_u8): Remove.
18405 (vshlq_m_r_s8): Remove.
18406 (vshlq_m_r_u16): Remove.
18407 (vshlq_m_r_s16): Remove.
18408 (vshlq_m_r_u32): Remove.
18409 (vshlq_m_r_s32): Remove.
18410 (vshlq_m_u8): Remove.
18411 (vshlq_m_s8): Remove.
18412 (vshlq_m_u16): Remove.
18413 (vshlq_m_s16): Remove.
18414 (vshlq_m_u32): Remove.
18415 (vshlq_m_s32): Remove.
18416 (vshlq_m_n_s8): Remove.
18417 (vshlq_m_n_s32): Remove.
18418 (vshlq_m_n_s16): Remove.
18419 (vshlq_m_n_u8): Remove.
18420 (vshlq_m_n_u32): Remove.
18421 (vshlq_m_n_u16): Remove.
18422 (vshlq_x_s8): Remove.
18423 (vshlq_x_s16): Remove.
18424 (vshlq_x_s32): Remove.
18425 (vshlq_x_u8): Remove.
18426 (vshlq_x_u16): Remove.
18427 (vshlq_x_u32): Remove.
18428 (vshlq_x_n_s8): Remove.
18429 (vshlq_x_n_s16): Remove.
18430 (vshlq_x_n_s32): Remove.
18431 (vshlq_x_n_u8): Remove.
18432 (vshlq_x_n_u16): Remove.
18433 (vshlq_x_n_u32): Remove.
18434 (__arm_vshlq_s8): Remove.
18435 (__arm_vshlq_s16): Remove.
18436 (__arm_vshlq_s32): Remove.
18437 (__arm_vshlq_u8): Remove.
18438 (__arm_vshlq_u16): Remove.
18439 (__arm_vshlq_u32): Remove.
18440 (__arm_vshlq_r_u8): Remove.
18441 (__arm_vshlq_n_u8): Remove.
18442 (__arm_vshlq_r_s8): Remove.
18443 (__arm_vshlq_n_s8): Remove.
18444 (__arm_vshlq_r_u16): Remove.
18445 (__arm_vshlq_n_u16): Remove.
18446 (__arm_vshlq_r_s16): Remove.
18447 (__arm_vshlq_n_s16): Remove.
18448 (__arm_vshlq_r_u32): Remove.
18449 (__arm_vshlq_n_u32): Remove.
18450 (__arm_vshlq_r_s32): Remove.
18451 (__arm_vshlq_n_s32): Remove.
18452 (__arm_vshlq_m_r_u8): Remove.
18453 (__arm_vshlq_m_r_s8): Remove.
18454 (__arm_vshlq_m_r_u16): Remove.
18455 (__arm_vshlq_m_r_s16): Remove.
18456 (__arm_vshlq_m_r_u32): Remove.
18457 (__arm_vshlq_m_r_s32): Remove.
18458 (__arm_vshlq_m_u8): Remove.
18459 (__arm_vshlq_m_s8): Remove.
18460 (__arm_vshlq_m_u16): Remove.
18461 (__arm_vshlq_m_s16): Remove.
18462 (__arm_vshlq_m_u32): Remove.
18463 (__arm_vshlq_m_s32): Remove.
18464 (__arm_vshlq_m_n_s8): Remove.
18465 (__arm_vshlq_m_n_s32): Remove.
18466 (__arm_vshlq_m_n_s16): Remove.
18467 (__arm_vshlq_m_n_u8): Remove.
18468 (__arm_vshlq_m_n_u32): Remove.
18469 (__arm_vshlq_m_n_u16): Remove.
18470 (__arm_vshlq_x_s8): Remove.
18471 (__arm_vshlq_x_s16): Remove.
18472 (__arm_vshlq_x_s32): Remove.
18473 (__arm_vshlq_x_u8): Remove.
18474 (__arm_vshlq_x_u16): Remove.
18475 (__arm_vshlq_x_u32): Remove.
18476 (__arm_vshlq_x_n_s8): Remove.
18477 (__arm_vshlq_x_n_s16): Remove.
18478 (__arm_vshlq_x_n_s32): Remove.
18479 (__arm_vshlq_x_n_u8): Remove.
18480 (__arm_vshlq_x_n_u16): Remove.
18481 (__arm_vshlq_x_n_u32): Remove.
18482 (__arm_vshlq): Remove.
18483 (__arm_vshlq_r): Remove.
18484 (__arm_vshlq_n): Remove.
18485 (__arm_vshlq_m_r): Remove.
18486 (__arm_vshlq_m): Remove.
18487 (__arm_vshlq_m_n): Remove.
18488 (__arm_vshlq_x): Remove.
18489 (__arm_vshlq_x_n): Remove.
18491 (vqshlq_r): Remove.
18492 (vqshlq_n): Remove.
18493 (vqshlq_m_r): Remove.
18494 (vqshlq_m_n): Remove.
18495 (vqshlq_m): Remove.
18496 (vqshlq_u8): Remove.
18497 (vqshlq_r_u8): Remove.
18498 (vqshlq_n_u8): Remove.
18499 (vqshlq_s8): Remove.
18500 (vqshlq_r_s8): Remove.
18501 (vqshlq_n_s8): Remove.
18502 (vqshlq_u16): Remove.
18503 (vqshlq_r_u16): Remove.
18504 (vqshlq_n_u16): Remove.
18505 (vqshlq_s16): Remove.
18506 (vqshlq_r_s16): Remove.
18507 (vqshlq_n_s16): Remove.
18508 (vqshlq_u32): Remove.
18509 (vqshlq_r_u32): Remove.
18510 (vqshlq_n_u32): Remove.
18511 (vqshlq_s32): Remove.
18512 (vqshlq_r_s32): Remove.
18513 (vqshlq_n_s32): Remove.
18514 (vqshlq_m_r_u8): Remove.
18515 (vqshlq_m_r_s8): Remove.
18516 (vqshlq_m_r_u16): Remove.
18517 (vqshlq_m_r_s16): Remove.
18518 (vqshlq_m_r_u32): Remove.
18519 (vqshlq_m_r_s32): Remove.
18520 (vqshlq_m_n_s8): Remove.
18521 (vqshlq_m_n_s32): Remove.
18522 (vqshlq_m_n_s16): Remove.
18523 (vqshlq_m_n_u8): Remove.
18524 (vqshlq_m_n_u32): Remove.
18525 (vqshlq_m_n_u16): Remove.
18526 (vqshlq_m_s8): Remove.
18527 (vqshlq_m_s32): Remove.
18528 (vqshlq_m_s16): Remove.
18529 (vqshlq_m_u8): Remove.
18530 (vqshlq_m_u32): Remove.
18531 (vqshlq_m_u16): Remove.
18532 (__arm_vqshlq_u8): Remove.
18533 (__arm_vqshlq_r_u8): Remove.
18534 (__arm_vqshlq_n_u8): Remove.
18535 (__arm_vqshlq_s8): Remove.
18536 (__arm_vqshlq_r_s8): Remove.
18537 (__arm_vqshlq_n_s8): Remove.
18538 (__arm_vqshlq_u16): Remove.
18539 (__arm_vqshlq_r_u16): Remove.
18540 (__arm_vqshlq_n_u16): Remove.
18541 (__arm_vqshlq_s16): Remove.
18542 (__arm_vqshlq_r_s16): Remove.
18543 (__arm_vqshlq_n_s16): Remove.
18544 (__arm_vqshlq_u32): Remove.
18545 (__arm_vqshlq_r_u32): Remove.
18546 (__arm_vqshlq_n_u32): Remove.
18547 (__arm_vqshlq_s32): Remove.
18548 (__arm_vqshlq_r_s32): Remove.
18549 (__arm_vqshlq_n_s32): Remove.
18550 (__arm_vqshlq_m_r_u8): Remove.
18551 (__arm_vqshlq_m_r_s8): Remove.
18552 (__arm_vqshlq_m_r_u16): Remove.
18553 (__arm_vqshlq_m_r_s16): Remove.
18554 (__arm_vqshlq_m_r_u32): Remove.
18555 (__arm_vqshlq_m_r_s32): Remove.
18556 (__arm_vqshlq_m_n_s8): Remove.
18557 (__arm_vqshlq_m_n_s32): Remove.
18558 (__arm_vqshlq_m_n_s16): Remove.
18559 (__arm_vqshlq_m_n_u8): Remove.
18560 (__arm_vqshlq_m_n_u32): Remove.
18561 (__arm_vqshlq_m_n_u16): Remove.
18562 (__arm_vqshlq_m_s8): Remove.
18563 (__arm_vqshlq_m_s32): Remove.
18564 (__arm_vqshlq_m_s16): Remove.
18565 (__arm_vqshlq_m_u8): Remove.
18566 (__arm_vqshlq_m_u32): Remove.
18567 (__arm_vqshlq_m_u16): Remove.
18568 (__arm_vqshlq): Remove.
18569 (__arm_vqshlq_r): Remove.
18570 (__arm_vqshlq_n): Remove.
18571 (__arm_vqshlq_m_r): Remove.
18572 (__arm_vqshlq_m_n): Remove.
18573 (__arm_vqshlq_m): Remove.
18575 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18577 * config/arm/arm-mve-builtins-functions.h (class
18578 unspec_mve_function_exact_insn_vshl): New.
18580 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18582 * config/arm/arm-mve-builtins-shapes.cc (binary_lshift_r): New.
18583 * config/arm/arm-mve-builtins-shapes.h (binary_lshift_r): New.
18585 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18587 * config/arm/arm-mve-builtins.cc (has_inactive_argument)
18588 (finish_opt_n_resolution): Handle MODE_r.
18589 * config/arm/arm-mve-builtins.def (r): New mode.
18591 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18593 * config/arm/arm-mve-builtins-shapes.cc (binary_lshift): New.
18594 * config/arm/arm-mve-builtins-shapes.h (binary_lshift): New.
18596 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18598 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITHOUT_N): New.
18600 * config/arm/arm-mve-builtins-base.def (vabdq): New.
18601 * config/arm/arm-mve-builtins-base.h (vabdq): New.
18602 * config/arm/arm_mve.h (vabdq): Remove.
18605 (vabdq_u8): Remove.
18606 (vabdq_s8): Remove.
18607 (vabdq_u16): Remove.
18608 (vabdq_s16): Remove.
18609 (vabdq_u32): Remove.
18610 (vabdq_s32): Remove.
18611 (vabdq_f16): Remove.
18612 (vabdq_f32): Remove.
18613 (vabdq_m_s8): Remove.
18614 (vabdq_m_s32): Remove.
18615 (vabdq_m_s16): Remove.
18616 (vabdq_m_u8): Remove.
18617 (vabdq_m_u32): Remove.
18618 (vabdq_m_u16): Remove.
18619 (vabdq_m_f32): Remove.
18620 (vabdq_m_f16): Remove.
18621 (vabdq_x_s8): Remove.
18622 (vabdq_x_s16): Remove.
18623 (vabdq_x_s32): Remove.
18624 (vabdq_x_u8): Remove.
18625 (vabdq_x_u16): Remove.
18626 (vabdq_x_u32): Remove.
18627 (vabdq_x_f16): Remove.
18628 (vabdq_x_f32): Remove.
18629 (__arm_vabdq_u8): Remove.
18630 (__arm_vabdq_s8): Remove.
18631 (__arm_vabdq_u16): Remove.
18632 (__arm_vabdq_s16): Remove.
18633 (__arm_vabdq_u32): Remove.
18634 (__arm_vabdq_s32): Remove.
18635 (__arm_vabdq_m_s8): Remove.
18636 (__arm_vabdq_m_s32): Remove.
18637 (__arm_vabdq_m_s16): Remove.
18638 (__arm_vabdq_m_u8): Remove.
18639 (__arm_vabdq_m_u32): Remove.
18640 (__arm_vabdq_m_u16): Remove.
18641 (__arm_vabdq_x_s8): Remove.
18642 (__arm_vabdq_x_s16): Remove.
18643 (__arm_vabdq_x_s32): Remove.
18644 (__arm_vabdq_x_u8): Remove.
18645 (__arm_vabdq_x_u16): Remove.
18646 (__arm_vabdq_x_u32): Remove.
18647 (__arm_vabdq_f16): Remove.
18648 (__arm_vabdq_f32): Remove.
18649 (__arm_vabdq_m_f32): Remove.
18650 (__arm_vabdq_m_f16): Remove.
18651 (__arm_vabdq_x_f16): Remove.
18652 (__arm_vabdq_x_f32): Remove.
18653 (__arm_vabdq): Remove.
18654 (__arm_vabdq_m): Remove.
18655 (__arm_vabdq_x): Remove.
18657 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18659 * config/arm/iterators.md (MVE_FP_M_BINARY): Add vabdq.
18660 (MVE_FP_VABDQ_ONLY): New.
18661 (mve_insn): Add vabd.
18662 * config/arm/mve.md (mve_vabdq_f<mode>): Move into ...
18663 (@mve_<mve_insn>q_f<mode>): ... this.
18664 (mve_vabdq_m_f<mode>): Remove.
18666 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18668 * config/arm/arm-mve-builtins-base.cc (vqrdmulhq): New.
18669 * config/arm/arm-mve-builtins-base.def (vqrdmulhq): New.
18670 * config/arm/arm-mve-builtins-base.h (vqrdmulhq): New.
18671 * config/arm/arm_mve.h (vqrdmulhq): Remove.
18672 (vqrdmulhq_m): Remove.
18673 (vqrdmulhq_s8): Remove.
18674 (vqrdmulhq_n_s8): Remove.
18675 (vqrdmulhq_s16): Remove.
18676 (vqrdmulhq_n_s16): Remove.
18677 (vqrdmulhq_s32): Remove.
18678 (vqrdmulhq_n_s32): Remove.
18679 (vqrdmulhq_m_n_s8): Remove.
18680 (vqrdmulhq_m_n_s32): Remove.
18681 (vqrdmulhq_m_n_s16): Remove.
18682 (vqrdmulhq_m_s8): Remove.
18683 (vqrdmulhq_m_s32): Remove.
18684 (vqrdmulhq_m_s16): Remove.
18685 (__arm_vqrdmulhq_s8): Remove.
18686 (__arm_vqrdmulhq_n_s8): Remove.
18687 (__arm_vqrdmulhq_s16): Remove.
18688 (__arm_vqrdmulhq_n_s16): Remove.
18689 (__arm_vqrdmulhq_s32): Remove.
18690 (__arm_vqrdmulhq_n_s32): Remove.
18691 (__arm_vqrdmulhq_m_n_s8): Remove.
18692 (__arm_vqrdmulhq_m_n_s32): Remove.
18693 (__arm_vqrdmulhq_m_n_s16): Remove.
18694 (__arm_vqrdmulhq_m_s8): Remove.
18695 (__arm_vqrdmulhq_m_s32): Remove.
18696 (__arm_vqrdmulhq_m_s16): Remove.
18697 (__arm_vqrdmulhq): Remove.
18698 (__arm_vqrdmulhq_m): Remove.
18700 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18702 * config/arm/iterators.md (MVE_SHIFT_M_R, MVE_SHIFT_M_N)
18703 (MVE_SHIFT_N, MVE_SHIFT_R): New.
18704 (mve_insn): Add vqshl, vshl.
18705 * config/arm/mve.md (mve_vqshlq_n_<supf><mode>)
18706 (mve_vshlq_n_<supf><mode>): Merge into ...
18707 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
18708 (mve_vqshlq_r_<supf><mode>, mve_vshlq_r_<supf><mode>): Merge into
18710 (@mve_<mve_insn>q_r_<supf><mode>): ... this.
18711 (mve_vqshlq_m_r_<supf><mode>, mve_vshlq_m_r_<supf><mode>): Merge
18713 (@mve_<mve_insn>q_m_r_<supf><mode>): ... this.
18714 (mve_vqshlq_m_n_<supf><mode>, mve_vshlq_m_n_<supf><mode>): Merge
18716 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
18717 * config/arm/vec-common.md (mve_vshlq_<supf><mode>): Transform
18719 (@mve_<mve_insn>q_<supf><mode>): ... this.
18721 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18723 * config/arm/arm-mve-builtins-base.cc (vqrshlq, vrshlq): New.
18724 * config/arm/arm-mve-builtins-base.def (vqrshlq, vrshlq): New.
18725 * config/arm/arm-mve-builtins-base.h (vqrshlq, vrshlq): New.
18726 * config/arm/arm-mve-builtins.cc (has_inactive_argument): Handle
18728 * config/arm/arm_mve.h (vrshlq): Remove.
18729 (vrshlq_m_n): Remove.
18730 (vrshlq_m): Remove.
18731 (vrshlq_x): Remove.
18732 (vrshlq_u8): Remove.
18733 (vrshlq_n_u8): Remove.
18734 (vrshlq_s8): Remove.
18735 (vrshlq_n_s8): Remove.
18736 (vrshlq_u16): Remove.
18737 (vrshlq_n_u16): Remove.
18738 (vrshlq_s16): Remove.
18739 (vrshlq_n_s16): Remove.
18740 (vrshlq_u32): Remove.
18741 (vrshlq_n_u32): Remove.
18742 (vrshlq_s32): Remove.
18743 (vrshlq_n_s32): Remove.
18744 (vrshlq_m_n_u8): Remove.
18745 (vrshlq_m_n_s8): Remove.
18746 (vrshlq_m_n_u16): Remove.
18747 (vrshlq_m_n_s16): Remove.
18748 (vrshlq_m_n_u32): Remove.
18749 (vrshlq_m_n_s32): Remove.
18750 (vrshlq_m_s8): Remove.
18751 (vrshlq_m_s32): Remove.
18752 (vrshlq_m_s16): Remove.
18753 (vrshlq_m_u8): Remove.
18754 (vrshlq_m_u32): Remove.
18755 (vrshlq_m_u16): Remove.
18756 (vrshlq_x_s8): Remove.
18757 (vrshlq_x_s16): Remove.
18758 (vrshlq_x_s32): Remove.
18759 (vrshlq_x_u8): Remove.
18760 (vrshlq_x_u16): Remove.
18761 (vrshlq_x_u32): Remove.
18762 (__arm_vrshlq_u8): Remove.
18763 (__arm_vrshlq_n_u8): Remove.
18764 (__arm_vrshlq_s8): Remove.
18765 (__arm_vrshlq_n_s8): Remove.
18766 (__arm_vrshlq_u16): Remove.
18767 (__arm_vrshlq_n_u16): Remove.
18768 (__arm_vrshlq_s16): Remove.
18769 (__arm_vrshlq_n_s16): Remove.
18770 (__arm_vrshlq_u32): Remove.
18771 (__arm_vrshlq_n_u32): Remove.
18772 (__arm_vrshlq_s32): Remove.
18773 (__arm_vrshlq_n_s32): Remove.
18774 (__arm_vrshlq_m_n_u8): Remove.
18775 (__arm_vrshlq_m_n_s8): Remove.
18776 (__arm_vrshlq_m_n_u16): Remove.
18777 (__arm_vrshlq_m_n_s16): Remove.
18778 (__arm_vrshlq_m_n_u32): Remove.
18779 (__arm_vrshlq_m_n_s32): Remove.
18780 (__arm_vrshlq_m_s8): Remove.
18781 (__arm_vrshlq_m_s32): Remove.
18782 (__arm_vrshlq_m_s16): Remove.
18783 (__arm_vrshlq_m_u8): Remove.
18784 (__arm_vrshlq_m_u32): Remove.
18785 (__arm_vrshlq_m_u16): Remove.
18786 (__arm_vrshlq_x_s8): Remove.
18787 (__arm_vrshlq_x_s16): Remove.
18788 (__arm_vrshlq_x_s32): Remove.
18789 (__arm_vrshlq_x_u8): Remove.
18790 (__arm_vrshlq_x_u16): Remove.
18791 (__arm_vrshlq_x_u32): Remove.
18792 (__arm_vrshlq): Remove.
18793 (__arm_vrshlq_m_n): Remove.
18794 (__arm_vrshlq_m): Remove.
18795 (__arm_vrshlq_x): Remove.
18797 (vqrshlq_m_n): Remove.
18798 (vqrshlq_m): Remove.
18799 (vqrshlq_u8): Remove.
18800 (vqrshlq_n_u8): Remove.
18801 (vqrshlq_s8): Remove.
18802 (vqrshlq_n_s8): Remove.
18803 (vqrshlq_u16): Remove.
18804 (vqrshlq_n_u16): Remove.
18805 (vqrshlq_s16): Remove.
18806 (vqrshlq_n_s16): Remove.
18807 (vqrshlq_u32): Remove.
18808 (vqrshlq_n_u32): Remove.
18809 (vqrshlq_s32): Remove.
18810 (vqrshlq_n_s32): Remove.
18811 (vqrshlq_m_n_u8): Remove.
18812 (vqrshlq_m_n_s8): Remove.
18813 (vqrshlq_m_n_u16): Remove.
18814 (vqrshlq_m_n_s16): Remove.
18815 (vqrshlq_m_n_u32): Remove.
18816 (vqrshlq_m_n_s32): Remove.
18817 (vqrshlq_m_s8): Remove.
18818 (vqrshlq_m_s32): Remove.
18819 (vqrshlq_m_s16): Remove.
18820 (vqrshlq_m_u8): Remove.
18821 (vqrshlq_m_u32): Remove.
18822 (vqrshlq_m_u16): Remove.
18823 (__arm_vqrshlq_u8): Remove.
18824 (__arm_vqrshlq_n_u8): Remove.
18825 (__arm_vqrshlq_s8): Remove.
18826 (__arm_vqrshlq_n_s8): Remove.
18827 (__arm_vqrshlq_u16): Remove.
18828 (__arm_vqrshlq_n_u16): Remove.
18829 (__arm_vqrshlq_s16): Remove.
18830 (__arm_vqrshlq_n_s16): Remove.
18831 (__arm_vqrshlq_u32): Remove.
18832 (__arm_vqrshlq_n_u32): Remove.
18833 (__arm_vqrshlq_s32): Remove.
18834 (__arm_vqrshlq_n_s32): Remove.
18835 (__arm_vqrshlq_m_n_u8): Remove.
18836 (__arm_vqrshlq_m_n_s8): Remove.
18837 (__arm_vqrshlq_m_n_u16): Remove.
18838 (__arm_vqrshlq_m_n_s16): Remove.
18839 (__arm_vqrshlq_m_n_u32): Remove.
18840 (__arm_vqrshlq_m_n_s32): Remove.
18841 (__arm_vqrshlq_m_s8): Remove.
18842 (__arm_vqrshlq_m_s32): Remove.
18843 (__arm_vqrshlq_m_s16): Remove.
18844 (__arm_vqrshlq_m_u8): Remove.
18845 (__arm_vqrshlq_m_u32): Remove.
18846 (__arm_vqrshlq_m_u16): Remove.
18847 (__arm_vqrshlq): Remove.
18848 (__arm_vqrshlq_m_n): Remove.
18849 (__arm_vqrshlq_m): Remove.
18851 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18853 * config/arm/iterators.md (MVE_RSHIFT_M_N, MVE_RSHIFT_N): New.
18854 (mve_insn): Add vqrshl, vrshl.
18855 * config/arm/mve.md (mve_vqrshlq_n_<supf><mode>)
18856 (mve_vrshlq_n_<supf><mode>): Merge into ...
18857 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
18858 (mve_vqrshlq_m_n_<supf><mode>, mve_vrshlq_m_n_<supf><mode>): Merge
18860 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
18862 2023-05-05 Christophe Lyon <christophe.lyon@arm.com>
18864 * config/arm/arm-mve-builtins-shapes.cc (binary_round_lshift): New.
18865 * config/arm/arm-mve-builtins-shapes.h (binary_round_lshift): New.
18867 2023-05-05 Juzhe-Zhong <juzhe.zhong@rivai.ai>
18870 * config/riscv/riscv-vsetvl.cc (avl_info::multiple_source_equal_p): Add
18871 denegrate PHI optmization.
18873 2023-05-05 Uros Bizjak <ubizjak@gmail.com>
18875 * config/i386/predicates.md (register_no_SP_operand):
18876 Rename from index_register_operand.
18877 (call_register_operand): Update for rename.
18878 * config/i386/i386.md (*lea<mode>_general_[1234]): Update for rename.
18880 2023-05-05 Tamar Christina <tamar.christina@arm.com>
18883 * Makefile.in (NUM_MATCH_SPLITS, MATCH_SPLITS_SEQ,
18884 GIMPLE_MATCH_PD_SEQ_SRC, GIMPLE_MATCH_PD_SEQ_O,
18885 GENERIC_MATCH_PD_SEQ_SRC, GENERIC_MATCH_PD_SEQ_O): New.
18886 (OBJS, MOSTLYCLEANFILES, .PRECIOUS): Use them.
18887 (s-match): Split into s-generic-match and s-gimple-match.
18888 * configure.ac (with-matchpd-partitions,
18889 DEFAULT_MATCHPD_PARTITIONS): New.
18890 * configure: Regenerate.
18892 2023-05-05 Tamar Christina <tamar.christina@arm.com>
18895 * genmatch.cc (emit_func, SIZED_BASED_CHUNKS, get_out_file): New.
18896 (decision_tree::gen): Accept list of files instead of single and update
18897 to write function definition to header and main file.
18898 (write_predicate): Likewise.
18899 (write_header): Emit pragmas and new includes.
18900 (main): Create file buffers and cleanup.
18901 (showUsage, write_header_includes): New.
18903 2023-05-05 Tamar Christina <tamar.christina@arm.com>
18906 * Makefile.in (OBJS): Add gimple-match-exports.o.
18907 * genmatch.cc (decision_tree::gen): Export gimple_gimplify helpers.
18908 * gimple-match-head.cc (gimple_simplify, gimple_resimplify1,
18909 gimple_resimplify2, gimple_resimplify3, gimple_resimplify4,
18910 gimple_resimplify5, constant_for_folding, convert_conditional_op,
18911 maybe_resimplify_conditional_op, gimple_match_op::resimplify,
18912 maybe_build_generic_op, build_call_internal, maybe_push_res_to_seq,
18913 do_valueize, try_conditional_simplification, gimple_extract,
18914 gimple_extract_op, canonicalize_code, commutative_binary_op_p,
18915 commutative_ternary_op_p, first_commutative_argument,
18916 associative_binary_op_p, directly_supported_p,
18917 get_conditional_internal_fn): Moved to gimple-match-exports.cc
18918 * gimple-match-exports.cc: New file.
18920 2023-05-05 Tamar Christina <tamar.christina@arm.com>
18923 * genmatch.cc (decision_tree::gen, write_predicate): Generate new
18925 (dt_simplify::gen_1): Use it.
18927 2023-05-05 Tamar Christina <tamar.christina@arm.com>
18930 * genmatch.cc (output_line_directive): Only emit commented directive
18933 2023-05-05 Tamar Christina <tamar.christina@arm.com>
18936 * genmatch.cc (dt_simplify::gen_1): Only emit labels if used.
18938 2023-05-05 Tobias Burnus <tobias@codesourcery.com>
18940 * config/gcn/gcn.cc (gcn_vectorize_builtin_vectorized_function): Remove
18941 unused in_mode/in_n variables.
18943 2023-05-05 Richard Biener <rguenther@suse.de>
18945 PR tree-optimization/109735
18946 * tree-vect-stmts.cc (vectorizable_operation): Perform
18947 conversion for POINTER_DIFF_EXPR unconditionally.
18949 2023-05-05 Uros Bizjak <ubizjak@gmail.com>
18951 * config/i386/mmx.md (mulv2si3): New expander.
18952 (*mulv2si3): New insn pattern.
18954 2023-05-05 Tobias Burnus <tobias@codesourcery.com>
18955 Thomas Schwinge <thomas@codesourcery.com>
18958 * config/nvptx/mkoffload.cc (process): Emit dummy procedure
18959 alongside reverse-offload function table to prevent NULL values
18960 of the function addresses.
18962 2023-05-05 Jakub Jelinek <jakub@redhat.com>
18964 * builtins.cc (do_mpfr_ckconv, do_mpc_ckconv): Fix comment typo,
18966 * fold-const-call.cc (do_mpfr_ckconv, do_mpc_ckconv): Likewise.
18968 2023-05-05 Andrew Pinski <apinski@marvell.com>
18970 PR tree-optimization/109732
18971 * tree-ssa-phiopt.cc (match_simplify_replacement): Fix the selection
18972 of the argtrue/argfalse.
18974 2023-05-05 Andrew Pinski <apinski@marvell.com>
18976 PR tree-optimization/109722
18977 * match.pd: Extend the `ABS<a> == 0` pattern
18978 to cover `ABSU<a> == 0` too.
18980 2023-05-04 Uros Bizjak <ubizjak@gmail.com>
18983 * config/i386/predicates.md (index_reg_operand): New predicate.
18984 * config/i386/i386.md (ashift to lea spliter): Use
18985 general_reg_operand and index_reg_operand predicates.
18987 2023-05-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18989 * config/aarch64/aarch64-simd.md (aarch64_<sur><addsub>hn2<mode>_insn_le):
18990 Rename and reimplement with RTL codes to...
18991 (aarch64_<optab>hn2<mode>_insn_le): .. This.
18992 (aarch64_r<optab>hn2<mode>_insn_le): New pattern.
18993 (aarch64_<sur><addsub>hn2<mode>_insn_be): Rename and reimplement with RTL
18995 (aarch64_<optab>hn2<mode>_insn_be): ... This.
18996 (aarch64_r<optab>hn2<mode>_insn_be): New pattern.
18997 (aarch64_<sur><addsub>hn2<mode>): Rename and adjust expander to...
18998 (aarch64_<optab>hn2<mode>): ... This.
18999 (aarch64_r<optab>hn2<mode>): New expander.
19000 * config/aarch64/iterators.md (UNSPEC_ADDHN, UNSPEC_RADDHN,
19001 UNSPEC_SUBHN, UNSPEC_RSUBHN): Delete unspecs.
19002 (ADDSUBHN): Delete.
19003 (sur): Remove handling of the above.
19004 (addsub): Likewise.
19006 2023-05-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19008 * config/aarch64/aarch64-simd.md (aarch64_<sur><addsub>hn<mode>_insn_le):
19010 (aarch64_<optab>hn<mode>_insn<vczle><vczbe>): New define_insn.
19011 (aarch64_<sur><addsub>hn<mode>_insn_be): Delete.
19012 (aarch64_r<optab>hn<mode>_insn<vczle><vczbe>): New define_insn.
19013 (aarch64_<sur><addsub>hn<mode>): Delete.
19014 (aarch64_<optab>hn<mode>): New define_expand.
19015 (aarch64_r<optab>hn<mode>): Likewise.
19016 * config/aarch64/predicates.md (aarch64_simd_raddsubhn_imm_vec):
19019 2023-05-04 Andrew Pinski <apinski@marvell.com>
19021 * tree-ssa-phiopt.cc (replace_phi_edge_with_variable): Handle
19022 diamond form bb with forwarder only empty blocks better.
19024 2023-05-04 Andrew Pinski <apinski@marvell.com>
19026 * tree-ssa-threadupdate.cc (copy_phi_arg_into_existing_phi): Move to ...
19027 * tree-cfg.cc (copy_phi_arg_into_existing_phi): Here and remove static.
19028 (gimple_duplicate_sese_tail): Use copy_phi_arg_into_existing_phi instead
19029 of an inline version of it.
19030 * tree-cfgcleanup.cc (remove_forwarder_block): Likewise.
19031 * tree-cfg.h (copy_phi_arg_into_existing_phi): New declaration.
19033 2023-05-04 Andrew Pinski <apinski@marvell.com>
19035 * tree-ssa-phiopt.cc (replace_phi_edge_with_variable): Change
19036 the default argument value for dce_ssa_names to nullptr.
19037 Check to make sure dce_ssa_names is a non-nullptr before
19038 calling simple_dce_from_worklist.
19040 2023-05-04 Uros Bizjak <ubizjak@gmail.com>
19042 * config/i386/predicates.md (index_register_operand): Reject
19043 arg_pointer_rtx, frame_pointer_rtx, stack_pointer_rtx and
19044 VIRTUAL_REGISTER_P operands. Allow subregs of memory before reload.
19045 (call_register_no_elim_operand): Rewrite as ...
19046 (call_register_operand): ... this.
19047 (call_insn_operand): Use call_register_operand predicate.
19049 2023-05-04 Richard Biener <rguenther@suse.de>
19051 PR tree-optimization/109721
19052 * tree-vect-stmts.cc (vectorizable_operation): Make sure
19053 to test word_mode for all !target_support_p operations.
19055 2023-05-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19058 * config/aarch64/aarch64-simd.md (aarch64_<su>aba<mode>): Rename to...
19059 (aarch64_<su>aba<mode><vczle><vczbe>): ... This.
19060 (aarch64_mla<mode>): Rename to...
19061 (aarch64_mla<mode><vczle><vczbe>): ... This.
19062 (*aarch64_mla_elt<mode>): Rename to...
19063 (*aarch64_mla_elt<mode><vczle><vczbe>): ... This.
19064 (*aarch64_mla_elt_<vswap_width_name><mode>): Rename to...
19065 (*aarch64_mla_elt_<vswap_width_name><mode><vczle><vczbe>): ... This.
19066 (aarch64_mla_n<mode>): Rename to...
19067 (aarch64_mla_n<mode><vczle><vczbe>): ... This.
19068 (aarch64_mls<mode>): Rename to...
19069 (aarch64_mls<mode><vczle><vczbe>): ... This.
19070 (*aarch64_mls_elt<mode>): Rename to...
19071 (*aarch64_mls_elt<mode><vczle><vczbe>): ... This.
19072 (*aarch64_mls_elt_<vswap_width_name><mode>): Rename to...
19073 (*aarch64_mls_elt_<vswap_width_name><mode><vczle><vczbe>): ... This.
19074 (aarch64_mls_n<mode>): Rename to...
19075 (aarch64_mls_n<mode><vczle><vczbe>): ... This.
19076 (fma<mode>4): Rename to...
19077 (fma<mode>4<vczle><vczbe>): ... This.
19078 (*aarch64_fma4_elt<mode>): Rename to...
19079 (*aarch64_fma4_elt<mode><vczle><vczbe>): ... This.
19080 (*aarch64_fma4_elt_<vswap_width_name><mode>): Rename to...
19081 (*aarch64_fma4_elt_<vswap_width_name><mode><vczle><vczbe>): ... This.
19082 (*aarch64_fma4_elt_from_dup<mode>): Rename to...
19083 (*aarch64_fma4_elt_from_dup<mode><vczle><vczbe>): ... This.
19084 (fnma<mode>4): Rename to...
19085 (fnma<mode>4<vczle><vczbe>): ... This.
19086 (*aarch64_fnma4_elt<mode>): Rename to...
19087 (*aarch64_fnma4_elt<mode><vczle><vczbe>): ... This.
19088 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Rename to...
19089 (*aarch64_fnma4_elt_<vswap_width_name><mode><vczle><vczbe>): ... This.
19090 (*aarch64_fnma4_elt_from_dup<mode>): Rename to...
19091 (*aarch64_fnma4_elt_from_dup<mode><vczle><vczbe>): ... This.
19092 (aarch64_simd_bsl<mode>_internal): Rename to...
19093 (aarch64_simd_bsl<mode>_internal<vczle><vczbe>): ... This.
19094 (*aarch64_simd_bsl<mode>_alt): Rename to...
19095 (*aarch64_simd_bsl<mode>_alt<vczle><vczbe>): ... This.
19097 2023-05-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19100 * config/aarch64/aarch64-simd.md (aarch64_<su>abd<mode>): Rename to...
19101 (aarch64_<su>abd<mode><vczle><vczbe>): ... This.
19102 (fabd<mode>3): Rename to...
19103 (fabd<mode>3<vczle><vczbe>): ... This.
19104 (aarch64_<optab>p<mode>): Rename to...
19105 (aarch64_<optab>p<mode><vczle><vczbe>): ... This.
19106 (aarch64_faddp<mode>): Rename to...
19107 (aarch64_faddp<mode><vczle><vczbe>): ... This.
19109 2023-05-04 Martin Liska <mliska@suse.cz>
19111 * gcov.cc (GCOV_JSON_FORMAT_VERSION): New definition.
19112 (print_version): Use it.
19113 (generate_results): Likewise.
19115 2023-05-04 Richard Biener <rguenther@suse.de>
19117 * tree-cfg.h (last_stmt): Rename to ...
19118 (last_nondebug_stmt): ... this.
19119 * tree-cfg.cc (last_stmt): Rename to ...
19120 (last_nondebug_stmt): ... this.
19121 (assign_discriminators): Adjust.
19122 (group_case_labels_stmt): Likewise.
19123 (gimple_can_duplicate_bb_p): Likewise.
19124 (execute_fixup_cfg): Likewise.
19125 * auto-profile.cc (afdo_propagate_circuit): Likewise.
19126 * gimple-range.cc (gimple_ranger::range_on_exit): Likewise.
19127 * omp-expand.cc (workshare_safe_to_combine_p): Likewise.
19128 (determine_parallel_type): Likewise.
19129 (adjust_context_and_scope): Likewise.
19130 (expand_task_call): Likewise.
19131 (remove_exit_barrier): Likewise.
19132 (expand_omp_taskreg): Likewise.
19133 (expand_omp_for_init_counts): Likewise.
19134 (expand_omp_for_init_vars): Likewise.
19135 (expand_omp_for_static_chunk): Likewise.
19136 (expand_omp_simd): Likewise.
19137 (expand_oacc_for): Likewise.
19138 (expand_omp_for): Likewise.
19139 (expand_omp_sections): Likewise.
19140 (expand_omp_atomic_fetch_op): Likewise.
19141 (expand_omp_atomic_cas): Likewise.
19142 (expand_omp_atomic): Likewise.
19143 (expand_omp_target): Likewise.
19144 (expand_omp): Likewise.
19145 (omp_make_gimple_edges): Likewise.
19146 * trans-mem.cc (tm_region_init): Likewise.
19147 * tree-inline.cc (redirect_all_calls): Likewise.
19148 * tree-parloops.cc (gen_parallel_loop): Likewise.
19149 * tree-ssa-loop-ch.cc (do_while_loop_p): Likewise.
19150 * tree-ssa-loop-ivcanon.cc (canonicalize_loop_induction_variables):
19152 * tree-ssa-loop-ivopts.cc (stmt_after_ip_normal_pos): Likewise.
19153 (may_eliminate_iv): Likewise.
19154 * tree-ssa-loop-manip.cc (standard_iv_increment_position): Likewise.
19155 * tree-ssa-loop-niter.cc (do_warn_aggressive_loop_optimizations):
19157 (estimate_numbers_of_iterations): Likewise.
19158 * tree-ssa-loop-split.cc (compute_added_num_insns): Likewise.
19159 * tree-ssa-loop-unswitch.cc (get_predicates_for_bb): Likewise.
19160 (set_predicates_for_bb): Likewise.
19161 (init_loop_unswitch_info): Likewise.
19162 (hoist_guard): Likewise.
19163 * tree-ssa-phiopt.cc (match_simplify_replacement): Likewise.
19164 (minmax_replacement): Likewise.
19165 * tree-ssa-reassoc.cc (update_range_test): Likewise.
19166 (optimize_range_tests_to_bit_test): Likewise.
19167 (optimize_range_tests_var_bound): Likewise.
19168 (optimize_range_tests): Likewise.
19169 (no_side_effect_bb): Likewise.
19170 (suitable_cond_bb): Likewise.
19171 (maybe_optimize_range_tests): Likewise.
19172 (reassociate_bb): Likewise.
19173 * tree-vrp.cc (rvrp_folder::pre_fold_bb): Likewise.
19175 2023-05-04 Jakub Jelinek <jakub@redhat.com>
19178 * config/i386/i386-features.cc (timode_scalar_chain::convert_insn):
19179 If src is REG, change its mode to V1TImode and call fix_debug_reg_uses
19180 for it only if it still has TImode. Don't decide whether to call
19181 fix_debug_reg_uses based on whether SRC is ever set or not.
19183 2023-05-04 Hans-Peter Nilsson <hp@axis.com>
19185 * config/cris/cris.cc (cris_split_constant): New function.
19186 * config/cris/cris.md (splitop): New iterator.
19187 (opsplit1): New define_peephole2.
19188 * config/cris/cris-protos.h (cris_split_constant): Declare.
19189 (cris_splittable_constant_p): New macro.
19191 2023-05-04 Hans-Peter Nilsson <hp@axis.com>
19193 * config/cris/cris.cc (TARGET_SPILL_CLASS): Define
19196 2023-05-04 Hans-Peter Nilsson <hp@axis.com>
19198 * config/cris/cris.cc (cris_side_effect_mode_ok): Use
19199 lra_in_progress, not reload_in_progress.
19200 * config/cris/cris.md ("movdi", "*addi_reload"): Ditto.
19201 * config/cris/constraints.md ("Q"): Ditto.
19203 2023-05-03 Andrew Pinski <apinski@marvell.com>
19205 * tree-ssa-dce.cc (simple_dce_from_worklist): Record
19206 stats on removed number of statements and phis.
19208 2023-05-03 Aldy Hernandez <aldyh@redhat.com>
19210 PR tree-optimization/109711
19211 * value-range.cc (irange::verify_range): Allow types of
19214 2023-05-03 Alexander Monakov <amonakov@ispras.ru>
19217 * calls.cc (can_implement_as_sibling_call_p): Reject calls
19218 to __sanitizer_cov_trace_pc.
19220 2023-05-03 Richard Sandiford <richard.sandiford@arm.com>
19223 * config/aarch64/aarch64.cc (aarch64_function_arg_alignment): Add
19224 a new ABI break parameter for GCC 14. Set it to the alignment
19225 of enums that have an underlying type. Take the true alignment
19226 of such enums from the TYPE_ALIGN of the underlying type's
19228 (aarch64_function_arg_boundary): Update accordingly.
19229 (aarch64_layout_arg, aarch64_gimplify_va_arg_expr): Likewise.
19230 Warn about ABI differences.
19232 2023-05-03 Richard Sandiford <richard.sandiford@arm.com>
19235 * config/aarch64/aarch64.cc (aarch64_function_arg_alignment): Rename
19236 ABI break variables to abi_break_gcc_9 and abi_break_gcc_13.
19237 (aarch64_layout_arg, aarch64_function_arg_boundary): Likewise.
19238 (aarch64_gimplify_va_arg_expr): Likewise.
19240 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
19242 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITH_M_N_NO_F)
19243 (FUNCTION_WITHOUT_N_NO_F, FUNCTION_WITH_M_N_NO_U_F): New.
19244 (vhaddq, vhsubq, vmulhq, vqaddq, vqsubq, vqdmulhq, vrhaddq)
19246 * config/arm/arm-mve-builtins-base.def (vhaddq, vhsubq, vmulhq)
19247 (vqaddq, vqsubq, vqdmulhq, vrhaddq, vrmulhq): New.
19248 * config/arm/arm-mve-builtins-base.h (vhaddq, vhsubq, vmulhq)
19249 (vqaddq, vqsubq, vqdmulhq, vrhaddq, vrmulhq): New.
19250 * config/arm/arm_mve.h (vhsubq): Remove.
19252 (vhaddq_m): Remove.
19253 (vhsubq_m): Remove.
19254 (vhaddq_x): Remove.
19255 (vhsubq_x): Remove.
19256 (vhsubq_u8): Remove.
19257 (vhsubq_n_u8): Remove.
19258 (vhaddq_u8): Remove.
19259 (vhaddq_n_u8): Remove.
19260 (vhsubq_s8): Remove.
19261 (vhsubq_n_s8): Remove.
19262 (vhaddq_s8): Remove.
19263 (vhaddq_n_s8): Remove.
19264 (vhsubq_u16): Remove.
19265 (vhsubq_n_u16): Remove.
19266 (vhaddq_u16): Remove.
19267 (vhaddq_n_u16): Remove.
19268 (vhsubq_s16): Remove.
19269 (vhsubq_n_s16): Remove.
19270 (vhaddq_s16): Remove.
19271 (vhaddq_n_s16): Remove.
19272 (vhsubq_u32): Remove.
19273 (vhsubq_n_u32): Remove.
19274 (vhaddq_u32): Remove.
19275 (vhaddq_n_u32): Remove.
19276 (vhsubq_s32): Remove.
19277 (vhsubq_n_s32): Remove.
19278 (vhaddq_s32): Remove.
19279 (vhaddq_n_s32): Remove.
19280 (vhaddq_m_n_s8): Remove.
19281 (vhaddq_m_n_s32): Remove.
19282 (vhaddq_m_n_s16): Remove.
19283 (vhaddq_m_n_u8): Remove.
19284 (vhaddq_m_n_u32): Remove.
19285 (vhaddq_m_n_u16): Remove.
19286 (vhaddq_m_s8): Remove.
19287 (vhaddq_m_s32): Remove.
19288 (vhaddq_m_s16): Remove.
19289 (vhaddq_m_u8): Remove.
19290 (vhaddq_m_u32): Remove.
19291 (vhaddq_m_u16): Remove.
19292 (vhsubq_m_n_s8): Remove.
19293 (vhsubq_m_n_s32): Remove.
19294 (vhsubq_m_n_s16): Remove.
19295 (vhsubq_m_n_u8): Remove.
19296 (vhsubq_m_n_u32): Remove.
19297 (vhsubq_m_n_u16): Remove.
19298 (vhsubq_m_s8): Remove.
19299 (vhsubq_m_s32): Remove.
19300 (vhsubq_m_s16): Remove.
19301 (vhsubq_m_u8): Remove.
19302 (vhsubq_m_u32): Remove.
19303 (vhsubq_m_u16): Remove.
19304 (vhaddq_x_n_s8): Remove.
19305 (vhaddq_x_n_s16): Remove.
19306 (vhaddq_x_n_s32): Remove.
19307 (vhaddq_x_n_u8): Remove.
19308 (vhaddq_x_n_u16): Remove.
19309 (vhaddq_x_n_u32): Remove.
19310 (vhaddq_x_s8): Remove.
19311 (vhaddq_x_s16): Remove.
19312 (vhaddq_x_s32): Remove.
19313 (vhaddq_x_u8): Remove.
19314 (vhaddq_x_u16): Remove.
19315 (vhaddq_x_u32): Remove.
19316 (vhsubq_x_n_s8): Remove.
19317 (vhsubq_x_n_s16): Remove.
19318 (vhsubq_x_n_s32): Remove.
19319 (vhsubq_x_n_u8): Remove.
19320 (vhsubq_x_n_u16): Remove.
19321 (vhsubq_x_n_u32): Remove.
19322 (vhsubq_x_s8): Remove.
19323 (vhsubq_x_s16): Remove.
19324 (vhsubq_x_s32): Remove.
19325 (vhsubq_x_u8): Remove.
19326 (vhsubq_x_u16): Remove.
19327 (vhsubq_x_u32): Remove.
19328 (__arm_vhsubq_u8): Remove.
19329 (__arm_vhsubq_n_u8): Remove.
19330 (__arm_vhaddq_u8): Remove.
19331 (__arm_vhaddq_n_u8): Remove.
19332 (__arm_vhsubq_s8): Remove.
19333 (__arm_vhsubq_n_s8): Remove.
19334 (__arm_vhaddq_s8): Remove.
19335 (__arm_vhaddq_n_s8): Remove.
19336 (__arm_vhsubq_u16): Remove.
19337 (__arm_vhsubq_n_u16): Remove.
19338 (__arm_vhaddq_u16): Remove.
19339 (__arm_vhaddq_n_u16): Remove.
19340 (__arm_vhsubq_s16): Remove.
19341 (__arm_vhsubq_n_s16): Remove.
19342 (__arm_vhaddq_s16): Remove.
19343 (__arm_vhaddq_n_s16): Remove.
19344 (__arm_vhsubq_u32): Remove.
19345 (__arm_vhsubq_n_u32): Remove.
19346 (__arm_vhaddq_u32): Remove.
19347 (__arm_vhaddq_n_u32): Remove.
19348 (__arm_vhsubq_s32): Remove.
19349 (__arm_vhsubq_n_s32): Remove.
19350 (__arm_vhaddq_s32): Remove.
19351 (__arm_vhaddq_n_s32): Remove.
19352 (__arm_vhaddq_m_n_s8): Remove.
19353 (__arm_vhaddq_m_n_s32): Remove.
19354 (__arm_vhaddq_m_n_s16): Remove.
19355 (__arm_vhaddq_m_n_u8): Remove.
19356 (__arm_vhaddq_m_n_u32): Remove.
19357 (__arm_vhaddq_m_n_u16): Remove.
19358 (__arm_vhaddq_m_s8): Remove.
19359 (__arm_vhaddq_m_s32): Remove.
19360 (__arm_vhaddq_m_s16): Remove.
19361 (__arm_vhaddq_m_u8): Remove.
19362 (__arm_vhaddq_m_u32): Remove.
19363 (__arm_vhaddq_m_u16): Remove.
19364 (__arm_vhsubq_m_n_s8): Remove.
19365 (__arm_vhsubq_m_n_s32): Remove.
19366 (__arm_vhsubq_m_n_s16): Remove.
19367 (__arm_vhsubq_m_n_u8): Remove.
19368 (__arm_vhsubq_m_n_u32): Remove.
19369 (__arm_vhsubq_m_n_u16): Remove.
19370 (__arm_vhsubq_m_s8): Remove.
19371 (__arm_vhsubq_m_s32): Remove.
19372 (__arm_vhsubq_m_s16): Remove.
19373 (__arm_vhsubq_m_u8): Remove.
19374 (__arm_vhsubq_m_u32): Remove.
19375 (__arm_vhsubq_m_u16): Remove.
19376 (__arm_vhaddq_x_n_s8): Remove.
19377 (__arm_vhaddq_x_n_s16): Remove.
19378 (__arm_vhaddq_x_n_s32): Remove.
19379 (__arm_vhaddq_x_n_u8): Remove.
19380 (__arm_vhaddq_x_n_u16): Remove.
19381 (__arm_vhaddq_x_n_u32): Remove.
19382 (__arm_vhaddq_x_s8): Remove.
19383 (__arm_vhaddq_x_s16): Remove.
19384 (__arm_vhaddq_x_s32): Remove.
19385 (__arm_vhaddq_x_u8): Remove.
19386 (__arm_vhaddq_x_u16): Remove.
19387 (__arm_vhaddq_x_u32): Remove.
19388 (__arm_vhsubq_x_n_s8): Remove.
19389 (__arm_vhsubq_x_n_s16): Remove.
19390 (__arm_vhsubq_x_n_s32): Remove.
19391 (__arm_vhsubq_x_n_u8): Remove.
19392 (__arm_vhsubq_x_n_u16): Remove.
19393 (__arm_vhsubq_x_n_u32): Remove.
19394 (__arm_vhsubq_x_s8): Remove.
19395 (__arm_vhsubq_x_s16): Remove.
19396 (__arm_vhsubq_x_s32): Remove.
19397 (__arm_vhsubq_x_u8): Remove.
19398 (__arm_vhsubq_x_u16): Remove.
19399 (__arm_vhsubq_x_u32): Remove.
19400 (__arm_vhsubq): Remove.
19401 (__arm_vhaddq): Remove.
19402 (__arm_vhaddq_m): Remove.
19403 (__arm_vhsubq_m): Remove.
19404 (__arm_vhaddq_x): Remove.
19405 (__arm_vhsubq_x): Remove.
19407 (vmulhq_m): Remove.
19408 (vmulhq_x): Remove.
19409 (vmulhq_u8): Remove.
19410 (vmulhq_s8): Remove.
19411 (vmulhq_u16): Remove.
19412 (vmulhq_s16): Remove.
19413 (vmulhq_u32): Remove.
19414 (vmulhq_s32): Remove.
19415 (vmulhq_m_s8): Remove.
19416 (vmulhq_m_s32): Remove.
19417 (vmulhq_m_s16): Remove.
19418 (vmulhq_m_u8): Remove.
19419 (vmulhq_m_u32): Remove.
19420 (vmulhq_m_u16): Remove.
19421 (vmulhq_x_s8): Remove.
19422 (vmulhq_x_s16): Remove.
19423 (vmulhq_x_s32): Remove.
19424 (vmulhq_x_u8): Remove.
19425 (vmulhq_x_u16): Remove.
19426 (vmulhq_x_u32): Remove.
19427 (__arm_vmulhq_u8): Remove.
19428 (__arm_vmulhq_s8): Remove.
19429 (__arm_vmulhq_u16): Remove.
19430 (__arm_vmulhq_s16): Remove.
19431 (__arm_vmulhq_u32): Remove.
19432 (__arm_vmulhq_s32): Remove.
19433 (__arm_vmulhq_m_s8): Remove.
19434 (__arm_vmulhq_m_s32): Remove.
19435 (__arm_vmulhq_m_s16): Remove.
19436 (__arm_vmulhq_m_u8): Remove.
19437 (__arm_vmulhq_m_u32): Remove.
19438 (__arm_vmulhq_m_u16): Remove.
19439 (__arm_vmulhq_x_s8): Remove.
19440 (__arm_vmulhq_x_s16): Remove.
19441 (__arm_vmulhq_x_s32): Remove.
19442 (__arm_vmulhq_x_u8): Remove.
19443 (__arm_vmulhq_x_u16): Remove.
19444 (__arm_vmulhq_x_u32): Remove.
19445 (__arm_vmulhq): Remove.
19446 (__arm_vmulhq_m): Remove.
19447 (__arm_vmulhq_x): Remove.
19450 (vqaddq_m): Remove.
19451 (vqsubq_m): Remove.
19452 (vqsubq_u8): Remove.
19453 (vqsubq_n_u8): Remove.
19454 (vqaddq_u8): Remove.
19455 (vqaddq_n_u8): Remove.
19456 (vqsubq_s8): Remove.
19457 (vqsubq_n_s8): Remove.
19458 (vqaddq_s8): Remove.
19459 (vqaddq_n_s8): Remove.
19460 (vqsubq_u16): Remove.
19461 (vqsubq_n_u16): Remove.
19462 (vqaddq_u16): Remove.
19463 (vqaddq_n_u16): Remove.
19464 (vqsubq_s16): Remove.
19465 (vqsubq_n_s16): Remove.
19466 (vqaddq_s16): Remove.
19467 (vqaddq_n_s16): Remove.
19468 (vqsubq_u32): Remove.
19469 (vqsubq_n_u32): Remove.
19470 (vqaddq_u32): Remove.
19471 (vqaddq_n_u32): Remove.
19472 (vqsubq_s32): Remove.
19473 (vqsubq_n_s32): Remove.
19474 (vqaddq_s32): Remove.
19475 (vqaddq_n_s32): Remove.
19476 (vqaddq_m_n_s8): Remove.
19477 (vqaddq_m_n_s32): Remove.
19478 (vqaddq_m_n_s16): Remove.
19479 (vqaddq_m_n_u8): Remove.
19480 (vqaddq_m_n_u32): Remove.
19481 (vqaddq_m_n_u16): Remove.
19482 (vqaddq_m_s8): Remove.
19483 (vqaddq_m_s32): Remove.
19484 (vqaddq_m_s16): Remove.
19485 (vqaddq_m_u8): Remove.
19486 (vqaddq_m_u32): Remove.
19487 (vqaddq_m_u16): Remove.
19488 (vqsubq_m_n_s8): Remove.
19489 (vqsubq_m_n_s32): Remove.
19490 (vqsubq_m_n_s16): Remove.
19491 (vqsubq_m_n_u8): Remove.
19492 (vqsubq_m_n_u32): Remove.
19493 (vqsubq_m_n_u16): Remove.
19494 (vqsubq_m_s8): Remove.
19495 (vqsubq_m_s32): Remove.
19496 (vqsubq_m_s16): Remove.
19497 (vqsubq_m_u8): Remove.
19498 (vqsubq_m_u32): Remove.
19499 (vqsubq_m_u16): Remove.
19500 (__arm_vqsubq_u8): Remove.
19501 (__arm_vqsubq_n_u8): Remove.
19502 (__arm_vqaddq_u8): Remove.
19503 (__arm_vqaddq_n_u8): Remove.
19504 (__arm_vqsubq_s8): Remove.
19505 (__arm_vqsubq_n_s8): Remove.
19506 (__arm_vqaddq_s8): Remove.
19507 (__arm_vqaddq_n_s8): Remove.
19508 (__arm_vqsubq_u16): Remove.
19509 (__arm_vqsubq_n_u16): Remove.
19510 (__arm_vqaddq_u16): Remove.
19511 (__arm_vqaddq_n_u16): Remove.
19512 (__arm_vqsubq_s16): Remove.
19513 (__arm_vqsubq_n_s16): Remove.
19514 (__arm_vqaddq_s16): Remove.
19515 (__arm_vqaddq_n_s16): Remove.
19516 (__arm_vqsubq_u32): Remove.
19517 (__arm_vqsubq_n_u32): Remove.
19518 (__arm_vqaddq_u32): Remove.
19519 (__arm_vqaddq_n_u32): Remove.
19520 (__arm_vqsubq_s32): Remove.
19521 (__arm_vqsubq_n_s32): Remove.
19522 (__arm_vqaddq_s32): Remove.
19523 (__arm_vqaddq_n_s32): Remove.
19524 (__arm_vqaddq_m_n_s8): Remove.
19525 (__arm_vqaddq_m_n_s32): Remove.
19526 (__arm_vqaddq_m_n_s16): Remove.
19527 (__arm_vqaddq_m_n_u8): Remove.
19528 (__arm_vqaddq_m_n_u32): Remove.
19529 (__arm_vqaddq_m_n_u16): Remove.
19530 (__arm_vqaddq_m_s8): Remove.
19531 (__arm_vqaddq_m_s32): Remove.
19532 (__arm_vqaddq_m_s16): Remove.
19533 (__arm_vqaddq_m_u8): Remove.
19534 (__arm_vqaddq_m_u32): Remove.
19535 (__arm_vqaddq_m_u16): Remove.
19536 (__arm_vqsubq_m_n_s8): Remove.
19537 (__arm_vqsubq_m_n_s32): Remove.
19538 (__arm_vqsubq_m_n_s16): Remove.
19539 (__arm_vqsubq_m_n_u8): Remove.
19540 (__arm_vqsubq_m_n_u32): Remove.
19541 (__arm_vqsubq_m_n_u16): Remove.
19542 (__arm_vqsubq_m_s8): Remove.
19543 (__arm_vqsubq_m_s32): Remove.
19544 (__arm_vqsubq_m_s16): Remove.
19545 (__arm_vqsubq_m_u8): Remove.
19546 (__arm_vqsubq_m_u32): Remove.
19547 (__arm_vqsubq_m_u16): Remove.
19548 (__arm_vqsubq): Remove.
19549 (__arm_vqaddq): Remove.
19550 (__arm_vqaddq_m): Remove.
19551 (__arm_vqsubq_m): Remove.
19552 (vqdmulhq): Remove.
19553 (vqdmulhq_m): Remove.
19554 (vqdmulhq_s8): Remove.
19555 (vqdmulhq_n_s8): Remove.
19556 (vqdmulhq_s16): Remove.
19557 (vqdmulhq_n_s16): Remove.
19558 (vqdmulhq_s32): Remove.
19559 (vqdmulhq_n_s32): Remove.
19560 (vqdmulhq_m_n_s8): Remove.
19561 (vqdmulhq_m_n_s32): Remove.
19562 (vqdmulhq_m_n_s16): Remove.
19563 (vqdmulhq_m_s8): Remove.
19564 (vqdmulhq_m_s32): Remove.
19565 (vqdmulhq_m_s16): Remove.
19566 (__arm_vqdmulhq_s8): Remove.
19567 (__arm_vqdmulhq_n_s8): Remove.
19568 (__arm_vqdmulhq_s16): Remove.
19569 (__arm_vqdmulhq_n_s16): Remove.
19570 (__arm_vqdmulhq_s32): Remove.
19571 (__arm_vqdmulhq_n_s32): Remove.
19572 (__arm_vqdmulhq_m_n_s8): Remove.
19573 (__arm_vqdmulhq_m_n_s32): Remove.
19574 (__arm_vqdmulhq_m_n_s16): Remove.
19575 (__arm_vqdmulhq_m_s8): Remove.
19576 (__arm_vqdmulhq_m_s32): Remove.
19577 (__arm_vqdmulhq_m_s16): Remove.
19578 (__arm_vqdmulhq): Remove.
19579 (__arm_vqdmulhq_m): Remove.
19581 (vrhaddq_m): Remove.
19582 (vrhaddq_x): Remove.
19583 (vrhaddq_u8): Remove.
19584 (vrhaddq_s8): Remove.
19585 (vrhaddq_u16): Remove.
19586 (vrhaddq_s16): Remove.
19587 (vrhaddq_u32): Remove.
19588 (vrhaddq_s32): Remove.
19589 (vrhaddq_m_s8): Remove.
19590 (vrhaddq_m_s32): Remove.
19591 (vrhaddq_m_s16): Remove.
19592 (vrhaddq_m_u8): Remove.
19593 (vrhaddq_m_u32): Remove.
19594 (vrhaddq_m_u16): Remove.
19595 (vrhaddq_x_s8): Remove.
19596 (vrhaddq_x_s16): Remove.
19597 (vrhaddq_x_s32): Remove.
19598 (vrhaddq_x_u8): Remove.
19599 (vrhaddq_x_u16): Remove.
19600 (vrhaddq_x_u32): Remove.
19601 (__arm_vrhaddq_u8): Remove.
19602 (__arm_vrhaddq_s8): Remove.
19603 (__arm_vrhaddq_u16): Remove.
19604 (__arm_vrhaddq_s16): Remove.
19605 (__arm_vrhaddq_u32): Remove.
19606 (__arm_vrhaddq_s32): Remove.
19607 (__arm_vrhaddq_m_s8): Remove.
19608 (__arm_vrhaddq_m_s32): Remove.
19609 (__arm_vrhaddq_m_s16): Remove.
19610 (__arm_vrhaddq_m_u8): Remove.
19611 (__arm_vrhaddq_m_u32): Remove.
19612 (__arm_vrhaddq_m_u16): Remove.
19613 (__arm_vrhaddq_x_s8): Remove.
19614 (__arm_vrhaddq_x_s16): Remove.
19615 (__arm_vrhaddq_x_s32): Remove.
19616 (__arm_vrhaddq_x_u8): Remove.
19617 (__arm_vrhaddq_x_u16): Remove.
19618 (__arm_vrhaddq_x_u32): Remove.
19619 (__arm_vrhaddq): Remove.
19620 (__arm_vrhaddq_m): Remove.
19621 (__arm_vrhaddq_x): Remove.
19623 (vrmulhq_m): Remove.
19624 (vrmulhq_x): Remove.
19625 (vrmulhq_u8): Remove.
19626 (vrmulhq_s8): Remove.
19627 (vrmulhq_u16): Remove.
19628 (vrmulhq_s16): Remove.
19629 (vrmulhq_u32): Remove.
19630 (vrmulhq_s32): Remove.
19631 (vrmulhq_m_s8): Remove.
19632 (vrmulhq_m_s32): Remove.
19633 (vrmulhq_m_s16): Remove.
19634 (vrmulhq_m_u8): Remove.
19635 (vrmulhq_m_u32): Remove.
19636 (vrmulhq_m_u16): Remove.
19637 (vrmulhq_x_s8): Remove.
19638 (vrmulhq_x_s16): Remove.
19639 (vrmulhq_x_s32): Remove.
19640 (vrmulhq_x_u8): Remove.
19641 (vrmulhq_x_u16): Remove.
19642 (vrmulhq_x_u32): Remove.
19643 (__arm_vrmulhq_u8): Remove.
19644 (__arm_vrmulhq_s8): Remove.
19645 (__arm_vrmulhq_u16): Remove.
19646 (__arm_vrmulhq_s16): Remove.
19647 (__arm_vrmulhq_u32): Remove.
19648 (__arm_vrmulhq_s32): Remove.
19649 (__arm_vrmulhq_m_s8): Remove.
19650 (__arm_vrmulhq_m_s32): Remove.
19651 (__arm_vrmulhq_m_s16): Remove.
19652 (__arm_vrmulhq_m_u8): Remove.
19653 (__arm_vrmulhq_m_u32): Remove.
19654 (__arm_vrmulhq_m_u16): Remove.
19655 (__arm_vrmulhq_x_s8): Remove.
19656 (__arm_vrmulhq_x_s16): Remove.
19657 (__arm_vrmulhq_x_s32): Remove.
19658 (__arm_vrmulhq_x_u8): Remove.
19659 (__arm_vrmulhq_x_u16): Remove.
19660 (__arm_vrmulhq_x_u32): Remove.
19661 (__arm_vrmulhq): Remove.
19662 (__arm_vrmulhq_m): Remove.
19663 (__arm_vrmulhq_x): Remove.
19665 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
19667 * config/arm/iterators.md (MVE_INT_SU_BINARY): New.
19668 (mve_insn): Add vabdq, vhaddq, vhsubq, vmulhq, vqaddq, vqdmulhq,
19669 vqrdmulhq, vqrshlq, vqshlq, vqsubq, vrhaddq, vrmulhq, vrshlq.
19670 (supf): Add VQDMULHQ_S, VQRDMULHQ_S.
19671 * config/arm/mve.md (mve_vabdq_<supf><mode>)
19672 (@mve_vhaddq_<supf><mode>, mve_vhsubq_<supf><mode>)
19673 (mve_vmulhq_<supf><mode>, mve_vqaddq_<supf><mode>)
19674 (mve_vqdmulhq_s<mode>, mve_vqrdmulhq_s<mode>)
19675 (mve_vqrshlq_<supf><mode>, mve_vqshlq_<supf><mode>)
19676 (mve_vqsubq_<supf><mode>, @mve_vrhaddq_<supf><mode>)
19677 (mve_vrmulhq_<supf><mode>, mve_vrshlq_<supf><mode>): Merge into
19679 (@mve_<mve_insn>q_<supf><mode>): ... this.
19680 * config/arm/vec-common.md (avg<mode>3_floor, uavg<mode>3_floor)
19681 (avg<mode>3_ceil, uavg<mode>3_ceil): Use gen_mve_q instead of
19682 gen_mve_vhaddq / gen_mve_vrhaddq.
19684 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
19686 * config/arm/iterators.md (MVE_INT_SU_M_N_BINARY): New.
19687 (mve_insn): Add vhaddq, vhsubq, vmlaq, vmlasq, vqaddq, vqdmlahq,
19688 vqdmlashq, vqdmulhq, vqrdmlahq, vqrdmlashq, vqrdmulhq, vqsubq.
19689 (supf): Add VQDMLAHQ_M_N_S, VQDMLASHQ_M_N_S, VQRDMLAHQ_M_N_S,
19690 VQRDMLASHQ_M_N_S, VQDMULHQ_M_N_S, VQRDMULHQ_M_N_S.
19691 * config/arm/mve.md (mve_vhaddq_m_n_<supf><mode>)
19692 (mve_vhsubq_m_n_<supf><mode>, mve_vmlaq_m_n_<supf><mode>)
19693 (mve_vmlasq_m_n_<supf><mode>, mve_vqaddq_m_n_<supf><mode>)
19694 (mve_vqdmlahq_m_n_s<mode>, mve_vqdmlashq_m_n_s<mode>)
19695 (mve_vqrdmlahq_m_n_s<mode>, mve_vqrdmlashq_m_n_s<mode>)
19696 (mve_vqsubq_m_n_<supf><mode>, mve_vqdmulhq_m_n_s<mode>)
19697 (mve_vqrdmulhq_m_n_s<mode>): Merge into ...
19698 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
19700 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
19702 * config/arm/iterators.md (MVE_INT_SU_N_BINARY): New.
19703 (mve_insn): Add vhaddq, vhsubq, vqaddq, vqdmulhq, vqrdmulhq,
19705 (supf): Add VQDMULHQ_N_S, VQRDMULHQ_N_S.
19706 * config/arm/mve.md (mve_vhaddq_n_<supf><mode>)
19707 (mve_vhsubq_n_<supf><mode>, mve_vqaddq_n_<supf><mode>)
19708 (mve_vqdmulhq_n_s<mode>, mve_vqrdmulhq_n_s<mode>)
19709 (mve_vqsubq_n_<supf><mode>): Merge into ...
19710 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
19712 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
19714 * config/arm/iterators.md (MVE_INT_SU_M_BINARY): New.
19715 (mve_insn): Add vabdq, vhaddq, vhsubq, vmaxq, vminq, vmulhq,
19716 vqaddq, vqdmladhq, vqdmladhxq, vqdmlsdhq, vqdmlsdhxq, vqdmulhq,
19717 vqrdmladhq, vqrdmladhxq, vqrdmlsdhq, vqrdmlsdhxq, vqrdmulhq,
19718 vqrshlq, vqshlq, vqsubq, vrhaddq, vrmulhq, vrshlq, vshlq.
19719 (supf): Add VQDMLADHQ_M_S, VQDMLADHXQ_M_S, VQDMLSDHQ_M_S,
19720 VQDMLSDHXQ_M_S, VQDMULHQ_M_S, VQRDMLADHQ_M_S, VQRDMLADHXQ_M_S,
19721 VQRDMLSDHQ_M_S, VQRDMLSDHXQ_M_S, VQRDMULHQ_M_S.
19722 * config/arm/mve.md (@mve_<mve_insn>q_m_<supf><mode>): New.
19723 (mve_vshlq_m_<supf><mode>): Merged into
19724 @mve_<mve_insn>q_m_<supf><mode>.
19725 (mve_vabdq_m_<supf><mode>): Likewise.
19726 (mve_vhaddq_m_<supf><mode>): Likewise.
19727 (mve_vhsubq_m_<supf><mode>): Likewise.
19728 (mve_vmaxq_m_<supf><mode>): Likewise.
19729 (mve_vminq_m_<supf><mode>): Likewise.
19730 (mve_vmulhq_m_<supf><mode>): Likewise.
19731 (mve_vqaddq_m_<supf><mode>): Likewise.
19732 (mve_vqrshlq_m_<supf><mode>): Likewise.
19733 (mve_vqshlq_m_<supf><mode>): Likewise.
19734 (mve_vqsubq_m_<supf><mode>): Likewise.
19735 (mve_vrhaddq_m_<supf><mode>): Likewise.
19736 (mve_vrmulhq_m_<supf><mode>): Likewise.
19737 (mve_vrshlq_m_<supf><mode>): Likewise.
19738 (mve_vqdmladhq_m_s<mode>): Likewise.
19739 (mve_vqdmladhxq_m_s<mode>): Likewise.
19740 (mve_vqdmlsdhq_m_s<mode>): Likewise.
19741 (mve_vqdmlsdhxq_m_s<mode>): Likewise.
19742 (mve_vqdmulhq_m_s<mode>): Likewise.
19743 (mve_vqrdmladhq_m_s<mode>): Likewise.
19744 (mve_vqrdmladhxq_m_s<mode>): Likewise.
19745 (mve_vqrdmlsdhq_m_s<mode>): Likewise.
19746 (mve_vqrdmlsdhxq_m_s<mode>): Likewise.
19747 (mve_vqrdmulhq_m_s<mode>): Likewise.
19749 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
19751 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITHOUT_M_N): New. (vcreateq): New.
19752 * config/arm/arm-mve-builtins-base.def (vcreateq): New.
19753 * config/arm/arm-mve-builtins-base.h (vcreateq): New.
19754 * config/arm/arm_mve.h (vcreateq_f16): Remove.
19755 (vcreateq_f32): Remove.
19756 (vcreateq_u8): Remove.
19757 (vcreateq_u16): Remove.
19758 (vcreateq_u32): Remove.
19759 (vcreateq_u64): Remove.
19760 (vcreateq_s8): Remove.
19761 (vcreateq_s16): Remove.
19762 (vcreateq_s32): Remove.
19763 (vcreateq_s64): Remove.
19764 (__arm_vcreateq_u8): Remove.
19765 (__arm_vcreateq_u16): Remove.
19766 (__arm_vcreateq_u32): Remove.
19767 (__arm_vcreateq_u64): Remove.
19768 (__arm_vcreateq_s8): Remove.
19769 (__arm_vcreateq_s16): Remove.
19770 (__arm_vcreateq_s32): Remove.
19771 (__arm_vcreateq_s64): Remove.
19772 (__arm_vcreateq_f16): Remove.
19773 (__arm_vcreateq_f32): Remove.
19775 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
19777 * config/arm/iterators.md (MVE_FP_CREATE_ONLY): New.
19778 (mve_insn): Add VCREATEQ_S, VCREATEQ_U, VCREATEQ_F.
19779 * config/arm/mve.md (mve_vcreateq_f<mode>): Rename into ...
19780 (@mve_<mve_insn>q_f<mode>): ... this.
19781 (mve_vcreateq_<supf><mode>): Rename into ...
19782 (@mve_<mve_insn>q_<supf><mode>): ... this.
19784 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
19786 * config/arm/arm-mve-builtins-shapes.cc (create): New.
19787 * config/arm/arm-mve-builtins-shapes.h: (create): New.
19789 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
19791 * config/arm/arm-mve-builtins-functions.h (class
19792 unspec_mve_function_exact_insn): New.
19794 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
19796 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITH_RTX_M_N_NO_N_F): New.
19798 * config/arm/arm-mve-builtins-base.def (vorrq): New.
19799 * config/arm/arm-mve-builtins-base.h (vorrq): New.
19800 * config/arm/arm-mve-builtins.cc
19801 (function_instance::has_inactive_argument): Handle vorrq.
19802 * config/arm/arm_mve.h (vorrq): Remove.
19803 (vorrq_m_n): Remove.
19806 (vorrq_u8): Remove.
19807 (vorrq_s8): Remove.
19808 (vorrq_u16): Remove.
19809 (vorrq_s16): Remove.
19810 (vorrq_u32): Remove.
19811 (vorrq_s32): Remove.
19812 (vorrq_n_u16): Remove.
19813 (vorrq_f16): Remove.
19814 (vorrq_n_s16): Remove.
19815 (vorrq_n_u32): Remove.
19816 (vorrq_f32): Remove.
19817 (vorrq_n_s32): Remove.
19818 (vorrq_m_n_s16): Remove.
19819 (vorrq_m_n_u16): Remove.
19820 (vorrq_m_n_s32): Remove.
19821 (vorrq_m_n_u32): Remove.
19822 (vorrq_m_s8): Remove.
19823 (vorrq_m_s32): Remove.
19824 (vorrq_m_s16): Remove.
19825 (vorrq_m_u8): Remove.
19826 (vorrq_m_u32): Remove.
19827 (vorrq_m_u16): Remove.
19828 (vorrq_m_f32): Remove.
19829 (vorrq_m_f16): Remove.
19830 (vorrq_x_s8): Remove.
19831 (vorrq_x_s16): Remove.
19832 (vorrq_x_s32): Remove.
19833 (vorrq_x_u8): Remove.
19834 (vorrq_x_u16): Remove.
19835 (vorrq_x_u32): Remove.
19836 (vorrq_x_f16): Remove.
19837 (vorrq_x_f32): Remove.
19838 (__arm_vorrq_u8): Remove.
19839 (__arm_vorrq_s8): Remove.
19840 (__arm_vorrq_u16): Remove.
19841 (__arm_vorrq_s16): Remove.
19842 (__arm_vorrq_u32): Remove.
19843 (__arm_vorrq_s32): Remove.
19844 (__arm_vorrq_n_u16): Remove.
19845 (__arm_vorrq_n_s16): Remove.
19846 (__arm_vorrq_n_u32): Remove.
19847 (__arm_vorrq_n_s32): Remove.
19848 (__arm_vorrq_m_n_s16): Remove.
19849 (__arm_vorrq_m_n_u16): Remove.
19850 (__arm_vorrq_m_n_s32): Remove.
19851 (__arm_vorrq_m_n_u32): Remove.
19852 (__arm_vorrq_m_s8): Remove.
19853 (__arm_vorrq_m_s32): Remove.
19854 (__arm_vorrq_m_s16): Remove.
19855 (__arm_vorrq_m_u8): Remove.
19856 (__arm_vorrq_m_u32): Remove.
19857 (__arm_vorrq_m_u16): Remove.
19858 (__arm_vorrq_x_s8): Remove.
19859 (__arm_vorrq_x_s16): Remove.
19860 (__arm_vorrq_x_s32): Remove.
19861 (__arm_vorrq_x_u8): Remove.
19862 (__arm_vorrq_x_u16): Remove.
19863 (__arm_vorrq_x_u32): Remove.
19864 (__arm_vorrq_f16): Remove.
19865 (__arm_vorrq_f32): Remove.
19866 (__arm_vorrq_m_f32): Remove.
19867 (__arm_vorrq_m_f16): Remove.
19868 (__arm_vorrq_x_f16): Remove.
19869 (__arm_vorrq_x_f32): Remove.
19870 (__arm_vorrq): Remove.
19871 (__arm_vorrq_m_n): Remove.
19872 (__arm_vorrq_m): Remove.
19873 (__arm_vorrq_x): Remove.
19875 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
19877 * config/arm/arm-mve-builtins-shapes.cc (binary_orrq): New.
19878 * config/arm/arm-mve-builtins-shapes.h (binary_orrq): New.
19879 * config/arm/arm-mve-builtins.cc (preds_m_or_none): Remove static.
19880 * config/arm/arm-mve-builtins.h (preds_m_or_none): Declare.
19882 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
19884 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITH_RTX_M): New.
19885 (vandq,veorq): New.
19886 * config/arm/arm-mve-builtins-base.def (vandq, veorq): New.
19887 * config/arm/arm-mve-builtins-base.h (vandq, veorq): New.
19888 * config/arm/arm_mve.h (vandq): Remove.
19891 (vandq_u8): Remove.
19892 (vandq_s8): Remove.
19893 (vandq_u16): Remove.
19894 (vandq_s16): Remove.
19895 (vandq_u32): Remove.
19896 (vandq_s32): Remove.
19897 (vandq_f16): Remove.
19898 (vandq_f32): Remove.
19899 (vandq_m_s8): Remove.
19900 (vandq_m_s32): Remove.
19901 (vandq_m_s16): Remove.
19902 (vandq_m_u8): Remove.
19903 (vandq_m_u32): Remove.
19904 (vandq_m_u16): Remove.
19905 (vandq_m_f32): Remove.
19906 (vandq_m_f16): Remove.
19907 (vandq_x_s8): Remove.
19908 (vandq_x_s16): Remove.
19909 (vandq_x_s32): Remove.
19910 (vandq_x_u8): Remove.
19911 (vandq_x_u16): Remove.
19912 (vandq_x_u32): Remove.
19913 (vandq_x_f16): Remove.
19914 (vandq_x_f32): Remove.
19915 (__arm_vandq_u8): Remove.
19916 (__arm_vandq_s8): Remove.
19917 (__arm_vandq_u16): Remove.
19918 (__arm_vandq_s16): Remove.
19919 (__arm_vandq_u32): Remove.
19920 (__arm_vandq_s32): Remove.
19921 (__arm_vandq_m_s8): Remove.
19922 (__arm_vandq_m_s32): Remove.
19923 (__arm_vandq_m_s16): Remove.
19924 (__arm_vandq_m_u8): Remove.
19925 (__arm_vandq_m_u32): Remove.
19926 (__arm_vandq_m_u16): Remove.
19927 (__arm_vandq_x_s8): Remove.
19928 (__arm_vandq_x_s16): Remove.
19929 (__arm_vandq_x_s32): Remove.
19930 (__arm_vandq_x_u8): Remove.
19931 (__arm_vandq_x_u16): Remove.
19932 (__arm_vandq_x_u32): Remove.
19933 (__arm_vandq_f16): Remove.
19934 (__arm_vandq_f32): Remove.
19935 (__arm_vandq_m_f32): Remove.
19936 (__arm_vandq_m_f16): Remove.
19937 (__arm_vandq_x_f16): Remove.
19938 (__arm_vandq_x_f32): Remove.
19939 (__arm_vandq): Remove.
19940 (__arm_vandq_m): Remove.
19941 (__arm_vandq_x): Remove.
19944 (veorq_u8): Remove.
19945 (veorq_s8): Remove.
19946 (veorq_u16): Remove.
19947 (veorq_s16): Remove.
19948 (veorq_u32): Remove.
19949 (veorq_s32): Remove.
19950 (veorq_f16): Remove.
19951 (veorq_f32): Remove.
19952 (veorq_m_s8): Remove.
19953 (veorq_m_s32): Remove.
19954 (veorq_m_s16): Remove.
19955 (veorq_m_u8): Remove.
19956 (veorq_m_u32): Remove.
19957 (veorq_m_u16): Remove.
19958 (veorq_m_f32): Remove.
19959 (veorq_m_f16): Remove.
19960 (veorq_x_s8): Remove.
19961 (veorq_x_s16): Remove.
19962 (veorq_x_s32): Remove.
19963 (veorq_x_u8): Remove.
19964 (veorq_x_u16): Remove.
19965 (veorq_x_u32): Remove.
19966 (veorq_x_f16): Remove.
19967 (veorq_x_f32): Remove.
19968 (__arm_veorq_u8): Remove.
19969 (__arm_veorq_s8): Remove.
19970 (__arm_veorq_u16): Remove.
19971 (__arm_veorq_s16): Remove.
19972 (__arm_veorq_u32): Remove.
19973 (__arm_veorq_s32): Remove.
19974 (__arm_veorq_m_s8): Remove.
19975 (__arm_veorq_m_s32): Remove.
19976 (__arm_veorq_m_s16): Remove.
19977 (__arm_veorq_m_u8): Remove.
19978 (__arm_veorq_m_u32): Remove.
19979 (__arm_veorq_m_u16): Remove.
19980 (__arm_veorq_x_s8): Remove.
19981 (__arm_veorq_x_s16): Remove.
19982 (__arm_veorq_x_s32): Remove.
19983 (__arm_veorq_x_u8): Remove.
19984 (__arm_veorq_x_u16): Remove.
19985 (__arm_veorq_x_u32): Remove.
19986 (__arm_veorq_f16): Remove.
19987 (__arm_veorq_f32): Remove.
19988 (__arm_veorq_m_f32): Remove.
19989 (__arm_veorq_m_f16): Remove.
19990 (__arm_veorq_x_f16): Remove.
19991 (__arm_veorq_x_f32): Remove.
19992 (__arm_veorq): Remove.
19993 (__arm_veorq_m): Remove.
19994 (__arm_veorq_x): Remove.
19996 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
19998 * config/arm/iterators.md (MVE_INT_M_BINARY_LOGIC)
19999 (MVE_FP_M_BINARY_LOGIC): New.
20000 (MVE_INT_M_N_BINARY_LOGIC): New.
20001 (MVE_INT_N_BINARY_LOGIC): New.
20002 (mve_insn): Add vand, veor, vorr, vbic.
20003 * config/arm/mve.md (mve_vandq_m_<supf><mode>)
20004 (mve_veorq_m_<supf><mode>, mve_vorrq_m_<supf><mode>)
20005 (mve_vbicq_m_<supf><mode>): Merge into ...
20006 (@mve_<mve_insn>q_m_<supf><mode>): ... this.
20007 (mve_vandq_m_f<mode>, mve_veorq_m_f<mode>, mve_vorrq_m_f<mode>)
20008 (mve_vbicq_m_f<mode>): Merge into ...
20009 (@mve_<mve_insn>q_m_f<mode>): ... this.
20010 (mve_vorrq_n_<supf><mode>)
20011 (mve_vbicq_n_<supf><mode>): Merge into ...
20012 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
20013 (mve_vorrq_m_n_<supf><mode>, mve_vbicq_m_n_<supf><mode>): Merge
20015 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
20017 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20019 * config/arm/arm-mve-builtins-shapes.cc (binary): New.
20020 * config/arm/arm-mve-builtins-shapes.h (binary): New.
20022 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20024 * config/arm/arm-mve-builtins-base.cc (FUNCTION_WITH_RTX_M_N):
20026 (vaddq, vmulq, vsubq): New.
20027 * config/arm/arm-mve-builtins-base.def (vaddq, vmulq, vsubq): New.
20028 * config/arm/arm-mve-builtins-base.h (vaddq, vmulq, vsubq): New.
20029 * config/arm/arm_mve.h (vaddq): Remove.
20032 (vaddq_n_u8): Remove.
20033 (vaddq_n_s8): Remove.
20034 (vaddq_n_u16): Remove.
20035 (vaddq_n_s16): Remove.
20036 (vaddq_n_u32): Remove.
20037 (vaddq_n_s32): Remove.
20038 (vaddq_n_f16): Remove.
20039 (vaddq_n_f32): Remove.
20040 (vaddq_m_n_s8): Remove.
20041 (vaddq_m_n_s32): Remove.
20042 (vaddq_m_n_s16): Remove.
20043 (vaddq_m_n_u8): Remove.
20044 (vaddq_m_n_u32): Remove.
20045 (vaddq_m_n_u16): Remove.
20046 (vaddq_m_s8): Remove.
20047 (vaddq_m_s32): Remove.
20048 (vaddq_m_s16): Remove.
20049 (vaddq_m_u8): Remove.
20050 (vaddq_m_u32): Remove.
20051 (vaddq_m_u16): Remove.
20052 (vaddq_m_f32): Remove.
20053 (vaddq_m_f16): Remove.
20054 (vaddq_m_n_f32): Remove.
20055 (vaddq_m_n_f16): Remove.
20056 (vaddq_s8): Remove.
20057 (vaddq_s16): Remove.
20058 (vaddq_s32): Remove.
20059 (vaddq_u8): Remove.
20060 (vaddq_u16): Remove.
20061 (vaddq_u32): Remove.
20062 (vaddq_f16): Remove.
20063 (vaddq_f32): Remove.
20064 (vaddq_x_s8): Remove.
20065 (vaddq_x_s16): Remove.
20066 (vaddq_x_s32): Remove.
20067 (vaddq_x_n_s8): Remove.
20068 (vaddq_x_n_s16): Remove.
20069 (vaddq_x_n_s32): Remove.
20070 (vaddq_x_u8): Remove.
20071 (vaddq_x_u16): Remove.
20072 (vaddq_x_u32): Remove.
20073 (vaddq_x_n_u8): Remove.
20074 (vaddq_x_n_u16): Remove.
20075 (vaddq_x_n_u32): Remove.
20076 (vaddq_x_f16): Remove.
20077 (vaddq_x_f32): Remove.
20078 (vaddq_x_n_f16): Remove.
20079 (vaddq_x_n_f32): Remove.
20080 (__arm_vaddq_n_u8): Remove.
20081 (__arm_vaddq_n_s8): Remove.
20082 (__arm_vaddq_n_u16): Remove.
20083 (__arm_vaddq_n_s16): Remove.
20084 (__arm_vaddq_n_u32): Remove.
20085 (__arm_vaddq_n_s32): Remove.
20086 (__arm_vaddq_m_n_s8): Remove.
20087 (__arm_vaddq_m_n_s32): Remove.
20088 (__arm_vaddq_m_n_s16): Remove.
20089 (__arm_vaddq_m_n_u8): Remove.
20090 (__arm_vaddq_m_n_u32): Remove.
20091 (__arm_vaddq_m_n_u16): Remove.
20092 (__arm_vaddq_m_s8): Remove.
20093 (__arm_vaddq_m_s32): Remove.
20094 (__arm_vaddq_m_s16): Remove.
20095 (__arm_vaddq_m_u8): Remove.
20096 (__arm_vaddq_m_u32): Remove.
20097 (__arm_vaddq_m_u16): Remove.
20098 (__arm_vaddq_s8): Remove.
20099 (__arm_vaddq_s16): Remove.
20100 (__arm_vaddq_s32): Remove.
20101 (__arm_vaddq_u8): Remove.
20102 (__arm_vaddq_u16): Remove.
20103 (__arm_vaddq_u32): Remove.
20104 (__arm_vaddq_x_s8): Remove.
20105 (__arm_vaddq_x_s16): Remove.
20106 (__arm_vaddq_x_s32): Remove.
20107 (__arm_vaddq_x_n_s8): Remove.
20108 (__arm_vaddq_x_n_s16): Remove.
20109 (__arm_vaddq_x_n_s32): Remove.
20110 (__arm_vaddq_x_u8): Remove.
20111 (__arm_vaddq_x_u16): Remove.
20112 (__arm_vaddq_x_u32): Remove.
20113 (__arm_vaddq_x_n_u8): Remove.
20114 (__arm_vaddq_x_n_u16): Remove.
20115 (__arm_vaddq_x_n_u32): Remove.
20116 (__arm_vaddq_n_f16): Remove.
20117 (__arm_vaddq_n_f32): Remove.
20118 (__arm_vaddq_m_f32): Remove.
20119 (__arm_vaddq_m_f16): Remove.
20120 (__arm_vaddq_m_n_f32): Remove.
20121 (__arm_vaddq_m_n_f16): Remove.
20122 (__arm_vaddq_f16): Remove.
20123 (__arm_vaddq_f32): Remove.
20124 (__arm_vaddq_x_f16): Remove.
20125 (__arm_vaddq_x_f32): Remove.
20126 (__arm_vaddq_x_n_f16): Remove.
20127 (__arm_vaddq_x_n_f32): Remove.
20128 (__arm_vaddq): Remove.
20129 (__arm_vaddq_m): Remove.
20130 (__arm_vaddq_x): Remove.
20134 (vmulq_u8): Remove.
20135 (vmulq_n_u8): Remove.
20136 (vmulq_s8): Remove.
20137 (vmulq_n_s8): Remove.
20138 (vmulq_u16): Remove.
20139 (vmulq_n_u16): Remove.
20140 (vmulq_s16): Remove.
20141 (vmulq_n_s16): Remove.
20142 (vmulq_u32): Remove.
20143 (vmulq_n_u32): Remove.
20144 (vmulq_s32): Remove.
20145 (vmulq_n_s32): Remove.
20146 (vmulq_n_f16): Remove.
20147 (vmulq_f16): Remove.
20148 (vmulq_n_f32): Remove.
20149 (vmulq_f32): Remove.
20150 (vmulq_m_n_s8): Remove.
20151 (vmulq_m_n_s32): Remove.
20152 (vmulq_m_n_s16): Remove.
20153 (vmulq_m_n_u8): Remove.
20154 (vmulq_m_n_u32): Remove.
20155 (vmulq_m_n_u16): Remove.
20156 (vmulq_m_s8): Remove.
20157 (vmulq_m_s32): Remove.
20158 (vmulq_m_s16): Remove.
20159 (vmulq_m_u8): Remove.
20160 (vmulq_m_u32): Remove.
20161 (vmulq_m_u16): Remove.
20162 (vmulq_m_f32): Remove.
20163 (vmulq_m_f16): Remove.
20164 (vmulq_m_n_f32): Remove.
20165 (vmulq_m_n_f16): Remove.
20166 (vmulq_x_s8): Remove.
20167 (vmulq_x_s16): Remove.
20168 (vmulq_x_s32): Remove.
20169 (vmulq_x_n_s8): Remove.
20170 (vmulq_x_n_s16): Remove.
20171 (vmulq_x_n_s32): Remove.
20172 (vmulq_x_u8): Remove.
20173 (vmulq_x_u16): Remove.
20174 (vmulq_x_u32): Remove.
20175 (vmulq_x_n_u8): Remove.
20176 (vmulq_x_n_u16): Remove.
20177 (vmulq_x_n_u32): Remove.
20178 (vmulq_x_f16): Remove.
20179 (vmulq_x_f32): Remove.
20180 (vmulq_x_n_f16): Remove.
20181 (vmulq_x_n_f32): Remove.
20182 (__arm_vmulq_u8): Remove.
20183 (__arm_vmulq_n_u8): Remove.
20184 (__arm_vmulq_s8): Remove.
20185 (__arm_vmulq_n_s8): Remove.
20186 (__arm_vmulq_u16): Remove.
20187 (__arm_vmulq_n_u16): Remove.
20188 (__arm_vmulq_s16): Remove.
20189 (__arm_vmulq_n_s16): Remove.
20190 (__arm_vmulq_u32): Remove.
20191 (__arm_vmulq_n_u32): Remove.
20192 (__arm_vmulq_s32): Remove.
20193 (__arm_vmulq_n_s32): Remove.
20194 (__arm_vmulq_m_n_s8): Remove.
20195 (__arm_vmulq_m_n_s32): Remove.
20196 (__arm_vmulq_m_n_s16): Remove.
20197 (__arm_vmulq_m_n_u8): Remove.
20198 (__arm_vmulq_m_n_u32): Remove.
20199 (__arm_vmulq_m_n_u16): Remove.
20200 (__arm_vmulq_m_s8): Remove.
20201 (__arm_vmulq_m_s32): Remove.
20202 (__arm_vmulq_m_s16): Remove.
20203 (__arm_vmulq_m_u8): Remove.
20204 (__arm_vmulq_m_u32): Remove.
20205 (__arm_vmulq_m_u16): Remove.
20206 (__arm_vmulq_x_s8): Remove.
20207 (__arm_vmulq_x_s16): Remove.
20208 (__arm_vmulq_x_s32): Remove.
20209 (__arm_vmulq_x_n_s8): Remove.
20210 (__arm_vmulq_x_n_s16): Remove.
20211 (__arm_vmulq_x_n_s32): Remove.
20212 (__arm_vmulq_x_u8): Remove.
20213 (__arm_vmulq_x_u16): Remove.
20214 (__arm_vmulq_x_u32): Remove.
20215 (__arm_vmulq_x_n_u8): Remove.
20216 (__arm_vmulq_x_n_u16): Remove.
20217 (__arm_vmulq_x_n_u32): Remove.
20218 (__arm_vmulq_n_f16): Remove.
20219 (__arm_vmulq_f16): Remove.
20220 (__arm_vmulq_n_f32): Remove.
20221 (__arm_vmulq_f32): Remove.
20222 (__arm_vmulq_m_f32): Remove.
20223 (__arm_vmulq_m_f16): Remove.
20224 (__arm_vmulq_m_n_f32): Remove.
20225 (__arm_vmulq_m_n_f16): Remove.
20226 (__arm_vmulq_x_f16): Remove.
20227 (__arm_vmulq_x_f32): Remove.
20228 (__arm_vmulq_x_n_f16): Remove.
20229 (__arm_vmulq_x_n_f32): Remove.
20230 (__arm_vmulq): Remove.
20231 (__arm_vmulq_m): Remove.
20232 (__arm_vmulq_x): Remove.
20236 (vsubq_n_f16): Remove.
20237 (vsubq_n_f32): Remove.
20238 (vsubq_u8): Remove.
20239 (vsubq_n_u8): Remove.
20240 (vsubq_s8): Remove.
20241 (vsubq_n_s8): Remove.
20242 (vsubq_u16): Remove.
20243 (vsubq_n_u16): Remove.
20244 (vsubq_s16): Remove.
20245 (vsubq_n_s16): Remove.
20246 (vsubq_u32): Remove.
20247 (vsubq_n_u32): Remove.
20248 (vsubq_s32): Remove.
20249 (vsubq_n_s32): Remove.
20250 (vsubq_f16): Remove.
20251 (vsubq_f32): Remove.
20252 (vsubq_m_s8): Remove.
20253 (vsubq_m_u8): Remove.
20254 (vsubq_m_s16): Remove.
20255 (vsubq_m_u16): Remove.
20256 (vsubq_m_s32): Remove.
20257 (vsubq_m_u32): Remove.
20258 (vsubq_m_n_s8): Remove.
20259 (vsubq_m_n_s32): Remove.
20260 (vsubq_m_n_s16): Remove.
20261 (vsubq_m_n_u8): Remove.
20262 (vsubq_m_n_u32): Remove.
20263 (vsubq_m_n_u16): Remove.
20264 (vsubq_m_f32): Remove.
20265 (vsubq_m_f16): Remove.
20266 (vsubq_m_n_f32): Remove.
20267 (vsubq_m_n_f16): Remove.
20268 (vsubq_x_s8): Remove.
20269 (vsubq_x_s16): Remove.
20270 (vsubq_x_s32): Remove.
20271 (vsubq_x_n_s8): Remove.
20272 (vsubq_x_n_s16): Remove.
20273 (vsubq_x_n_s32): Remove.
20274 (vsubq_x_u8): Remove.
20275 (vsubq_x_u16): Remove.
20276 (vsubq_x_u32): Remove.
20277 (vsubq_x_n_u8): Remove.
20278 (vsubq_x_n_u16): Remove.
20279 (vsubq_x_n_u32): Remove.
20280 (vsubq_x_f16): Remove.
20281 (vsubq_x_f32): Remove.
20282 (vsubq_x_n_f16): Remove.
20283 (vsubq_x_n_f32): Remove.
20284 (__arm_vsubq_u8): Remove.
20285 (__arm_vsubq_n_u8): Remove.
20286 (__arm_vsubq_s8): Remove.
20287 (__arm_vsubq_n_s8): Remove.
20288 (__arm_vsubq_u16): Remove.
20289 (__arm_vsubq_n_u16): Remove.
20290 (__arm_vsubq_s16): Remove.
20291 (__arm_vsubq_n_s16): Remove.
20292 (__arm_vsubq_u32): Remove.
20293 (__arm_vsubq_n_u32): Remove.
20294 (__arm_vsubq_s32): Remove.
20295 (__arm_vsubq_n_s32): Remove.
20296 (__arm_vsubq_m_s8): Remove.
20297 (__arm_vsubq_m_u8): Remove.
20298 (__arm_vsubq_m_s16): Remove.
20299 (__arm_vsubq_m_u16): Remove.
20300 (__arm_vsubq_m_s32): Remove.
20301 (__arm_vsubq_m_u32): Remove.
20302 (__arm_vsubq_m_n_s8): Remove.
20303 (__arm_vsubq_m_n_s32): Remove.
20304 (__arm_vsubq_m_n_s16): Remove.
20305 (__arm_vsubq_m_n_u8): Remove.
20306 (__arm_vsubq_m_n_u32): Remove.
20307 (__arm_vsubq_m_n_u16): Remove.
20308 (__arm_vsubq_x_s8): Remove.
20309 (__arm_vsubq_x_s16): Remove.
20310 (__arm_vsubq_x_s32): Remove.
20311 (__arm_vsubq_x_n_s8): Remove.
20312 (__arm_vsubq_x_n_s16): Remove.
20313 (__arm_vsubq_x_n_s32): Remove.
20314 (__arm_vsubq_x_u8): Remove.
20315 (__arm_vsubq_x_u16): Remove.
20316 (__arm_vsubq_x_u32): Remove.
20317 (__arm_vsubq_x_n_u8): Remove.
20318 (__arm_vsubq_x_n_u16): Remove.
20319 (__arm_vsubq_x_n_u32): Remove.
20320 (__arm_vsubq_n_f16): Remove.
20321 (__arm_vsubq_n_f32): Remove.
20322 (__arm_vsubq_f16): Remove.
20323 (__arm_vsubq_f32): Remove.
20324 (__arm_vsubq_m_f32): Remove.
20325 (__arm_vsubq_m_f16): Remove.
20326 (__arm_vsubq_m_n_f32): Remove.
20327 (__arm_vsubq_m_n_f16): Remove.
20328 (__arm_vsubq_x_f16): Remove.
20329 (__arm_vsubq_x_f32): Remove.
20330 (__arm_vsubq_x_n_f16): Remove.
20331 (__arm_vsubq_x_n_f32): Remove.
20332 (__arm_vsubq): Remove.
20333 (__arm_vsubq_m): Remove.
20334 (__arm_vsubq_x): Remove.
20335 * config/arm/arm_mve_builtins.def (vsubq_u, vsubq_s, vsubq_f):
20337 (vmulq_u, vmulq_s, vmulq_f): Remove.
20338 * config/arm/mve.md (mve_vsubq_<supf><mode>): Remove.
20339 (mve_vmulq_<supf><mode>): Remove.
20341 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20343 * config/arm/iterators.md (MVE_INT_BINARY_RTX, MVE_INT_M_BINARY)
20344 (MVE_INT_M_N_BINARY, MVE_INT_N_BINARY, MVE_FP_M_BINARY)
20345 (MVE_FP_M_N_BINARY, MVE_FP_N_BINARY, mve_addsubmul, mve_insn): New
20347 * config/arm/mve.md
20348 (mve_vsubq_n_f<mode>, mve_vaddq_n_f<mode>, mve_vmulq_n_f<mode>):
20350 (@mve_<mve_insn>q_n_f<mode>): ... this.
20351 (mve_vaddq_n_<supf><mode>, mve_vmulq_n_<supf><mode>)
20352 (mve_vsubq_n_<supf><mode>): Factorize into ...
20353 (@mve_<mve_insn>q_n_<supf><mode>): ... this.
20354 (mve_vaddq<mode>, mve_vmulq<mode>, mve_vsubq<mode>): Factorize
20356 (mve_<mve_addsubmul>q<mode>): ... this.
20357 (mve_vaddq_f<mode>, mve_vmulq_f<mode>, mve_vsubq_f<mode>):
20359 (mve_<mve_addsubmul>q_f<mode>): ... this.
20360 (mve_vaddq_m_<supf><mode>, mve_vmulq_m_<supf><mode>)
20361 (mve_vsubq_m_<supf><mode>): Factorize into ...
20362 (@mve_<mve_insn>q_m_<supf><mode>): ... this,
20363 (mve_vaddq_m_n_<supf><mode>, mve_vmulq_m_n_<supf><mode>)
20364 (mve_vsubq_m_n_<supf><mode>): Factorize into ...
20365 (@mve_<mve_insn>q_m_n_<supf><mode>): ... this.
20366 (mve_vaddq_m_f<mode>, mve_vmulq_m_f<mode>, mve_vsubq_m_f<mode>):
20368 (@mve_<mve_insn>q_m_f<mode>): ... this.
20369 (mve_vaddq_m_n_f<mode>, mve_vmulq_m_n_f<mode>)
20370 (mve_vsubq_m_n_f<mode>): Factorize into ...
20371 (@mve_<mve_insn>q_m_n_f<mode>): ... this.
20373 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20375 * config/arm/arm-mve-builtins-functions.h (class
20376 unspec_based_mve_function_base): New.
20377 (class unspec_based_mve_function_exact_insn): New.
20379 2023-05-03 Christophe Lyon <christophe.lyon@arm.com>
20381 * config/arm/arm-mve-builtins-shapes.cc (binary_opt_n): New.
20382 * config/arm/arm-mve-builtins-shapes.h (binary_opt_n): New.
20384 2023-05-03 Murray Steele <murray.steele@arm.com>
20385 Christophe Lyon <christophe.lyon@arm.com>
20387 * config/arm/arm-mve-builtins-base.cc (class
20388 vuninitializedq_impl): New.
20389 * config/arm/arm-mve-builtins-base.def (vuninitializedq): New.
20390 * config/arm/arm-mve-builtins-base.h (vuninitializedq): New
20392 * config/arm/arm-mve-builtins-shapes.cc (inherent): New.
20393 * config/arm/arm-mve-builtins-shapes.h (inherent): New
20395 * config/arm/arm_mve_types.h (__arm_vuninitializedq): Move to ...
20396 * config/arm/arm_mve.h (__arm_vuninitializedq): ... here.
20397 (__arm_vuninitializedq_u8): Remove.
20398 (__arm_vuninitializedq_u16): Remove.
20399 (__arm_vuninitializedq_u32): Remove.
20400 (__arm_vuninitializedq_u64): Remove.
20401 (__arm_vuninitializedq_s8): Remove.
20402 (__arm_vuninitializedq_s16): Remove.
20403 (__arm_vuninitializedq_s32): Remove.
20404 (__arm_vuninitializedq_s64): Remove.
20405 (__arm_vuninitializedq_f16): Remove.
20406 (__arm_vuninitializedq_f32): Remove.
20408 2023-05-03 Murray Steele <murray.steele@arm.com>
20409 Christophe Lyon <christophe.lyon@arm.com>
20411 * config/arm/arm-mve-builtins-base.cc (vreinterpretq_impl): New class.
20412 * config/arm/arm-mve-builtins-base.def: Define vreinterpretq.
20413 * config/arm/arm-mve-builtins-base.h (vreinterpretq): New declaration.
20414 * config/arm/arm-mve-builtins-shapes.cc (parse_element_type): New function.
20415 (parse_type): Likewise.
20416 (parse_signature): Likewise.
20417 (build_one): Likewise.
20418 (build_all): Likewise.
20419 (overloaded_base): New struct.
20420 (unary_convert_def): Likewise.
20421 * config/arm/arm-mve-builtins-shapes.h (unary_convert): Declare.
20422 * config/arm/arm-mve-builtins.cc (TYPES_reinterpret_signed1): New
20424 (TYPES_reinterpret_unsigned1): Likewise.
20425 (TYPES_reinterpret_integer): Likewise.
20426 (TYPES_reinterpret_integer1): Likewise.
20427 (TYPES_reinterpret_float1): Likewise.
20428 (TYPES_reinterpret_float): Likewise.
20429 (reinterpret_integer): New.
20430 (reinterpret_float): New.
20431 (handle_arm_mve_h): Register builtins.
20432 * config/arm/arm_mve.h (vreinterpretq_s16): Remove.
20433 (vreinterpretq_s32): Likewise.
20434 (vreinterpretq_s64): Likewise.
20435 (vreinterpretq_s8): Likewise.
20436 (vreinterpretq_u16): Likewise.
20437 (vreinterpretq_u32): Likewise.
20438 (vreinterpretq_u64): Likewise.
20439 (vreinterpretq_u8): Likewise.
20440 (vreinterpretq_f16): Likewise.
20441 (vreinterpretq_f32): Likewise.
20442 (vreinterpretq_s16_s32): Likewise.
20443 (vreinterpretq_s16_s64): Likewise.
20444 (vreinterpretq_s16_s8): Likewise.
20445 (vreinterpretq_s16_u16): Likewise.
20446 (vreinterpretq_s16_u32): Likewise.
20447 (vreinterpretq_s16_u64): Likewise.
20448 (vreinterpretq_s16_u8): Likewise.
20449 (vreinterpretq_s32_s16): Likewise.
20450 (vreinterpretq_s32_s64): Likewise.
20451 (vreinterpretq_s32_s8): Likewise.
20452 (vreinterpretq_s32_u16): Likewise.
20453 (vreinterpretq_s32_u32): Likewise.
20454 (vreinterpretq_s32_u64): Likewise.
20455 (vreinterpretq_s32_u8): Likewise.
20456 (vreinterpretq_s64_s16): Likewise.
20457 (vreinterpretq_s64_s32): Likewise.
20458 (vreinterpretq_s64_s8): Likewise.
20459 (vreinterpretq_s64_u16): Likewise.
20460 (vreinterpretq_s64_u32): Likewise.
20461 (vreinterpretq_s64_u64): Likewise.
20462 (vreinterpretq_s64_u8): Likewise.
20463 (vreinterpretq_s8_s16): Likewise.
20464 (vreinterpretq_s8_s32): Likewise.
20465 (vreinterpretq_s8_s64): Likewise.
20466 (vreinterpretq_s8_u16): Likewise.
20467 (vreinterpretq_s8_u32): Likewise.
20468 (vreinterpretq_s8_u64): Likewise.
20469 (vreinterpretq_s8_u8): Likewise.
20470 (vreinterpretq_u16_s16): Likewise.
20471 (vreinterpretq_u16_s32): Likewise.
20472 (vreinterpretq_u16_s64): Likewise.
20473 (vreinterpretq_u16_s8): Likewise.
20474 (vreinterpretq_u16_u32): Likewise.
20475 (vreinterpretq_u16_u64): Likewise.
20476 (vreinterpretq_u16_u8): Likewise.
20477 (vreinterpretq_u32_s16): Likewise.
20478 (vreinterpretq_u32_s32): Likewise.
20479 (vreinterpretq_u32_s64): Likewise.
20480 (vreinterpretq_u32_s8): Likewise.
20481 (vreinterpretq_u32_u16): Likewise.
20482 (vreinterpretq_u32_u64): Likewise.
20483 (vreinterpretq_u32_u8): Likewise.
20484 (vreinterpretq_u64_s16): Likewise.
20485 (vreinterpretq_u64_s32): Likewise.
20486 (vreinterpretq_u64_s64): Likewise.
20487 (vreinterpretq_u64_s8): Likewise.
20488 (vreinterpretq_u64_u16): Likewise.
20489 (vreinterpretq_u64_u32): Likewise.
20490 (vreinterpretq_u64_u8): Likewise.
20491 (vreinterpretq_u8_s16): Likewise.
20492 (vreinterpretq_u8_s32): Likewise.
20493 (vreinterpretq_u8_s64): Likewise.
20494 (vreinterpretq_u8_s8): Likewise.
20495 (vreinterpretq_u8_u16): Likewise.
20496 (vreinterpretq_u8_u32): Likewise.
20497 (vreinterpretq_u8_u64): Likewise.
20498 (vreinterpretq_s32_f16): Likewise.
20499 (vreinterpretq_s32_f32): Likewise.
20500 (vreinterpretq_u16_f16): Likewise.
20501 (vreinterpretq_u16_f32): Likewise.
20502 (vreinterpretq_u32_f16): Likewise.
20503 (vreinterpretq_u32_f32): Likewise.
20504 (vreinterpretq_u64_f16): Likewise.
20505 (vreinterpretq_u64_f32): Likewise.
20506 (vreinterpretq_u8_f16): Likewise.
20507 (vreinterpretq_u8_f32): Likewise.
20508 (vreinterpretq_f16_f32): Likewise.
20509 (vreinterpretq_f16_s16): Likewise.
20510 (vreinterpretq_f16_s32): Likewise.
20511 (vreinterpretq_f16_s64): Likewise.
20512 (vreinterpretq_f16_s8): Likewise.
20513 (vreinterpretq_f16_u16): Likewise.
20514 (vreinterpretq_f16_u32): Likewise.
20515 (vreinterpretq_f16_u64): Likewise.
20516 (vreinterpretq_f16_u8): Likewise.
20517 (vreinterpretq_f32_f16): Likewise.
20518 (vreinterpretq_f32_s16): Likewise.
20519 (vreinterpretq_f32_s32): Likewise.
20520 (vreinterpretq_f32_s64): Likewise.
20521 (vreinterpretq_f32_s8): Likewise.
20522 (vreinterpretq_f32_u16): Likewise.
20523 (vreinterpretq_f32_u32): Likewise.
20524 (vreinterpretq_f32_u64): Likewise.
20525 (vreinterpretq_f32_u8): Likewise.
20526 (vreinterpretq_s16_f16): Likewise.
20527 (vreinterpretq_s16_f32): Likewise.
20528 (vreinterpretq_s64_f16): Likewise.
20529 (vreinterpretq_s64_f32): Likewise.
20530 (vreinterpretq_s8_f16): Likewise.
20531 (vreinterpretq_s8_f32): Likewise.
20532 (__arm_vreinterpretq_f16): Likewise.
20533 (__arm_vreinterpretq_f32): Likewise.
20534 (__arm_vreinterpretq_s16): Likewise.
20535 (__arm_vreinterpretq_s32): Likewise.
20536 (__arm_vreinterpretq_s64): Likewise.
20537 (__arm_vreinterpretq_s8): Likewise.
20538 (__arm_vreinterpretq_u16): Likewise.
20539 (__arm_vreinterpretq_u32): Likewise.
20540 (__arm_vreinterpretq_u64): Likewise.
20541 (__arm_vreinterpretq_u8): Likewise.
20542 * config/arm/arm_mve_types.h (__arm_vreinterpretq_s16_s32): Remove.
20543 (__arm_vreinterpretq_s16_s64): Likewise.
20544 (__arm_vreinterpretq_s16_s8): Likewise.
20545 (__arm_vreinterpretq_s16_u16): Likewise.
20546 (__arm_vreinterpretq_s16_u32): Likewise.
20547 (__arm_vreinterpretq_s16_u64): Likewise.
20548 (__arm_vreinterpretq_s16_u8): Likewise.
20549 (__arm_vreinterpretq_s32_s16): Likewise.
20550 (__arm_vreinterpretq_s32_s64): Likewise.
20551 (__arm_vreinterpretq_s32_s8): Likewise.
20552 (__arm_vreinterpretq_s32_u16): Likewise.
20553 (__arm_vreinterpretq_s32_u32): Likewise.
20554 (__arm_vreinterpretq_s32_u64): Likewise.
20555 (__arm_vreinterpretq_s32_u8): Likewise.
20556 (__arm_vreinterpretq_s64_s16): Likewise.
20557 (__arm_vreinterpretq_s64_s32): Likewise.
20558 (__arm_vreinterpretq_s64_s8): Likewise.
20559 (__arm_vreinterpretq_s64_u16): Likewise.
20560 (__arm_vreinterpretq_s64_u32): Likewise.
20561 (__arm_vreinterpretq_s64_u64): Likewise.
20562 (__arm_vreinterpretq_s64_u8): Likewise.
20563 (__arm_vreinterpretq_s8_s16): Likewise.
20564 (__arm_vreinterpretq_s8_s32): Likewise.
20565 (__arm_vreinterpretq_s8_s64): Likewise.
20566 (__arm_vreinterpretq_s8_u16): Likewise.
20567 (__arm_vreinterpretq_s8_u32): Likewise.
20568 (__arm_vreinterpretq_s8_u64): Likewise.
20569 (__arm_vreinterpretq_s8_u8): Likewise.
20570 (__arm_vreinterpretq_u16_s16): Likewise.
20571 (__arm_vreinterpretq_u16_s32): Likewise.
20572 (__arm_vreinterpretq_u16_s64): Likewise.
20573 (__arm_vreinterpretq_u16_s8): Likewise.
20574 (__arm_vreinterpretq_u16_u32): Likewise.
20575 (__arm_vreinterpretq_u16_u64): Likewise.
20576 (__arm_vreinterpretq_u16_u8): Likewise.
20577 (__arm_vreinterpretq_u32_s16): Likewise.
20578 (__arm_vreinterpretq_u32_s32): Likewise.
20579 (__arm_vreinterpretq_u32_s64): Likewise.
20580 (__arm_vreinterpretq_u32_s8): Likewise.
20581 (__arm_vreinterpretq_u32_u16): Likewise.
20582 (__arm_vreinterpretq_u32_u64): Likewise.
20583 (__arm_vreinterpretq_u32_u8): Likewise.
20584 (__arm_vreinterpretq_u64_s16): Likewise.
20585 (__arm_vreinterpretq_u64_s32): Likewise.
20586 (__arm_vreinterpretq_u64_s64): Likewise.
20587 (__arm_vreinterpretq_u64_s8): Likewise.
20588 (__arm_vreinterpretq_u64_u16): Likewise.
20589 (__arm_vreinterpretq_u64_u32): Likewise.
20590 (__arm_vreinterpretq_u64_u8): Likewise.
20591 (__arm_vreinterpretq_u8_s16): Likewise.
20592 (__arm_vreinterpretq_u8_s32): Likewise.
20593 (__arm_vreinterpretq_u8_s64): Likewise.
20594 (__arm_vreinterpretq_u8_s8): Likewise.
20595 (__arm_vreinterpretq_u8_u16): Likewise.
20596 (__arm_vreinterpretq_u8_u32): Likewise.
20597 (__arm_vreinterpretq_u8_u64): Likewise.
20598 (__arm_vreinterpretq_s32_f16): Likewise.
20599 (__arm_vreinterpretq_s32_f32): Likewise.
20600 (__arm_vreinterpretq_s16_f16): Likewise.
20601 (__arm_vreinterpretq_s16_f32): Likewise.
20602 (__arm_vreinterpretq_s64_f16): Likewise.
20603 (__arm_vreinterpretq_s64_f32): Likewise.
20604 (__arm_vreinterpretq_s8_f16): Likewise.
20605 (__arm_vreinterpretq_s8_f32): Likewise.
20606 (__arm_vreinterpretq_u16_f16): Likewise.
20607 (__arm_vreinterpretq_u16_f32): Likewise.
20608 (__arm_vreinterpretq_u32_f16): Likewise.
20609 (__arm_vreinterpretq_u32_f32): Likewise.
20610 (__arm_vreinterpretq_u64_f16): Likewise.
20611 (__arm_vreinterpretq_u64_f32): Likewise.
20612 (__arm_vreinterpretq_u8_f16): Likewise.
20613 (__arm_vreinterpretq_u8_f32): Likewise.
20614 (__arm_vreinterpretq_f16_f32): Likewise.
20615 (__arm_vreinterpretq_f16_s16): Likewise.
20616 (__arm_vreinterpretq_f16_s32): Likewise.
20617 (__arm_vreinterpretq_f16_s64): Likewise.
20618 (__arm_vreinterpretq_f16_s8): Likewise.
20619 (__arm_vreinterpretq_f16_u16): Likewise.
20620 (__arm_vreinterpretq_f16_u32): Likewise.
20621 (__arm_vreinterpretq_f16_u64): Likewise.
20622 (__arm_vreinterpretq_f16_u8): Likewise.
20623 (__arm_vreinterpretq_f32_f16): Likewise.
20624 (__arm_vreinterpretq_f32_s16): Likewise.
20625 (__arm_vreinterpretq_f32_s32): Likewise.
20626 (__arm_vreinterpretq_f32_s64): Likewise.
20627 (__arm_vreinterpretq_f32_s8): Likewise.
20628 (__arm_vreinterpretq_f32_u16): Likewise.
20629 (__arm_vreinterpretq_f32_u32): Likewise.
20630 (__arm_vreinterpretq_f32_u64): Likewise.
20631 (__arm_vreinterpretq_f32_u8): Likewise.
20632 (__arm_vreinterpretq_s16): Likewise.
20633 (__arm_vreinterpretq_s32): Likewise.
20634 (__arm_vreinterpretq_s64): Likewise.
20635 (__arm_vreinterpretq_s8): Likewise.
20636 (__arm_vreinterpretq_u16): Likewise.
20637 (__arm_vreinterpretq_u32): Likewise.
20638 (__arm_vreinterpretq_u64): Likewise.
20639 (__arm_vreinterpretq_u8): Likewise.
20640 (__arm_vreinterpretq_f16): Likewise.
20641 (__arm_vreinterpretq_f32): Likewise.
20642 * config/arm/mve.md (@arm_mve_reinterpret<mode>): New pattern.
20643 * config/arm/unspecs.md: (REINTERPRET): New unspec.
20645 2023-05-03 Murray Steele <murray.steele@arm.com>
20646 Christophe Lyon <christophe.lyon@arm.com>
20647 Christophe Lyon <christophe.lyon@arm.com
20649 * config.gcc: Add arm-mve-builtins-base.o and
20650 arm-mve-builtins-shapes.o to extra_objs.
20651 * config/arm/arm-builtins.cc (arm_builtin_decl): Handle MVE builtin
20653 (arm_expand_builtin): Likewise
20654 (arm_check_builtin_call): Likewise
20655 (arm_describe_resolver): Likewise.
20656 * config/arm/arm-builtins.h (enum resolver_ident): Add
20658 * config/arm/arm-c.cc (arm_pragma_arm): Handle new pragma.
20659 (arm_resolve_overloaded_builtin): Handle MVE builtins.
20660 (arm_register_target_pragmas): Register arm_check_builtin_call.
20661 * config/arm/arm-mve-builtins.cc (class registered_function): New
20663 (struct registered_function_hasher): New struct.
20664 (pred_suffixes): New table.
20665 (mode_suffixes): New table.
20666 (type_suffix_info): New table.
20667 (TYPES_float16): New.
20668 (TYPES_all_float): New.
20669 (TYPES_integer_8): New.
20670 (TYPES_integer_8_16): New.
20671 (TYPES_integer_16_32): New.
20672 (TYPES_integer_32): New.
20673 (TYPES_signed_16_32): New.
20674 (TYPES_signed_32): New.
20675 (TYPES_all_signed): New.
20676 (TYPES_all_unsigned): New.
20677 (TYPES_all_integer): New.
20678 (TYPES_all_integer_with_64): New.
20679 (DEF_VECTOR_TYPE): New.
20680 (DEF_DOUBLE_TYPE): New.
20681 (DEF_MVE_TYPES_ARRAY): New.
20682 (all_integer): New.
20683 (all_integer_with_64): New.
20687 (all_unsigned): New.
20689 (integer_8_16): New.
20690 (integer_16_32): New.
20692 (signed_16_32): New.
20694 (register_vector_type): Use void_type_node for mve.fp-only types when
20695 mve.fp is not enabled.
20696 (register_builtin_tuple_types): Likewise.
20697 (handle_arm_mve_h): New function..
20698 (matches_type_p): Likewise..
20699 (report_out_of_range): Likewise.
20700 (report_not_enum): Likewise.
20701 (report_missing_float): Likewise.
20702 (report_non_ice): Likewise.
20703 (check_requires_float): Likewise.
20704 (function_instance::hash): Likewise
20705 (function_instance::call_properties): Likewise.
20706 (function_instance::reads_global_state_p): Likewise.
20707 (function_instance::modifies_global_state_p): Likewise.
20708 (function_instance::could_trap_p): Likewise.
20709 (function_instance::has_inactive_argument): Likewise.
20710 (registered_function_hasher::hash): Likewise.
20711 (registered_function_hasher::equal): Likewise.
20712 (function_builder::function_builder): Likewise.
20713 (function_builder::~function_builder): Likewise.
20714 (function_builder::append_name): Likewise.
20715 (function_builder::finish_name): Likewise.
20716 (function_builder::get_name): Likewise.
20717 (add_attribute): Likewise.
20718 (function_builder::get_attributes): Likewise.
20719 (function_builder::add_function): Likewise.
20720 (function_builder::add_unique_function): Likewise.
20721 (function_builder::add_overloaded_function): Likewise.
20722 (function_builder::add_overloaded_functions): Likewise.
20723 (function_builder::register_function_group): Likewise.
20724 (function_call_info::function_call_info): Likewise.
20725 (function_resolver::function_resolver): Likewise.
20726 (function_resolver::get_vector_type): Likewise.
20727 (function_resolver::get_scalar_type_name): Likewise.
20728 (function_resolver::get_argument_type): Likewise.
20729 (function_resolver::scalar_argument_p): Likewise.
20730 (function_resolver::report_no_such_form): Likewise.
20731 (function_resolver::lookup_form): Likewise.
20732 (function_resolver::resolve_to): Likewise.
20733 (function_resolver::infer_vector_or_tuple_type): Likewise.
20734 (function_resolver::infer_vector_type): Likewise.
20735 (function_resolver::require_vector_or_scalar_type): Likewise.
20736 (function_resolver::require_vector_type): Likewise.
20737 (function_resolver::require_matching_vector_type): Likewise.
20738 (function_resolver::require_derived_vector_type): Likewise.
20739 (function_resolver::require_derived_scalar_type): Likewise.
20740 (function_resolver::require_integer_immediate): Likewise.
20741 (function_resolver::require_scalar_type): Likewise.
20742 (function_resolver::check_num_arguments): Likewise.
20743 (function_resolver::check_gp_argument): Likewise.
20744 (function_resolver::finish_opt_n_resolution): Likewise.
20745 (function_resolver::resolve_unary): Likewise.
20746 (function_resolver::resolve_unary_n): Likewise.
20747 (function_resolver::resolve_uniform): Likewise.
20748 (function_resolver::resolve_uniform_opt_n): Likewise.
20749 (function_resolver::resolve): Likewise.
20750 (function_checker::function_checker): Likewise.
20751 (function_checker::argument_exists_p): Likewise.
20752 (function_checker::require_immediate): Likewise.
20753 (function_checker::require_immediate_enum): Likewise.
20754 (function_checker::require_immediate_range): Likewise.
20755 (function_checker::check): Likewise.
20756 (gimple_folder::gimple_folder): Likewise.
20757 (gimple_folder::fold): Likewise.
20758 (function_expander::function_expander): Likewise.
20759 (function_expander::direct_optab_handler): Likewise.
20760 (function_expander::get_fallback_value): Likewise.
20761 (function_expander::get_reg_target): Likewise.
20762 (function_expander::add_output_operand): Likewise.
20763 (function_expander::add_input_operand): Likewise.
20764 (function_expander::add_integer_operand): Likewise.
20765 (function_expander::generate_insn): Likewise.
20766 (function_expander::use_exact_insn): Likewise.
20767 (function_expander::use_unpred_insn): Likewise.
20768 (function_expander::use_pred_x_insn): Likewise.
20769 (function_expander::use_cond_insn): Likewise.
20770 (function_expander::map_to_rtx_codes): Likewise.
20771 (function_expander::expand): Likewise.
20772 (resolve_overloaded_builtin): Likewise.
20773 (check_builtin_call): Likewise.
20774 (gimple_fold_builtin): Likewise.
20775 (expand_builtin): Likewise.
20776 (gt_ggc_mx): Likewise.
20777 (gt_pch_nx): Likewise.
20778 (gt_pch_nx): Likewise.
20779 * config/arm/arm-mve-builtins.def(s8): Define new type suffix.
20790 (offset): New mode.
20791 * config/arm/arm-mve-builtins.h (MAX_TUPLE_SIZE): New constant.
20792 (CP_READ_FPCR): Likewise.
20793 (CP_RAISE_FP_EXCEPTIONS): Likewise.
20794 (CP_READ_MEMORY): Likewise.
20795 (CP_WRITE_MEMORY): Likewise.
20796 (enum units_index): New enum.
20797 (enum predication_index): New.
20798 (enum type_class_index): New.
20799 (enum mode_suffix_index): New enum.
20800 (enum type_suffix_index): New.
20801 (struct mode_suffix_info): New struct.
20802 (struct type_suffix_info): New.
20803 (struct function_group_info): Likewise.
20804 (class function_instance): Likewise.
20805 (class registered_function): Likewise.
20806 (class function_builder): Likewise.
20807 (class function_call_info): Likewise.
20808 (class function_resolver): Likewise.
20809 (class function_checker): Likewise.
20810 (class gimple_folder): Likewise.
20811 (class function_expander): Likewise.
20812 (get_mve_pred16_t): Likewise.
20813 (find_mode_suffix): New function.
20814 (class function_base): Likewise.
20815 (class function_shape): Likewise.
20816 (function_instance::operator==): New function.
20817 (function_instance::operator!=): Likewise.
20818 (function_instance::vectors_per_tuple): Likewise.
20819 (function_instance::mode_suffix): Likewise.
20820 (function_instance::type_suffix): Likewise.
20821 (function_instance::scalar_type): Likewise.
20822 (function_instance::vector_type): Likewise.
20823 (function_instance::tuple_type): Likewise.
20824 (function_instance::vector_mode): Likewise.
20825 (function_call_info::function_returns_void_p): Likewise.
20826 (function_base::call_properties): Likewise.
20827 * config/arm/arm-protos.h (enum arm_builtin_class): Add
20829 (handle_arm_mve_h): New.
20830 (resolve_overloaded_builtin): New.
20831 (check_builtin_call): New.
20832 (gimple_fold_builtin): New.
20833 (expand_builtin): New.
20834 * config/arm/arm.cc (TARGET_GIMPLE_FOLD_BUILTIN): Define as
20835 arm_gimple_fold_builtin.
20836 (arm_gimple_fold_builtin): New function.
20837 * config/arm/arm_mve.h: Use new arm_mve.h pragma.
20838 * config/arm/predicates.md (arm_any_register_operand): New predicate.
20839 * config/arm/t-arm: (arm-mve-builtins.o): Add includes.
20840 (arm-mve-builtins-shapes.o): New target.
20841 (arm-mve-builtins-base.o): New target.
20842 * config/arm/arm-mve-builtins-base.cc: New file.
20843 * config/arm/arm-mve-builtins-base.def: New file.
20844 * config/arm/arm-mve-builtins-base.h: New file.
20845 * config/arm/arm-mve-builtins-functions.h: New file.
20846 * config/arm/arm-mve-builtins-shapes.cc: New file.
20847 * config/arm/arm-mve-builtins-shapes.h: New file.
20849 2023-05-03 Murray Steele <murray.steele@arm.com>
20850 Christophe Lyon <christophe.lyon@arm.com>
20851 Christophe Lyon <christophe.lyon@arm.com>
20853 * config/arm/arm-builtins.cc (arm_general_add_builtin_function):
20855 (arm_init_builtin): Use arm_general_add_builtin_function instead
20856 of arm_add_builtin_function.
20857 (arm_init_acle_builtins): Likewise.
20858 (arm_init_mve_builtins): Likewise.
20859 (arm_init_crypto_builtins): Likewise.
20860 (arm_init_builtins): Likewise.
20861 (arm_general_builtin_decl): New function.
20862 (arm_builtin_decl): Defer to numberspace-specialized functions.
20863 (arm_expand_builtin_args): Rename into arm_general_expand_builtin_args.
20864 (arm_expand_builtin_1): Rename into arm_general_expand_builtin_1 and ...
20865 (arm_general_expand_builtin_1): ... specialize for general builtins.
20866 (arm_expand_acle_builtin): Use arm_general_expand_builtin
20867 instead of arm_expand_builtin.
20868 (arm_expand_mve_builtin): Likewise.
20869 (arm_expand_neon_builtin): Likewise.
20870 (arm_expand_vfp_builtin): Likewise.
20871 (arm_general_expand_builtin): New function.
20872 (arm_expand_builtin): Specialize for general builtins.
20873 (arm_general_check_builtin_call): New function.
20874 (arm_check_builtin_call): Specialize for general builtins.
20875 (arm_describe_resolver): Validate numberspace.
20876 (arm_cde_end_args): Likewise.
20877 * config/arm/arm-protos.h (enum arm_builtin_class): New enum.
20878 (ARM_BUILTIN_SHIFT, ARM_BUILTIN_CLASS): New constants.
20880 2023-05-03 Martin Liska <mliska@suse.cz>
20883 * config/riscv/sync.md: Add gcc_unreachable to a switch.
20885 2023-05-03 Richard Biener <rguenther@suse.de>
20887 * tree-ssa-loop-split.cc (split_at_bb_p): Avoid last_stmt.
20888 (patch_loop_exit): Likewise.
20889 (connect_loops): Likewise.
20890 (split_loop): Likewise.
20891 (control_dep_semi_invariant_p): Likewise.
20892 (do_split_loop_on_cond): Likewise.
20893 (split_loop_on_cond): Likewise.
20894 * tree-ssa-loop-unswitch.cc (find_unswitching_predicates_for_bb):
20896 (simplify_loop_version): Likewise.
20897 (evaluate_bbs): Likewise.
20898 (find_loop_guard): Likewise.
20899 (clean_up_after_unswitching): Likewise.
20900 * tree-ssa-math-opts.cc (maybe_optimize_guarding_check):
20902 (optimize_spaceship): Take a gcond * argument, avoid
20904 (math_opts_dom_walker::after_dom_children): Adjust call to
20905 optimize_spaceship.
20906 * tree-vrp.cc (maybe_set_nonzero_bits): Avoid last_stmt.
20907 * value-pointer-equiv.cc (pointer_equiv_analyzer::visit_edge):
20910 2023-05-03 Andreas Schwab <schwab@suse.de>
20912 * config/riscv/linux.h (LIB_SPEC): Don't redefine.
20914 2023-05-03 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
20916 * config/riscv/riscv-vector-builtins-bases.cc (fold_fault_load):
20918 (class vlseg): New class.
20919 (class vsseg): Ditto.
20920 (class vlsseg): Ditto.
20921 (class vssseg): Ditto.
20922 (class seg_indexed_load): Ditto.
20923 (class seg_indexed_store): Ditto.
20924 (class vlsegff): Ditto.
20926 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
20927 * config/riscv/riscv-vector-builtins-functions.def (vlseg):
20937 * config/riscv/riscv-vector-builtins-shapes.cc (struct
20938 seg_loadstore_def): Ditto.
20939 (struct seg_indexed_loadstore_def): Ditto.
20940 (struct seg_fault_load_def): Ditto.
20942 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
20943 * config/riscv/riscv-vector-builtins.cc
20944 (function_builder::append_nf): New function.
20945 * config/riscv/riscv-vector-builtins.def (vfloat32m1x2_t):
20946 Change ptr from double into float.
20947 (vfloat32m1x3_t): Ditto.
20948 (vfloat32m1x4_t): Ditto.
20949 (vfloat32m1x5_t): Ditto.
20950 (vfloat32m1x6_t): Ditto.
20951 (vfloat32m1x7_t): Ditto.
20952 (vfloat32m1x8_t): Ditto.
20953 (vfloat32m2x2_t): Ditto.
20954 (vfloat32m2x3_t): Ditto.
20955 (vfloat32m2x4_t): Ditto.
20956 (vfloat32m4x2_t): Ditto.
20957 * config/riscv/riscv-vector-builtins.h: Add segment intrinsics.
20958 * config/riscv/riscv-vsetvl.cc (fault_first_load_p): Adapt for
20960 * config/riscv/riscv.md: Add segment instructions.
20961 * config/riscv/vector-iterators.md: Support segment intrinsics.
20962 * config/riscv/vector.md (@pred_unit_strided_load<mode>): New
20964 (@pred_unit_strided_store<mode>): Ditto.
20965 (@pred_strided_load<mode>): Ditto.
20966 (@pred_strided_store<mode>): Ditto.
20967 (@pred_fault_load<mode>): Ditto.
20968 (@pred_indexed_<order>load<V1T:mode><V1I:mode>): Ditto.
20969 (@pred_indexed_<order>load<V2T:mode><V2I:mode>): Ditto.
20970 (@pred_indexed_<order>load<V4T:mode><V4I:mode>): Ditto.
20971 (@pred_indexed_<order>load<V8T:mode><V8I:mode>): Ditto.
20972 (@pred_indexed_<order>load<V16T:mode><V16I:mode>): Ditto.
20973 (@pred_indexed_<order>load<V32T:mode><V32I:mode>): Ditto.
20974 (@pred_indexed_<order>load<V64T:mode><V64I:mode>): Ditto.
20975 (@pred_indexed_<order>store<V1T:mode><V1I:mode>): Ditto.
20976 (@pred_indexed_<order>store<V2T:mode><V2I:mode>): Ditto.
20977 (@pred_indexed_<order>store<V4T:mode><V4I:mode>): Ditto.
20978 (@pred_indexed_<order>store<V8T:mode><V8I:mode>): Ditto.
20979 (@pred_indexed_<order>store<V16T:mode><V16I:mode>): Ditto.
20980 (@pred_indexed_<order>store<V32T:mode><V32I:mode>): Ditto.
20981 (@pred_indexed_<order>store<V64T:mode><V64I:mode>): Ditto.
20983 2023-05-03 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
20985 * config/riscv/genrvv-type-indexer.cc (valid_type): Adapt for
20986 tuple type support.
20988 (floattype): Ditto.
20990 * config/riscv/riscv-vector-builtins-bases.cc: Ditto.
20991 * config/riscv/riscv-vector-builtins-functions.def (vset): Add
20993 (vget): Add tuple type vget.
20994 * config/riscv/riscv-vector-builtins-types.def
20995 (DEF_RVV_TUPLE_OPS): New macro.
20996 (vint8mf8x2_t): Ditto.
20997 (vuint8mf8x2_t): Ditto.
20998 (vint8mf8x3_t): Ditto.
20999 (vuint8mf8x3_t): Ditto.
21000 (vint8mf8x4_t): Ditto.
21001 (vuint8mf8x4_t): Ditto.
21002 (vint8mf8x5_t): Ditto.
21003 (vuint8mf8x5_t): Ditto.
21004 (vint8mf8x6_t): Ditto.
21005 (vuint8mf8x6_t): Ditto.
21006 (vint8mf8x7_t): Ditto.
21007 (vuint8mf8x7_t): Ditto.
21008 (vint8mf8x8_t): Ditto.
21009 (vuint8mf8x8_t): Ditto.
21010 (vint8mf4x2_t): Ditto.
21011 (vuint8mf4x2_t): Ditto.
21012 (vint8mf4x3_t): Ditto.
21013 (vuint8mf4x3_t): Ditto.
21014 (vint8mf4x4_t): Ditto.
21015 (vuint8mf4x4_t): Ditto.
21016 (vint8mf4x5_t): Ditto.
21017 (vuint8mf4x5_t): Ditto.
21018 (vint8mf4x6_t): Ditto.
21019 (vuint8mf4x6_t): Ditto.
21020 (vint8mf4x7_t): Ditto.
21021 (vuint8mf4x7_t): Ditto.
21022 (vint8mf4x8_t): Ditto.
21023 (vuint8mf4x8_t): Ditto.
21024 (vint8mf2x2_t): Ditto.
21025 (vuint8mf2x2_t): Ditto.
21026 (vint8mf2x3_t): Ditto.
21027 (vuint8mf2x3_t): Ditto.
21028 (vint8mf2x4_t): Ditto.
21029 (vuint8mf2x4_t): Ditto.
21030 (vint8mf2x5_t): Ditto.
21031 (vuint8mf2x5_t): Ditto.
21032 (vint8mf2x6_t): Ditto.
21033 (vuint8mf2x6_t): Ditto.
21034 (vint8mf2x7_t): Ditto.
21035 (vuint8mf2x7_t): Ditto.
21036 (vint8mf2x8_t): Ditto.
21037 (vuint8mf2x8_t): Ditto.
21038 (vint8m1x2_t): Ditto.
21039 (vuint8m1x2_t): Ditto.
21040 (vint8m1x3_t): Ditto.
21041 (vuint8m1x3_t): Ditto.
21042 (vint8m1x4_t): Ditto.
21043 (vuint8m1x4_t): Ditto.
21044 (vint8m1x5_t): Ditto.
21045 (vuint8m1x5_t): Ditto.
21046 (vint8m1x6_t): Ditto.
21047 (vuint8m1x6_t): Ditto.
21048 (vint8m1x7_t): Ditto.
21049 (vuint8m1x7_t): Ditto.
21050 (vint8m1x8_t): Ditto.
21051 (vuint8m1x8_t): Ditto.
21052 (vint8m2x2_t): Ditto.
21053 (vuint8m2x2_t): Ditto.
21054 (vint8m2x3_t): Ditto.
21055 (vuint8m2x3_t): Ditto.
21056 (vint8m2x4_t): Ditto.
21057 (vuint8m2x4_t): Ditto.
21058 (vint8m4x2_t): Ditto.
21059 (vuint8m4x2_t): Ditto.
21060 (vint16mf4x2_t): Ditto.
21061 (vuint16mf4x2_t): Ditto.
21062 (vint16mf4x3_t): Ditto.
21063 (vuint16mf4x3_t): Ditto.
21064 (vint16mf4x4_t): Ditto.
21065 (vuint16mf4x4_t): Ditto.
21066 (vint16mf4x5_t): Ditto.
21067 (vuint16mf4x5_t): Ditto.
21068 (vint16mf4x6_t): Ditto.
21069 (vuint16mf4x6_t): Ditto.
21070 (vint16mf4x7_t): Ditto.
21071 (vuint16mf4x7_t): Ditto.
21072 (vint16mf4x8_t): Ditto.
21073 (vuint16mf4x8_t): Ditto.
21074 (vint16mf2x2_t): Ditto.
21075 (vuint16mf2x2_t): Ditto.
21076 (vint16mf2x3_t): Ditto.
21077 (vuint16mf2x3_t): Ditto.
21078 (vint16mf2x4_t): Ditto.
21079 (vuint16mf2x4_t): Ditto.
21080 (vint16mf2x5_t): Ditto.
21081 (vuint16mf2x5_t): Ditto.
21082 (vint16mf2x6_t): Ditto.
21083 (vuint16mf2x6_t): Ditto.
21084 (vint16mf2x7_t): Ditto.
21085 (vuint16mf2x7_t): Ditto.
21086 (vint16mf2x8_t): Ditto.
21087 (vuint16mf2x8_t): Ditto.
21088 (vint16m1x2_t): Ditto.
21089 (vuint16m1x2_t): Ditto.
21090 (vint16m1x3_t): Ditto.
21091 (vuint16m1x3_t): Ditto.
21092 (vint16m1x4_t): Ditto.
21093 (vuint16m1x4_t): Ditto.
21094 (vint16m1x5_t): Ditto.
21095 (vuint16m1x5_t): Ditto.
21096 (vint16m1x6_t): Ditto.
21097 (vuint16m1x6_t): Ditto.
21098 (vint16m1x7_t): Ditto.
21099 (vuint16m1x7_t): Ditto.
21100 (vint16m1x8_t): Ditto.
21101 (vuint16m1x8_t): Ditto.
21102 (vint16m2x2_t): Ditto.
21103 (vuint16m2x2_t): Ditto.
21104 (vint16m2x3_t): Ditto.
21105 (vuint16m2x3_t): Ditto.
21106 (vint16m2x4_t): Ditto.
21107 (vuint16m2x4_t): Ditto.
21108 (vint16m4x2_t): Ditto.
21109 (vuint16m4x2_t): Ditto.
21110 (vint32mf2x2_t): Ditto.
21111 (vuint32mf2x2_t): Ditto.
21112 (vint32mf2x3_t): Ditto.
21113 (vuint32mf2x3_t): Ditto.
21114 (vint32mf2x4_t): Ditto.
21115 (vuint32mf2x4_t): Ditto.
21116 (vint32mf2x5_t): Ditto.
21117 (vuint32mf2x5_t): Ditto.
21118 (vint32mf2x6_t): Ditto.
21119 (vuint32mf2x6_t): Ditto.
21120 (vint32mf2x7_t): Ditto.
21121 (vuint32mf2x7_t): Ditto.
21122 (vint32mf2x8_t): Ditto.
21123 (vuint32mf2x8_t): Ditto.
21124 (vint32m1x2_t): Ditto.
21125 (vuint32m1x2_t): Ditto.
21126 (vint32m1x3_t): Ditto.
21127 (vuint32m1x3_t): Ditto.
21128 (vint32m1x4_t): Ditto.
21129 (vuint32m1x4_t): Ditto.
21130 (vint32m1x5_t): Ditto.
21131 (vuint32m1x5_t): Ditto.
21132 (vint32m1x6_t): Ditto.
21133 (vuint32m1x6_t): Ditto.
21134 (vint32m1x7_t): Ditto.
21135 (vuint32m1x7_t): Ditto.
21136 (vint32m1x8_t): Ditto.
21137 (vuint32m1x8_t): Ditto.
21138 (vint32m2x2_t): Ditto.
21139 (vuint32m2x2_t): Ditto.
21140 (vint32m2x3_t): Ditto.
21141 (vuint32m2x3_t): Ditto.
21142 (vint32m2x4_t): Ditto.
21143 (vuint32m2x4_t): Ditto.
21144 (vint32m4x2_t): Ditto.
21145 (vuint32m4x2_t): Ditto.
21146 (vint64m1x2_t): Ditto.
21147 (vuint64m1x2_t): Ditto.
21148 (vint64m1x3_t): Ditto.
21149 (vuint64m1x3_t): Ditto.
21150 (vint64m1x4_t): Ditto.
21151 (vuint64m1x4_t): Ditto.
21152 (vint64m1x5_t): Ditto.
21153 (vuint64m1x5_t): Ditto.
21154 (vint64m1x6_t): Ditto.
21155 (vuint64m1x6_t): Ditto.
21156 (vint64m1x7_t): Ditto.
21157 (vuint64m1x7_t): Ditto.
21158 (vint64m1x8_t): Ditto.
21159 (vuint64m1x8_t): Ditto.
21160 (vint64m2x2_t): Ditto.
21161 (vuint64m2x2_t): Ditto.
21162 (vint64m2x3_t): Ditto.
21163 (vuint64m2x3_t): Ditto.
21164 (vint64m2x4_t): Ditto.
21165 (vuint64m2x4_t): Ditto.
21166 (vint64m4x2_t): Ditto.
21167 (vuint64m4x2_t): Ditto.
21168 (vfloat32mf2x2_t): Ditto.
21169 (vfloat32mf2x3_t): Ditto.
21170 (vfloat32mf2x4_t): Ditto.
21171 (vfloat32mf2x5_t): Ditto.
21172 (vfloat32mf2x6_t): Ditto.
21173 (vfloat32mf2x7_t): Ditto.
21174 (vfloat32mf2x8_t): Ditto.
21175 (vfloat32m1x2_t): Ditto.
21176 (vfloat32m1x3_t): Ditto.
21177 (vfloat32m1x4_t): Ditto.
21178 (vfloat32m1x5_t): Ditto.
21179 (vfloat32m1x6_t): Ditto.
21180 (vfloat32m1x7_t): Ditto.
21181 (vfloat32m1x8_t): Ditto.
21182 (vfloat32m2x2_t): Ditto.
21183 (vfloat32m2x3_t): Ditto.
21184 (vfloat32m2x4_t): Ditto.
21185 (vfloat32m4x2_t): Ditto.
21186 (vfloat64m1x2_t): Ditto.
21187 (vfloat64m1x3_t): Ditto.
21188 (vfloat64m1x4_t): Ditto.
21189 (vfloat64m1x5_t): Ditto.
21190 (vfloat64m1x6_t): Ditto.
21191 (vfloat64m1x7_t): Ditto.
21192 (vfloat64m1x8_t): Ditto.
21193 (vfloat64m2x2_t): Ditto.
21194 (vfloat64m2x3_t): Ditto.
21195 (vfloat64m2x4_t): Ditto.
21196 (vfloat64m4x2_t): Ditto.
21197 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_TUPLE_OPS):
21199 (DEF_RVV_TYPE_INDEX): Ditto.
21200 (rvv_arg_type_info::get_tuple_subpart_type): New function.
21201 (DEF_RVV_TUPLE_TYPE): New macro.
21202 * config/riscv/riscv-vector-builtins.def (DEF_RVV_TYPE_INDEX):
21203 Adapt for tuple vget/vset support.
21204 (vint8mf4_t): Ditto.
21205 (vuint8mf4_t): Ditto.
21206 (vint8mf2_t): Ditto.
21207 (vuint8mf2_t): Ditto.
21208 (vint8m1_t): Ditto.
21209 (vuint8m1_t): Ditto.
21210 (vint8m2_t): Ditto.
21211 (vuint8m2_t): Ditto.
21212 (vint8m4_t): Ditto.
21213 (vuint8m4_t): Ditto.
21214 (vint8m8_t): Ditto.
21215 (vuint8m8_t): Ditto.
21216 (vint16mf4_t): Ditto.
21217 (vuint16mf4_t): Ditto.
21218 (vint16mf2_t): Ditto.
21219 (vuint16mf2_t): Ditto.
21220 (vint16m1_t): Ditto.
21221 (vuint16m1_t): Ditto.
21222 (vint16m2_t): Ditto.
21223 (vuint16m2_t): Ditto.
21224 (vint16m4_t): Ditto.
21225 (vuint16m4_t): Ditto.
21226 (vint16m8_t): Ditto.
21227 (vuint16m8_t): Ditto.
21228 (vint32mf2_t): Ditto.
21229 (vuint32mf2_t): Ditto.
21230 (vint32m1_t): Ditto.
21231 (vuint32m1_t): Ditto.
21232 (vint32m2_t): Ditto.
21233 (vuint32m2_t): Ditto.
21234 (vint32m4_t): Ditto.
21235 (vuint32m4_t): Ditto.
21236 (vint32m8_t): Ditto.
21237 (vuint32m8_t): Ditto.
21238 (vint64m1_t): Ditto.
21239 (vuint64m1_t): Ditto.
21240 (vint64m2_t): Ditto.
21241 (vuint64m2_t): Ditto.
21242 (vint64m4_t): Ditto.
21243 (vuint64m4_t): Ditto.
21244 (vint64m8_t): Ditto.
21245 (vuint64m8_t): Ditto.
21246 (vfloat32mf2_t): Ditto.
21247 (vfloat32m1_t): Ditto.
21248 (vfloat32m2_t): Ditto.
21249 (vfloat32m4_t): Ditto.
21250 (vfloat32m8_t): Ditto.
21251 (vfloat64m1_t): Ditto.
21252 (vfloat64m2_t): Ditto.
21253 (vfloat64m4_t): Ditto.
21254 (vfloat64m8_t): Ditto.
21255 (tuple_subpart): Add tuple subpart base type.
21256 * config/riscv/riscv-vector-builtins.h (struct
21257 rvv_arg_type_info): Ditto.
21258 (tuple_type_field): New function.
21260 2023-05-03 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
21262 * config/riscv/riscv-modes.def (RVV_TUPLE_MODES): New macro.
21263 (RVV_TUPLE_PARTIAL_MODES): Ditto.
21264 * config/riscv/riscv-protos.h (riscv_v_ext_tuple_mode_p): New
21267 (get_subpart_mode): Ditto.
21268 (get_tuple_mode): Ditto.
21269 (expand_tuple_move): Ditto.
21270 * config/riscv/riscv-v.cc (ENTRY): New macro.
21271 (TUPLE_ENTRY): Ditto.
21272 (get_nf): New function.
21273 (get_subpart_mode): Ditto.
21274 (get_tuple_mode): Ditto.
21275 (expand_tuple_move): Ditto.
21276 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_TUPLE_TYPE):
21278 (register_tuple_type): New function
21279 * config/riscv/riscv-vector-builtins.def (DEF_RVV_TUPLE_TYPE):
21281 (vint8mf8x2_t): New macro.
21282 (vuint8mf8x2_t): Ditto.
21283 (vint8mf8x3_t): Ditto.
21284 (vuint8mf8x3_t): Ditto.
21285 (vint8mf8x4_t): Ditto.
21286 (vuint8mf8x4_t): Ditto.
21287 (vint8mf8x5_t): Ditto.
21288 (vuint8mf8x5_t): Ditto.
21289 (vint8mf8x6_t): Ditto.
21290 (vuint8mf8x6_t): Ditto.
21291 (vint8mf8x7_t): Ditto.
21292 (vuint8mf8x7_t): Ditto.
21293 (vint8mf8x8_t): Ditto.
21294 (vuint8mf8x8_t): Ditto.
21295 (vint8mf4x2_t): Ditto.
21296 (vuint8mf4x2_t): Ditto.
21297 (vint8mf4x3_t): Ditto.
21298 (vuint8mf4x3_t): Ditto.
21299 (vint8mf4x4_t): Ditto.
21300 (vuint8mf4x4_t): Ditto.
21301 (vint8mf4x5_t): Ditto.
21302 (vuint8mf4x5_t): Ditto.
21303 (vint8mf4x6_t): Ditto.
21304 (vuint8mf4x6_t): Ditto.
21305 (vint8mf4x7_t): Ditto.
21306 (vuint8mf4x7_t): Ditto.
21307 (vint8mf4x8_t): Ditto.
21308 (vuint8mf4x8_t): Ditto.
21309 (vint8mf2x2_t): Ditto.
21310 (vuint8mf2x2_t): Ditto.
21311 (vint8mf2x3_t): Ditto.
21312 (vuint8mf2x3_t): Ditto.
21313 (vint8mf2x4_t): Ditto.
21314 (vuint8mf2x4_t): Ditto.
21315 (vint8mf2x5_t): Ditto.
21316 (vuint8mf2x5_t): Ditto.
21317 (vint8mf2x6_t): Ditto.
21318 (vuint8mf2x6_t): Ditto.
21319 (vint8mf2x7_t): Ditto.
21320 (vuint8mf2x7_t): Ditto.
21321 (vint8mf2x8_t): Ditto.
21322 (vuint8mf2x8_t): Ditto.
21323 (vint8m1x2_t): Ditto.
21324 (vuint8m1x2_t): Ditto.
21325 (vint8m1x3_t): Ditto.
21326 (vuint8m1x3_t): Ditto.
21327 (vint8m1x4_t): Ditto.
21328 (vuint8m1x4_t): Ditto.
21329 (vint8m1x5_t): Ditto.
21330 (vuint8m1x5_t): Ditto.
21331 (vint8m1x6_t): Ditto.
21332 (vuint8m1x6_t): Ditto.
21333 (vint8m1x7_t): Ditto.
21334 (vuint8m1x7_t): Ditto.
21335 (vint8m1x8_t): Ditto.
21336 (vuint8m1x8_t): Ditto.
21337 (vint8m2x2_t): Ditto.
21338 (vuint8m2x2_t): Ditto.
21339 (vint8m2x3_t): Ditto.
21340 (vuint8m2x3_t): Ditto.
21341 (vint8m2x4_t): Ditto.
21342 (vuint8m2x4_t): Ditto.
21343 (vint8m4x2_t): Ditto.
21344 (vuint8m4x2_t): Ditto.
21345 (vint16mf4x2_t): Ditto.
21346 (vuint16mf4x2_t): Ditto.
21347 (vint16mf4x3_t): Ditto.
21348 (vuint16mf4x3_t): Ditto.
21349 (vint16mf4x4_t): Ditto.
21350 (vuint16mf4x4_t): Ditto.
21351 (vint16mf4x5_t): Ditto.
21352 (vuint16mf4x5_t): Ditto.
21353 (vint16mf4x6_t): Ditto.
21354 (vuint16mf4x6_t): Ditto.
21355 (vint16mf4x7_t): Ditto.
21356 (vuint16mf4x7_t): Ditto.
21357 (vint16mf4x8_t): Ditto.
21358 (vuint16mf4x8_t): Ditto.
21359 (vint16mf2x2_t): Ditto.
21360 (vuint16mf2x2_t): Ditto.
21361 (vint16mf2x3_t): Ditto.
21362 (vuint16mf2x3_t): Ditto.
21363 (vint16mf2x4_t): Ditto.
21364 (vuint16mf2x4_t): Ditto.
21365 (vint16mf2x5_t): Ditto.
21366 (vuint16mf2x5_t): Ditto.
21367 (vint16mf2x6_t): Ditto.
21368 (vuint16mf2x6_t): Ditto.
21369 (vint16mf2x7_t): Ditto.
21370 (vuint16mf2x7_t): Ditto.
21371 (vint16mf2x8_t): Ditto.
21372 (vuint16mf2x8_t): Ditto.
21373 (vint16m1x2_t): Ditto.
21374 (vuint16m1x2_t): Ditto.
21375 (vint16m1x3_t): Ditto.
21376 (vuint16m1x3_t): Ditto.
21377 (vint16m1x4_t): Ditto.
21378 (vuint16m1x4_t): Ditto.
21379 (vint16m1x5_t): Ditto.
21380 (vuint16m1x5_t): Ditto.
21381 (vint16m1x6_t): Ditto.
21382 (vuint16m1x6_t): Ditto.
21383 (vint16m1x7_t): Ditto.
21384 (vuint16m1x7_t): Ditto.
21385 (vint16m1x8_t): Ditto.
21386 (vuint16m1x8_t): Ditto.
21387 (vint16m2x2_t): Ditto.
21388 (vuint16m2x2_t): Ditto.
21389 (vint16m2x3_t): Ditto.
21390 (vuint16m2x3_t): Ditto.
21391 (vint16m2x4_t): Ditto.
21392 (vuint16m2x4_t): Ditto.
21393 (vint16m4x2_t): Ditto.
21394 (vuint16m4x2_t): Ditto.
21395 (vint32mf2x2_t): Ditto.
21396 (vuint32mf2x2_t): Ditto.
21397 (vint32mf2x3_t): Ditto.
21398 (vuint32mf2x3_t): Ditto.
21399 (vint32mf2x4_t): Ditto.
21400 (vuint32mf2x4_t): Ditto.
21401 (vint32mf2x5_t): Ditto.
21402 (vuint32mf2x5_t): Ditto.
21403 (vint32mf2x6_t): Ditto.
21404 (vuint32mf2x6_t): Ditto.
21405 (vint32mf2x7_t): Ditto.
21406 (vuint32mf2x7_t): Ditto.
21407 (vint32mf2x8_t): Ditto.
21408 (vuint32mf2x8_t): Ditto.
21409 (vint32m1x2_t): Ditto.
21410 (vuint32m1x2_t): Ditto.
21411 (vint32m1x3_t): Ditto.
21412 (vuint32m1x3_t): Ditto.
21413 (vint32m1x4_t): Ditto.
21414 (vuint32m1x4_t): Ditto.
21415 (vint32m1x5_t): Ditto.
21416 (vuint32m1x5_t): Ditto.
21417 (vint32m1x6_t): Ditto.
21418 (vuint32m1x6_t): Ditto.
21419 (vint32m1x7_t): Ditto.
21420 (vuint32m1x7_t): Ditto.
21421 (vint32m1x8_t): Ditto.
21422 (vuint32m1x8_t): Ditto.
21423 (vint32m2x2_t): Ditto.
21424 (vuint32m2x2_t): Ditto.
21425 (vint32m2x3_t): Ditto.
21426 (vuint32m2x3_t): Ditto.
21427 (vint32m2x4_t): Ditto.
21428 (vuint32m2x4_t): Ditto.
21429 (vint32m4x2_t): Ditto.
21430 (vuint32m4x2_t): Ditto.
21431 (vint64m1x2_t): Ditto.
21432 (vuint64m1x2_t): Ditto.
21433 (vint64m1x3_t): Ditto.
21434 (vuint64m1x3_t): Ditto.
21435 (vint64m1x4_t): Ditto.
21436 (vuint64m1x4_t): Ditto.
21437 (vint64m1x5_t): Ditto.
21438 (vuint64m1x5_t): Ditto.
21439 (vint64m1x6_t): Ditto.
21440 (vuint64m1x6_t): Ditto.
21441 (vint64m1x7_t): Ditto.
21442 (vuint64m1x7_t): Ditto.
21443 (vint64m1x8_t): Ditto.
21444 (vuint64m1x8_t): Ditto.
21445 (vint64m2x2_t): Ditto.
21446 (vuint64m2x2_t): Ditto.
21447 (vint64m2x3_t): Ditto.
21448 (vuint64m2x3_t): Ditto.
21449 (vint64m2x4_t): Ditto.
21450 (vuint64m2x4_t): Ditto.
21451 (vint64m4x2_t): Ditto.
21452 (vuint64m4x2_t): Ditto.
21453 (vfloat32mf2x2_t): Ditto.
21454 (vfloat32mf2x3_t): Ditto.
21455 (vfloat32mf2x4_t): Ditto.
21456 (vfloat32mf2x5_t): Ditto.
21457 (vfloat32mf2x6_t): Ditto.
21458 (vfloat32mf2x7_t): Ditto.
21459 (vfloat32mf2x8_t): Ditto.
21460 (vfloat32m1x2_t): Ditto.
21461 (vfloat32m1x3_t): Ditto.
21462 (vfloat32m1x4_t): Ditto.
21463 (vfloat32m1x5_t): Ditto.
21464 (vfloat32m1x6_t): Ditto.
21465 (vfloat32m1x7_t): Ditto.
21466 (vfloat32m1x8_t): Ditto.
21467 (vfloat32m2x2_t): Ditto.
21468 (vfloat32m2x3_t): Ditto.
21469 (vfloat32m2x4_t): Ditto.
21470 (vfloat32m4x2_t): Ditto.
21471 (vfloat64m1x2_t): Ditto.
21472 (vfloat64m1x3_t): Ditto.
21473 (vfloat64m1x4_t): Ditto.
21474 (vfloat64m1x5_t): Ditto.
21475 (vfloat64m1x6_t): Ditto.
21476 (vfloat64m1x7_t): Ditto.
21477 (vfloat64m1x8_t): Ditto.
21478 (vfloat64m2x2_t): Ditto.
21479 (vfloat64m2x3_t): Ditto.
21480 (vfloat64m2x4_t): Ditto.
21481 (vfloat64m4x2_t): Ditto.
21482 * config/riscv/riscv-vector-builtins.h (DEF_RVV_TUPLE_TYPE):
21484 * config/riscv/riscv-vector-switch.def (TUPLE_ENTRY): Ditto.
21485 * config/riscv/riscv.cc (riscv_v_ext_tuple_mode_p): New
21487 (TUPLE_ENTRY): Ditto.
21488 (riscv_v_ext_mode_p): New function.
21489 (riscv_v_adjust_nunits): Add tuple mode adjustment.
21490 (riscv_classify_address): Ditto.
21491 (riscv_binary_cost): Ditto.
21492 (riscv_rtx_costs): Ditto.
21493 (riscv_secondary_memory_needed): Ditto.
21494 (riscv_hard_regno_nregs): Ditto.
21495 (riscv_hard_regno_mode_ok): Ditto.
21496 (riscv_vector_mode_supported_p): Ditto.
21497 (riscv_regmode_natural_size): Ditto.
21498 (riscv_array_mode): New function.
21499 (TARGET_ARRAY_MODE): New target hook.
21500 * config/riscv/riscv.md: Add tuple modes.
21501 * config/riscv/vector-iterators.md: Ditto.
21502 * config/riscv/vector.md (mov<mode>): Add tuple modes data
21504 (*mov<VT:mode>_<P:mode>): Ditto.
21506 2023-05-03 Richard Biener <rguenther@suse.de>
21508 * cse.cc (cse_insn): Track an equivalence to the destination
21509 separately and delay using src_related for it.
21511 2023-05-03 Richard Biener <rguenther@suse.de>
21513 * cse.cc (HASH): Turn into inline function and mix
21514 in another HASH_SHIFT bits.
21515 (SAFE_HASH): Likewise.
21517 2023-05-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21520 * config/aarch64/aarch64-simd.md (aarch64_<sur>h<addsub><mode>): Rename to...
21521 (aarch64_<sur>h<addsub><mode><vczle><vczbe>): ... This.
21523 2023-05-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21526 * config/aarch64/aarch64-simd.md (add<mode>3): Rename to...
21527 (add<mode>3<vczle><vczbe>): ... This.
21528 (sub<mode>3): Rename to...
21529 (sub<mode>3<vczle><vczbe>): ... This.
21530 (mul<mode>3): Rename to...
21531 (mul<mode>3<vczle><vczbe>): ... This.
21532 (*div<mode>3): Rename to...
21533 (*div<mode>3<vczle><vczbe>): ... This.
21534 (neg<mode>2): Rename to...
21535 (neg<mode>2<vczle><vczbe>): ... This.
21536 (abs<mode>2): Rename to...
21537 (abs<mode>2<vczle><vczbe>): ... This.
21538 (<frint_pattern><mode>2): Rename to...
21539 (<frint_pattern><mode>2<vczle><vczbe>): ... This.
21540 (<fmaxmin><mode>3): Rename to...
21541 (<fmaxmin><mode>3<vczle><vczbe>): ... This.
21542 (*sqrt<mode>2): Rename to...
21543 (*sqrt<mode>2<vczle><vczbe>): ... This.
21545 2023-05-03 Kito Cheng <kito.cheng@sifive.com>
21547 * doc/md.texi (RISC-V): Add vr, vm, vd constarint.
21549 2023-05-03 Martin Liska <mliska@suse.cz>
21551 PR tree-optimization/109693
21552 * value-range-storage.cc (vrange_allocator::vrange_allocator):
21553 Remove unused field.
21554 * value-range-storage.h: Likewise.
21556 2023-05-02 Andrew Pinski <apinski@marvell.com>
21558 * tree-ssa-phiopt.cc (move_stmt): New function.
21559 (match_simplify_replacement): Use move_stmt instead
21560 of the inlined version.
21562 2023-05-02 Andrew Pinski <apinski@marvell.com>
21564 * match.pd (a != 0 ? CLRSB(a) : CST -> CLRSB(a)): New
21567 2023-05-02 Andrew Pinski <apinski@marvell.com>
21569 PR tree-optimization/109702
21570 * match.pd: Fix "a != 0 ? FUNC(a) : CST" patterns
21571 for FUNC of POPCOUNT BSWAP FFS PARITY CLZ and CTZ.
21573 2023-05-02 Andrew Pinski <apinski@marvell.com>
21576 * config/aarch64/aarch64.md (*cmov<mode>_insn_m1): New
21577 insn_and_split pattern.
21579 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
21581 * config/riscv/sync.md (atomic_load<mode>): Implement atomic
21584 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
21586 * config/riscv/sync.md (mem_thread_fence_1): Change fence
21587 depending on the given memory model.
21589 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
21591 * config/riscv/riscv-protos.h (riscv_union_memmodels): Expose
21592 riscv_union_memmodels function to sync.md.
21593 * config/riscv/riscv.cc (riscv_union_memmodels): Add function to
21594 get the union of two memmodels in sync.md.
21595 (riscv_print_operand): Add %I and %J flags that output the
21596 optimal LR/SC flag bits for a given memory model.
21597 * config/riscv/sync.md: Remove static .aqrl bits on LR op/.rl
21598 bits on SC op and replace with optimized %I, %J flags.
21600 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
21602 * config/riscv/riscv.cc
21603 (riscv_memmodel_needs_amo_release): Change function name.
21604 (riscv_print_operand): Remove unneeded %F case.
21605 * config/riscv/sync.md: Remove unneeded fences.
21607 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
21610 * config/riscv/sync.md (atomic_store<mode>): Use simple store
21611 instruction in combination with fence(s).
21613 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
21615 * config/riscv/riscv.cc (riscv_print_operand): Change behavior
21616 of %A to include release bits.
21618 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
21620 * config/riscv/sync.md (atomic_cas_value_strong<mode>): Change
21621 FENCE/LR.aq/SC.aq into sequentially consistent LR.aqrl/SC.rl
21624 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
21626 * config/riscv/sync.md: Change LR.aq/SC.rl pairs into
21627 sequentially consistent LR.aqrl/SC.rl pairs.
21629 2023-05-02 Patrick O'Neill <patrick@rivosinc.com>
21631 * config/riscv/riscv.cc: Remove MEMMODEL_SYNC_* cases and
21632 sanitize memmodel input with memmodel_base.
21634 2023-05-02 Yanzhang Wang <yanzhang.wang@intel.com>
21635 Pan Li <pan2.li@intel.com>
21638 * config/riscv/vector-iterators.md: Support VNx2HI and VNX4DI when MIN_VLEN >= 128.
21640 2023-05-02 Romain Naour <romain.naour@gmail.com>
21642 * config/riscv/genrvv-type-indexer.cc: Use log2 from the C header, without
21645 2023-05-02 Martin Liska <mliska@suse.cz>
21647 * doc/invoke.texi: Update documentation based on param.opt file.
21649 2023-05-02 Richard Biener <rguenther@suse.de>
21651 PR tree-optimization/109672
21652 * tree-vect-stmts.cc (vectorizable_operation): For plus,
21653 minus and negate always check the vector mode is word mode.
21655 2023-05-01 Andrew Pinski <apinski@marvell.com>
21657 * tree-ssa-phiopt.cc: Update comment about
21658 how the transformation are implemented.
21660 2023-05-01 Jeff Law <jlaw@ventanamicro>
21662 * config/stormy16/stormy16.cc (TARGET_LRA_P): Remove defintion.
21664 2023-05-01 Jeff Law <jlaw@ventanamicro>
21666 * config/cris/cris.cc (TARGET_LRA_P): Remove.
21667 * config/epiphany/epiphany.cc (TARGET_LRA_P): Remove.
21668 * config/iq2000/iq2000.cc (TARGET_LRA_P): Remove.
21669 * config/m32r/m32r.cc (TARGET_LRA_P): Remove.
21670 * config/microblaze/microblaze.cc (TARGET_LRA_P): Remove.
21671 * config/mmix/mmix.cc (TARGET_LRA_P): Remove.
21673 2023-05-01 Rasmus Villemoes <rasmus.villemoes@prevas.dk>
21675 * print-tree.h (PRINT_DECL_REMAP_DEBUG): New flag.
21676 * print-tree.cc (print_decl_identifier): Implement it.
21677 * toplev.cc (output_stack_usage_1): Use it.
21679 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
21681 * value-range.h (class int_range): Remove gt_ggc_mx and gt_pch_nx
21684 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
21686 * value-range.h (irange::set_nonzero): Inline.
21688 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
21690 * gimple-range-op.cc (cfn_ffs::fold_range): Use the correct
21692 * gimple-ssa-warn-alloca.cc (alloca_call_type): Use <2> for
21693 invalid_range, as it is an inverse range.
21694 * tree-vrp.cc (find_case_label_range): Avoid trees.
21695 * value-range.cc (irange::irange_set): Delete.
21696 (irange::irange_set_1bit_anti_range): Delete.
21697 (irange::irange_set_anti_range): Delete.
21698 (irange::set): Cleanup.
21699 * value-range.h (class irange): Remove irange_set,
21700 irange_set_anti_range, irange_set_1bit_anti_range.
21701 (irange::set_undefined): Remove set to m_type.
21703 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
21705 * range-op.cc (update_known_bitmask): Adjust for irange containing
21706 wide_ints internally.
21707 * tree-ssanames.cc (set_nonzero_bits): Same.
21708 * tree-ssanames.h (set_nonzero_bits): Same.
21709 * value-range-storage.cc (irange_storage::set_irange): Same.
21710 (irange_storage::get_irange): Same.
21711 * value-range.cc (irange::operator=): Same.
21712 (irange::irange_set): Same.
21713 (irange::irange_set_1bit_anti_range): Same.
21714 (irange::irange_set_anti_range): Same.
21715 (irange::set): Same.
21716 (irange::verify_range): Same.
21717 (irange::contains_p): Same.
21718 (irange::irange_single_pair_union): Same.
21719 (irange::union_): Same.
21720 (irange::irange_contains_p): Same.
21721 (irange::intersect): Same.
21722 (irange::invert): Same.
21723 (irange::set_range_from_nonzero_bits): Same.
21724 (irange::set_nonzero_bits): Same.
21725 (mask_to_wi): Same.
21726 (irange::intersect_nonzero_bits): Same.
21727 (irange::union_nonzero_bits): Same.
21730 (tree_range): Same.
21731 (range_tests_strict_enum): Same.
21732 (range_tests_misc): Same.
21733 (range_tests_nonzero_bits): Same.
21734 * value-range.h (irange::type): Same.
21735 (irange::varying_compatible_p): Same.
21736 (irange::irange): Same.
21737 (int_range::int_range): Same.
21738 (irange::set_undefined): Same.
21739 (irange::set_varying): Same.
21740 (irange::lower_bound): Same.
21741 (irange::upper_bound): Same.
21743 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
21745 * gimple-range-fold.cc (tree_lower_bound): Delete.
21746 (tree_upper_bound): Delete.
21747 (vrp_val_max): Delete.
21748 (vrp_val_min): Delete.
21749 (fold_using_range::range_of_ssa_name_with_loop_info): Call
21750 range_of_var_in_loop.
21751 * vr-values.cc (valid_value_p): Delete.
21752 (fix_overflow): Delete.
21753 (get_scev_info): New.
21754 (bounds_of_var_in_loop): Refactor into...
21755 (induction_variable_may_overflow_p): ...this,
21756 (range_from_loop_direction): ...and this,
21757 (range_of_var_in_loop): ...and this.
21758 * vr-values.h (bounds_of_var_in_loop): Delete.
21759 (range_of_var_in_loop): New.
21761 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
21763 * gimple-range-fold.cc (adjust_pointer_diff_expr): Rewrite with
21765 (vrp_val_max): New.
21766 (vrp_val_min): New.
21767 * gimple-range-op.cc (cfn_strlen::fold_range): Use irange_val_*.
21768 * range-op.cc (max_limit): Same.
21770 (plus_minus_ranges): Same.
21771 (operator_rshift::op1_range): Same.
21772 (operator_cast::inside_domain_p): Same.
21773 * value-range.cc (vrp_val_is_max): Delete.
21774 (vrp_val_is_min): Delete.
21775 (range_tests_misc): Use irange_val_*.
21776 * value-range.h (vrp_val_is_min): Delete.
21777 (vrp_val_is_max): Delete.
21778 (vrp_val_max): Delete.
21779 (irange_val_min): New.
21780 (vrp_val_min): Delete.
21781 (irange_val_max): New.
21782 * vr-values.cc (check_for_binary_op_overflow): Use irange_val_*.
21784 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
21786 * fold-const.cc (expr_not_equal_to): Convert to irange wide_int API.
21787 * gimple-fold.cc (size_must_be_zero_p): Same.
21788 * gimple-loop-versioning.cc
21789 (loop_versioning::prune_loop_conditions): Same.
21790 * gimple-range-edge.cc (gcond_edge_range): Same.
21791 (gimple_outgoing_range::calc_switch_ranges): Same.
21792 * gimple-range-fold.cc (adjust_imagpart_expr): Same.
21793 (adjust_realpart_expr): Same.
21794 (fold_using_range::range_of_address): Same.
21795 (fold_using_range::relation_fold_and_or): Same.
21796 * gimple-range-gori.cc (gori_compute::gori_compute): Same.
21797 (range_is_either_true_or_false): Same.
21798 * gimple-range-op.cc (cfn_toupper_tolower::get_letter_range): Same.
21799 (cfn_clz::fold_range): Same.
21800 (cfn_ctz::fold_range): Same.
21801 * gimple-range-tests.cc (class test_expr_eval): Same.
21802 * gimple-ssa-warn-alloca.cc (alloca_call_type): Same.
21803 * ipa-cp.cc (ipa_value_range_from_jfunc): Same.
21804 (propagate_vr_across_jump_function): Same.
21805 (decide_whether_version_node): Same.
21806 * ipa-prop.cc (ipa_get_value_range): Same.
21807 * ipa-prop.h (ipa_range_set_and_normalize): Same.
21808 * range-op.cc (get_shift_range): Same.
21809 (value_range_from_overflowed_bounds): Same.
21810 (value_range_with_overflow): Same.
21811 (create_possibly_reversed_range): Same.
21812 (equal_op1_op2_relation): Same.
21813 (not_equal_op1_op2_relation): Same.
21814 (lt_op1_op2_relation): Same.
21815 (le_op1_op2_relation): Same.
21816 (gt_op1_op2_relation): Same.
21817 (ge_op1_op2_relation): Same.
21818 (operator_mult::op1_range): Same.
21819 (operator_exact_divide::op1_range): Same.
21820 (operator_lshift::op1_range): Same.
21821 (operator_rshift::op1_range): Same.
21822 (operator_cast::op1_range): Same.
21823 (operator_logical_and::fold_range): Same.
21824 (set_nonzero_range_from_mask): Same.
21825 (operator_bitwise_or::op1_range): Same.
21826 (operator_bitwise_xor::op1_range): Same.
21827 (operator_addr_expr::fold_range): Same.
21828 (pointer_plus_operator::wi_fold): Same.
21829 (pointer_or_operator::op1_range): Same.
21836 (range_op_cast_tests): Same.
21837 (range_op_lshift_tests): Same.
21838 (range_op_rshift_tests): Same.
21839 (range_op_bitwise_and_tests): Same.
21840 (range_relational_tests): Same.
21841 * range.cc (range_zero): Same.
21842 (range_nonzero): Same.
21843 * range.h (range_true): Same.
21844 (range_false): Same.
21845 (range_true_and_false): Same.
21846 * tree-data-ref.cc (split_constant_offset_1): Same.
21847 * tree-ssa-loop-ch.cc (entry_loop_condition_is_static): Same.
21848 * tree-ssa-loop-unswitch.cc (struct unswitch_predicate): Same.
21849 (find_unswitching_predicates_for_bb): Same.
21850 * tree-ssa-phiopt.cc (value_replacement): Same.
21851 * tree-ssa-threadbackward.cc
21852 (back_threader::find_taken_edge_cond): Same.
21853 * tree-ssanames.cc (ssa_name_has_boolean_range): Same.
21854 * tree-vrp.cc (find_case_label_range): Same.
21855 * value-query.cc (range_query::get_tree_range): Same.
21856 * value-range.cc (irange::set_nonnegative): Same.
21857 (frange::contains_p): Same.
21858 (frange::singleton_p): Same.
21859 (frange::internal_singleton_p): Same.
21860 (irange::irange_set): Same.
21861 (irange::irange_set_1bit_anti_range): Same.
21862 (irange::irange_set_anti_range): Same.
21863 (irange::set): Same.
21864 (irange::operator==): Same.
21865 (irange::singleton_p): Same.
21866 (irange::contains_p): Same.
21867 (irange::set_range_from_nonzero_bits): Same.
21868 (DEFINE_INT_RANGE_INSTANCE): Same.
21878 (range_uint128): New.
21879 (range_uchar): New.
21881 (build_range3): Convert to irange wide_int API.
21882 (range_tests_irange3): Same.
21883 (range_tests_int_range_max): Same.
21884 (range_tests_strict_enum): Same.
21885 (range_tests_misc): Same.
21886 (range_tests_nonzero_bits): Same.
21887 (range_tests_nan): Same.
21888 (range_tests_signed_zeros): Same.
21889 * value-range.h (Value_Range::Value_Range): Same.
21890 (irange::set): Same.
21891 (irange::nonzero_p): Same.
21892 (irange::contains_p): Same.
21893 (range_includes_zero_p): Same.
21894 (irange::set_nonzero): Same.
21895 (irange::set_zero): Same.
21896 (contains_zero_p): Same.
21897 (frange::contains_p): Same.
21899 (simplify_using_ranges::op_with_boolean_value_range_p): Same.
21900 (bounds_of_var_in_loop): Same.
21901 (simplify_using_ranges::legacy_fold_cond_overflow): Same.
21903 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
21905 * value-range.cc (irange::irange_union): Rename to...
21906 (irange::union_): ...this.
21907 (irange::irange_intersect): Rename to...
21908 (irange::intersect): ...this.
21909 * value-range.h (irange::union_): Delete.
21910 (irange::intersect): Delete.
21912 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
21914 * vr-values.cc (bounds_of_var_in_loop): Convert to irange API.
21916 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
21918 * vr-values.cc (check_for_binary_op_overflow): Tidy up by using
21920 (compare_ranges): Delete.
21921 (compare_range_with_value): Delete.
21922 (bounds_of_var_in_loop): Tidy up by using ranger API.
21923 (simplify_using_ranges::fold_cond_with_ops): Cleanup and rename
21924 from vrp_evaluate_conditional_warnv_with_ops_using_ranges.
21925 (simplify_using_ranges::legacy_fold_cond_overflow): Remove
21926 strict_overflow_p and only_ranges.
21927 (simplify_using_ranges::legacy_fold_cond): Adjust call to
21928 legacy_fold_cond_overflow.
21929 (simplify_using_ranges::simplify_abs_using_ranges): Adjust for
21931 (range_fits_type_p): Rename value_range to irange.
21932 * vr-values.h (range_fits_type_p): Adjust prototype.
21934 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
21936 * value-range.cc (irange::irange_set_anti_range): Remove uses of
21937 tree_lower_bound and tree_upper_bound.
21938 (irange::verify_range): Same.
21939 (irange::operator==): Same.
21940 (irange::singleton_p): Same.
21941 * value-range.h (irange::tree_lower_bound): Delete.
21942 (irange::tree_upper_bound): Delete.
21943 (irange::lower_bound): Delete.
21944 (irange::upper_bound): Delete.
21945 (irange::zero_p): Remove uses of tree_lower_bound and
21948 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
21950 * tree-ssa-loop-niter.cc (refine_value_range_using_guard): Remove
21952 (determine_value_range): Same.
21953 (record_nonwrapping_iv): Same.
21954 (infer_loop_bounds_from_signedness): Same.
21955 (scev_var_range_cant_overflow): Same.
21956 * tree-vrp.cc (operand_less_p): Delete.
21957 * tree-vrp.h (operand_less_p): Delete.
21958 * value-range.cc (get_legacy_range): Remove uses of deprecated API.
21959 (irange::value_inside_range): Delete.
21960 * value-range.h (vrange::kind): Delete.
21961 (irange::num_pairs): Remove check of m_kind.
21962 (irange::min): Delete.
21963 (irange::max): Delete.
21965 2023-05-01 Aldy Hernandez <aldyh@redhat.com>
21967 * gimple-fold.cc (maybe_fold_comparisons_from_match_pd): Adjust
21968 for vrange_storage.
21969 * gimple-range-cache.cc (sbr_vector::sbr_vector): Same.
21970 (sbr_vector::grow): Same.
21971 (sbr_vector::set_bb_range): Same.
21972 (sbr_vector::get_bb_range): Same.
21973 (sbr_sparse_bitmap::sbr_sparse_bitmap): Same.
21974 (sbr_sparse_bitmap::set_bb_range): Same.
21975 (sbr_sparse_bitmap::get_bb_range): Same.
21976 (block_range_cache::block_range_cache): Same.
21977 (ssa_global_cache::ssa_global_cache): Same.
21978 (ssa_global_cache::get_global_range): Same.
21979 (ssa_global_cache::set_global_range): Same.
21980 * gimple-range-cache.h: Same.
21981 * gimple-range-edge.cc
21982 (gimple_outgoing_range::gimple_outgoing_range): Same.
21983 (gimple_outgoing_range::switch_edge_range): Same.
21984 (gimple_outgoing_range::calc_switch_ranges): Same.
21985 * gimple-range-edge.h: Same.
21986 * gimple-range-infer.cc
21987 (infer_range_manager::infer_range_manager): Same.
21988 (infer_range_manager::get_nonzero): Same.
21989 (infer_range_manager::maybe_adjust_range): Same.
21990 (infer_range_manager::add_range): Same.
21991 * gimple-range-infer.h: Rename obstack_vrange_allocator to
21993 * tree-core.h (struct irange_storage_slot): Remove.
21994 (struct tree_ssa_name): Remove irange_info and frange_info. Make
21995 range_info a pointer to vrange_storage.
21996 * tree-ssanames.cc (range_info_fits_p): Adjust for vrange_storage.
21997 (range_info_alloc): Same.
21998 (range_info_free): Same.
21999 (range_info_get_range): Same.
22000 (range_info_set_range): Same.
22001 (get_nonzero_bits): Same.
22002 * value-query.cc (get_ssa_name_range_info): Same.
22003 * value-range-storage.cc (class vrange_internal_alloc): New.
22004 (class vrange_obstack_alloc): New.
22005 (class vrange_ggc_alloc): New.
22006 (vrange_allocator::vrange_allocator): New.
22007 (vrange_allocator::~vrange_allocator): New.
22008 (vrange_storage::alloc_slot): New.
22009 (vrange_allocator::alloc): New.
22010 (vrange_allocator::free): New.
22011 (vrange_allocator::clone): New.
22012 (vrange_allocator::clone_varying): New.
22013 (vrange_allocator::clone_undefined): New.
22014 (vrange_storage::alloc): New.
22015 (vrange_storage::set_vrange): Remove slot argument.
22016 (vrange_storage::get_vrange): Same.
22017 (vrange_storage::fits_p): Same.
22018 (vrange_storage::equal_p): New.
22019 (irange_storage::write_lengths_address): New.
22020 (irange_storage::lengths_address): New.
22021 (irange_storage_slot::alloc_slot): Remove.
22022 (irange_storage::alloc): New.
22023 (irange_storage_slot::irange_storage_slot): Remove.
22024 (irange_storage::irange_storage): New.
22025 (write_wide_int): New.
22026 (irange_storage_slot::set_irange): Remove.
22027 (irange_storage::set_irange): New.
22028 (read_wide_int): New.
22029 (irange_storage_slot::get_irange): Remove.
22030 (irange_storage::get_irange): New.
22031 (irange_storage_slot::size): Remove.
22032 (irange_storage::equal_p): New.
22033 (irange_storage_slot::num_wide_ints_needed): Remove.
22034 (irange_storage::size): New.
22035 (irange_storage_slot::fits_p): Remove.
22036 (irange_storage::fits_p): New.
22037 (irange_storage_slot::dump): Remove.
22038 (irange_storage::dump): New.
22039 (frange_storage_slot::alloc_slot): Remove.
22040 (frange_storage::alloc): New.
22041 (frange_storage_slot::set_frange): Remove.
22042 (frange_storage::set_frange): New.
22043 (frange_storage_slot::get_frange): Remove.
22044 (frange_storage::get_frange): New.
22045 (frange_storage_slot::fits_p): Remove.
22046 (frange_storage::equal_p): New.
22047 (frange_storage::fits_p): New.
22048 (ggc_vrange_allocator): New.
22049 (ggc_alloc_vrange_storage): New.
22050 * value-range-storage.h (class vrange_storage): Rewrite.
22051 (class irange_storage): Rewrite.
22052 (class frange_storage): Rewrite.
22053 (class obstack_vrange_allocator): Remove.
22054 (class ggc_vrange_allocator): Remove.
22055 (vrange_allocator::alloc_vrange): Remove.
22056 (vrange_allocator::alloc_irange): Remove.
22057 (vrange_allocator::alloc_frange): Remove.
22058 (ggc_alloc_vrange_storage): New.
22059 * value-range.h (class irange): Rename vrange_allocator to
22061 (class frange): Same.
22063 2023-04-30 Roger Sayle <roger@nextmovesoftware.com>
22065 * config/stormy16/stormy16.md (neghi2): Rewrite pattern using
22066 inc to avoid clobbering the carry flag.
22068 2023-04-30 Andrew Pinski <apinski@marvell.com>
22070 * match.pd: Add patterns for "a != 0 ? FUNC(a) : CST"
22071 for FUNC of POPCOUNT BSWAP FFS PARITY CLZ and CTZ.
22073 2023-04-30 Andrew Pinski <apinski@marvell.com>
22075 * tree-ssa-phiopt.cc (empty_bb_or_one_feeding_into_p):
22076 Allow some builtin/internal function calls which
22077 are known not to trap/throw.
22078 (phiopt_worker::match_simplify_replacement):
22079 Use name instead of getting the lhs again.
22081 2023-04-30 Joakim Nohlgård <joakim@nohlgard.se>
22083 * configure: Regenerate.
22084 * configure.ac: Use ld -r in the check for HAVE_LD_RO_RW_SECTION_MIXING
22086 2023-04-29 Hans-Peter Nilsson <hp@axis.com>
22088 * reload1.cc (emit_insn_if_valid_for_reload_1): Rename from
22089 emit_insn_if_valid_for_reload.
22090 (emit_insn_if_valid_for_reload): Call new helper, and if a SET fails
22091 to be recognized, also try emitting a parallel that clobbers
22092 TARGET_FLAGS_REGNUM, as applicable.
22094 2023-04-29 Roger Sayle <roger@nextmovesoftware.com>
22096 * config/stormy16/stormy16.md (neghi2): Convert from a define_expand
22098 (*rotatehi_1): New define_insn for efficient 2 insn sequence.
22099 (*rotatehi_8, *rotaterthi_8): New define_insn to emit a swpb.
22101 2023-04-29 Roger Sayle <roger@nextmovesoftware.com>
22103 * config/stormy16/stormy16.md (any_lshift): New code iterator.
22104 (any_or_plus): Likewise.
22105 (any_rotate): Likewise.
22106 (*<any_lshift>_and_internal): New define_insn_and_split to
22107 recognize a logical shift followed by an AND, and split it
22108 again after reload.
22109 (*swpn): New define_insn matching xstormy16's swpn.
22110 (*swpn_zext): New define_insn recognizing swpn followed by
22111 zero_extendqihi2, i.e. with the high byte set to zero.
22112 (*swpn_sext): Likewise, for swpn followed by cbw.
22113 (*swpn_sext_2): Likewise, for an alternate RTL form.
22114 (*swpn_zext_ior): A pre-reload splitter so that an swpn+zext+ior
22115 sequence is split in the correct place to recognize the *swpn_zext
22116 followed by any_or_plus (ior, xor or plus) instruction.
22118 2023-04-29 Mikael Pettersson <mikpelinux@gmail.com>
22121 * config.gcc (vax-*-linux*): Add glibc-stdint.h.
22122 (lm32-*-uclinux*): Likewise.
22124 2023-04-29 Fei Gao <gaofei@eswincomputing.com>
22126 * config/riscv/riscv.cc (riscv_avoid_save_libcall): helper function
22127 for riscv_use_save_libcall.
22128 (riscv_use_save_libcall): call riscv_avoid_save_libcall.
22129 (riscv_compute_frame_info): restructure to decouple stack allocation
22130 for rv32e w/o save-restore.
22132 2023-04-28 Eugene Rozenfeld <erozen@microsoft.com>
22134 * doc/install.texi: Fix documentation typo
22136 2023-04-28 Matevos Mehrabyan <matevosmehrabyan@gmail.com>
22138 * config/riscv/iterators.md (only_div, paired_mod): New iterators.
22139 (u): Add div/udiv cases.
22140 * config/riscv/riscv-protos.h (riscv_use_divmod_expander): Prototype.
22141 * config/riscv/riscv.cc (struct riscv_tune_param): Add field for
22143 (rocket_tune_info, sifive_7_tune_info): Initialize new field.
22144 (thead_c906_tune_info): Likewise.
22145 (optimize_size_tune_info): Likewise.
22146 (riscv_use_divmod_expander): New function.
22147 * config/riscv/riscv.md (<u>divmod<mode>4): New expander.
22149 2023-04-28 Karen Sargsyan <karen1999411@gmail.com>
22151 * config/riscv/bitmanip.md: Added clmulr instruction.
22152 * config/riscv/riscv-builtins.cc (AVAIL): Add new.
22153 * config/riscv/riscv.md: (UNSPEC_CLMULR): Add new unspec type.
22155 * config/riscv/riscv-cmo.def: Added built-in function for clmulr.
22156 * config/riscv/crypto.md: Move clmul[h] instructions to bitmanip.md.
22157 * config/riscv/riscv-scalar-crypto.def: Move clmul[h] built-in
22158 functions to riscv-cmo.def.
22159 * config/riscv/generic.md: Add clmul to list of instructions
22160 using the generic_imul reservation.
22162 2023-04-28 Jivan Hakobyan <jivanhakobyan9@gmail.com>
22164 * config/riscv/bitmanip.md: Added expanders for minu/maxu instructions
22166 2023-04-28 Andrew Pinski <apinski@marvell.com>
22168 PR tree-optimization/100958
22169 * tree-ssa-phiopt.cc (two_value_replacement): Remove.
22170 (pass_phiopt::execute): Don't call two_value_replacement.
22171 * match.pd (a !=/== CST1 ? CST2 : CST3): Add pattern to
22172 handle what two_value_replacement did.
22174 2023-04-28 Andrew Pinski <apinski@marvell.com>
22176 * match.pd: Add patterns for
22177 "(A CMP B) ? MIN/MAX<A, C> : MIN/MAX <B, C>".
22179 2023-04-28 Andrew Pinski <apinski@marvell.com>
22181 * match.pd: Factor out the deciding the min/max from
22182 the "(cond (cmp (convert1? x) c1) (convert2? x) c2)"
22184 * fold-const.cc (minmax_from_comparison): this new function.
22185 * fold-const.h (minmax_from_comparison): New prototype.
22187 2023-04-28 Roger Sayle <roger@nextmovesoftware.com>
22189 PR rtl-optimization/109476
22190 * lower-subreg.cc: Include explow.h for force_reg.
22191 (find_decomposable_shift_zext): Pass an additional SPEED_P argument.
22192 If decomposing a suitable LSHIFTRT and we're not splitting
22193 ZERO_EXTEND (based on the current SPEED_P), then use a ZERO_EXTEND
22194 instead of setting a high part SUBREG to zero, which helps combine.
22195 (decompose_multiword_subregs): Update call to resolve_shift_zext.
22197 2023-04-28 Richard Biener <rguenther@suse.de>
22199 * tree-vect-data-refs.cc (vect_analyze_data_refs): Always
22201 * tree-vect-stmts.cc (vect_model_store_cost): Pass in the
22202 gather-scatter info and cost emulated scatters accordingly.
22203 (get_load_store_type): Support emulated scatters.
22204 (vectorizable_store): Likewise. Emulate them by extracting
22205 scalar offsets and data, doing scalar stores.
22207 2023-04-28 Richard Biener <rguenther@suse.de>
22209 * config/i386/i386.cc (ix86_vector_costs::add_stmt_cost):
22210 Tame down element extracts and scalar loads for gather/scatter
22211 similar to elementwise strided accesses.
22213 2023-04-28 Pan Li <pan2.li@intel.com>
22214 kito-cheng <kito.cheng@sifive.com>
22216 * config/riscv/vector.md: Add new define split to perform
22217 the simplification.
22219 2023-04-28 Richard Biener <rguenther@suse.de>
22222 * ipa-param-manipulation.cc
22223 (ipa_param_body_adjustments::modify_expression): Allow
22224 conversion of a register to a non-register type. Elide
22225 conversions inside BIT_FIELD_REFs.
22227 2023-04-28 Richard Biener <rguenther@suse.de>
22229 PR tree-optimization/109644
22230 * tree-cfg.cc (verify_types_in_gimple_reference): Check
22231 register constraints on the outermost VIEW_CONVERT_EXPR
22232 only. Do not allow register or invariant bases on
22233 multi-level or possibly variable index handled components.
22235 2023-04-28 Richard Biener <rguenther@suse.de>
22237 * gimplify.cc (gimplify_compound_lval): When there's a
22238 non-register type produced by one of the handled component
22239 operations make sure we get a non-register base.
22241 2023-04-28 Richard Biener <rguenther@suse.de>
22243 PR tree-optimization/108752
22244 * tree-vect-generic.cc (build_replicated_const): Rename
22245 to build_replicated_int_cst and move to tree.{h,cc}.
22246 (do_plus_minus): Adjust.
22247 (do_negate): Likewise.
22248 * tree-vect-stmts.cc (vectorizable_operation): Emit emulated
22249 arithmetic vector operations in lowered form.
22250 * tree.h (build_replicated_int_cst): Declare.
22251 * tree.cc (build_replicated_int_cst): Moved from
22252 tree-vect-generic.cc build_replicated_const.
22254 2023-04-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22257 * config/aarch64/aarch64-simd.md (aarch64_rbit<mode>): Rename to...
22258 (aarch64_rbit<mode><vczle><vczbe>): ... This.
22259 (neg<mode>2): Rename to...
22260 (neg<mode>2<vczle><vczbe>): ... This.
22261 (abs<mode>2): Rename to...
22262 (abs<mode>2<vczle><vczbe>): ... This.
22263 (aarch64_abs<mode>): Rename to...
22264 (aarch64_abs<mode><vczle><vczbe>): ... This.
22265 (one_cmpl<mode>2): Rename to...
22266 (one_cmpl<mode>2<vczle><vczbe>): ... This.
22267 (clrsb<mode>2): Rename to...
22268 (clrsb<mode>2<vczle><vczbe>): ... This.
22269 (clz<mode>2): Rename to...
22270 (clz<mode>2<vczle><vczbe>): ... This.
22271 (popcount<mode>2): Rename to...
22272 (popcount<mode>2<vczle><vczbe>): ... This.
22274 2023-04-28 Jakub Jelinek <jakub@redhat.com>
22276 * gimple-range-op.cc (class cfn_sqrt): New type.
22277 (op_cfn_sqrt): New variable.
22278 (gimple_range_op_handler::maybe_builtin_call): Handle
22279 CASE_CFN_SQRT{,_FN}.
22281 2023-04-28 Aldy Hernandez <aldyh@redhat.com>
22282 Jakub Jelinek <jakub@redhat.com>
22284 * value-range.h (frange_nextafter): Declare.
22285 * gimple-range-op.cc (class cfn_sincos): New.
22286 (op_cfn_sin, op_cfn_cos): New variables.
22287 (gimple_range_op_handler::maybe_builtin_call): Handle
22288 CASE_CFN_{SIN,COS}{,_FN}.
22290 2023-04-28 Jakub Jelinek <jakub@redhat.com>
22292 * target.def (libm_function_max_error): New target hook.
22293 * doc/tm.texi.in (TARGET_LIBM_FUNCTION_MAX_ERROR): Add.
22294 * doc/tm.texi: Regenerated.
22295 * targhooks.h (default_libm_function_max_error,
22296 glibc_linux_libm_function_max_error): Declare.
22297 * targhooks.cc: Include case-cfn-macros.h.
22298 (default_libm_function_max_error,
22299 glibc_linux_libm_function_max_error): New functions.
22300 * config/linux.h (TARGET_LIBM_FUNCTION_MAX_ERROR): Redefine.
22301 * config/linux-protos.h (linux_libm_function_max_error): Declare.
22302 * config/linux.cc: Include target.h and targhooks.h.
22303 (linux_libm_function_max_error): New function.
22304 * config/arc/arc.cc: Include targhooks.h and case-cfn-macros.h.
22305 (arc_libm_function_max_error): New function.
22306 (TARGET_LIBM_FUNCTION_MAX_ERROR): Redefine.
22307 * config/i386/i386.cc (ix86_libc_has_fast_function): Formatting fix.
22308 (ix86_libm_function_max_error): New function.
22309 (TARGET_LIBM_FUNCTION_MAX_ERROR): Redefine.
22310 * config/rs6000/rs6000-protos.h
22311 (rs6000_linux_libm_function_max_error): Declare.
22312 * config/rs6000/rs6000-linux.cc: Include target.h, targhooks.h, tree.h
22313 and case-cfn-macros.h.
22314 (rs6000_linux_libm_function_max_error): New function.
22315 * config/rs6000/linux.h (TARGET_LIBM_FUNCTION_MAX_ERROR): Redefine.
22316 * config/rs6000/linux64.h (TARGET_LIBM_FUNCTION_MAX_ERROR): Redefine.
22317 * config/or1k/or1k.cc: Include targhooks.h and case-cfn-macros.h.
22318 (or1k_libm_function_max_error): New function.
22319 (TARGET_LIBM_FUNCTION_MAX_ERROR): Redefine.
22321 2023-04-28 Alexandre Oliva <oliva@adacore.com>
22323 * gimple-harden-conditionals.cc (insert_edge_check_and_trap):
22324 Move detach value calls...
22325 (pass_harden_conditional_branches::execute): ... here.
22326 (pass_harden_compares::execute): Detach values before
22329 2023-04-27 Andrew Stubbs <ams@codesourcery.com>
22331 * config/gcn/gcn-valu.md (cmul<conj_op><mode>3): Use gcn_gen_undef.
22332 (cml<addsub_as><mode>4): Likewise.
22333 (vec_addsub<mode>3): Likewise.
22334 (cadd<rot><mode>3): Likewise.
22335 (vec_fmaddsub<mode>4): Likewise.
22336 (vec_fmsubadd<mode>4): Likewise, and use sub for the odd lanes.
22338 2023-04-27 Andrew Pinski <apinski@marvell.com>
22340 * tree-ssa-phiopt.cc (phiopt_early_allow): Allow for
22341 up to 2 min/max expressions in the sequence/match code.
22343 2023-04-27 Andrew Pinski <apinski@marvell.com>
22345 * rtlanal.cc (may_trap_p_1): Treat SMIN/SMAX similar as
22347 * tree-eh.cc (operation_could_trap_helper_p): Treate
22348 MIN_EXPR/MAX_EXPR similar as other comparisons.
22350 2023-04-27 Andrew Pinski <apinski@marvell.com>
22352 * tree-ssa-phiopt.cc (cond_store_replacement): Remove
22354 (cond_if_else_store_replacement): Likewise.
22355 (get_non_trapping): Likewise.
22356 (store_elim_worker): Move into ...
22357 (pass_cselim::execute): This.
22359 2023-04-27 Andrew Pinski <apinski@marvell.com>
22361 * tree-ssa-phiopt.cc (two_value_replacement): Remove
22363 (match_simplify_replacement): Likewise.
22364 (factor_out_conditional_conversion): Likewise.
22365 (value_replacement): Likewise.
22366 (minmax_replacement): Likewise.
22367 (spaceship_replacement): Likewise.
22368 (cond_removal_in_builtin_zero_pattern): Likewise.
22369 (hoist_adjacent_loads): Likewise.
22370 (tree_ssa_phiopt_worker): Move into ...
22371 (pass_phiopt::execute): this.
22373 2023-04-27 Andrew Pinski <apinski@marvell.com>
22375 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker): Remove
22376 do_store_elim argument and split that part out to ...
22377 (store_elim_worker): This new function.
22378 (pass_cselim::execute): Call store_elim_worker.
22379 (pass_phiopt::execute): Update call to tree_ssa_phiopt_worker.
22381 2023-04-27 Jan Hubicka <jh@suse.cz>
22383 * cfgloopmanip.h (unloop_loops): Export.
22384 * tree-ssa-loop-ch.cc (ch_base::copy_headers): Unloop loops
22385 that no longer loop.
22386 * tree-ssa-loop-ivcanon.cc (unloop_loops): Export; do not free
22387 vectors of loops to unloop.
22388 (canonicalize_induction_variables): Free vectors here.
22389 (tree_unroll_loops_completely): Free vectors here.
22391 2023-04-27 Richard Biener <rguenther@suse.de>
22393 PR tree-optimization/109170
22394 * gimple-range-op.cc (gimple_range_op_handler::maybe_builtin_call):
22395 Handle __builtin_expect and similar via cfn_pass_through_arg1
22396 and inspecting the calls fnspec.
22397 * builtins.cc (builtin_fnspec): Handle BUILT_IN_EXPECT
22398 and BUILT_IN_EXPECT_WITH_PROBABILITY.
22400 2023-04-27 Alexandre Oliva <oliva@adacore.com>
22402 * genmultilib: Use CONFIG_SHELL to run sub-scripts.
22404 2023-04-27 Aldy Hernandez <aldyh@redhat.com>
22406 PR tree-optimization/109639
22407 * ipa-cp.cc (ipa_value_range_from_jfunc): Normalize range.
22408 (propagate_vr_across_jump_function): Same.
22409 * ipa-fnsummary.cc (evaluate_conditions_for_known_args): Same.
22410 * ipa-prop.h (ipa_range_set_and_normalize): New.
22411 * value-range.cc (irange::set): Assert min and max are INTEGER_CST.
22413 2023-04-27 Richard Biener <rguenther@suse.de>
22415 * match.pd (BIT_FIELD_REF CONSTRUCTOR@0 @1 @2): Do not
22416 create a CTOR operand in the result when simplifying GIMPLE.
22418 2023-04-27 Richard Biener <rguenther@suse.de>
22420 * gimplify.cc (gimplify_compound_lval): When the base
22421 gimplified to a register make sure to split up chains
22424 2023-04-27 Richard Biener <rguenther@suse.de>
22427 * ipa-param-manipulation.h
22428 (ipa_param_body_adjustments::modify_expression): Add extra_stmts
22430 * ipa-param-manipulation.cc
22431 (ipa_param_body_adjustments::modify_expression): Likewise.
22432 When we need a conversion and the replacement is a register
22433 split the conversion out.
22434 (ipa_param_body_adjustments::modify_assignment): Pass
22435 extra_stmts to RHS modify_expression.
22437 2023-04-27 Jonathan Wakely <jwakely@redhat.com>
22439 * doc/extend.texi (Zero Length): Describe example.
22441 2023-04-27 Richard Biener <rguenther@suse.de>
22443 PR tree-optimization/109594
22444 * tree-ssa.cc (non_rewritable_mem_ref_base): Constrain
22445 what we rewrite to a register based on the above.
22447 2023-04-26 Patrick O'Neill <patrick@rivosinc.com>
22449 * config/riscv/riscv.cc: Fix whitespace.
22450 * config/riscv/sync.md: Fix whitespace.
22452 2023-04-26 Andrew MacLeod <amacleod@redhat.com>
22454 PR tree-optimization/108697
22455 * gimple-range-cache.cc (ssa_global_cache::clear_range): Do
22456 not clear the vector on an out of range query.
22457 (ssa_cache::dump): Use dump_range_query instead of get_range.
22458 (ssa_cache::dump_range_query): New.
22459 (ssa_lazy_cache::dump_range_query): New.
22460 (ssa_lazy_cache::set_range): New.
22461 * gimple-range-cache.h (ssa_cache::dump_range_query): New.
22462 (class ssa_lazy_cache): New.
22463 (ssa_lazy_cache::ssa_lazy_cache): New.
22464 (ssa_lazy_cache::~ssa_lazy_cache): New.
22465 (ssa_lazy_cache::get_range): New.
22466 (ssa_lazy_cache::clear_range): New.
22467 (ssa_lazy_cache::clear): New.
22468 (ssa_lazy_cache::dump): New.
22469 * gimple-range-path.cc (path_range_query::path_range_query): Do
22470 not allocate a ssa_cache object nor has_cache bitmap.
22471 (path_range_query::~path_range_query): Do not free objects.
22472 (path_range_query::clear_cache): Remove.
22473 (path_range_query::get_cache): Adjust.
22474 (path_range_query::set_cache): Remove.
22475 (path_range_query::dump): Don't call through a pointer.
22476 (path_range_query::internal_range_of_expr): Set cache directly.
22477 (path_range_query::reset_path): Clear cache directly.
22478 (path_range_query::ssa_range_in_phi): Fold with globals only.
22479 (path_range_query::compute_ranges_in_phis): Simply set range.
22480 (path_range_query::compute_ranges_in_block): Call cache directly.
22481 * gimple-range-path.h (class path_range_query): Replace bitmap
22482 and cache pointer with lazy cache object.
22483 * gimple-range.h (class assume_query): Use ssa_lazy_cache.
22485 2023-04-26 Andrew MacLeod <amacleod@redhat.com>
22487 * gimple-range-cache.cc (ssa_cache::ssa_cache): Rename.
22488 (ssa_cache::~ssa_cache): Rename.
22489 (ssa_cache::has_range): New.
22490 (ssa_cache::get_range): Rename.
22491 (ssa_cache::set_range): Rename.
22492 (ssa_cache::clear_range): Rename.
22493 (ssa_cache::clear): Rename.
22494 (ssa_cache::dump): Rename and use get_range.
22495 (ranger_cache::get_global_range): Use get_range and set_range.
22496 (ranger_cache::range_of_def): Use get_range.
22497 * gimple-range-cache.h (class ssa_cache): Rename class and methods.
22498 (class ranger_cache): Use ssa_cache.
22499 * gimple-range-path.cc (path_range_query::path_range_query): Use
22501 (path_range_query::get_cache): Use get_range.
22502 (path_range_query::set_cache): Use set_range.
22503 * gimple-range-path.h (class path_range_query): Use ssa_cache.
22504 * gimple-range.cc (assume_query::assume_range_p): Use get_range.
22505 (assume_query::range_of_expr): Use get_range.
22506 (assume_query::assume_query): Use set_range.
22507 (assume_query::calculate_op): Use get_range and set_range.
22508 * gimple-range.h (class assume_query): Use ssa_cache.
22510 2023-04-26 Andrew MacLeod <amacleod@redhat.com>
22512 * gimple-range-cache.cc (sbr_vector::sbr_vector): Add parameter
22513 and local to optionally zero memory.
22514 (br_vector::grow): Only zero memory if flag is set.
22515 (class sbr_lazy_vector): New.
22516 (sbr_lazy_vector::sbr_lazy_vector): New.
22517 (sbr_lazy_vector::set_bb_range): New.
22518 (sbr_lazy_vector::get_bb_range): New.
22519 (sbr_lazy_vector::bb_range_p): New.
22520 (block_range_cache::set_bb_range): Check flags and Use sbr_lazy_vector.
22521 * gimple-range-gori.cc (gori_map::calculate_gori): Use
22522 param_vrp_switch_limit.
22523 (gori_compute::gori_compute): Use param_vrp_switch_limit.
22524 * params.opt (vrp_sparse_threshold): Rename from evrp_sparse_threshold.
22525 (vrp_switch_limit): Rename from evrp_switch_limit.
22526 (vrp_vector_threshold): New.
22528 2023-04-26 Andrew MacLeod <amacleod@redhat.com>
22530 * value-relation.cc (dom_oracle::query_relation): Check early for lack
22532 * value-relation.h (equiv_oracle::has_equiv_p): New.
22534 2023-04-26 Andrew MacLeod <amacleod@redhat.com>
22536 PR tree-optimization/109417
22537 * gimple-range-gori.cc (range_def_chain::register_dependency):
22538 Save the ssa version number, not the pointer.
22539 (gori_compute::may_recompute_p): No need to check if a dependency
22540 is in the free list.
22541 * gimple-range-gori.h (class range_def_chain): Change ssa1 and ssa2
22542 fields to be unsigned int instead of trees.
22543 (ange_def_chain::depend1): Adjust.
22544 (ange_def_chain::depend2): Adjust.
22545 * gimple-range.h: Include "ssa.h" to inline ssa_name().
22547 2023-04-26 David Edelsohn <dje.gcc@gmail.com>
22549 * config/rs6000/aix72.h (TARGET_DEFAULT): Use ISA_2_6_MASKS_SERVER.
22550 * config/rs6000/aix73.h (TARGET_DEFAULT): Use ISA_2_7_MASKS_SERVER.
22551 (PROCESSOR_DEFAULT): Use PROCESSOR_POWER8.
22553 2023-04-26 Patrick O'Neill <patrick@rivosinc.com>
22556 * config/riscv/riscv-protos.h: Add helper function stubs.
22557 * config/riscv/riscv.cc: Add helper functions for subword masking.
22558 * config/riscv/riscv.opt: Add command-line flags -minline-atomics and
22559 -mno-inline-atomics.
22560 * config/riscv/sync.md: Add masking logic and inline asm for fetch_and_op,
22561 fetch_and_nand, CAS, and exchange ops.
22562 * doc/invoke.texi: Add blurb regarding new command-line flags
22563 -minline-atomics and -mno-inline-atomics.
22565 2023-04-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22567 * config/aarch64/aarch64-simd.md (aarch64_rshrn2<mode>_insn_le):
22568 Reimplement using standard RTL codes instead of unspec.
22569 (aarch64_rshrn2<mode>_insn_be): Likewise.
22570 (aarch64_rshrn2<mode>): Adjust for the above.
22571 * config/aarch64/aarch64.md (UNSPEC_RSHRN): Delete.
22573 2023-04-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22575 * config/aarch64/aarch64-simd.md (aarch64_rshrn<mode>_insn_le): Reimplement
22576 with standard RTL codes instead of an UNSPEC.
22577 (aarch64_rshrn<mode>_insn_be): Likewise.
22578 (aarch64_rshrn<mode>): Adjust for the above.
22579 * config/aarch64/predicates.md (aarch64_simd_rshrn_imm_vec): Define.
22581 2023-04-26 Pan Li <pan2.li@intel.com>
22582 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
22584 * config/riscv/riscv.cc (riscv_classify_address): Allow
22585 const0_rtx for the RVV load/store.
22587 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22589 * range-op.cc (range_op_cast_tests): Remove legacy support.
22590 * value-range-storage.h (vrange_allocator::alloc_irange): Same.
22591 * value-range.cc (irange::operator=): Same.
22592 (get_legacy_range): Same.
22593 (irange::copy_legacy_to_multi_range): Delete.
22594 (irange::copy_to_legacy): Delete.
22595 (irange::irange_set_anti_range): Delete.
22596 (irange::set): Remove legacy support.
22597 (irange::verify_range): Same.
22598 (irange::legacy_lower_bound): Delete.
22599 (irange::legacy_upper_bound): Delete.
22600 (irange::legacy_equal_p): Delete.
22601 (irange::operator==): Remove legacy support.
22602 (irange::singleton_p): Same.
22603 (irange::value_inside_range): Same.
22604 (irange::contains_p): Same.
22605 (intersect_ranges): Delete.
22606 (irange::legacy_intersect): Delete.
22607 (union_ranges): Delete.
22608 (irange::legacy_union): Delete.
22609 (irange::legacy_verbose_union_): Delete.
22610 (irange::legacy_verbose_intersect): Delete.
22611 (irange::irange_union): Remove legacy support.
22612 (irange::irange_intersect): Same.
22613 (irange::intersect): Same.
22614 (irange::invert): Same.
22615 (ranges_from_anti_range): Delete.
22616 (gt_pch_nx): Adjust for legacy removal.
22618 (range_tests_legacy): Delete.
22619 (range_tests_misc): Adjust for legacy removal.
22620 (range_tests): Same.
22621 * value-range.h (class irange): Same.
22622 (irange::legacy_mode_p): Delete.
22623 (ranges_from_anti_range): Delete.
22624 (irange::nonzero_p): Adjust for legacy removal.
22625 (irange::lower_bound): Same.
22626 (irange::upper_bound): Same.
22627 (irange::union_): Same.
22628 (irange::intersect): Same.
22629 (irange::set_nonzero): Same.
22630 (irange::set_zero): Same.
22631 * vr-values.cc (simplify_using_ranges::legacy_fold_cond_overflow): Same.
22633 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22635 * value-range.cc (irange::copy_legacy_to_multi_range): Rewrite use
22636 of range_has_numeric_bounds_p with irange API.
22637 (range_has_numeric_bounds_p): Delete.
22638 * value-range.h (range_has_numeric_bounds_p): Delete.
22640 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22642 * tree-data-ref.cc (compute_distributive_range): Replace uses of
22643 range_int_cst_p with irange API.
22644 * tree-ssa-strlen.cc (get_range_strlen_dynamic): Same.
22645 * tree-vrp.h (range_int_cst_p): Delete.
22646 * vr-values.cc (check_for_binary_op_overflow): Replace usees of
22647 range_int_cst_p with irange API.
22648 (vr_set_zero_nonzero_bits): Same.
22649 (range_fits_type_p): Same.
22650 (simplify_using_ranges::simplify_casted_cond): Same.
22651 * tree-vrp.cc (range_int_cst_p): Remove.
22653 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22655 * tree-ssa-strlen.cc (compare_nonzero_chars): Convert to wide_ints.
22657 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22659 * builtins.cc (expand_builtin_strnlen): Rewrite deprecated irange
22660 API uses to new API.
22661 * gimple-predicate-analysis.cc (find_var_cmp_const): Same.
22662 * internal-fn.cc (get_min_precision): Same.
22664 * tree-affine.cc (expr_to_aff_combination): Same.
22665 * tree-data-ref.cc (dr_step_indicator): Same.
22666 * tree-dfa.cc (get_ref_base_and_extent): Same.
22667 * tree-scalar-evolution.cc (iv_can_overflow_p): Same.
22668 * tree-ssa-phiopt.cc (two_value_replacement): Same.
22669 * tree-ssa-pre.cc (insert_into_preds_of_block): Same.
22670 * tree-ssa-reassoc.cc (optimize_range_tests_to_bit_test): Same.
22671 * tree-ssa-strlen.cc (compare_nonzero_chars): Same.
22672 * tree-switch-conversion.cc (bit_test_cluster::emit): Same.
22673 * tree-vect-patterns.cc (vect_recog_divmod_pattern): Same.
22674 * tree.cc (get_range_pos_neg): Same.
22676 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22678 * ipa-prop.cc (ipa_print_node_jump_functions_for_edge): Use
22679 vrange::dump instead of ad-hoc dumper.
22680 * tree-ssa-strlen.cc (dump_strlen_info): Same.
22681 * value-range-pretty-print.cc (visit): Pass TDF_NOUID to
22684 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22686 * range-op.cc (operator_cast::op1_range): Use
22687 create_possibly_reversed_range.
22688 (operator_bitwise_and::simple_op1_range_solver): Same.
22689 * value-range.cc (swap_out_of_order_endpoints): Delete.
22690 (irange::set): Remove call to swap_out_of_order_endpoints.
22692 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22694 * builtins.cc (determine_block_size): Convert use of legacy API to
22696 * gimple-array-bounds.cc (check_out_of_bounds_and_warn): Same.
22697 (array_bounds_checker::check_array_ref): Same.
22698 * gimple-ssa-warn-restrict.cc
22699 (builtin_memref::extend_offset_range): Same.
22700 * ipa-cp.cc (ipcp_store_vr_results): Same.
22701 * ipa-fnsummary.cc (set_switch_stmt_execution_predicate): Same.
22702 * ipa-prop.cc (struct ipa_vr_ggc_hash_traits): Same.
22703 (ipa_write_jump_function): Same.
22704 * pointer-query.cc (get_size_range): Same.
22705 * tree-data-ref.cc (split_constant_offset): Same.
22706 * tree-ssa-strlen.cc (get_range): Same.
22707 (maybe_diag_stxncpy_trunc): Same.
22708 (strlen_pass::get_len_or_size): Same.
22709 (strlen_pass::count_nonzero_bytes_addr): Same.
22710 * tree-vect-patterns.cc (vect_get_range_info): Same.
22711 * value-range.cc (irange::maybe_anti_range): Remove.
22712 (get_legacy_range): New.
22713 (irange::copy_to_legacy): Use get_legacy_range.
22714 (ranges_from_anti_range): Same.
22715 * value-range.h (class irange): Remove maybe_anti_range.
22716 (get_legacy_range): New.
22717 * vr-values.cc (check_for_binary_op_overflow): Convert use of
22718 legacy API to get_legacy_range.
22719 (compare_ranges): Same.
22720 (compare_range_with_value): Same.
22721 (bounds_of_var_in_loop): Same.
22722 (find_case_label_ranges): Same.
22723 (simplify_using_ranges::simplify_switch_using_ranges): Same.
22725 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22727 * value-range-pretty-print.cc (vrange_printer::visit): Remove
22729 * value-range.cc (irange::constant_p): Remove.
22730 (irange::get_nonzero_bits_from_range): Remove constant_p use.
22731 * value-range.h (class irange): Remove constant_p.
22732 (irange::num_pairs): Remove constant_p use.
22734 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22736 * value-range.cc (irange::copy_legacy_to_multi_range): Remove
22738 (irange::set): Same.
22739 (irange::legacy_lower_bound): Same.
22740 (irange::legacy_upper_bound): Same.
22741 (irange::contains_p): Same.
22742 (range_tests_legacy): Same.
22743 (irange::normalize_addresses): Remove.
22744 (irange::normalize_symbolics): Remove.
22745 (irange::symbolic_p): Remove.
22746 * value-range.h (class irange): Remove symbolic_p,
22747 normalize_symbolics, and normalize_addresses.
22748 * vr-values.cc (simplify_using_ranges::two_valued_val_range_p):
22749 Remove symbolics support.
22751 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22753 * value-range.cc (irange::may_contain_p): Remove.
22754 * value-range.h (range_includes_zero_p): Rewrite may_contain_p
22755 usage with contains_p.
22756 * vr-values.cc (compare_range_with_value): Same.
22758 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22760 * tree-vrp.cc (supported_types_p): Remove.
22761 (defined_ranges_p): Remove.
22762 (range_fold_binary_expr): Remove.
22763 (range_fold_unary_expr): Remove.
22764 * tree-vrp.h (range_fold_unary_expr): Remove.
22765 (range_fold_binary_expr): Remove.
22767 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22769 * ipa-cp.cc (ipa_vr_operation_and_type_effects): Convert to ranger API.
22770 (ipa_value_range_from_jfunc): Same.
22771 (propagate_vr_across_jump_function): Same.
22772 * ipa-fnsummary.cc (evaluate_conditions_for_known_args): Same.
22773 * ipa-prop.cc (ipa_compute_jump_functions_for_edge): Same.
22774 * vr-values.cc (bounds_of_var_in_loop): Same.
22776 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22778 * gimple-array-bounds.cc (array_bounds_checker::get_value_range):
22779 Add irange argument.
22780 (check_out_of_bounds_and_warn): Remove check for vr.
22781 (array_bounds_checker::check_array_ref): Remove pointer qualifier
22782 for vr and adjust accordingly.
22783 * gimple-array-bounds.h (get_value_range): Add irange argument.
22784 * value-query.cc (class equiv_allocator): Delete.
22785 (range_query::get_value_range): Delete.
22786 (range_query::range_query): Remove allocator access.
22787 (range_query::~range_query): Same.
22788 * value-query.h (get_value_range): Delete.
22790 (simplify_using_ranges::op_with_boolean_value_range_p): Remove
22791 call to get_value_range.
22792 (check_for_binary_op_overflow): Same.
22793 (simplify_using_ranges::legacy_fold_cond_overflow): Same.
22794 (simplify_using_ranges::simplify_abs_using_ranges): Same.
22795 (simplify_using_ranges::simplify_cond_using_ranges_1): Same.
22796 (simplify_using_ranges::simplify_casted_cond): Same.
22797 (simplify_using_ranges::simplify_switch_using_ranges): Same.
22798 (simplify_using_ranges::two_valued_val_range_p): Same.
22800 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22803 (simplify_using_ranges::vrp_evaluate_conditional_warnv_with_ops):
22805 (simplify_using_ranges::legacy_fold_cond_overflow): ...this.
22806 (simplify_using_ranges::vrp_visit_cond_stmt): Rename to...
22807 (simplify_using_ranges::legacy_fold_cond): ...this.
22808 (simplify_using_ranges::fold_cond): Rename
22809 vrp_evaluate_conditional_warnv_with_ops to
22810 legacy_fold_cond_overflow.
22811 * vr-values.h (class vr_values): Replace vrp_visit_cond_stmt and
22812 vrp_evaluate_conditional_warnv_with_ops with legacy_fold_cond and
22813 legacy_fold_cond_overflow respectively.
22815 2023-04-26 Aldy Hernandez <aldyh@redhat.com>
22817 * vr-values.cc (get_vr_for_comparison): Remove.
22818 (compare_name_with_value): Same.
22819 (vrp_evaluate_conditional_warnv_with_ops): Remove calls to
22820 compare_name_with_value.
22821 * vr-values.h: Remove compare_name_with_value.
22822 Remove get_vr_for_comparison.
22824 2023-04-26 Roger Sayle <roger@nextmovesoftware.com>
22826 * config/stormy16/stormy16.md (bswaphi2): New define_insn.
22827 (bswapsi2): New define_insn.
22828 (swaphi): New define_insn to exchange two registers (swpw).
22829 (define_peephole2): Recognize exchange of registers as swaphi.
22831 2023-04-26 Richard Biener <rguenther@suse.de>
22833 * gimple-range-path.cc (path_range_query::compute_outgoing_relations):
22835 * ipa-pure-const.cc (pass_nothrow::execute): Likewise.
22836 * predict.cc (apply_return_prediction): Likewise.
22837 * sese.cc (set_ifsese_condition): Likewise. Simplify.
22838 * tree-cfg.cc (assert_unreachable_fallthru_edge_p): Avoid last_stmt.
22839 (make_edges_bb): Likewise.
22840 (make_cond_expr_edges): Likewise.
22841 (end_recording_case_labels): Likewise.
22842 (make_gimple_asm_edges): Likewise.
22843 (cleanup_dead_labels): Likewise.
22844 (group_case_labels): Likewise.
22845 (gimple_can_merge_blocks_p): Likewise.
22846 (gimple_merge_blocks): Likewise.
22847 (find_taken_edge): Likewise. Also handle empty fallthru blocks.
22848 (gimple_duplicate_sese_tail): Avoid last_stmt.
22849 (find_loop_dist_alias): Likewise.
22850 (gimple_block_ends_with_condjump_p): Likewise.
22851 (gimple_purge_dead_eh_edges): Likewise.
22852 (gimple_purge_dead_abnormal_call_edges): Likewise.
22853 (pass_warn_function_return::execute): Likewise.
22854 (execute_fixup_cfg): Likewise.
22855 * tree-eh.cc (redirect_eh_edge_1): Likewise.
22856 (pass_lower_resx::execute): Likewise.
22857 (pass_lower_eh_dispatch::execute): Likewise.
22858 (cleanup_empty_eh): Likewise.
22859 * tree-if-conv.cc (if_convertible_bb_p): Likewise.
22860 (predicate_bbs): Likewise.
22861 (ifcvt_split_critical_edges): Likewise.
22862 * tree-loop-distribution.cc (create_edge_for_control_dependence):
22864 (loop_distribution::transform_reduction_loop): Likewise.
22865 * tree-parloops.cc (transform_to_exit_first_loop_alt): Likewise.
22866 (try_transform_to_exit_first_loop_alt): Likewise.
22867 (transform_to_exit_first_loop): Likewise.
22868 (create_parallel_loop): Likewise.
22869 * tree-scalar-evolution.cc (get_loop_exit_condition): Likewise.
22870 * tree-ssa-dce.cc (mark_last_stmt_necessary): Likewise.
22871 (eliminate_unnecessary_stmts): Likewise.
22873 (dom_opt_dom_walker::set_global_ranges_from_unreachable_edges):
22875 * tree-ssa-ifcombine.cc (ifcombine_ifandif): Likewise.
22876 (pass_tree_ifcombine::execute): Likewise.
22877 * tree-ssa-loop-ch.cc (entry_loop_condition_is_static): Likewise.
22878 (should_duplicate_loop_header_p): Likewise.
22879 * tree-ssa-loop-ivcanon.cc (create_canonical_iv): Likewise.
22880 (tree_estimate_loop_size): Likewise.
22881 (try_unroll_loop_completely): Likewise.
22882 * tree-ssa-loop-ivopts.cc (tree_ssa_iv_optimize_loop): Likewise.
22883 * tree-ssa-loop-manip.cc (ip_normal_pos): Likewise.
22884 (canonicalize_loop_ivs): Likewise.
22885 * tree-ssa-loop-niter.cc (determine_value_range): Likewise.
22886 (bound_difference): Likewise.
22887 (number_of_iterations_popcount): Likewise.
22888 (number_of_iterations_cltz): Likewise.
22889 (number_of_iterations_cltz_complement): Likewise.
22890 (simplify_using_initial_conditions): Likewise.
22891 (number_of_iterations_exit_assumptions): Likewise.
22892 (loop_niter_by_eval): Likewise.
22893 (estimate_numbers_of_iterations): Likewise.
22895 2023-04-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
22897 * config/riscv/vector.md: Refine vmadc/vmsbc RA constraint.
22899 2023-04-26 Kewen Lin <linkw@linux.ibm.com>
22902 * config/rs6000/rs6000-builtins.def
22903 (__builtin_vsx_scalar_cmp_exp_qp_eq, __builtin_vsx_scalar_cmp_exp_qp_gt
22904 __builtin_vsx_scalar_cmp_exp_qp_lt,
22905 __builtin_vsx_scalar_cmp_exp_qp_unordered): Move from stanza ieee128-hw
22908 2023-04-26 Kewen Lin <linkw@linux.ibm.com>
22911 * config/rs6000/altivec.md (sldoi_to_mov<mode>): Replace predicate
22912 easy_vector_constant with const_vector_each_byte_same, add
22913 handlings in preparation for !easy_vector_constant, and update
22914 VECTOR_UNIT_ALTIVEC_OR_VSX_P with VECTOR_MEM_ALTIVEC_OR_VSX_P.
22915 * config/rs6000/predicates.md (const_vector_each_byte_same): New
22918 2023-04-26 Juzhe-Zhong <juzhe.zhong@rivai.ai>
22920 * config/riscv/vector.md (*pred_cmp<mode>_merge_tie_mask): New pattern.
22921 (*pred_ltge<mode>_merge_tie_mask): Ditto.
22922 (*pred_cmp<mode>_scalar_merge_tie_mask): Ditto.
22923 (*pred_eqne<mode>_scalar_merge_tie_mask): Ditto.
22924 (*pred_cmp<mode>_extended_scalar_merge_tie_mask): Ditto.
22925 (*pred_eqne<mode>_extended_scalar_merge_tie_mask): Ditto.
22926 (*pred_cmp<mode>_narrow_merge_tie_mask): Ditto.
22928 2023-04-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
22930 * config/riscv/vector.md: Fix redundant vmv1r.v.
22932 2023-04-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
22934 * config/riscv/vector.md: Fix RA constraint.
22936 2023-04-26 Pan Li <pan2.li@intel.com>
22939 * tree-ssa-sccvn.cc (vn_reference_eq): add type vector subparts
22940 check for vn_reference equal.
22942 2023-04-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
22944 * config/riscv/riscv-opts.h (enum riscv_autovec_preference_enum): Add enum for
22945 auto-vectorization preference.
22946 (enum riscv_autovec_lmul_enum): Add enum for choosing LMUL of RVV
22947 auto-vectorization.
22948 * config/riscv/riscv.opt: Add compile option for RVV auto-vectorization.
22950 2023-04-26 Jivan Hakobyan <jivanhakobyan9@gmail.com>
22952 * config/riscv/bitmanip.md: Updated predicates of bclri<mode>_nottwobits
22953 and bclridisi_nottwobits patterns.
22954 * config/riscv/predicates.md: (not_uimm_extra_bit_or_nottwobits): Adjust
22955 predicate to avoid splitting arith constants.
22956 (const_nottwobits_not_arith_operand): New predicate.
22958 2023-04-25 Hans-Peter Nilsson <hp@axis.com>
22960 * recog.cc (peep2_attempt, peep2_update_life): Correct
22961 head-comment description of parameter match_len.
22963 2023-04-25 Vineet Gupta <vineetg@rivosinc.com>
22965 * config/riscv/riscv.md: riscv_move_integer() drop in_splitter arg.
22966 riscv_split_symbol() drop in_splitter arg.
22967 * config/riscv/riscv.cc: riscv_move_integer() drop in_splitter arg.
22968 riscv_split_symbol() drop in_splitter arg.
22969 riscv_force_temporary() drop in_splitter arg.
22970 * config/riscv/riscv-protos.h: riscv_move_integer() drop in_splitter arg.
22971 riscv_split_symbol() drop in_splitter arg.
22973 2023-04-25 Eric Botcazou <ebotcazou@adacore.com>
22975 * tree-ssa.cc (insert_debug_temp_for_var_def): Do not create
22976 superfluous debug temporaries for single GIMPLE assignments.
22978 2023-04-25 Richard Biener <rguenther@suse.de>
22980 PR tree-optimization/109609
22981 * attr-fnspec.h (arg_max_access_size_given_by_arg_p):
22983 * tree-ssa-alias.cc (check_fnspec): Correctly interpret
22984 the size given by arg_max_access_size_given_by_arg_p as
22985 maximum, not exact, size.
22987 2023-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22990 * config/aarch64/aarch64-simd.md (orn<mode>3): Rename to...
22991 (orn<mode>3<vczle><vczbe>): ... This.
22992 (bic<mode>3): Rename to...
22993 (bic<mode>3<vczle><vczbe>): ... This.
22994 (<su><maxmin><mode>3): Rename to...
22995 (<su><maxmin><mode>3<vczle><vczbe>): ... This.
22997 2023-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22999 * config/aarch64/aarch64-simd.md (<su_optab>div<mode>3): New define_expand.
23000 * config/aarch64/iterators.md (VQDIV): New mode iterator.
23001 (vnx2di): New mode attribute.
23003 2023-04-25 Richard Biener <rguenther@suse.de>
23005 PR rtl-optimization/109585
23006 * tree-ssa-alias.cc (aliasing_component_refs_p): Fix typo.
23008 2023-04-25 Jakub Jelinek <jakub@redhat.com>
23011 * config/rs6000/rs6000.cc (rs6000_is_valid_rotate_dot_mask): For
23012 !TARGET_64BIT, don't return true if UINTVAL (mask) << (63 - nb)
23013 is larger than signed int maximum.
23015 2023-04-25 Martin Liska <mliska@suse.cz>
23017 * doc/gcov.texi: Document the new "calls" field and document
23018 the API bump. Mention also "block_ids" for lines.
23019 * gcov.cc (output_intermediate_json_line): Output info about
23020 calls and extend branches as well.
23021 (generate_results): Bump version to 2.
23022 (output_line_details): Use block ID instead of a non-sensual
23025 2023-04-25 Roger Sayle <roger@nextmovesoftware.com>
23027 * config/stormy16/stormy16.md (zero_extendqihi2): Restore/fix
23028 length attribute for the first (memory operand) alternative.
23030 2023-04-25 Victor Do Nascimento <victor.donascimento@arm.com>
23032 * config/aarch64/aarch64-simd.md(aarch64_simd_stp<mode>): New.
23033 * config/aarch64/constraints.md: Make "Umn" relaxed memory
23035 * config/aarch64/iterators.md(ldpstp_vel_sz): New.
23037 2023-04-25 Aldy Hernandez <aldyh@redhat.com>
23039 * value-range.cc (frange::set): Adjust constructor.
23040 * value-range.h (nan_state::nan_state): Replace default
23041 constructor with one taking an argument.
23043 2023-04-25 Aldy Hernandez <aldyh@redhat.com>
23045 * ipa-cp.cc (ipa_range_contains_p): New.
23046 (decide_whether_version_node): Use it.
23048 2023-04-24 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
23050 * tree-ssa-forwprop.cc (is_combined_permutation_identity): Try to
23051 simplify two successive VEC_PERM_EXPRs with same VLA mask,
23052 where mask chooses elements in reverse order.
23054 2023-04-24 Andrew Pinski <apinski@marvell.com>
23056 * tree-ssa-phiopt.cc (match_simplify_replacement): Add new arguments
23057 and support diamond shaped basic block form.
23058 (tree_ssa_phiopt_worker): Update call to match_simplify_replacement
23060 2023-04-24 Andrew Pinski <apinski@marvell.com>
23062 * tree-ssa-phiopt.cc (empty_bb_or_one_feeding_into_p):
23063 Instead of calling last_and_only_stmt, look for the last statement
23066 2023-04-24 Andrew Pinski <apinski@marvell.com>
23068 * tree-ssa-phiopt.cc (empty_bb_or_one_feeding_into_p):
23070 (match_simplify_replacement): Call
23071 empty_bb_or_one_feeding_into_p instead of doing it inline.
23073 2023-04-24 Andrew Pinski <apinski@marvell.com>
23075 PR tree-optimization/68894
23076 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker): Remove the
23077 continue for the do_hoist_loads diamond case.
23079 2023-04-24 Andrew Pinski <apinski@marvell.com>
23081 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker): Rearrange
23082 code for better code readability.
23084 2023-04-24 Andrew Pinski <apinski@marvell.com>
23086 PR tree-optimization/109604
23087 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker): Move the
23088 diamond form check from ...
23089 (minmax_replacement): Here.
23091 2023-04-24 Patrick Palka <ppalka@redhat.com>
23093 * tree.cc (strip_array_types): Don't define here.
23094 (is_typedef_decl): Don't define here.
23095 (typedef_variant_p): Don't define here.
23096 * tree.h (strip_array_types): Define here.
23097 (is_typedef_decl): Define here.
23098 (typedef_variant_p): Define here.
23100 2023-04-24 Frederik Harwath <frederik@codesourcery.com>
23102 * doc/generic.texi (OpenMP): Add != to allowed
23103 conditions and state that vars can be unsigned.
23104 * tree.def (OMP_FOR): Likewise.
23106 2023-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23108 * config/aarch64/aarch64-simd.md (mulv2di3): New expander.
23110 2023-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23112 * doc/install.texi: Consistently use Solaris rather than Solaris 2.
23113 Remove explicit Solaris 11 references.
23115 (Options specification, --with-gnu-as): as and gas always differ
23117 Remove /usr/ccs/bin reference.
23118 (Installing GCC: Binaries, Solaris (SPARC, Intel)): Remove.
23119 (i?86-*-solaris2*): Merge assembler, linker recommendations ...
23120 (*-*-solaris2*): ... here.
23121 Update bundled GCC versions.
23122 Don't refer to pre-built binaries.
23123 Remove /bin/sh warning.
23124 Update assembler, linker recommendations.
23125 Document GNAT bootstrap compiler.
23126 (sparc-sun-solaris2*): Remove non-UltraSPARC reference.
23127 (sparc64-*-solaris2*): Move content...
23128 (sparcv9-*-solaris2*): ...here.
23129 Add GDC for 64-bit bootstrap compilers.
23131 2023-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23134 * config/aarch64/aarch64-sve.md (<optab><mode>3): Handle TARGET_SVE2 MUL
23136 * config/aarch64/aarch64-sve2.md (*aarch64_mul_unpredicated_<mode>): New
23139 2023-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23141 * config/aarch64/aarch64-simd.md (aarch64_<sur>abal2<mode>): Rename to...
23142 (aarch64_<su>abal2<mode>_insn): ... This. Use RTL codes instead of unspec.
23143 (aarch64_<su>abal2<mode>): New define_expand.
23144 * config/aarch64/aarch64.cc (aarch64_abd_rtx_p): New function.
23145 (aarch64_rtx_costs): Handle ABD rtxes.
23146 * config/aarch64/aarch64.md (UNSPEC_SABAL2, UNSPEC_UABAL2): Delete.
23147 * config/aarch64/iterators.md (ABAL2): Delete.
23148 (sur): Remove handling of UNSPEC_UABAL2 and UNSPEC_SABAL2.
23150 2023-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23152 * config/aarch64/aarch64-simd.md (aarch64_<sur>abal<mode>): Rename to...
23153 (aarch64_<su>abal<mode>): ... This. Use RTL codes instead of unspec.
23154 (<sur>sadv16qi): Rename to...
23155 (<su>sadv16qi): ... This. Adjust for the above.
23156 * config/aarch64/aarch64-sve.md (<sur>sad<vsi2qi>): Rename to...
23157 (<su>sad<vsi2qi>): ... This. Adjust for the above.
23158 * config/aarch64/aarch64.md (UNSPEC_SABAL, UNSPEC_UABAL): Delete.
23159 * config/aarch64/iterators.md (ABAL): Delete.
23160 (sur): Remove handling of UNSPEC_SABAL and UNSPEC_UABAL.
23162 2023-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23164 * config/aarch64/aarch64-simd.md (aarch64_<sur>abdl2<mode>): Rename to...
23165 (aarch64_<su>abdl2<mode>_insn): ... This. Use RTL codes instead of unspec.
23166 (aarch64_<su>abdl2<mode>): New define_expand.
23167 * config/aarch64/aarch64.md (UNSPEC_SABDL2, UNSPEC_UABDL2): Delete.
23168 * config/aarch64/iterators.md (ABDL2): Delete.
23169 (sur): Remove handling of UNSPEC_SABDL2 and UNSPEC_UABDL2.
23171 2023-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23173 * config/aarch64/aarch64-simd.md (aarch64_<sur>abdl<mode>): Rename to...
23174 (aarch64_<su>abdl<mode>): ... This. Use standard RTL ops instead of
23176 * config/aarch64/aarch64.md (UNSPEC_SABDL, UNSPEC_UABDL): Delete.
23177 * config/aarch64/iterators.md (ABDL): Delete.
23178 (sur): Remove handling of UNSPEC_SABDL and UNSPEC_UABDL.
23180 2023-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23182 * config/aarch64/aarch64-simd.md
23183 (*aarch64_<su>addlv<VDQV_L:mode>_ze<GPI:mode>): New pattern.
23185 2023-04-24 Richard Biener <rguenther@suse.de>
23187 * gimple-ssa-split-paths.cc (is_feasible_trace): Avoid
23189 * graphite-scop-detection.cc (single_pred_cond_non_loop_exit):
23191 * ipa-fnsummary.cc (set_cond_stmt_execution_predicate): Likewise.
23192 (set_switch_stmt_execution_predicate): Likewise.
23193 (phi_result_unknown_predicate): Likewise.
23194 * ipa-prop.cc (compute_complex_ancestor_jump_func): Likewise.
23195 (ipa_analyze_indirect_call_uses): Likewise.
23196 * predict.cc (predict_iv_comparison): Likewise.
23197 (predict_extra_loop_exits): Likewise.
23198 (predict_loops): Likewise.
23199 (tree_predict_by_opcode): Likewise.
23200 * gimple-predicate-analysis.cc (predicate::init_from_control_deps):
23202 * gimple-pretty-print.cc (dump_implicit_edges): Likewise.
23203 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker): Likewise.
23204 (replace_phi_edge_with_variable): Likewise.
23205 (two_value_replacement): Likewise.
23206 (value_replacement): Likewise.
23207 (minmax_replacement): Likewise.
23208 (spaceship_replacement): Likewise.
23209 (cond_removal_in_builtin_zero_pattern): Likewise.
23210 * tree-ssa-reassoc.cc (maybe_optimize_range_tests): Likewise.
23211 * tree-ssa-sccvn.cc (vn_phi_eq): Likewise.
23212 (vn_phi_lookup): Likewise.
23213 (vn_phi_insert): Likewise.
23214 * tree-ssa-structalias.cc (compute_points_to_sets): Likewise.
23215 * tree-ssa-threadbackward.cc (back_threader::maybe_thread_block):
23217 (back_threader_profitability::possibly_profitable_path_p):
23219 * tree-ssa-threadedge.cc (jump_threader::thread_outgoing_edges):
23221 * tree-switch-conversion.cc (pass_convert_switch::execute):
23223 (pass_lower_switch<O0>::execute): Likewise.
23224 * tree-tailcall.cc (tree_optimize_tail_calls_1): Likewise.
23225 * tree-vect-loop-manip.cc (vect_loop_versioning): Likewise.
23226 * tree-vect-slp.cc (vect_slp_function): Likewise.
23227 * tree-vect-stmts.cc (cfun_returns): Likewise.
23228 * tree-vectorizer.cc (vect_loop_vectorized_call): Likewise.
23229 (vect_loop_dist_alias_call): Likewise.
23231 2023-04-24 Richard Biener <rguenther@suse.de>
23233 * cfgcleanup.cc (outgoing_edges_match): Use FORWARDER_BLOCK_P.
23235 2023-04-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
23237 * config/riscv/riscv-vsetvl.cc
23238 (vector_infos_manager::all_avail_in_compatible_p): New function.
23239 (pass_vsetvl::refine_vsetvls): Optimize vsetvls.
23240 * config/riscv/riscv-vsetvl.h: New function.
23242 2023-04-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
23244 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::pre_vsetvl): Add function
23245 comment for cleanup_insns.
23247 2023-04-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
23249 * config/riscv/vector-iterators.md: New unspec to refine fault first load pattern.
23250 * config/riscv/vector.md: Refine fault first load pattern to erase avl from instructions
23251 with the fault first load property.
23253 2023-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23255 * config/aarch64/aarch64-simd.md (aarch64_float_truncate_lo_): Rename to...
23256 (aarch64_float_truncate_lo_<mode><vczle><vczbe>): ... This.
23258 2023-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23261 * config/aarch64/aarch64-simd.md (aarch64_addp<mode>): Rename to...
23262 (aarch64_addp<mode><vczle><vczbe>): ... This.
23264 2023-04-23 Roger Sayle <roger@nextmovesoftware.com>
23266 * config/stormy16/stormy16.cc (xstormy16_rtx_costs): Rewrite to
23267 provide reasonable values for common arithmetic operations and
23268 immediate operands (in several machine modes).
23270 2023-04-23 Roger Sayle <roger@nextmovesoftware.com>
23272 * config/stormy16/stormy16.cc (xstormy16_print_operand): Add %h
23273 format specifier to output high_part register name of SImode reg.
23274 * config/stormy16/stormy16.md (extendhisi2): New define_insn.
23275 (zero_extendqihi2): Fix lengths, consistent formatting and add
23276 "and Rx,#255" alternative, for documentation purposes.
23277 (zero_extendhisi2): New define_insn.
23279 2023-04-23 Roger Sayle <roger@nextmovesoftware.com>
23281 * config/stormy16/stormy16.cc (xstormy16_output_shift): Implement
23282 SImode shifts by two by performing a single bit SImode shift twice.
23284 2023-04-23 Aldy Hernandez <aldyh@redhat.com>
23286 PR tree-optimization/109593
23287 * value-range.cc (frange::operator==): Handle NANs.
23289 2023-04-23 liuhongt <hongtao.liu@intel.com>
23291 PR rtl-optimization/108707
23292 * ira-costs.cc (scan_one_insn): Use NO_REGS instead of
23293 GENERAL_REGS when preferred reg_class is not known.
23295 2023-04-22 Andrew Pinski <apinski@marvell.com>
23297 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker):
23298 Change the code around slightly to move diamond
23299 handling for do_store_elim/do_hoist_loads out of
23302 2023-04-22 Andrew Pinski <apinski@marvell.com>
23304 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker):
23305 Remove check on empty_block_p.
23307 2023-04-22 Jakub Jelinek <jakub@redhat.com>
23309 PR bootstrap/109589
23310 * system.h (class auto_mpz): Workaround PR62101 bug in GCC 4.8 and 4.9.
23311 * realmpfr.h (class auto_mpfr): Likewise.
23313 2023-04-22 Jakub Jelinek <jakub@redhat.com>
23315 PR tree-optimization/109583
23316 * match.pd (fneg/fadd simplify): Don't call related_vector_mode
23317 if vec_mode is not VECTOR_MODE_P.
23319 2023-04-22 Jan Hubicka <hubicka@ucw.cz>
23320 Ondrej Kubanek <kubanek0ondrej@gmail.com>
23322 * cfgloopmanip.h (adjust_loop_info_after_peeling): Declare.
23323 * tree-ssa-loop-ch.cc (ch_base::copy_headers): Fix updating of
23324 loop profile and bounds after header duplication.
23325 * tree-ssa-loop-ivcanon.cc (adjust_loop_info_after_peeling):
23326 Break out from try_peel_loop; fix handling of 0 iterations.
23327 (try_peel_loop): Use adjust_loop_info_after_peeling.
23329 2023-04-21 Andrew MacLeod <amacleod@redhat.com>
23331 PR tree-optimization/109546
23332 * tree-vrp.cc (remove_unreachable::remove_and_update_globals): Do
23333 not fold conditions with ADDR_EXPR early.
23335 2023-04-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23337 * config/aarch64/aarch64.md (aarch64_umax<mode>3_insn): Delete.
23338 (umax<mode>3): Emit raw UMAX RTL instead of going through gen_ function
23340 (<optab><mode>3): New define_expand for MAXMIN_NOUMAX codes.
23341 (*aarch64_<optab><mode>3_zero): Define.
23342 (*aarch64_<optab><mode>3_cssc): Likewise.
23343 * config/aarch64/iterators.md (maxminand): New code attribute.
23345 2023-04-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23348 * config/aarch64/aarch64-opts.h (enum aarch64_tp_reg): Define.
23349 * config/aarch64/aarch64-protos.h (aarch64_output_load_tp):
23351 * config/aarch64/aarch64.cc (aarch64_tpidr_register): Declare.
23352 (aarch64_override_options_internal): Handle the above.
23353 (aarch64_output_load_tp): New function.
23354 * config/aarch64/aarch64.md (aarch64_load_tp_hard): Call
23355 aarch64_output_load_tp.
23356 * config/aarch64/aarch64.opt (aarch64_tp_reg): Define enum.
23357 (mtp=): New option.
23358 * doc/invoke.texi (AArch64 Options): Document -mtp=.
23360 2023-04-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23363 * config/aarch64/aarch64-simd.md (add_vec_concat_subst_le): Define.
23364 (add_vec_concat_subst_be): Likewise.
23367 (add<mode>3): Rename to...
23368 (add<mode>3<vczle><vczbe>): ... This.
23369 (sub<mode>3): Rename to...
23370 (sub<mode>3<vczle><vczbe>): ... This.
23371 (mul<mode>3): Rename to...
23372 (mul<mode>3<vczle><vczbe>): ... This.
23373 (and<mode>3): Rename to...
23374 (and<mode>3<vczle><vczbe>): ... This.
23375 (ior<mode>3): Rename to...
23376 (ior<mode>3<vczle><vczbe>): ... This.
23377 (xor<mode>3): Rename to...
23378 (xor<mode>3<vczle><vczbe>): ... This.
23379 * config/aarch64/iterators.md (VDZ): Define.
23381 2023-04-21 Patrick Palka <ppalka@redhat.com>
23383 * tree.cc (walk_tree_1): Avoid repeatedly dereferencing tp
23386 2023-04-21 Jan Hubicka <jh@suse.cz>
23388 * tree-ssa-loop-ch.cc (ch_base::copy_headers): Fix previous
23391 2023-04-21 Vineet Gupta <vineetg@rivosinc.com>
23393 * expmed.h (x_shift*_cost): convert to int [speed][mode][shift].
23394 (shift*_cost_ptr ()): Access x_shift*_cost array directly.
23396 2023-04-21 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
23398 * config/aarch64/aarch64.cc (aarch64_simd_dup_constant): Use
23399 force_reg instead of copy_to_mode_reg.
23400 (aarch64_expand_vector_init): Likewise.
23402 2023-04-21 Uroš Bizjak <ubizjak@gmail.com>
23404 * config/i386/i386.h (REG_OK_FOR_INDEX_P, REG_OK_FOR_BASE_P): Remove.
23405 (REG_OK_FOR_INDEX_NONSTRICT_P, REG_OK_FOR_BASE_NONSTRICT_P): Ditto.
23406 (REG_OK_FOR_INDEX_STRICT_P, REG_OK_FOR_BASE_STRICT_P): Ditto.
23407 (FIRST_INDEX_REG, LAST_INDEX_REG): New defines.
23408 (LEGACY_INDEX_REG_P, LEGACY_INDEX_REGNO_P): New macros.
23409 (INDEX_REG_P, INDEX_REGNO_P): Ditto.
23410 (REGNO_OK_FOR_INDEX_P): Use INDEX_REGNO_P predicates.
23411 (REGNO_OK_FOR_INDEX_NONSTRICT_P): New macro.
23412 (EG_OK_FOR_BASE_NONSTRICT_P): Ditto.
23413 * config/i386/predicates.md (index_register_operand):
23414 Use REGNO_OK_FOR_INDEX_P and REGNO_OK_FOR_INDEX_NONSTRICT_P macros.
23415 * config/i386/i386.cc (ix86_legitimate_address_p): Use
23416 REGNO_OK_FOR_BASE_P, REGNO_OK_FOR_BASE_NONSTRICT_P,
23417 REGNO_OK_FOR_INDEX_P and REGNO_OK_FOR_INDEX_NONSTRICT_P macros.
23419 2023-04-21 Jan Hubicka <hubicka@ucw.cz>
23420 Ondrej Kubanek <kubanek0ondrej@gmail.com>
23422 * tree-ssa-loop-ch.cc (ch_base::copy_headers): Update loop header and
23425 2023-04-21 Richard Biener <rguenther@suse.de>
23427 * is-a.h (safe_is_a): New.
23429 2023-04-21 Richard Biener <rguenther@suse.de>
23431 * gimple-iterator.h (gimple_stmt_iterator::operator*): Add.
23432 (gphi_iterator::operator*): Likewise.
23434 2023-04-21 Jan Hubicka <hubicka@ucw.cz>
23435 Michal Jires <michal@jires.eu>
23437 * ipa-inline.cc (class inline_badness): New class.
23438 (edge_heap_t, edge_heap_node_t): Use inline_badness for badness instead
23440 (update_edge_key): Update.
23441 (lookup_recursive_calls): Likewise.
23442 (recursive_inlining): Likewise.
23443 (add_new_edges_to_heap): Likewise.
23444 (inline_small_functions): Likewise.
23446 2023-04-21 Jan Hubicka <hubicka@ucw.cz>
23448 * ipa-devirt.cc (odr_types_equivalent_p): Cleanup warned checks.
23450 2023-04-21 Richard Biener <rguenther@suse.de>
23452 PR tree-optimization/109573
23453 * tree-vect-loop.cc (vectorizable_live_operation): Allow
23454 unhandled SSA copy as well. Demote assert to checking only.
23456 2023-04-21 Richard Biener <rguenther@suse.de>
23458 * df-core.cc (df_analyze): Compute RPO on the reverse graph
23459 for DF_BACKWARD problems.
23460 (loop_post_order_compute): Rename to ...
23461 (loop_rev_post_order_compute): ... this, compute a RPO.
23462 (loop_inverted_post_order_compute): Rename to ...
23463 (loop_inverted_rev_post_order_compute): ... this, compute a RPO.
23464 (df_analyze_loop): Use RPO on the forward graph for DF_FORWARD
23465 problems, RPO on the inverted graph for DF_BACKWARD.
23467 2023-04-21 Richard Biener <rguenther@suse.de>
23469 * cfganal.h (inverted_rev_post_order_compute): Rename
23471 (inverted_post_order_compute): ... this. Add struct function
23472 argument, change allocation to a C array.
23473 * cfganal.cc (inverted_rev_post_order_compute): Likewise.
23474 * lcm.cc (compute_antinout_edge): Adjust.
23475 * lra-lives.cc (lra_create_live_ranges_1): Likewise.
23476 * tree-ssa-dce.cc (remove_dead_stmt): Likewise.
23477 * tree-ssa-pre.cc (compute_antic): Likewise.
23479 2023-04-21 Richard Biener <rguenther@suse.de>
23481 * df.h (df_d::postorder_inverted): Change back to int *,
23483 * df-core.cc (rest_of_handle_df_finish): Adjust.
23484 (df_analyze_1): Likewise.
23485 (df_analyze): For DF_FORWARD problems use RPO on the forward
23487 (loop_inverted_post_order_compute): Adjust API.
23488 (df_analyze_loop): Adjust.
23489 (df_get_n_blocks): Likewise.
23490 (df_get_postorder): Likewise.
23492 2023-04-21 Juzhe-Zhong <juzhe.zhong@rivai.ai>
23495 * config/riscv/riscv-vsetvl.cc
23496 (vector_infos_manager::all_empty_predecessor_p): New function.
23497 (pass_vsetvl::backward_demand_fusion): Ditto.
23498 * config/riscv/riscv-vsetvl.h: Ditto.
23500 2023-04-21 Robin Dapp <rdapp@ventanamicro.com>
23503 * config/riscv/generic.md: Change standard names to insn names.
23505 2023-04-21 Richard Biener <rguenther@suse.de>
23507 * lcm.cc (compute_antinout_edge): Use RPO on the inverted graph.
23508 (compute_laterin): Use RPO.
23509 (compute_available): Likewise.
23511 2023-04-21 Peng Fan <fanpeng@loongson.cn>
23513 * config/loongarch/gnu-user.h (MUSL_DYNAMIC_LINKER): Redefine.
23515 2023-04-21 Juzhe-Zhong <juzhe.zhong@rivai.ai>
23518 * config/riscv/riscv-vsetvl.cc (local_eliminate_vsetvl_insn): New function.
23519 (vector_insn_info::skip_avl_compatible_p): Ditto.
23520 (vector_insn_info::merge): Remove default value.
23521 (pass_vsetvl::compute_local_backward_infos): Ditto.
23522 (pass_vsetvl::cleanup_insns): Add local vsetvl elimination.
23523 * config/riscv/riscv-vsetvl.h: Ditto.
23525 2023-04-20 Alejandro Colomar <alx.manpages@gmail.com>
23527 * doc/extend.texi (Common Function Attributes): Remove duplicate
23530 2023-04-20 Andrew MacLeod <amacleod@redhat.com>
23532 PR tree-optimization/109564
23533 * gimple-range-fold.cc (fold_using_range::range_of_phi): Do no ignore
23534 UNDEFINED range names when deciding if all PHI arguments are the same,
23536 2023-04-20 Jakub Jelinek <jakub@redhat.com>
23538 PR tree-optimization/109011
23539 * tree-vect-patterns.cc (vect_recog_ctz_ffs_pattern): Use
23540 .CTZ (X) = .POPCOUNT ((X - 1) & ~X) in preference to
23541 .CTZ (X) = PREC - .POPCOUNT (X | -X).
23543 2023-04-20 Vladimir N. Makarov <vmakarov@redhat.com>
23545 * lra-constraints.cc (match_reload): Exclude some hard regs for
23546 multi-reg inout reload pseudos used in asm in different mode.
23548 2023-04-20 Uros Bizjak <ubizjak@gmail.com>
23550 * config/arm/arm.cc (thumb1_legitimate_address_p):
23551 Use VIRTUAL_REGISTER_P predicate.
23552 (arm_eliminable_register): Ditto.
23553 * config/avr/avr.md (push<mode>_1): Ditto.
23554 * config/bfin/predicates.md (register_no_elim_operand): Ditto.
23555 * config/h8300/predicates.md (register_no_sp_elim_operand): Ditto.
23556 * config/i386/predicates.md (register_no_elim_operand): Ditto.
23557 * config/iq2000/predicates.md (call_insn_operand): Ditto.
23558 * config/microblaze/microblaze.h (CALL_INSN_OP): Ditto.
23560 2023-04-20 Uros Bizjak <ubizjak@gmail.com>
23563 * config/i386/predicates.md (extract_operator): New predicate.
23564 * config/i386/i386.md (any_extract): Remove code iterator.
23565 (*cmpqi_ext<mode>_1_mem_rex64): Use extract_operator predicate.
23566 (*cmpqi_ext<mode>_1): Ditto.
23567 (*cmpqi_ext<mode>_2): Ditto.
23568 (*cmpqi_ext<mode>_3_mem_rex64): Ditto.
23569 (*cmpqi_ext<mode>_3): Ditto.
23570 (*cmpqi_ext<mode>_4): Ditto.
23571 (*extzvqi_mem_rex64): Ditto.
23573 (*insvqi_2): Ditto.
23574 (*extendqi<SWI24:mode>_ext_1): Ditto.
23575 (*addqi_ext<mode>_0): Ditto.
23576 (*addqi_ext<mode>_1): Ditto.
23577 (*addqi_ext<mode>_2): Ditto.
23578 (*subqi_ext<mode>_0): Ditto.
23579 (*subqi_ext<mode>_2): Ditto.
23580 (*testqi_ext<mode>_1): Ditto.
23581 (*testqi_ext<mode>_2): Ditto.
23582 (*andqi_ext<mode>_0): Ditto.
23583 (*andqi_ext<mode>_1): Ditto.
23584 (*andqi_ext<mode>_1_cc): Ditto.
23585 (*andqi_ext<mode>_2): Ditto.
23586 (*<any_or:code>qi_ext<mode>_0): Ditto.
23587 (*<any_or:code>qi_ext<mode>_1): Ditto.
23588 (*<any_or:code>qi_ext<mode>_2): Ditto.
23589 (*xorqi_ext<mode>_1_cc): Ditto.
23590 (*negqi_ext<mode>_2): Ditto.
23591 (*ashlqi_ext<mode>_2): Ditto.
23592 (*<any_shiftrt:insn>qi_ext<mode>_2): Ditto.
23594 2023-04-20 Raphael Zinsly <rzinsly@ventanamicro.com>
23597 * config/riscv/bitmanip.md (clz, ctz, pcnt, min, max patterns): Use
23598 <bitmanip_insn> as the type to allow for fine grained control of
23599 scheduling these insns.
23600 * config/riscv/generic.md (generic_alu): Add bitmanip, clz, ctz, pcnt,
23602 * config/riscv/riscv.md (type attribute): Add types for clz, ctz,
23603 pcnt, signed and unsigned min/max.
23605 2023-04-20 Juzhe-Zhong <juzhe.zhong@rivai.ai>
23606 kito-cheng <kito.cheng@sifive.com>
23608 * config/riscv/riscv.h (enum reg_class): Fix RVV register order.
23610 2023-04-20 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
23611 kito-cheng <kito.cheng@sifive.com>
23614 * config/riscv/riscv-vsetvl.cc (count_regno_occurrences): New function.
23615 (pass_vsetvl::cleanup_insns): Fix bug.
23617 2023-04-20 Andrew Stubbs <ams@codesourcery.com>
23619 * config/gcn/gcn-valu.md (vnsi, VnSI): Add scalar modes.
23620 (ldexp<mode>3): Delete.
23621 (ldexp<mode>3<exec>): Change "B" to "A".
23623 2023-04-20 Jakub Jelinek <jakub@redhat.com>
23624 Jonathan Wakely <jwakely@redhat.com>
23626 * tree.h (built_in_function_equal_p): New helper function.
23627 (fndecl_built_in_p): Turn into variadic template to support
23628 1 or more built_in_function arguments.
23629 * builtins.cc (fold_builtin_expect): Use 3 argument fndecl_built_in_p.
23630 * gimplify.cc (goa_stabilize_expr): Likewise.
23631 * cgraphclones.cc (cgraph_node::create_clone): Likewise.
23632 * ipa-fnsummary.cc (compute_fn_summary): Likewise.
23633 * omp-low.cc (setjmp_or_longjmp_p): Likewise.
23634 * cgraph.cc (cgraph_edge::redirect_call_stmt_to_callee,
23635 cgraph_update_edges_for_call_stmt_node,
23636 cgraph_edge::verify_corresponds_to_fndecl,
23637 cgraph_node::verify_node): Likewise.
23638 * tree-stdarg.cc (optimize_va_list_gpr_fpr_size): Likewise.
23639 * gimple-ssa-warn-access.cc (matching_alloc_calls_p): Likewise.
23640 * ipa-prop.cc (try_make_edge_direct_virtual_call): Likewise.
23642 2023-04-20 Jakub Jelinek <jakub@redhat.com>
23644 PR tree-optimization/109011
23645 * tree-vect-patterns.cc (vect_recog_ctz_ffs_pattern): New function.
23646 (vect_recog_popcount_clz_ctz_ffs_pattern): Move vect_pattern_detected
23647 call later. Don't punt for IFN_CTZ or IFN_FFS if it doesn't have
23648 direct optab support, but has instead IFN_CLZ, IFN_POPCOUNT or
23649 for IFN_FFS IFN_CTZ support, use vect_recog_ctz_ffs_pattern for that
23651 (vect_vect_recog_func_ptrs): Add ctz_ffs entry.
23653 2023-04-20 Richard Biener <rguenther@suse.de>
23655 * df-core.cc (rest_of_handle_df_initialize): Remove
23656 computation of df->postorder, df->postorder_inverted and
23659 2023-04-20 Haochen Jiang <haochen.jiang@intel.com>
23661 * common/config/i386/i386-common.cc
23662 (OPTION_MASK_ISA2_AVX_UNSET): Add OPTION_MASK_ISA2_VAES_UNSET.
23663 (ix86_handle_option): Set AVX flag for VAES.
23664 * config/i386/i386-builtins.cc (ix86_init_mmx_sse_builtins):
23665 Add OPTION_MASK_ISA2_VAES_UNSET.
23666 (def_builtin): Share builtin between AES and VAES.
23667 * config/i386/i386-expand.cc (ix86_check_builtin_isa_match):
23669 * config/i386/i386.md (aes): New isa attribute.
23670 * config/i386/sse.md (aesenc): Add pattern for VAES with xmm.
23671 (aesenclast): Ditto.
23673 (aesdeclast): Ditto.
23674 * config/i386/vaesintrin.h: Remove redundant avx target push.
23675 * config/i386/wmmintrin.h (_mm_aesdec_si128): Change to macro.
23676 (_mm_aesdeclast_si128): Ditto.
23677 (_mm_aesenc_si128): Ditto.
23678 (_mm_aesenclast_si128): Ditto.
23680 2023-04-20 Hu, Lin1 <lin1.hu@intel.com>
23682 * config/i386/avx2intrin.h
23683 (_MM_REDUCE_OPERATOR_BASIC_EPI16): New macro.
23684 (_MM_REDUCE_OPERATOR_MAX_MIN_EP16): Ditto.
23685 (_MM256_REDUCE_OPERATOR_BASIC_EPI16): Ditto.
23686 (_MM256_REDUCE_OPERATOR_MAX_MIN_EP16): Ditto.
23687 (_MM_REDUCE_OPERATOR_BASIC_EPI8): Ditto.
23688 (_MM_REDUCE_OPERATOR_MAX_MIN_EP8): Ditto.
23689 (_MM256_REDUCE_OPERATOR_BASIC_EPI8): Ditto.
23690 (_MM256_REDUCE_OPERATOR_MAX_MIN_EP8): Ditto.
23691 (_mm_reduce_add_epi16): New instrinsics.
23692 (_mm_reduce_mul_epi16): Ditto.
23693 (_mm_reduce_and_epi16): Ditto.
23694 (_mm_reduce_or_epi16): Ditto.
23695 (_mm_reduce_max_epi16): Ditto.
23696 (_mm_reduce_max_epu16): Ditto.
23697 (_mm_reduce_min_epi16): Ditto.
23698 (_mm_reduce_min_epu16): Ditto.
23699 (_mm256_reduce_add_epi16): Ditto.
23700 (_mm256_reduce_mul_epi16): Ditto.
23701 (_mm256_reduce_and_epi16): Ditto.
23702 (_mm256_reduce_or_epi16): Ditto.
23703 (_mm256_reduce_max_epi16): Ditto.
23704 (_mm256_reduce_max_epu16): Ditto.
23705 (_mm256_reduce_min_epi16): Ditto.
23706 (_mm256_reduce_min_epu16): Ditto.
23707 (_mm_reduce_add_epi8): Ditto.
23708 (_mm_reduce_mul_epi8): Ditto.
23709 (_mm_reduce_and_epi8): Ditto.
23710 (_mm_reduce_or_epi8): Ditto.
23711 (_mm_reduce_max_epi8): Ditto.
23712 (_mm_reduce_max_epu8): Ditto.
23713 (_mm_reduce_min_epi8): Ditto.
23714 (_mm_reduce_min_epu8): Ditto.
23715 (_mm256_reduce_add_epi8): Ditto.
23716 (_mm256_reduce_mul_epi8): Ditto.
23717 (_mm256_reduce_and_epi8): Ditto.
23718 (_mm256_reduce_or_epi8): Ditto.
23719 (_mm256_reduce_max_epi8): Ditto.
23720 (_mm256_reduce_max_epu8): Ditto.
23721 (_mm256_reduce_min_epi8): Ditto.
23722 (_mm256_reduce_min_epu8): Ditto.
23723 * config/i386/avx512vlbwintrin.h:
23724 (_mm_mask_reduce_add_epi16): Ditto.
23725 (_mm_mask_reduce_mul_epi16): Ditto.
23726 (_mm_mask_reduce_and_epi16): Ditto.
23727 (_mm_mask_reduce_or_epi16): Ditto.
23728 (_mm_mask_reduce_max_epi16): Ditto.
23729 (_mm_mask_reduce_max_epu16): Ditto.
23730 (_mm_mask_reduce_min_epi16): Ditto.
23731 (_mm_mask_reduce_min_epu16): Ditto.
23732 (_mm256_mask_reduce_add_epi16): Ditto.
23733 (_mm256_mask_reduce_mul_epi16): Ditto.
23734 (_mm256_mask_reduce_and_epi16): Ditto.
23735 (_mm256_mask_reduce_or_epi16): Ditto.
23736 (_mm256_mask_reduce_max_epi16): Ditto.
23737 (_mm256_mask_reduce_max_epu16): Ditto.
23738 (_mm256_mask_reduce_min_epi16): Ditto.
23739 (_mm256_mask_reduce_min_epu16): Ditto.
23740 (_mm_mask_reduce_add_epi8): Ditto.
23741 (_mm_mask_reduce_mul_epi8): Ditto.
23742 (_mm_mask_reduce_and_epi8): Ditto.
23743 (_mm_mask_reduce_or_epi8): Ditto.
23744 (_mm_mask_reduce_max_epi8): Ditto.
23745 (_mm_mask_reduce_max_epu8): Ditto.
23746 (_mm_mask_reduce_min_epi8): Ditto.
23747 (_mm_mask_reduce_min_epu8): Ditto.
23748 (_mm256_mask_reduce_add_epi8): Ditto.
23749 (_mm256_mask_reduce_mul_epi8): Ditto.
23750 (_mm256_mask_reduce_and_epi8): Ditto.
23751 (_mm256_mask_reduce_or_epi8): Ditto.
23752 (_mm256_mask_reduce_max_epi8): Ditto.
23753 (_mm256_mask_reduce_max_epu8): Ditto.
23754 (_mm256_mask_reduce_min_epi8): Ditto.
23755 (_mm256_mask_reduce_min_epu8): Ditto.
23757 2023-04-20 Haochen Jiang <haochen.jiang@intel.com>
23759 * common/config/i386/i386-common.cc
23760 (OPTION_MASK_ISA_VPCLMULQDQ_SET):
23761 Add OPTION_MASK_ISA_PCLMUL_SET and OPTION_MASK_ISA_AVX_SET.
23762 (OPTION_MASK_ISA_AVX_UNSET):
23763 Add OPTION_MASK_ISA_VPCLMULQDQ_UNSET.
23764 (OPTION_MASK_ISA_PCLMUL_UNSET): Ditto.
23765 * config/i386/i386.md (vpclmulqdqvl): New.
23766 * config/i386/sse.md (pclmulqdq): Add evex encoding.
23767 * config/i386/vpclmulqdqintrin.h: Remove redudant avx target
23770 2023-04-20 Haochen Jiang <haochen.jiang@intel.com>
23772 * config/i386/avx512vlbwintrin.h
23773 (_mm_mask_blend_epi16): Remove __OPTIMIZE__ wrapper.
23774 (_mm_mask_blend_epi8): Ditto.
23775 (_mm256_mask_blend_epi16): Ditto.
23776 (_mm256_mask_blend_epi8): Ditto.
23777 * config/i386/avx512vlintrin.h
23778 (_mm256_mask_blend_pd): Ditto.
23779 (_mm256_mask_blend_ps): Ditto.
23780 (_mm256_mask_blend_epi64): Ditto.
23781 (_mm256_mask_blend_epi32): Ditto.
23782 (_mm_mask_blend_pd): Ditto.
23783 (_mm_mask_blend_ps): Ditto.
23784 (_mm_mask_blend_epi64): Ditto.
23785 (_mm_mask_blend_epi32): Ditto.
23786 * config/i386/sse.md (VF_AVX512BWHFBF16): Removed.
23787 (VF_AVX512HFBFVL): Move it before the first usage.
23788 (<avx512>_blendm<mode>): Change iterator from VF_AVX512BWHFBF16
23789 to VF_AVX512HFBFVL.
23791 2023-04-20 Haochen Jiang <haochen.jiang@intel.com>
23793 * common/config/i386/i386-common.cc
23794 (OPTION_MASK_ISA_AVX512VBMI2_SET): Change OPTION_MASK_ISA_AVX512F_SET
23795 to OPTION_MASK_ISA_AVX512BW_SET.
23796 (OPTION_MASK_ISA_AVX512F_UNSET):
23797 Remove OPTION_MASK_ISA_AVX512VBMI2_UNSET.
23798 (OPTION_MASK_ISA_AVX512BW_UNSET):
23799 Add OPTION_MASK_ISA_AVX512VBMI2_UNSET.
23800 * config/i386/avx512vbmi2intrin.h: Do not push avx512bw.
23801 * config/i386/avx512vbmi2vlintrin.h: Ditto.
23802 * config/i386/i386-builtin.def: Remove OPTION_MASK_ISA_AVX512BW.
23803 * config/i386/sse.md (VI12_AVX512VLBW): Removed.
23804 (VI12_VI48F_AVX512VLBW): Rename to VI12_VI48F_AVX512VL.
23805 (compress<mode>_mask): Change iterator from VI12_AVX512VLBW to
23807 (compressstore<mode>_mask): Ditto.
23808 (expand<mode>_mask): Ditto.
23809 (expand<mode>_maskz): Ditto.
23810 (*expand<mode>_mask): Change iterator from VI12_VI48F_AVX512VLBW to
23811 VI12_VI48F_AVX512VL.
23813 2023-04-20 Haochen Jiang <haochen.jiang@intel.com>
23815 * common/config/i386/i386-common.cc
23816 (OPTION_MASK_ISA_AVX512BITALG_SET):
23817 Change OPTION_MASK_ISA_AVX512F_SET
23818 to OPTION_MASK_ISA_AVX512BW_SET.
23819 (OPTION_MASK_ISA_AVX512F_UNSET):
23820 Remove OPTION_MASK_ISA_AVX512BITALG_SET.
23821 (OPTION_MASK_ISA_AVX512BW_UNSET):
23822 Add OPTION_MASK_ISA_AVX512BITALG_SET.
23823 * config/i386/avx512bitalgintrin.h: Do not push avx512bw.
23824 * config/i386/i386-builtin.def:
23825 Remove redundant OPTION_MASK_ISA_AVX512BW.
23826 * config/i386/sse.md (VI1_AVX512VLBW): Removed.
23827 (avx512vl_vpshufbitqmb<mode><mask_scalar_merge_name>):
23828 Change the iterator from VI1_AVX512VLBW to VI1_AVX512VL.
23830 2023-04-20 Haochen Jiang <haochen.jiang@intel.com>
23832 * config/i386/i386-expand.cc
23833 (ix86_check_builtin_isa_match): Correct wrong comments.
23834 Add a new macro SHARE_BUILTIN and refactor the current if
23837 2023-04-20 Mo, Zewei <zewei.mo@intel.com>
23839 * config/i386/cpuid.h: Open a new section for Extended Features
23840 Leaf (%eax == 7, %ecx == 0) and Extended Features Sub-leaf (%eax == 7,
23843 2023-04-20 Hu, Lin1 <lin1.hu@intel.com>
23845 * config/i386/sse.md: Modify insn vperm{i,f}
23848 2023-04-19 Max Filippov <jcmvbkbc@gmail.com>
23850 * config/xtensa/xtensa-opts.h: New header.
23851 * config/xtensa/xtensa.h (STRICT_ALIGNMENT): Redefine as
23852 xtensa_strict_align.
23853 * config/xtensa/xtensa.cc (xtensa_option_override): When
23854 -m[no-]strict-align is not specified in the command line set
23855 xtensa_strict_align to 0 if the hardware supports both unaligned
23856 loads and stores or to 1 otherwise.
23857 * config/xtensa/xtensa.opt (mstrict-align): New option.
23858 * doc/invoke.texi (Xtensa Options): Document -m[no-]strict-align.
23860 2023-04-19 Max Filippov <jcmvbkbc@gmail.com>
23862 * config/xtensa/xtensa-dynconfig.cc (xtensa_get_config_v4): New
23865 2023-04-19 Andrew Pinski <apinski@marvell.com>
23867 * config/i386/i386.md (*movsicc_noc_zext_1): New pattern.
23869 2023-04-19 Juzhe-Zhong <juzhe.zhong@rivai.ai>
23871 * config/riscv/riscv-modes.def (FLOAT_MODE): Add chunk 128 support.
23872 (VECTOR_BOOL_MODE): Ditto.
23873 (ADJUST_NUNITS): Ditto.
23874 (ADJUST_ALIGNMENT): Ditto.
23875 (ADJUST_BYTESIZE): Ditto.
23876 (ADJUST_PRECISION): Ditto.
23877 (RVV_MODES): Ditto.
23878 (VECTOR_MODE_WITH_PREFIX): Ditto.
23879 * config/riscv/riscv-v.cc (ENTRY): Ditto.
23880 (get_vlmul): Ditto.
23881 (get_ratio): Ditto.
23882 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_TYPE): Ditto.
23883 * config/riscv/riscv-vector-builtins.def (DEF_RVV_TYPE): Ditto.
23884 (vbool64_t): Ditto.
23885 (vbool32_t): Ditto.
23886 (vbool16_t): Ditto.
23891 (vint8mf8_t): Ditto.
23892 (vuint8mf8_t): Ditto.
23893 (vint8mf4_t): Ditto.
23894 (vuint8mf4_t): Ditto.
23895 (vint8mf2_t): Ditto.
23896 (vuint8mf2_t): Ditto.
23897 (vint8m1_t): Ditto.
23898 (vuint8m1_t): Ditto.
23899 (vint8m2_t): Ditto.
23900 (vuint8m2_t): Ditto.
23901 (vint8m4_t): Ditto.
23902 (vuint8m4_t): Ditto.
23903 (vint8m8_t): Ditto.
23904 (vuint8m8_t): Ditto.
23905 (vint16mf4_t): Ditto.
23906 (vuint16mf4_t): Ditto.
23907 (vint16mf2_t): Ditto.
23908 (vuint16mf2_t): Ditto.
23909 (vint16m1_t): Ditto.
23910 (vuint16m1_t): Ditto.
23911 (vint16m2_t): Ditto.
23912 (vuint16m2_t): Ditto.
23913 (vint16m4_t): Ditto.
23914 (vuint16m4_t): Ditto.
23915 (vint16m8_t): Ditto.
23916 (vuint16m8_t): Ditto.
23917 (vint32mf2_t): Ditto.
23918 (vuint32mf2_t): Ditto.
23919 (vint32m1_t): Ditto.
23920 (vuint32m1_t): Ditto.
23921 (vint32m2_t): Ditto.
23922 (vuint32m2_t): Ditto.
23923 (vint32m4_t): Ditto.
23924 (vuint32m4_t): Ditto.
23925 (vint32m8_t): Ditto.
23926 (vuint32m8_t): Ditto.
23927 (vint64m1_t): Ditto.
23928 (vuint64m1_t): Ditto.
23929 (vint64m2_t): Ditto.
23930 (vuint64m2_t): Ditto.
23931 (vint64m4_t): Ditto.
23932 (vuint64m4_t): Ditto.
23933 (vint64m8_t): Ditto.
23934 (vuint64m8_t): Ditto.
23935 (vfloat32mf2_t): Ditto.
23936 (vfloat32m1_t): Ditto.
23937 (vfloat32m2_t): Ditto.
23938 (vfloat32m4_t): Ditto.
23939 (vfloat32m8_t): Ditto.
23940 (vfloat64m1_t): Ditto.
23941 (vfloat64m2_t): Ditto.
23942 (vfloat64m4_t): Ditto.
23943 (vfloat64m8_t): Ditto.
23944 * config/riscv/riscv-vector-switch.def (ENTRY): Ditto.
23945 * config/riscv/riscv.cc (riscv_legitimize_poly_move): Ditto.
23946 (riscv_convert_vector_bits): Ditto.
23947 * config/riscv/riscv.md:
23948 * config/riscv/vector-iterators.md:
23949 * config/riscv/vector.md
23950 (@pred_indexed_<order>store<VNX32_QH:mode><VNX32_QHI:mode>): Ditto.
23951 (@pred_indexed_<order>store<VNX32_QHS:mode><VNX32_QHSI:mode>): Ditto.
23952 (@pred_indexed_<order>store<VNX64_Q:mode><VNX64_Q:mode>): Ditto.
23953 (@pred_indexed_<order>store<VNX64_QH:mode><VNX64_QHI:mode>): Ditto.
23954 (@pred_indexed_<order>store<VNX128_Q:mode><VNX128_Q:mode>): Ditto.
23955 (@pred_reduc_<reduc><mode><vlmul1_zve64>): Ditto.
23956 (@pred_widen_reduc_plus<v_su><mode><vwlmul1_zve64>): Ditto.
23957 (@pred_reduc_plus<order><mode><vlmul1_zve64>): Ditto.
23958 (@pred_widen_reduc_plus<order><mode><vwlmul1_zve64>): Ditto.
23960 2023-04-19 Pan Li <pan2.li@intel.com>
23962 * simplify-rtx.cc (simplify_context::simplify_binary_operation_1):
23963 Align IOR (A | (~A) -> -1) optimization MODE_CLASS condition to AND.
23965 2023-04-19 Uros Bizjak <ubizjak@gmail.com>
23969 * config/i386/i386.md (*cmpqi_ext<mode>_1_mem_rex64): New insn pattern.
23970 (*cmpqi_ext<mode>_1): Use nonimmediate_operand predicate
23971 for operand 0. Use any_extract code iterator.
23972 (*cmpqi_ext<mode>_1 peephole2): New peephole2 pattern.
23973 (*cmpqi_ext<mode>_2): Use any_extract code iterator.
23974 (*cmpqi_ext<mode>_3_mem_rex64): New insn pattern.
23975 (*cmpqi_ext<mode>_1): Use general_operand predicate
23976 for operand 1. Use any_extract code iterator.
23977 (*cmpqi_ext<mode>_3 peephole2): New peephole2 pattern.
23978 (*cmpqi_ext<mode>_4): Use any_extract code iterator.
23980 2023-04-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23982 * config/aarch64/aarch64-simd.md (aarch64_saddw2<mode>): Delete.
23983 (aarch64_uaddw2<mode>): Delete.
23984 (aarch64_ssubw2<mode>): Delete.
23985 (aarch64_usubw2<mode>): Delete.
23986 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>): New define_expand.
23988 2023-04-19 Richard Biener <rguenther@suse.de>
23990 * tree-ssa-structalias.cc (do_ds_constraint): Use
23991 solve_add_graph_edge.
23993 2023-04-19 Richard Biener <rguenther@suse.de>
23995 * tree-ssa-structalias.cc (solve_add_graph_edge): New function,
23997 (do_sd_constraint): ... here.
23999 2023-04-19 Richard Biener <rguenther@suse.de>
24001 * tree-cfg.cc (gimple_can_merge_blocks_p): Remove condition
24002 rejecting the merge when A contains only a non-local label.
24004 2023-04-19 Uros Bizjak <ubizjak@gmail.com>
24006 * rtl.h (VIRTUAL_REGISTER_P): New predicate.
24007 (VIRTUAL_REGISTER_NUM_P): Ditto.
24008 (REGNO_PTR_FRAME_P): Use VIRTUAL_REGISTER_NUM_P predicate.
24009 * expr.cc (force_operand): Use VIRTUAL_REGISTER_P predicate.
24010 * function.cc (instantiate_decl_rtl): Ditto.
24011 * rtlanal.cc (rtx_addr_can_trap_p_1): Ditto.
24012 (nonzero_address_p): Ditto.
24013 (refers_to_regno_p): Use VIRTUAL_REGISTER_NUM_P predicate.
24015 2023-04-19 Aldy Hernandez <aldyh@redhat.com>
24017 * value-range.h (Value_Range::Value_Range): Avoid pointer sharing.
24019 2023-04-19 Richard Biener <rguenther@suse.de>
24021 * system.h (auto_mpz::operator->()): New.
24022 * realmpfr.h (auto_mpfr::operator->()): New.
24023 * builtins.cc (do_mpfr_lgamma_r): Use auto_mpfr.
24024 * real.cc (real_from_string): Likewise.
24025 (dconst_e_ptr): Likewise.
24026 (dconst_sqrt2_ptr): Likewise.
24027 * tree-ssa-loop-niter.cc (refine_value_range_using_guard):
24029 (bound_difference_of_offsetted_base): Likewise.
24030 (number_of_iterations_ne): Likewise.
24031 (number_of_iterations_lt_to_ne): Likewise.
24032 * ubsan.cc: Include realmpfr.h.
24033 (ubsan_instrument_float_cast): Use auto_mpfr.
24035 2023-04-19 Richard Biener <rguenther@suse.de>
24037 * tree-ssa-structalias.cc (solve_graph): Remove self-copy
24038 edges, remove edges from escaped after special-casing them.
24040 2023-04-19 Richard Biener <rguenther@suse.de>
24042 * tree-ssa-structalias.cc (do_sd_constraint): Fixup escape
24045 2023-04-19 Richard Biener <rguenther@suse.de>
24047 * tree-ssa-structalias.cc (do_sd_constraint): Do not write
24048 to the LHS varinfo solution member.
24050 2023-04-19 Richard Biener <rguenther@suse.de>
24052 * tree-ssa-structalias.cc (topo_visit): Look at the real
24053 destination of edges.
24055 2023-04-19 Richard Biener <rguenther@suse.de>
24057 PR tree-optimization/44794
24058 * tree-ssa-loop-manip.cc (tree_transform_and_unroll_loop):
24059 If an epilogue loop is required set its iteration upper bound.
24061 2023-04-19 Xi Ruoyao <xry111@xry111.site>
24064 * config/loongarch/loongarch-protos.h
24065 (loongarch_expand_block_move): Add a parameter as alignment RTX.
24066 * config/loongarch/loongarch.h:
24067 (LARCH_MAX_MOVE_BYTES_PER_LOOP_ITER): Remove.
24068 (LARCH_MAX_MOVE_BYTES_STRAIGHT): Remove.
24069 (LARCH_MAX_MOVE_OPS_PER_LOOP_ITER): Define.
24070 (LARCH_MAX_MOVE_OPS_STRAIGHT): Define.
24071 (MOVE_RATIO): Use LARCH_MAX_MOVE_OPS_PER_LOOP_ITER instead of
24072 LARCH_MAX_MOVE_BYTES_PER_LOOP_ITER.
24073 * config/loongarch/loongarch.cc (loongarch_expand_block_move):
24074 Take the alignment from the parameter, but set it to
24075 UNITS_PER_WORD if !TARGET_STRICT_ALIGN. Limit the length of
24076 straight-line implementation with LARCH_MAX_MOVE_OPS_STRAIGHT
24077 instead of LARCH_MAX_MOVE_BYTES_STRAIGHT.
24078 (loongarch_block_move_straight): When there are left-over bytes,
24079 half the mode size instead of falling back to byte mode at once.
24080 (loongarch_block_move_loop): Limit the length of loop body with
24081 LARCH_MAX_MOVE_OPS_PER_LOOP_ITER instead of
24082 LARCH_MAX_MOVE_BYTES_PER_LOOP_ITER.
24083 * config/loongarch/loongarch.md (cpymemsi): Pass the alignment
24084 to loongarch_expand_block_move.
24086 2023-04-19 Xi Ruoyao <xry111@xry111.site>
24088 * config/loongarch/loongarch.cc
24089 (loongarch_setup_incoming_varargs): Don't save more GARs than
24090 cfun->va_list_gpr_size / UNITS_PER_WORD.
24092 2023-04-19 Richard Biener <rguenther@suse.de>
24094 * tree-ssa-loop-manip.cc (determine_exit_conditions): Fix
24095 no epilogue condition.
24097 2023-04-19 Richard Biener <rguenther@suse.de>
24099 * gimple.h (gimple_assign_load): Outline...
24100 * gimple.cc (gimple_assign_load): ... here. Avoid
24101 get_base_address and instead just strip the outermost
24102 handled component, treating a remaining handled component
24105 2023-04-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24107 * config/aarch64/aarch64-simd-builtins.def (neg): Delete builtins
24109 * config/aarch64/arm_fp16.h (vnegh_f16): Reimplement using normal negation.
24111 2023-04-19 Jakub Jelinek <jakub@redhat.com>
24113 PR tree-optimization/109011
24114 * tree-vect-patterns.cc (vect_recog_popcount_pattern): Rename to ...
24115 (vect_recog_popcount_clz_ctz_ffs_pattern): ... this. Handle also
24116 CLZ, CTZ and FFS. Remove vargs variable, use
24117 gimple_build_call_internal rather than gimple_build_call_internal_vec.
24118 (vect_vect_recog_func_ptrs): Adjust popcount entry.
24120 2023-04-19 Jakub Jelinek <jakub@redhat.com>
24123 * dse.cc (replace_read): If read_reg is a SUBREG of a word mode
24124 REG, for WORD_REGISTER_OPERATIONS copy SUBREG_REG of it into
24125 a new REG rather than the SUBREG.
24127 2023-04-19 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
24129 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_set_zero<mode>):
24132 2023-04-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24135 * config/aarch64/aarch64.cc (aarch64_rtx_costs): Merge ASHIFT and
24136 ROTATE, ROTATERT, LSHIFTRT, ASHIFTRT cases. Handle subregs in op1.
24138 2023-04-19 Richard Biener <rguenther@suse.de>
24140 PR rtl-optimization/109237
24141 * cse.cc (insn_live_p): Remove NEXT_INSN walk, instead check
24142 TREE_VISITED on INSN_VAR_LOCATION_DECL.
24143 (delete_trivially_dead_insns): Maintain TREE_VISITED on
24144 active debug bind INSN_VAR_LOCATION_DECL.
24146 2023-04-19 Richard Biener <rguenther@suse.de>
24148 PR rtl-optimization/109237
24149 * cfgcleanup.cc (bb_is_just_return): Walk insns backwards.
24151 2023-04-19 Christophe Lyon <christophe.lyon@arm.com>
24153 * doc/install.texi (enable-decimal-float): Add AArch64.
24155 2023-04-19 liuhongt <hongtao.liu@intel.com>
24157 PR rtl-optimization/109351
24158 * ira.cc (setup_class_subset_and_memory_move_costs): Check
24159 hard_regno_mode_ok before setting lowest memory move cost for
24160 the mode with different reg classes.
24162 2023-04-18 Jason Merrill <jason@redhat.com>
24164 * doc/invoke.texi: Remove stray @gol.
24166 2023-04-18 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
24168 * ifcvt.cc (cond_move_process_if_block): Consider the result of
24169 targetm.noce_conversion_profitable_p() when replacing the original
24170 sequence with the converted one.
24172 2023-04-18 Mark Harmstone <mark@harmstone.com>
24174 * common.opt (gcodeview): Add new option.
24175 * gcc.cc (driver_handle_option); Handle OPT_gcodeview.
24176 * opts.cc (command_handle_option): Similarly.
24177 * doc/invoke.texi: Add documentation for -gcodeview.
24179 2023-04-18 Andrew Pinski <apinski@marvell.com>
24181 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker): Remove declaration.
24182 (make_pass_phiopt): Make execute out of line.
24183 (tree_ssa_cs_elim): Move code into ...
24184 (pass_cselim::execute): here.
24186 2023-04-18 Sam James <sam@gentoo.org>
24188 * system.h: Drop unused INCLUDE_PTHREAD_H.
24190 2023-04-18 Kevin Lee <kevinl@rivosinc.com>
24192 * tree-vect-data-refs.cc (vect_grouped_store_supported): Add new
24195 2023-04-18 Sinan Lin <sinan.lin@linux.alibaba.com>
24197 * config/riscv/bitmanip.md (rotr<mode>3 expander): Enable for ZBKB.
24198 (bswapdi2, bswapsi2): Similarly.
24200 2023-04-18 Uros Bizjak <ubizjak@gmail.com>
24203 * config/i386/i386-builtin.def (__builtin_ia32_insertps128):
24204 Use CODE_FOR_sse4_1_insertps_v4sf.
24205 * config/i386/i386-expand.cc (expand_vec_perm_insertps): New.
24206 (expand_vec_perm_1): Call expand_vec_per_insertps.
24207 * config/i386/i386.md ("unspec"): Declare UNSPEC_INSERTPS here.
24208 * config/i386/mmx.md (mmxscalarmode): New mode attribute.
24209 (@sse4_1_insertps_<mode>): New insn pattern.
24210 * config/i386/sse.md (@sse4_1_insertps_<mode>): Macroize insn
24211 pattern from sse4_1_insertps using VI4F_128 mode iterator.
24213 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
24215 * value-range.cc (gt_ggc_mx): New.
24217 * value-range.h (class vrange): Add GTY marker.
24218 (class frange): Same.
24219 (gt_ggc_mx): Remove.
24220 (gt_pch_nx): Remove.
24222 2023-04-18 Victor L. Do Nascimento <victor.donascimento@arm.com>
24224 * lra-constraints.cc (constraint_unique): New.
24225 (process_address_1): Apply constraint_unique test.
24226 * recog.cc (constrain_operands): Allow relaxed memory
24229 2023-04-18 Kito Cheng <kito.cheng@sifive.com>
24231 * doc/extend.texi (Target Builtins): Add RISC-V Vector
24233 (RISC-V Vector Intrinsics): Document GCC implemented which
24234 version of RISC-V vector intrinsics and its reference.
24236 2023-04-18 Richard Biener <rguenther@suse.de>
24238 PR middle-end/108786
24239 * bitmap.h (bitmap_clear_first_set_bit): New.
24240 * bitmap.cc (bitmap_first_set_bit_worker): Rename from
24241 bitmap_first_set_bit and add optional clearing of the bit.
24242 (bitmap_first_set_bit): Wrap bitmap_first_set_bit_worker.
24243 (bitmap_clear_first_set_bit): Likewise.
24244 * df-core.cc (df_worklist_dataflow_doublequeue): Use
24245 bitmap_clear_first_set_bit.
24246 * graphite-scop-detection.cc (scop_detection::merge_sese):
24248 * sanopt.cc (sanitize_asan_mark_unpoison): Likewise.
24249 (sanitize_asan_mark_poison): Likewise.
24250 * tree-cfgcleanup.cc (cleanup_tree_cfg_noloop): Likewise.
24251 * tree-into-ssa.cc (rewrite_blocks): Likewise.
24252 * tree-ssa-dce.cc (simple_dce_from_worklist): Likewise.
24253 * tree-ssa-sccvn.cc (do_rpo_vn_1): Likewise.
24255 2023-04-18 Richard Biener <rguenther@suse.de>
24257 * tree-ssa-structalias.cc (dump_sa_stats): Split out from...
24258 (dump_sa_points_to_info): ... this function.
24259 (compute_points_to_sets): Guard large dumps with TDF_DETAILS,
24260 and call dump_sa_stats guarded with TDF_STATS.
24261 (ipa_pta_execute): Likewise.
24262 (compute_may_aliases): Guard dump_alias_info with
24263 TDF_DETAILS|TDF_ALIAS.
24265 2023-04-18 Andrew Pinski <apinski@marvell.com>
24267 * tree-ssa-phiopt.cc (gimple_simplify_phiopt): Dump
24268 the expression that is being tried when TDF_FOLDING
24270 (phiopt_worker::match_simplify_replacement): Dump
24271 the sequence which was created by gimple_simplify_phiopt
24272 when TDF_FOLDING is true.
24274 2023-04-18 Andrew Pinski <apinski@marvell.com>
24276 * tree-ssa-phiopt.cc (match_simplify_replacement):
24277 Simplify code that does the movement slightly.
24279 2023-04-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24281 * config/aarch64/aarch64.md (@aarch64_rev16<mode>): Change to
24283 (rev16<mode>2): Rename to...
24284 (aarch64_rev16<mode>2_alt1): ... This.
24285 (rev16<mode>2_alt): Rename to...
24286 (*aarch64_rev16<mode>2_alt2): ... This.
24288 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
24290 * emit-rtl.cc (init_emit_once): Initialize dconstm0.
24291 * gimple-range-op.cc (class cfn_signbit): Remove dconstm0
24293 * range-op-float.cc (zero_range): Use dconstm0.
24294 (zero_to_inf_range): Same.
24295 * real.h (dconstm0): New.
24296 * value-range.cc (frange::flush_denormals_to_zero): Use dconstm0.
24297 (frange::set_zero): Do not declare dconstm0.
24299 2023-04-18 Richard Biener <rguenther@suse.de>
24301 * system.h (class auto_mpz): New,
24302 * realmpfr.h (class auto_mpfr): Likewise.
24303 * fold-const-call.cc (do_mpfr_arg1): Use auto_mpfr.
24304 (do_mpfr_arg2): Likewise.
24305 * tree-ssa-loop-niter.cc (bound_difference): Use auto_mpz;
24307 2023-04-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24309 * config/aarch64/aarch64-builtins.cc (aarch64_init_simd_intrinsics): Take
24310 builtin flags from intrinsic data rather than hardcoded FLAG_AUTO_FP.
24312 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
24314 * value-range.cc (frange::operator==): Adjust for NAN.
24315 (range_tests_nan): Remove some NAN tests.
24317 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
24319 * inchash.cc (hash::add_real_value): New.
24320 * inchash.h (class hash): Add add_real_value.
24321 * value-range.cc (add_vrange): New.
24322 * value-range.h (inchash::add_vrange): New.
24324 2023-04-18 Richard Biener <rguenther@suse.de>
24326 PR tree-optimization/109539
24327 * gimple-ssa-warn-access.cc (pass_waccess::check_pointer_uses):
24328 Re-implement pointer relatedness for PHIs.
24330 2023-04-18 Andrew Stubbs <ams@codesourcery.com>
24332 * config/gcn/gcn-valu.md (SV_SFDF): New iterator.
24333 (SV_FP): New iterator.
24334 (scalar_mode, SCALAR_MODE): Add identity mappings for scalar modes.
24335 (recip<mode>2): Unify the two patterns using SV_FP.
24336 (div_scale<mode><exec_vcc>): New insn.
24337 (div_fmas<mode><exec>): New insn.
24338 (div_fixup<mode><exec>): New insn.
24339 (div<mode>3): Unify the two expanders and rewrite using hardfp.
24340 * config/gcn/gcn.cc (gcn_md_reorg): Support "vccwait" attribute.
24341 * config/gcn/gcn.md (unspec): Add UNSPEC_DIV_SCALE, UNSPEC_DIV_FMAS,
24342 and UNSPEC_DIV_FIXUP.
24343 (vccwait): New attribute.
24345 2023-04-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24347 * config/aarch64/aarch64.cc (aarch64_validate_mcpu): Add hint to use -march
24348 if the argument matches that.
24350 2023-04-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24352 * config/aarch64/atomics.md
24353 (*aarch64_atomic_load<ALLX:mode>_rcpc_zext):
24354 Use SD_HSDI for destination mode iterator.
24356 2023-04-18 Jin Ma <jinma@linux.alibaba.com>
24358 * common/config/riscv/riscv-common.cc (multi_letter_subset_rank): Swap the order
24359 of z-extensions and s-extensions.
24360 (riscv_subset_list::parse): Likewise.
24362 2023-04-18 Jakub Jelinek <jakub@redhat.com>
24364 PR tree-optimization/109240
24365 * match.pd (fneg/fadd): Rewrite such that it handles both plus as
24366 first vec_perm operand and minus as second using fneg/fadd and
24367 minus as first vec_perm operand and plus as second using fneg/fsub.
24369 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
24371 * data-streamer.cc (bp_pack_real_value): New.
24372 (bp_unpack_real_value): New.
24373 * data-streamer.h (bp_pack_real_value): New.
24374 (bp_unpack_real_value): New.
24375 * tree-streamer-in.cc (unpack_ts_real_cst_value_fields): Use
24376 bp_unpack_real_value.
24377 * tree-streamer-out.cc (pack_ts_real_cst_value_fields): Use
24378 bp_pack_real_value.
24380 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
24382 * wide-int.h (WIDE_INT_MAX_HWIS): New.
24383 (class fixed_wide_int_storage): Use it.
24384 (trailing_wide_ints <N>::set_precision): Use it.
24385 (trailing_wide_ints <N>::extra_size): Use it.
24387 2023-04-18 Xi Ruoyao <xry111@xry111.site>
24389 * config/loongarch/loongarch-protos.h
24390 (loongarch_addu16i_imm12_operand_p): New function prototype.
24391 (loongarch_split_plus_constant): Likewise.
24392 * config/loongarch/loongarch.cc
24393 (loongarch_addu16i_imm12_operand_p): New function.
24394 (loongarch_split_plus_constant): Likewise.
24395 * config/loongarch/loongarch.h (ADDU16I_OPERAND): New macro.
24396 (DUAL_IMM12_OPERAND): Likewise.
24397 (DUAL_ADDU16I_OPERAND): Likewise.
24398 * config/loongarch/constraints.md (La, Lb, Lc, Ld, Le): New
24400 * config/loongarch/predicates.md (const_dual_imm12_operand): New
24402 (const_addu16i_operand): Likewise.
24403 (const_addu16i_imm12_di_operand): Likewise.
24404 (const_addu16i_imm12_si_operand): Likewise.
24405 (plus_di_operand): Likewise.
24406 (plus_si_operand): Likewise.
24407 (plus_si_extend_operand): Likewise.
24408 * config/loongarch/loongarch.md (add<mode>3): Convert to
24409 define_insn_and_split. Use plus_<mode>_operand predicate
24410 instead of arith_operand. Add alternatives for La, Lb, Lc, Ld,
24411 and Le constraints.
24412 (*addsi3_extended): Convert to define_insn_and_split. Use
24413 plus_si_extend_operand instead of arith_operand. Add
24414 alternatives for La and Le alternatives.
24416 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
24418 * value-range.h (Value_Range::Value_Range): New.
24419 (Value_Range::contains_p): New.
24421 2023-04-18 Aldy Hernandez <aldyh@redhat.com>
24423 * value-range.h (class vrange): Make m_discriminator const.
24424 (class irange): Make m_max_ranges const. Adjust constructors
24426 (class unsupported_range): Construct vrange appropriately.
24427 (class frange): Same.
24429 2023-04-18 Lulu Cheng <chenglulu@loongson.cn>
24431 * config/loongarch/loongarch.h (LOGICAL_OP_NON_SHORT_CIRCUIT): Remove the macro
24434 2023-04-18 Lulu Cheng <chenglulu@loongson.cn>
24436 * doc/extend.texi: Add section for LoongArch Base Built-in functions.
24438 2023-04-18 Fei Gao <gaofei@eswincomputing.com>
24440 * config/riscv/riscv.cc (riscv_first_stack_step): Make codes more
24442 (riscv_expand_epilogue): Likewise.
24444 2023-04-17 Fei Gao <gaofei@eswincomputing.com>
24446 * config/riscv/riscv.cc (riscv_expand_prologue): Consider save-restore in
24448 (riscv_expand_epilogue): Consider save-restore in stack deallocation.
24450 2023-04-17 Andrew Pinski <apinski@marvell.com>
24452 * tree-ssa-phiopt.cc (gate_hoist_loads): Remove
24455 2023-04-17 Aldy Hernandez <aldyh@redhat.com>
24457 * gimple-ssa-warn-alloca.cc (pass_walloca::execute): Do not export
24460 2023-04-17 Fei Gao <gaofei@eswincomputing.com>
24462 * config/riscv/riscv.cc (riscv_first_stack_step): Add a new function
24463 parameter remaining_size.
24464 (riscv_compute_frame_info): Adapt new riscv_first_stack_step interface.
24465 (riscv_expand_prologue): Likewise.
24466 (riscv_expand_epilogue): Likewise.
24468 2023-04-17 Feng Wang <wangfeng@eswincomputing.com>
24470 * config/riscv/bitmanip.md (rotrsi3_sext): Support generating
24471 roriw for constant counts.
24472 * rtl.h (reverse_rotate_by_imm_p): Add function declartion
24473 * simplify-rtx.cc (reverse_rotate_by_imm_p): New function.
24474 (simplify_context::simplify_binary_operation_1): Use it.
24475 * expmed.cc (expand_shift_1): Likewise.
24477 2023-04-17 Martin Jambor <mjambor@suse.cz>
24481 * cgraph.h (symtab_node::find_reference): Add parameter use_type.
24482 * ipa-prop.h (ipa_pass_through_data): New flag refdesc_decremented.
24483 (ipa_zap_jf_refdesc): New function.
24484 (ipa_get_jf_pass_through_refdesc_decremented): Likewise.
24485 (ipa_set_jf_pass_through_refdesc_decremented): Likewise.
24486 * ipa-cp.cc (ipcp_discover_new_direct_edges): Provide a value for
24487 the new parameter of find_reference.
24488 (adjust_references_in_caller): Likewise. Make sure the constant jump
24489 function is not used to decrement a refdec counter again. Only
24490 decrement refdesc counters when the pass_through jump function allows
24491 it. Added a detailed dump when decrementing refdesc counters.
24492 * ipa-prop.cc (ipa_print_node_jump_functions_for_edge): Dump new flag.
24493 (ipa_set_jf_simple_pass_through): Initialize the new flag.
24494 (ipa_set_jf_unary_pass_through): Likewise.
24495 (ipa_set_jf_arith_pass_through): Likewise.
24496 (remove_described_reference): Provide a value for the new parameter of
24498 (update_jump_functions_after_inlining): Zap refdesc of new jfunc if
24499 the previous pass_through had a flag mandating that we do so.
24500 (propagate_controlled_uses): Likewise. Only decrement refdesc
24501 counters when the pass_through jump function allows it.
24502 (ipa_edge_args_sum_t::duplicate): Provide a value for the new
24503 parameter of find_reference.
24504 (ipa_write_jump_function): Assert the new flag does not have to be
24506 * symtab.cc (symtab_node::find_reference): Add parameter use_type, use
24509 2023-04-17 Philipp Tomsich <philipp.tomsich@vrull.eu>
24510 Di Zhao <di.zhao@amperecomputing.com>
24512 * config/aarch64/aarch64-tuning-flags.def (AARCH64_EXTRA_TUNING_OPTION):
24513 Add AARCH64_EXTRA_TUNE_NO_LDP_COMBINE.
24514 * config/aarch64/aarch64.cc (aarch64_operands_ok_for_ldpstp):
24515 Check for the above tuning option when processing loads.
24517 2023-04-17 Richard Biener <rguenther@suse.de>
24519 PR tree-optimization/109524
24520 * tree-vrp.cc (remove_unreachable::m_list): Change to a
24521 vector of pairs of block indices.
24522 (remove_unreachable::maybe_register_block): Adjust.
24523 (remove_unreachable::remove_and_update_globals): Likewise.
24524 Deal with removed blocks.
24526 2023-04-16 Jeff Law <jlaw@ventanamicro>
24529 * config/riscv/riscv.cc (riscv_expand_conditional_move): For
24530 TARGET_SFB_ALU, force the true arm into a register.
24532 2023-04-15 John David Anglin <danglin@gcc.gnu.org>
24535 * config/pa/pa-protos.h (pa_function_arg_size): Update prototype.
24536 * config/pa/pa.cc (pa_function_arg): Return NULL_RTX if argument
24538 (pa_arg_partial_bytes): Don't call pa_function_arg_size twice.
24539 (pa_function_arg_size): Change return type to int. Return zero
24540 for arguments larger than 1 GB. Update comments.
24542 2023-04-15 Jakub Jelinek <jakub@redhat.com>
24544 PR tree-optimization/109154
24545 * tree-if-conv.cc (predicate_scalar_phi): For complex PHIs, emit just
24546 args_len - 1 COND_EXPRs rather than args_len. Formatting fix.
24548 2023-04-15 Jason Merrill <jason@redhat.com>
24551 * gimple-ssa-warn-access.cc (pass_waccess::check_dangling_stores):
24552 Overhaul lhs_ref.ref analysis.
24554 2023-04-14 Richard Biener <rguenther@suse.de>
24556 PR tree-optimization/109502
24557 * tree-vect-stmts.cc (vectorizable_assignment): Fix
24558 check for conversion between mask and non-mask types.
24560 2023-04-14 Jeff Law <jlaw@ventanamicro.com>
24561 Jakub Jelinek <jakub@redhat.com>
24565 * combine.cc (simplify_and_const_int_1): Compute nonzero_bits in
24566 word_mode rather than mode if WORD_REGISTER_OPERATIONS and mode is
24567 smaller than word_mode.
24568 * simplify-rtx.cc (simplify_context::simplify_binary_operation_1)
24569 <case AND>: Likewise.
24571 2023-04-14 Jakub Jelinek <jakub@redhat.com>
24573 * loop-iv.cc (iv_number_of_iterations): Use gen_int_mode instead
24576 2023-04-13 Andrew MacLeod <amacleod@redhat.com>
24578 PR tree-optimization/108139
24579 PR tree-optimization/109462
24580 * gimple-range-cache.cc (ranger_cache::fill_block_cache): Remove
24581 equivalency check for PHI nodes.
24582 * gimple-range-fold.cc (fold_using_range::range_of_phi): Ensure def
24583 does not dominate single-arg equivalency edges.
24585 2023-04-13 Richard Sandiford <richard.sandiford@arm.com>
24588 * config/aarch64/aarch64.cc (aarch64_function_arg_alignment): Do
24589 not trust TYPE_ALIGN for pointer types; use POINTER_SIZE instead.
24591 2023-04-13 Richard Biener <rguenther@suse.de>
24593 PR tree-optimization/109491
24594 * tree-ssa-sccvn.cc (expressions_equal_p): Restore the
24595 NULL operands test.
24597 2023-04-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
24600 * config/riscv/riscv-vector-builtins-types.def (vint8mf8_t): Fix predicate.
24601 (vint16mf4_t): Ditto.
24602 (vint32mf2_t): Ditto.
24603 (vint64m1_t): Ditto.
24604 (vint64m2_t): Ditto.
24605 (vint64m4_t): Ditto.
24606 (vint64m8_t): Ditto.
24607 (vuint8mf8_t): Ditto.
24608 (vuint16mf4_t): Ditto.
24609 (vuint32mf2_t): Ditto.
24610 (vuint64m1_t): Ditto.
24611 (vuint64m2_t): Ditto.
24612 (vuint64m4_t): Ditto.
24613 (vuint64m8_t): Ditto.
24614 (vfloat32mf2_t): Ditto.
24615 (vbool64_t): Ditto.
24616 * config/riscv/riscv-vector-builtins.cc (register_builtin_type): Add comments.
24617 (register_vector_type): Ditto.
24618 (check_required_extensions): Fix condition.
24619 * config/riscv/riscv-vector-builtins.h (RVV_REQUIRE_ZVE64): Remove it.
24620 (RVV_REQUIRE_ELEN_64): New define.
24621 (RVV_REQUIRE_MIN_VLEN_64): Ditto.
24622 * config/riscv/riscv-vector-switch.def (TARGET_VECTOR_FP32): Remove it.
24623 (TARGET_VECTOR_FP64): Ditto.
24624 (ENTRY): Fix predicate.
24625 * config/riscv/vector-iterators.md: Fix predicate.
24627 2023-04-12 Jakub Jelinek <jakub@redhat.com>
24629 PR tree-optimization/109410
24630 * tree-ssa-reassoc.cc (build_and_add_sum): Split edge from entry
24631 block if first statement of the function is a call to returns_twice
24634 2023-04-12 Jakub Jelinek <jakub@redhat.com>
24637 * config/i386/i386.cc: Include rtl-error.h.
24638 (ix86_print_operand): For z modifier warning, use warning_for_asm
24639 if this_is_asm_operands. For Z modifier errors, use %c and code
24640 instead of hardcoded Z.
24642 2023-04-12 Costas Argyris <costas.argyris@gmail.com>
24644 * config/i386/x-mingw32-utf8: Remove extrataneous $@
24646 2023-04-12 Andrew MacLeod <amacleod@redhat.com>
24648 PR tree-optimization/109462
24649 * gimple-range-cache.cc (ranger_cache::fill_block_cache): Don't
24650 check for equivalences if NAME is a phi node.
24652 2023-04-12 Richard Biener <rguenther@suse.de>
24654 PR tree-optimization/109473
24655 * tree-vect-loop.cc (vect_create_epilog_for_reduction):
24656 Convert scalar result to the computation type before performing
24657 the reduction adjustment.
24659 2023-04-12 Richard Biener <rguenther@suse.de>
24661 PR tree-optimization/109469
24662 * tree-vect-slp.cc (vect_slp_function): Skip region starts with
24663 a returns-twice call.
24665 2023-04-12 Richard Biener <rguenther@suse.de>
24667 PR tree-optimization/109434
24668 * tree-ssa-dse.cc (initialize_ao_ref_for_dse): Properly
24669 handle possibly throwing calls when processing the LHS
24670 and may-defs are not OK.
24672 2023-04-11 Lin Sinan <mynameisxiaou@gmail.com>
24674 * config/riscv/predicates.md (uimm_extra_bit_or_twobits): Adjust
24675 predicate to avoid splitting arith constants.
24677 2023-04-11 Yanzhang Wang <yanzhang.wang@intel.com>
24678 Pan Li <pan2.li@intel.com>
24679 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
24680 Kito Cheng <kito.cheng@sifive.com>
24683 * config/riscv/riscv-protos.h (emit_hard_vlmax_vsetvl): New.
24684 * config/riscv/riscv-v.cc (emit_hard_vlmax_vsetvl): New.
24685 (emit_vlmax_vsetvl): Use emit_hard_vlmax_vsetvl.
24686 * config/riscv/riscv.cc (vector_zero_call_used_regs): New.
24687 (riscv_zero_call_used_regs): New.
24688 (TARGET_ZERO_CALL_USED_REGS): New.
24690 2023-04-11 Martin Liska <mliska@suse.cz>
24693 * opts.cc (finish_options): Drop also
24694 x_flag_var_tracking_assignments.
24696 2023-04-11 Andre Vieira <andre.simoesdiasvieira@arm.com>
24698 PR tree-optimization/108888
24699 * tree-if-conv.cc (predicate_statements): Fix gimple call check.
24701 2023-04-11 Haochen Gui <guihaoc@gcc.gnu.org>
24704 * config/rs6000/vsx.md (vsx_sign_extend_qi_<mode>): Rename to...
24705 (vsx_sign_extend_v16qi_<mode>): ... this.
24706 (vsx_sign_extend_hi_<mode>): Rename to...
24707 (vsx_sign_extend_v8hi_<mode>): ... this.
24708 (vsx_sign_extend_si_v2di): Rename to...
24709 (vsx_sign_extend_v4si_v2di): ... this.
24710 (vsignextend_qi_<mode>): Remove.
24711 (vsignextend_hi_<mode>): Remove.
24712 (vsignextend_si_v2di): Remove.
24713 (vsignextend_v2di_v1ti): Remove.
24714 (*xxspltib_<mode>_split): Replace gen_vsx_sign_extend_qi_v2di with
24715 gen_vsx_sign_extend_v16qi_v2di and gen_vsx_sign_extend_qi_v4si
24716 with gen_vsx_sign_extend_v16qi_v4si.
24717 * config/rs6000/rs6000.md (split for DI constant generation):
24718 Replace gen_vsx_sign_extend_qi_si with gen_vsx_sign_extend_v16qi_si.
24719 (split for HSDI constant generation): Replace gen_vsx_sign_extend_qi_di
24720 with gen_vsx_sign_extend_v16qi_di and gen_vsx_sign_extend_qi_si
24721 with gen_vsx_sign_extend_v16qi_si.
24722 * config/rs6000/rs6000-builtins.def (__builtin_altivec_vsignextsb2d):
24723 Set bif-pattern to vsx_sign_extend_v16qi_v2di.
24724 (__builtin_altivec_vsignextsb2w): Set bif-pattern to
24725 vsx_sign_extend_v16qi_v4si.
24726 (__builtin_altivec_visgnextsh2d): Set bif-pattern to
24727 vsx_sign_extend_v8hi_v2di.
24728 (__builtin_altivec_vsignextsh2w): Set bif-pattern to
24729 vsx_sign_extend_v8hi_v4si.
24730 (__builtin_altivec_vsignextsw2d): Set bif-pattern to
24731 vsx_sign_extend_si_v2di.
24732 (__builtin_altivec_vsignext): Set bif-pattern to
24733 vsx_sign_extend_v2di_v1ti.
24734 * config/rs6000/rs6000-builtin.cc (lxvrse_expand_builtin): Replace
24735 gen_vsx_sign_extend_qi_v2di with gen_vsx_sign_extend_v16qi_v2di,
24736 gen_vsx_sign_extend_hi_v2di with gen_vsx_sign_extend_v8hi_v2di and
24737 gen_vsx_sign_extend_si_v2di with gen_vsx_sign_extend_v4si_v2di.
24739 2023-04-10 Michael Meissner <meissner@linux.ibm.com>
24742 * config/rs6000/vsx.md (vsx_fmav4sf4): Do not generate vmaddfp.
24743 (vsx_nfmsv4sf4): Do not generate vnmsubfp.
24745 2023-04-10 Haochen Jiang <haochen.jiang@intel.com>
24747 * config/i386/i386.h (PTA_GRANITERAPIDS): Add PTA_AMX_COMPLEX.
24749 2023-04-10 Haochen Jiang <haochen.jiang@intel.com>
24751 * common/config/i386/cpuinfo.h (get_available_features):
24752 Detect AMX-COMPLEX.
24753 * common/config/i386/i386-common.cc
24754 (OPTION_MASK_ISA2_AMX_COMPLEX_SET,
24755 OPTION_MASK_ISA2_AMX_COMPLEX_UNSET): New.
24756 (ix86_handle_option): Handle -mamx-complex.
24757 * common/config/i386/i386-cpuinfo.h (enum processor_features):
24758 Add FEATURE_AMX_COMPLEX.
24759 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
24761 * config.gcc: Add amxcomplexintrin.h.
24762 * config/i386/cpuid.h (bit_AMX_COMPLEX): New.
24763 * config/i386/i386-c.cc (ix86_target_macros_internal): Define
24765 * config/i386/i386-isa.def (AMX_COMPLEX): Add DEF_PTA(AMX_COMPLEX).
24766 * config/i386/i386-options.cc (ix86_valid_target_attribute_inner_p):
24767 Handle amx-complex.
24768 * config/i386/i386.opt: Add option -mamx-complex.
24769 * config/i386/immintrin.h: Include amxcomplexintrin.h.
24770 * doc/extend.texi: Document amx-complex.
24771 * doc/invoke.texi: Document -mamx-complex.
24772 * doc/sourcebuild.texi: Document target amx-complex.
24773 * config/i386/amxcomplexintrin.h: New file.
24775 2023-04-08 Jakub Jelinek <jakub@redhat.com>
24777 PR tree-optimization/109392
24778 * tree-vect-generic.cc (tree_vec_extract): Handle failure
24779 of maybe_push_res_to_seq better.
24781 2023-04-08 Jakub Jelinek <jakub@redhat.com>
24783 * Makefile.in (CORETYPES_H): Depend on align.h, poly-int.h and
24785 (SYSTEM_H): Depend on $(HASHTAB_H).
24786 * config/riscv/t-riscv (build/genrvv-type-indexer.o): Remove unused
24787 dependency on $(RTL_BASE_H), remove redundant dependency on
24790 2023-04-06 Richard Earnshaw <rearnsha@arm.com>
24793 * config/arm/arm.cc (arm_effective_regno): New function.
24794 (mve_vector_mem_operand): Use it.
24796 2023-04-06 Andrew MacLeod <amacleod@redhat.com>
24798 PR tree-optimization/109417
24799 * gimple-range-gori.cc (gori_compute::may_recompute_p): Check if
24800 dependency is in SSA_NAME_FREE_LIST.
24802 2023-04-06 Andrew Pinski <apinski@marvell.com>
24804 PR tree-optimization/109427
24805 * params.opt (-param=vect-induction-float=):
24806 Fix option attribute typo for IntegerRange.
24808 2023-04-05 Jeff Law <jlaw@ventanamicro>
24811 * combine.cc (combine_instructions): Force re-recognition when
24812 after restoring the body of an insn to its original form.
24814 2023-04-05 Martin Jambor <mjambor@suse.cz>
24817 * ipa-sra.cc (zap_useless_ipcp_results): New function.
24818 (process_isra_node_results): Call it.
24820 2023-04-05 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
24822 * config/riscv/vector.md: Fix incorrect operand order.
24824 2023-04-05 Juzhe-Zhong <juzhe.zhong@rivai.ai>
24826 * config/riscv/riscv-vsetvl.cc
24827 (pass_vsetvl::compute_local_backward_infos): Update user vsetvl in local
24830 2023-04-05 Li Xu <xuli1@eswincomputing.com>
24832 * config/riscv/riscv-vector-builtins.def: Fix typo.
24833 * config/riscv/riscv.cc (riscv_dwarf_poly_indeterminate_value): Ditto.
24834 * config/riscv/vector-iterators.md: Ditto.
24836 2023-04-04 Hans-Peter Nilsson <hp@axis.com>
24838 * doc/md.texi (Including Patterns): Fix page break.
24840 2023-04-04 Jakub Jelinek <jakub@redhat.com>
24842 PR tree-optimization/109386
24843 * range-op-float.cc (foperator_lt::op1_range, foperator_lt::op2_range,
24844 foperator_le::op1_range, foperator_le::op2_range,
24845 foperator_gt::op1_range, foperator_gt::op2_range,
24846 foperator_ge::op1_range, foperator_ge::op2_range): Make r varying for
24847 BRS_FALSE case even if the other op is maybe_isnan, not just
24849 (foperator_unordered_lt::op1_range, foperator_unordered_lt::op2_range,
24850 foperator_unordered_le::op1_range, foperator_unordered_le::op2_range,
24851 foperator_unordered_gt::op1_range, foperator_unordered_gt::op2_range,
24852 foperator_unordered_ge::op1_range, foperator_unordered_ge::op2_range):
24853 Make r varying for BRS_TRUE case even if the other op is maybe_isnan,
24854 not just known_isnan.
24856 2023-04-04 Marek Polacek <polacek@redhat.com>
24858 PR sanitizer/109107
24859 * fold-const.cc (fold_binary_loc): Use TYPE_OVERFLOW_SANITIZED
24861 * match.pd: Use TYPE_OVERFLOW_SANITIZED.
24863 2023-04-04 Stam Markianos-Wright <stam.markianos-wright@arm.com>
24865 * config/arm/mve.md (mve_vcvtq_n_to_f_<supf><mode>): Swap operands.
24866 (mve_vcreateq_f<mode>): Swap operands.
24868 2023-04-04 Andrew Stubbs <ams@codesourcery.com>
24870 * config/gcn/gcn-valu.md (one_cmpl<mode>2<exec>): New.
24872 2023-04-04 Jakub Jelinek <jakub@redhat.com>
24875 * common/config/riscv/riscv-common.cc (riscv_subset_list::parse):
24876 Reword diagnostics about zfinx conflict with f, formatting fixes.
24878 2023-04-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24880 * config/sol2.h (LIB_SPEC): Don't link with -lpthread.
24882 2023-04-04 Richard Biener <rguenther@suse.de>
24884 PR tree-optimization/109304
24885 * tree-profile.cc (tree_profiling): Use symtab node
24886 availability to decide whether to skip adjusting calls.
24887 Do not adjust calls to internal functions.
24889 2023-04-04 Kewen Lin <linkw@linux.ibm.com>
24892 * config/rs6000/rs6000.cc (rs6000_expand_vector_set_var_p9): Fix gen
24893 function for permutation control vector by considering big endianness.
24895 2023-04-04 Kewen Lin <linkw@linux.ibm.com>
24898 * config/rs6000/altivec.md (*p9v_parity<mode>2): Rename to ...
24899 (rs6000_vprtyb<mode>2): ... this.
24900 * config/rs6000/rs6000-builtins.def (VPRTYBD): Replace parityv2di2 with
24901 rs6000_vprtybv2di2.
24902 (VPRTYBW): Replace parityv4si2 with rs6000_vprtybv4si2.
24903 (VPRTYBQ): Replace parityv1ti2 with rs6000_vprtybv1ti2.
24904 * config/rs6000/vector.md (parity<mode>2 with VEC_IP): Expand with
24905 popcountv16qi2 and the corresponding rs6000_vprtyb<mode>2.
24907 2023-04-04 Hans-Peter Nilsson <hp@axis.com>
24908 Sandra Loosemore <sandra@codesourcery.com>
24910 * doc/md.texi (Insn Splitting): Tweak wording for readability.
24912 2023-04-03 Martin Jambor <mjambor@suse.cz>
24915 * ipa-prop.cc (determine_known_aggregate_parts): Check that the
24916 offset + size will be representable in unsigned int.
24918 2023-04-03 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24920 * configure.ac (ZSTD_LIB): Move before zstd.h check.
24921 Unset gcc_cv_header_zstd_h without libzstd.
24922 * configure: Regenerate.
24924 2023-04-03 Martin Liska <mliska@suse.cz>
24926 * doc/invoke.texi: Document new param.
24928 2023-04-03 Cupertino Miranda <cupertino.miranda@oracle.com>
24930 * doc/sourcebuild.texi (const_volatile_readonly_section): Document
24931 new check_effective_target function.
24933 2023-04-03 Li Xu <xuli1@eswincomputing.com>
24935 * config/riscv/riscv-vector-builtins.def (vuint32m8_t): Fix typo.
24936 (vfloat32m8_t): Likewise
24938 2023-04-03 liuhongt <hongtao.liu@intel.com>
24940 * doc/md.texi: Document signbitm2.
24942 2023-04-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
24943 kito-cheng <kito.cheng@sifive.com>
24945 * config/riscv/vector.md: Fix RA constraint.
24947 2023-04-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
24949 * config/riscv/riscv-protos.h (gen_avl_for_scalar_move): New function.
24950 * config/riscv/riscv-v.cc (gen_avl_for_scalar_move): New function.
24951 * config/riscv/vector.md: Fix scalar move bug.
24953 2023-04-01 Jakub Jelinek <jakub@redhat.com>
24955 * range-op-float.cc (foperator_equal::fold_range): If at least
24956 one of the op ranges is not singleton and neither is NaN and all
24957 4 bounds are zero, return [1, 1].
24958 (foperator_not_equal::fold_range): In the same case return [0, 0].
24960 2023-04-01 Jakub Jelinek <jakub@redhat.com>
24962 * range-op-float.cc (foperator_equal::fold_range): Perform the
24963 non-singleton handling regardless of maybe_isnan (op1, op2).
24964 (foperator_not_equal::fold_range): Likewise.
24965 (foperator_lt::fold_range, foperator_le::fold_range,
24966 foperator_gt::fold_range, foperator_ge::fold_range): Perform the
24967 real_* comparison check which results in range_false (type)
24968 even if maybe_isnan (op1, op2). Simplify.
24969 (foperator_ltgt): New class.
24970 (fop_ltgt): New variable.
24971 (floating_op_table::floating_op_table): Handle LTGT_EXPR using
24974 2023-04-01 Jakub Jelinek <jakub@redhat.com>
24977 * builtins.cc (apply_args_size): If targetm.calls.get_raw_arg_mode
24978 returns VOIDmode, handle it like if the register isn't used for
24979 passing arguments at all.
24980 (apply_result_size): If targetm.calls.get_raw_result_mode returns
24981 VOIDmode, handle it like if the register isn't used for returning
24983 * target.def (get_raw_result_mode, get_raw_arg_mode): Document what it
24984 means to return VOIDmode.
24985 * doc/tm.texi: Regenerated.
24986 * config/aarch64/aarch64.cc (aarch64_function_value_regno_p): Return
24987 TARGET_SVE for P0_REGNUM.
24988 (aarch64_function_arg_regno_p): Also return true for p0-p3.
24989 (aarch64_get_reg_raw_mode): Return VOIDmode for PR_REGNUM_P regs.
24991 2023-03-31 Vladimir N. Makarov <vmakarov@redhat.com>
24993 * lra-constraints.cc: (combine_reload_insn): New function.
24995 2023-03-31 Jakub Jelinek <jakub@redhat.com>
24997 PR tree-optimization/91645
24998 * range-op-float.cc (foperator_unordered_lt::fold_range,
24999 foperator_unordered_le::fold_range,
25000 foperator_unordered_gt::fold_range,
25001 foperator_unordered_ge::fold_range,
25002 foperator_unordered_equal::fold_range): Call the ordered
25003 fold_range on ranges with cleared NaNs.
25004 * value-query.cc (range_query::get_tree_range): Handle also
25005 COMPARISON_CLASS_P trees.
25007 2023-03-31 Kito Cheng <kito.cheng@sifive.com>
25008 Andrew Pinski <pinskia@gmail.com>
25011 * config/riscv/t-riscv: Add missing dependencies.
25013 2023-03-31 liuhongt <hongtao.liu@intel.com>
25015 * config/i386/i386.cc (inline_memory_move_cost): Return 100
25016 for MASK_REGS when MODE_SIZE > 8.
25018 2023-03-31 liuhongt <hongtao.liu@intel.com>
25021 * config/i386/i386-builtin.def (BDESC): Adjust icode name from
25022 ufloat/ufix to floatuns/fixuns.
25023 * config/i386/i386-expand.cc
25024 (ix86_expand_vector_convert_uns_vsivsf): Adjust comments.
25025 * config/i386/sse.md
25026 (ufloat<sseintvecmodelower><mode>2<mask_name><round_name>):
25028 (<mask_codefor>floatuns<sseintvecmodelower><mode>2<mask_name><round_name>):.. this.
25029 (<mask_codefor><avx512>_ufix_notrunc<sf2simodelower><mode><mask_name><round_name>):
25031 (<mask_codefor><avx512>_fixuns_notrunc<sf2simodelower><mode><mask_name><round_name>):
25033 (<fixsuffix>fix_truncv16sfv16si2<mask_name><round_saeonly_name>):
25035 (fix<fixunssuffix>_truncv16sfv16si2<mask_name><round_saeonly_name>):.. this.
25036 (ufloat<si2dfmodelower><mode>2<mask_name>): Renamed to ..
25037 (floatuns<si2dfmodelower><mode>2<mask_name>): .. this.
25038 (ufloatv2siv2df2<mask_name>): Renamed to ..
25039 (<mask_codefor>floatunsv2siv2df2<mask_name>): .. this.
25040 (ufix_notrunc<mode><si2dfmodelower>2<mask_name><round_name>):
25042 (fixuns_notrunc<mode><si2dfmodelower>2<mask_name><round_name>):
25044 (ufix_notruncv2dfv2si2): Renamed to ..
25045 (fixuns_notruncv2dfv2si2):.. this.
25046 (ufix_notruncv2dfv2si2_mask): Renamed to ..
25047 (fixuns_notruncv2dfv2si2_mask): .. this.
25048 (*ufix_notruncv2dfv2si2_mask_1): Renamed to ..
25049 (*fixuns_notruncv2dfv2si2_mask_1): .. this.
25050 (ufix_truncv2dfv2si2): Renamed to ..
25051 (*fixuns_truncv2dfv2si2): .. this.
25052 (ufix_truncv2dfv2si2_mask): Renamed to ..
25053 (fixuns_truncv2dfv2si2_mask): .. this.
25054 (*ufix_truncv2dfv2si2_mask_1): Renamed to ..
25055 (*fixuns_truncv2dfv2si2_mask_1): .. this.
25056 (ufix_truncv4dfv4si2<mask_name>): Renamed to ..
25057 (fixuns_truncv4dfv4si2<mask_name>): .. this.
25058 (ufix_notrunc<mode><sseintvecmodelower>2<mask_name><round_name>):
25060 (fixuns_notrunc<mode><sseintvecmodelower>2<mask_name><round_name>):
25062 (ufix_trunc<mode><sseintvecmodelower>2<mask_name>): Renamed to ..
25063 (<mask_codefor>fixuns_trunc<mode><sseintvecmodelower>2<mask_name>):
25066 2023-03-30 Andrew MacLeod <amacleod@redhat.com>
25068 PR tree-optimization/109154
25069 * gimple-range-gori.cc (gori_compute::may_recompute_p): Add depth limit.
25070 * gimple-range-gori.h (may_recompute_p): Add depth param.
25071 * params.opt (ranger-recompute-depth): New param.
25073 2023-03-30 Jason Merrill <jason@redhat.com>
25077 * cgraph.h: Move reset() from cgraph_node to symtab_node.
25078 * cgraphunit.cc (symtab_node::reset): Adjust. Also call
25079 remove_from_same_comdat_group.
25081 2023-03-30 Richard Biener <rguenther@suse.de>
25083 PR tree-optimization/107561
25084 * gimple-ssa-warn-access.cc (get_size_range): Add flags
25085 argument and pass it on.
25086 (check_access): When querying for the size range pass
25087 SR_ALLOW_ZERO when the known destination size is zero.
25089 2023-03-30 Richard Biener <rguenther@suse.de>
25091 PR tree-optimization/109342
25092 * tree-ssa-sccvn.cc (vn_nary_op_get_predicated_value): New
25093 overload for edge. When that edge is a backedge use
25094 dominated_by_p directly.
25096 2023-03-30 liuhongt <hongtao.liu@intel.com>
25098 * config/i386/i386-expand.cc (expand_vec_perm_blend): Generate
25099 vpblendd instead of vpblendw for V4SI under avx2.
25101 2023-03-29 Hans-Peter Nilsson <hp@axis.com>
25103 * config/cris/cris.cc (cris_rtx_costs) [CONST_INT]: Return 0
25104 for many quick operands, for register-sized modes.
25106 2023-03-29 Jiawei <jiawei@iscas.ac.cn>
25108 * common/config/riscv/riscv-common.cc (riscv_subset_list::parse):
25111 2023-03-29 Martin Liska <mliska@suse.cz>
25113 PR bootstrap/109310
25114 * configure.ac: Emit a warning for deprecated option
25115 --enable-link-mutex.
25116 * configure: Regenerate.
25118 2023-03-29 Richard Biener <rguenther@suse.de>
25120 PR tree-optimization/109331
25121 * tree-ssa-forwprop.cc (pass_forwprop::execute): When we
25122 discover a taken edge make sure to cleanup the CFG.
25124 2023-03-29 Richard Biener <rguenther@suse.de>
25126 PR tree-optimization/109327
25127 * tree-ssa-forwprop.cc (pass_forwprop::execute): Deal with
25128 already removed stmts when draining to_remove.
25130 2023-03-29 Richard Biener <rguenther@suse.de>
25133 * dwarf2out.cc (lookup_type_die): Reset TREE_ASM_WRITTEN
25134 so we can re-create the DIE for the type if required.
25136 2023-03-29 Jakub Jelinek <jakub@redhat.com>
25137 Richard Biener <rguenther@suse.de>
25139 PR tree-optimization/109301
25140 * tree-ssa-math-opts.cc (pass_data_cse_sincos): Change
25141 properties_provided from PROP_gimple_opt_math to 0.
25142 (pass_data_expand_powcabs): Change properties_provided from 0 to
25143 PROP_gimple_opt_math.
25145 2023-03-29 Richard Biener <rguenther@suse.de>
25147 PR tree-optimization/109154
25148 * tree-if-conv.cc (gen_phi_arg_condition): Handle single
25149 inverted condition specially by inverting at the caller.
25150 (gen_phi_arg_condition): Swap COND_EXPR arms if requested.
25152 2023-03-28 David Malcolm <dmalcolm@redhat.com>
25155 * diagnostic-show-locus.cc (column_range::column_range): Factor
25156 out assertion conditional into...
25157 (column_range::valid_p): ...this new function.
25158 (line_corrections::add_hint): Don't attempt to consolidate hints
25159 if it would lead to invalid column_range instances.
25161 2023-03-28 Kito Cheng <kito.cheng@sifive.com>
25164 * config/riscv/riscv-c.cc (riscv_ext_version_value): New.
25165 (riscv_cpu_cpp_builtins): Define __riscv_v_intrinsic and
25168 2023-03-28 Alexander Monakov <amonakov@ispras.ru>
25170 PR rtl-optimization/109187
25171 * haifa-sched.cc (autopref_rank_for_schedule): Avoid use of overflowing
25172 subtraction in three-way comparison.
25174 2023-03-28 Andrew MacLeod <amacleod@redhat.com>
25176 PR tree-optimization/109265
25177 PR tree-optimization/109274
25178 * gimple-range-gori.cc (gori_compute::compute_operand_range): Do
25179 not create a relation record is op1 and op2 are the same symbol.
25180 (gori_compute::compute_operand1_range): Pass op1 == op2 to the
25181 handler for this stmt, but create a new record only if this statement
25182 generates a relation based on the ranges.
25183 (gori_compute::compute_operand2_range): Ditto.
25184 * value-relation.h (value_relation::set_relation): Always create the
25185 record that is requested.
25187 2023-03-28 Richard Biener <rguenther@suse.de>
25189 PR tree-optimization/107087
25190 * tree-ssa-forwprop.cc (pass_forwprop::execute): Track
25191 executable regions to avoid useless work and to better
25192 propagate degenerate PHIs.
25194 2023-03-28 Costas Argyris <costas.argyris@gmail.com>
25196 * config/i386/x-mingw32-utf8: update comments.
25198 2023-03-28 Richard Sandiford <richard.sandiford@arm.com>
25201 * config/aarch64/aarch64-protos.h (aarch64_vector_load_decl): Declare.
25202 * config/aarch64/aarch64.h (machine_function::vector_load_decls): New
25204 * config/aarch64/aarch64-builtins.cc (aarch64_record_vector_load_arg):
25206 (aarch64_general_gimple_fold_builtin): Delay folding of vld1 until
25207 after inlining. Record which decls are loaded from. Fix handling
25208 of vops for loads and stores.
25209 * config/aarch64/aarch64.cc (aarch64_vector_load_decl): New function.
25210 (aarch64_accesses_vector_load_decl_p): Likewise.
25211 (aarch64_vector_costs::m_stores_to_vector_load_decl): New member
25213 (aarch64_vector_costs::add_stmt_cost): If the function has a vld1
25214 that loads from a decl, treat vector stores to those decls as
25216 (aarch64_vector_costs::finish_cost): ...and in that case,
25217 if the vector code does nothing more than a store, give the
25218 prologue a zero cost as well.
25220 2023-03-28 Richard Biener <rguenther@suse.de>
25223 PR tree-optimization/108129
25224 * genmatch.cc (lower_for): For (match ...) delay
25225 substituting into the match operator if possible.
25226 (dt_operand::gen_gimple_expr): For user_id look at the
25227 first substitute for determining how to access operands.
25228 (dt_operand::gen_generic_expr): Likewise.
25229 (dt_node::gen_kids): Properly sort user_ids according
25230 to their substitutes.
25231 (dt_node::gen_kids_1): Code-generate user_id matching.
25233 2023-03-28 Jakub Jelinek <jakub@redhat.com>
25234 Jonathan Wakely <jwakely@redhat.com>
25236 * gcov-tool.cc (do_merge, do_merge_stream, do_rewrite, do_overlap):
25237 Use subcommand rather than sub-command in function comments.
25239 2023-03-28 Jakub Jelinek <jakub@redhat.com>
25241 PR tree-optimization/109154
25242 * value-range.h (frange::flush_denormals_to_zero): Make it public
25243 rather than private.
25244 * value-range.cc (frange::set): Don't call flush_denormals_to_zero
25246 * range-op-float.cc (range_operator_float::fold_range): Call
25247 flush_denormals_to_zero.
25249 2023-03-28 Jakub Jelinek <jakub@redhat.com>
25251 PR middle-end/106190
25252 * sanopt.cc (pass_sanopt::execute): Return TODO_cleanup_cfg if any
25253 of the IFN_{UB,HWA,A}SAN_* internal fns are lowered.
25255 2023-03-28 Jakub Jelinek <jakub@redhat.com>
25257 * range-op-float.cc (float_widen_lhs_range): Use pass get_nan_state
25258 as 4th argument to set to avoid clear_nan and union_ calls.
25260 2023-03-28 Jakub Jelinek <jakub@redhat.com>
25263 * config/i386/i386.cc (assign_386_stack_local): For DImode
25264 with SLOT_FLOATxFDI_387 and -m32 -mpreferred-stack-boundary=2 pass
25265 align 32 rather than 0 to assign_stack_local.
25267 2023-03-28 Eric Botcazou <ebotcazou@adacore.com>
25270 * config/sparc/sparc.cc (sparc_expand_vcond): Call signed_condition
25271 on operand #3 to get the final condition code. Use std::swap.
25272 * config/sparc/sparc.md (vcondv8qiv8qi): New VIS 4 expander.
25273 (fucmp<gcond:code>8<P:mode>_vis): Move around.
25274 (fpcmpu<gcond:code><GCM:gcm_name><P:mode>_vis): Likewise.
25275 (vcondu<GCM:mode><GCM:mode>): New VIS 4 expander.
25277 2023-03-28 Eric Botcazou <ebotcazou@adacore.com>
25279 * doc/gm2.texi: Add missing Next, Previous and Top fields to most
25280 top-level sections.
25282 2023-03-28 Costas Argyris <costas.argyris@gmail.com>
25284 * config.host: Pull in i386/x-mingw32-utf8 Makefile
25285 fragment and reference utf8rc-mingw32.o explicitly
25287 * config/i386/sym-mingw32.cc: prevent name mangling of
25289 * config/i386/x-mingw32-utf8: Make utf8rc-mingw32.o
25290 depend on manifest file explicitly.
25292 2023-03-28 Richard Biener <rguenther@suse.de>
25295 2023-03-27 Richard Biener <rguenther@suse.de>
25297 PR rtl-optimization/109237
25298 * cfgcleanup.cc (bb_is_just_return): Walk insns backwards.
25300 2023-03-28 Richard Biener <rguenther@suse.de>
25302 * common.opt (gdwarf): Remove Negative(gdwarf-).
25304 2023-03-28 Richard Biener <rguenther@suse.de>
25306 * common.opt (gdwarf): Add RejectNegative.
25307 (gdwarf-): Likewise.
25311 2023-03-28 Hans-Peter Nilsson <hp@axis.com>
25313 * config/cris/constraints.md ("T"): Correct to
25314 define_memory_constraint.
25316 2023-03-28 Hans-Peter Nilsson <hp@axis.com>
25318 * config/cris/cris.md (BW2): New mode-iterator.
25319 (lra_szext_decomposed, lra_szext_decomposed_indirect_with_offset): New
25322 2023-03-28 Hans-Peter Nilsson <hp@axis.com>
25324 * config/cris/cris.md ("*add<mode>3_addi"): Improve to bail only
25325 for possible eliminable compares.
25327 2023-03-28 Hans-Peter Nilsson <hp@axis.com>
25329 * config/cris/constraints.md ("R"): Remove unused constraint.
25331 2023-03-27 Jonathan Wakely <jwakely@redhat.com>
25333 PR gcov-profile/109297
25334 * gcov-tool.cc (merge_usage): Fix "subcomand" typo.
25335 (merge_stream_usage): Likewise.
25336 (overlap_usage): Likewise.
25338 2023-03-27 Christoph Müllner <christoph.muellner@vrull.eu>
25341 * config/riscv/thead.md: Add missing mode specifiers.
25343 2023-03-27 Philipp Tomsich <philipp.tomsich@vrull.eu>
25344 Jiangning Liu <jiangning.liu@amperecomputing.com>
25345 Manolis Tsamis <manolis.tsamis@vrull.eu>
25347 * config/aarch64/aarch64.cc: Update vector costs for ampere1.
25349 2023-03-27 Richard Biener <rguenther@suse.de>
25351 PR rtl-optimization/109237
25352 * cfgcleanup.cc (bb_is_just_return): Walk insns backwards.
25354 2023-03-27 Richard Biener <rguenther@suse.de>
25357 * lto-wrapper.cc (run_gcc): Parse alternate debug options
25358 as well, they always enable debug.
25360 2023-03-27 Kewen Lin <linkw@linux.ibm.com>
25363 * config/rs6000/emmintrin.h (_mm_bslli_si128): Move the implementation
25365 (_mm_slli_si128): ... here. Change to call _mm_bslli_si128 directly.
25367 2023-03-27 Kewen Lin <linkw@linux.ibm.com>
25370 * config/rs6000/emmintrin.h (_mm_bslli_si128): Check __N is not less
25371 than zero when calling vec_sld.
25372 (_mm_bsrli_si128): Return __A if __N is zero, check __N is bigger than
25373 zero when calling vec_sld.
25374 (_mm_slli_si128): Return __A if _imm5 is zero, check _imm5 is bigger
25375 than zero when calling vec_sld.
25377 2023-03-27 Sandra Loosemore <sandra@codesourcery.com>
25379 * doc/generic.texi (OpenMP): Document OMP_SIMD, OMP_DISTRIBUTE,
25380 OMP_TASKLOOP, and OMP_LOOP with OMP_FOR. Document how collapsed
25381 loops are represented and which fields are vectors. Add
25382 documentation for OMP_FOR_PRE_BODY field. Document internal
25383 form of non-rectangular loops and OMP_FOR_NON_RECTANGULAR.
25384 * tree.def (OMP_FOR): Make documentation consistent with the
25385 Texinfo manual, to fill some gaps and correct errors.
25387 2023-03-26 Andreas Schwab <schwab@linux-m68k.org>
25390 * config/m68k/m68k.h (FINAL_PRESCAN_INSN): Define.
25391 * config/m68k/m68k.cc (m68k_final_prescan_insn): Define.
25392 (handle_move_double): Call it before handle_movsi.
25393 * config/m68k/m68k-protos.h: Declare it.
25395 2023-03-26 Jakub Jelinek <jakub@redhat.com>
25397 PR tree-optimization/109230
25398 * match.pd (fneg/fadd simplify): Verify also odd permutation indexes.
25400 2023-03-26 Jakub Jelinek <jakub@redhat.com>
25403 * predict.cc (compute_function_frequency): Don't call
25404 warn_function_cold if function already has cold attribute.
25406 2023-03-26 Gerald Pfeifer <gerald@pfeifer.com>
25408 * doc/install.texi: Remove anachronistic note
25409 related to languages built and separate source tarballs.
25411 2023-03-25 David Malcolm <dmalcolm@redhat.com>
25414 * diagnostic-format-sarif.cc (read_until_eof): Delete.
25415 (maybe_read_file): Delete.
25416 (sarif_builder::maybe_make_artifact_content_object): Use
25417 get_source_file_content rather than maybe_read_file.
25418 Reject it if it's not valid UTF-8.
25419 * input.cc (file_cache_slot::get_full_file_content): New.
25420 (get_source_file_content): New.
25421 (selftest::check_cpp_valid_utf8_p): New.
25422 (selftest::test_cpp_valid_utf8_p): New.
25423 (selftest::input_cc_tests): Call selftest::test_cpp_valid_utf8_p.
25424 * input.h (get_source_file_content): New prototype.
25426 2023-03-24 David Malcolm <dmalcolm@redhat.com>
25428 * doc/analyzer.texi (Debugging the Analyzer): Add notes on useful
25430 (Special Functions for Debugging the Analyzer): Convert to a
25431 table, and rewrite in places.
25432 (Other Debugging Techniques): Add notes on how to compare two
25433 different exploded graphs.
25435 2023-03-24 David Malcolm <dmalcolm@redhat.com>
25438 * json.cc: Update comments to indicate that we now preserve
25439 insertion order of keys within objects.
25440 (object::print): Traverse keys in insertion order.
25441 (object::set): Preserve insertion order of keys.
25442 (selftest::test_writing_objects): Add an additional key to verify
25443 that we preserve insertion order.
25444 * json.h (object::m_keys): New field.
25446 2023-03-24 Andrew MacLeod <amacleod@redhat.com>
25448 PR tree-optimization/109238
25449 * gimple-range-cache.cc (ranger_cache::resolve_dom): Ignore
25450 predecessors which this block dominates.
25452 2023-03-24 Richard Biener <rguenther@suse.de>
25454 PR tree-optimization/106912
25455 * tree-profile.cc (tree_profiling): Update stmts only when
25456 profiling or testing coverage. Make sure to update calls
25457 fntype, stripping 'const' there.
25459 2023-03-24 Jakub Jelinek <jakub@redhat.com>
25461 PR middle-end/109258
25462 * builtins.cc (inline_expand_builtin_bytecmp): Return NULL_RTX early
25463 if target == const0_rtx.
25465 2023-03-24 Alexandre Oliva <oliva@adacore.com>
25467 * doc/sourcebuild.texi (weak_undefined, posix_memalign):
25468 Document options and effective targets.
25470 2023-03-24 Costas Argyris <costas.argyris@gmail.com>
25472 * config/i386/x-mingw32-utf8: Make HOST_EXTRA_OBJS_SYMBOL
25475 2023-03-23 Pat Haugen <pthaugen@linux.ibm.com>
25477 * config/rs6000/rs6000.md (*mod<mode>3, umod<mode>3): Add
25478 non-earlyclobber alternative.
25480 2023-03-23 Andrew Pinski <apinski@marvell.com>
25483 * fold-const.cc (maybe_lvalue_p): Treat COMPOUND_LITERAL_EXPR
25486 2023-03-23 Richard Biener <rguenther@suse.de>
25488 PR tree-optimization/107569
25489 * tree-ssa-sccvn.cc (eliminate_dom_walker::eliminate_stmt):
25490 Do not push SSA names with zero uses as available leader.
25491 (process_bb): Likewise.
25493 2023-03-23 Richard Biener <rguenther@suse.de>
25495 PR tree-optimization/109262
25496 * tree-ssa-forwprop.cc (pass_forwprop::execute): When
25497 combining a piecewise complex load avoid touching loads
25498 that throw internally. Use fun, not cfun throughout.
25500 2023-03-23 Jakub Jelinek <jakub@redhat.com>
25502 * value-range.cc (irange::irange_union, irange::intersect): Fix
25503 comment spelling bugs.
25504 * gimple-range-trace.cc (range_tracer::do_header): Likewise.
25505 * gimple-range-trace.h: Likewise.
25506 * gimple-range-edge.cc: Likewise.
25507 (gimple_outgoing_range_stmt_p,
25508 gimple_outgoing_range::switch_edge_range,
25509 gimple_outgoing_range::edge_range_p): Likewise.
25510 * gimple-range.cc (gimple_ranger::prefill_stmt_dependencies,
25511 gimple_ranger::fold_stmt, gimple_ranger::register_transitive_infer,
25512 assume_query::assume_query, assume_query::calculate_phi): Likewise.
25513 * gimple-range-edge.h: Likewise.
25514 * value-range.h (Value_Range::set, Value_Range::lower_bound,
25515 Value_Range::upper_bound, frange::set_undefined): Likewise.
25516 * gimple-range-gori.h (range_def_chain::depend, gori_map::m_outgoing,
25517 gori_compute): Likewise.
25518 * gimple-range-fold.h (fold_using_range): Likewise.
25519 * gimple-range-path.cc (path_range_query::compute_ranges_in_phis):
25521 * gimple-range-gori.cc (range_def_chain::in_chain_p,
25522 range_def_chain::dump, gori_map::calculate_gori,
25523 gori_compute::compute_operand_range_switch,
25524 gori_compute::logical_combine, gori_compute::refine_using_relation,
25525 gori_compute::compute_operand1_range, gori_compute::may_recompute_p):
25527 * gimple-range.h: Likewise.
25528 (enable_ranger): Likewise.
25529 * range-op.h (empty_range_varying): Likewise.
25530 * value-query.h (value_query): Likewise.
25531 * gimple-range-cache.cc (block_range_cache::set_bb_range,
25532 block_range_cache::dump, ssa_global_cache::clear_global_range,
25533 temporal_cache::temporal_value, temporal_cache::current_p,
25534 ranger_cache::range_of_def, ranger_cache::propagate_updated_value,
25535 ranger_cache::range_from_dom, ranger_cache::register_inferred_value):
25537 * gimple-range-fold.cc (fur_edge::get_phi_operand,
25538 fur_stmt::get_operand, gimple_range_adjustment,
25539 fold_using_range::range_of_phi,
25540 fold_using_range::relation_fold_and_or): Likewise.
25541 * value-range-storage.h (irange_storage_slot::MAX_INTS): Likewise.
25542 * value-query.cc (range_query::value_of_expr,
25543 range_query::value_on_edge, range_query::query_relation): Likewise.
25544 * tree-vrp.cc (remove_unreachable::remove_and_update_globals,
25545 intersect_range_with_nonzero_bits): Likewise.
25546 * gimple-range-infer.cc (gimple_infer_range::check_assume_func,
25547 exit_range): Likewise.
25548 * value-relation.h: Likewise.
25549 (equiv_oracle, relation_trio::relation_trio, value_relation,
25550 value_relation::value_relation, pe_min): Likewise.
25551 * range-op-float.cc (range_operator_float::rv_fold,
25552 frange_arithmetic, foperator_unordered_equal::op1_range,
25553 foperator_div::rv_fold): Likewise.
25554 * gimple-range-op.cc (cfn_clz::fold_range): Likewise.
25555 * value-relation.cc (equiv_oracle::query_relation,
25556 equiv_oracle::register_equiv, equiv_oracle::add_equiv_to_block,
25557 value_relation::apply_transitive, relation_chain_head::find_relation,
25558 dom_oracle::query_relation, dom_oracle::find_relation_block,
25559 dom_oracle::find_relation_dom, path_oracle::register_equiv): Likewise.
25560 * range-op.cc (range_operator::wi_fold_in_parts_equiv,
25561 create_possibly_reversed_range, adjust_op1_for_overflow,
25562 operator_mult::wi_fold, operator_exact_divide::op1_range,
25563 operator_cast::lhs_op1_relation, operator_cast::fold_pair,
25564 operator_cast::fold_range, operator_abs::wi_fold, range_op_cast_tests,
25565 range_op_lshift_tests): Likewise.
25567 2023-03-23 Andrew Stubbs <ams@codesourcery.com>
25569 * config/gcn/gcn.cc (gcn_class_max_nregs): Handle vectors in SGPRs.
25570 (move_callee_saved_registers): Detect the bug condition early.
25572 2023-03-23 Andrew Stubbs <ams@codesourcery.com>
25574 * config/gcn/gcn-protos.h (gcn_stepped_zero_int_parallel_p): New.
25575 * config/gcn/gcn-valu.md (V_1REG_ALT): New.
25577 (vec_extract<V_1REG:mode><V_1REG_ALT:mode>_nop): New.
25578 (vec_extract<V_2REG:mode><V_2REG_ALT:mode>_nop): New.
25579 (vec_extract<V_ALL:mode><V_ALL_ALT:mode>): Use new patterns.
25580 * config/gcn/gcn.cc (gcn_stepped_zero_int_parallel_p): New.
25581 * config/gcn/predicates.md (ascending_zero_int_parallel): New.
25583 2023-03-23 Jakub Jelinek <jakub@redhat.com>
25585 PR tree-optimization/109176
25586 * tree-vect-generic.cc (expand_vector_condition): If a has
25587 vector boolean type and is a comparison, also check if both
25588 the comparison and VEC_COND_EXPR could be successfully expanded
25591 2023-03-23 Pan Li <pan2.li@intel.com>
25592 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
25596 * config/riscv/riscv-modes.def (ADJUST_BYTESIZE): Adjust size
25597 for vector mask modes.
25598 * config/riscv/riscv.cc (riscv_v_adjust_bytesize): New.
25599 * config/riscv/riscv.h (riscv_v_adjust_bytesize): New.
25601 2023-03-23 Songhe Zhu <zhusonghe@eswincomputing.com>
25603 * config/riscv/multilib-generator: Adjusting the loop of 'alt' in 'alts'.
25605 2023-03-23 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
25608 * config/riscv/riscv-protos.h (emit_vlmax_vsetvl): Define as global.
25609 (emit_vlmax_op): Ditto.
25610 * config/riscv/riscv-v.cc (get_sew): New function.
25611 (emit_vlmax_vsetvl): Adapt function.
25612 (emit_pred_op): Ditto.
25613 (emit_vlmax_op): Ditto.
25614 (emit_nonvlmax_op): Ditto.
25615 (legitimize_move): Fix LRA ICE.
25616 (gen_no_side_effects_vsetvl_rtx): Adapt function.
25617 * config/riscv/vector.md (@mov<V_FRACT:mode><P:mode>_lra): New pattern.
25618 (@mov<VB:mode><P:mode>_lra): Ditto.
25619 (*mov<V_FRACT:mode><P:mode>_lra): Ditto.
25620 (*mov<VB:mode><P:mode>_lra): Ditto.
25622 2023-03-23 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
25625 * config/riscv/riscv-vector-builtins-bases.cc (class vlenb): Add
25626 __riscv_vlenb support.
25628 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
25629 * config/riscv/riscv-vector-builtins-functions.def (vlenb): Ditto.
25630 * config/riscv/riscv-vector-builtins-shapes.cc (struct vlenb_def): Ditto.
25632 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
25633 * config/riscv/riscv-vector-builtins.cc: Ditto.
25635 2023-03-23 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
25636 kito-cheng <kito.cheng@sifive.com>
25638 * config/riscv/riscv-vsetvl.cc (reg_available_p): Fix bugs.
25639 (pass_vsetvl::compute_local_backward_infos): Fix bugs.
25640 (pass_vsetvl::need_vsetvl): Fix bugs.
25641 (pass_vsetvl::backward_demand_fusion): Fix bugs.
25642 (pass_vsetvl::demand_fusion): Fix bugs.
25643 (eliminate_insn): Fix bugs.
25644 (insert_vsetvl): Ditto.
25645 (pass_vsetvl::emit_local_forward_vsetvls): Ditto.
25646 * config/riscv/riscv-vsetvl.h (enum vsetvl_type): Ditto.
25647 * config/riscv/vector.md: Ditto.
25649 2023-03-23 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
25650 kito-cheng <kito.cheng@sifive.com>
25652 * config/riscv/riscv-vector-builtins-bases.cc: Fix ternary bug.
25653 * config/riscv/vector-iterators.md (nmsac): Ditto.
25659 * config/riscv/vector.md (@pred_mul_<optab><mode>): Ditto.
25660 (@pred_mul_plus<mode>): Ditto.
25661 (*pred_madd<mode>): Ditto.
25662 (*pred_macc<mode>): Ditto.
25663 (*pred_mul_plus<mode>): Ditto.
25664 (@pred_mul_plus<mode>_scalar): Ditto.
25665 (*pred_madd<mode>_scalar): Ditto.
25666 (*pred_macc<mode>_scalar): Ditto.
25667 (*pred_mul_plus<mode>_scalar): Ditto.
25668 (*pred_madd<mode>_extended_scalar): Ditto.
25669 (*pred_macc<mode>_extended_scalar): Ditto.
25670 (*pred_mul_plus<mode>_extended_scalar): Ditto.
25671 (@pred_minus_mul<mode>): Ditto.
25672 (*pred_<madd_nmsub><mode>): Ditto.
25673 (*pred_nmsub<mode>): Ditto.
25674 (*pred_<macc_nmsac><mode>): Ditto.
25675 (*pred_nmsac<mode>): Ditto.
25676 (*pred_mul_<optab><mode>): Ditto.
25677 (*pred_minus_mul<mode>): Ditto.
25678 (@pred_mul_<optab><mode>_scalar): Ditto.
25679 (@pred_minus_mul<mode>_scalar): Ditto.
25680 (*pred_<madd_nmsub><mode>_scalar): Ditto.
25681 (*pred_nmsub<mode>_scalar): Ditto.
25682 (*pred_<macc_nmsac><mode>_scalar): Ditto.
25683 (*pred_nmsac<mode>_scalar): Ditto.
25684 (*pred_mul_<optab><mode>_scalar): Ditto.
25685 (*pred_minus_mul<mode>_scalar): Ditto.
25686 (*pred_<madd_nmsub><mode>_extended_scalar): Ditto.
25687 (*pred_nmsub<mode>_extended_scalar): Ditto.
25688 (*pred_<macc_nmsac><mode>_extended_scalar): Ditto.
25689 (*pred_nmsac<mode>_extended_scalar): Ditto.
25690 (*pred_mul_<optab><mode>_extended_scalar): Ditto.
25691 (*pred_minus_mul<mode>_extended_scalar): Ditto.
25692 (*pred_<madd_msub><mode>): Ditto.
25693 (*pred_<macc_msac><mode>): Ditto.
25694 (*pred_<madd_msub><mode>_scalar): Ditto.
25695 (*pred_<macc_msac><mode>_scalar): Ditto.
25696 (@pred_neg_mul_<optab><mode>): Ditto.
25697 (@pred_mul_neg_<optab><mode>): Ditto.
25698 (*pred_<nmadd_msub><mode>): Ditto.
25699 (*pred_<nmsub_nmadd><mode>): Ditto.
25700 (*pred_<nmacc_msac><mode>): Ditto.
25701 (*pred_<nmsac_nmacc><mode>): Ditto.
25702 (*pred_neg_mul_<optab><mode>): Ditto.
25703 (*pred_mul_neg_<optab><mode>): Ditto.
25704 (@pred_neg_mul_<optab><mode>_scalar): Ditto.
25705 (@pred_mul_neg_<optab><mode>_scalar): Ditto.
25706 (*pred_<nmadd_msub><mode>_scalar): Ditto.
25707 (*pred_<nmsub_nmadd><mode>_scalar): Ditto.
25708 (*pred_<nmacc_msac><mode>_scalar): Ditto.
25709 (*pred_<nmsac_nmacc><mode>_scalar): Ditto.
25710 (*pred_neg_mul_<optab><mode>_scalar): Ditto.
25711 (*pred_mul_neg_<optab><mode>_scalar): Ditto.
25712 (@pred_widen_neg_mul_<optab><mode>): Ditto.
25713 (@pred_widen_mul_neg_<optab><mode>): Ditto.
25714 (@pred_widen_neg_mul_<optab><mode>_scalar): Ditto.
25715 (@pred_widen_mul_neg_<optab><mode>_scalar): Ditto.
25717 2023-03-23 liuhongt <hongtao.liu@intel.com>
25719 * builtins.cc (builtin_memset_read_str): Replace
25720 targetm.gen_memset_scratch_rtx with gen_reg_rtx.
25721 (builtin_memset_gen_str): Ditto.
25722 * config/i386/i386-expand.cc
25723 (ix86_convert_const_wide_int_to_broadcast): Replace
25724 ix86_gen_scratch_sse_rtx with gen_reg_rtx.
25725 (ix86_expand_vector_move): Ditto.
25726 * config/i386/i386-protos.h (ix86_gen_scratch_sse_rtx):
25728 * config/i386/i386.cc (ix86_gen_scratch_sse_rtx): Removed.
25729 (TARGET_GEN_MEMSET_SCRATCH_RTX): Removed.
25730 * doc/tm.texi: Remove TARGET_GEN_MEMSET_SCRATCH_RTX.
25731 * doc/tm.texi.in: Ditto.
25732 * target.def: Ditto.
25734 2023-03-22 Vladimir N. Makarov <vmakarov@redhat.com>
25736 * lra.cc (lra): Do not repeat inheritance and live range splitting
25737 when asm error is found.
25739 2023-03-22 Andrew Jenner <andrew@codesourcery.com>
25741 * config/gcn/gcn-protos.h (gcn_expand_dpp_swap_pairs_insn)
25742 (gcn_expand_dpp_distribute_even_insn)
25743 (gcn_expand_dpp_distribute_odd_insn): Declare.
25744 * config/gcn/gcn-valu.md (@dpp_swap_pairs<mode>)
25745 (@dpp_distribute_even<mode>, @dpp_distribute_odd<mode>)
25746 (cmul<conj_op><mode>3, cml<addsub_as><mode>4, vec_addsub<mode>3)
25747 (cadd<rot><mode>3, vec_fmaddsub<mode>4, vec_fmsubadd<mode>4)
25748 (fms<mode>4<exec>, fms<mode>4_negop2<exec>, fms<mode>4)
25749 (fms<mode>4_negop2): New patterns.
25750 * config/gcn/gcn.cc (gcn_expand_dpp_swap_pairs_insn)
25751 (gcn_expand_dpp_distribute_even_insn)
25752 (gcn_expand_dpp_distribute_odd_insn): New functions.
25753 * config/gcn/gcn.md: Add entries to unspec enum.
25755 2023-03-22 Aldy Hernandez <aldyh@redhat.com>
25757 PR tree-optimization/109008
25758 * value-range.cc (frange::set): Add nan_state argument.
25759 * value-range.h (class nan_state): New.
25760 (frange::get_nan_state): New.
25762 2023-03-22 Martin Liska <mliska@suse.cz>
25764 * configure: Regenerate.
25766 2023-03-21 Joseph Myers <joseph@codesourcery.com>
25768 * stor-layout.cc (finalize_type_size): Copy TYPE_TYPELESS_STORAGE
25771 2023-03-21 Andrew MacLeod <amacleod@redhat.com>
25773 PR tree-optimization/109192
25774 * gimple-range-gori.cc (gori_compute::compute_operand_range):
25775 Terminate gori calculations if a relation is not relevant.
25776 * value-relation.h (value_relation::set_relation): Allow
25777 equality between op1 and op2 if they are the same.
25779 2023-03-21 Richard Biener <rguenther@suse.de>
25781 PR tree-optimization/109219
25782 * tree-vect-loop.cc (vectorizable_reduction): Check
25783 slp_node, not STMT_SLP_TYPE.
25784 * tree-vect-stmts.cc (vectorizable_condition): Likewise.
25785 * tree-vect-slp.cc (vect_slp_analyze_node_operations_1):
25786 Remove assertion on STMT_SLP_TYPE.
25788 2023-03-21 Jakub Jelinek <jakub@redhat.com>
25790 PR tree-optimization/109215
25791 * tree.h (enum special_array_member): Adjust comments for int_0
25793 * tree.cc (component_ref_sam_type): Clear zero_elts if memtype
25794 has zero sized element type and the array has variable number of
25795 elements or constant one or more elements.
25796 (component_ref_size): Adjust comments, formatting fix.
25798 2023-03-21 Arsen Arsenović <arsen@aarsen.me>
25800 * configure.ac: Add check for the Texinfo 6.8
25801 CONTENTS_OUTPUT_LOCATION customization variable and set it if
25803 * configure: Regenerate.
25804 * Makefile.in (MAKEINFO_TOC_INLINE_FLAG): New variable. Set by
25805 configure.ac to -c CONTENTS_OUTPUT_LOCATION=inline if
25806 CONTENTS_OUTPUT_LOCATION support is detected, empty otherwise.
25807 ($(build_htmldir)/%/index.html): Pass MAKEINFO_TOC_INLINE_FLAG.
25809 2023-03-21 Arsen Arsenović <arsen@aarsen.me>
25811 * doc/extend.texi: Associate use_hazard_barrier_return index
25812 entry with its attribute.
25813 * doc/invoke.texi: Associate -fcanon-prefix-map index entry with
25816 2023-03-21 Arsen Arsenović <arsen@aarsen.me>
25818 * doc/implement-c.texi: Remove usage of @gol.
25819 * doc/invoke.texi: Ditto.
25820 * doc/sourcebuild.texi: Ditto.
25821 * doc/include/gcc-common.texi: Remove @gol. In new Makeinfo and
25822 texinfo.tex versions, the bug it was working around appears to
25825 2023-03-21 Arsen Arsenović <arsen@aarsen.me>
25827 * doc/include/texinfo.tex: Update to 2023-01-17.19.
25829 2023-03-21 Arsen Arsenović <arsen@aarsen.me>
25831 * doc/include/gcc-common.texi: Add @defbuiltin{,x} and
25832 @enddefbuiltin for defining built-in functions.
25833 * doc/extend.texi: Apply @defbuiltin{,x} to many, but not all,
25834 places where it should be used.
25836 2023-03-21 Arsen Arsenović <arsen@aarsen.me>
25838 * doc/extend.texi (Formatted Output Function Checking): New
25839 subsection for grouping together printf et al.
25840 (Exception handling) Fix missing @ sign before copyright
25841 header, which lead to the copyright line leaking into
25842 '(gcc)Exception handling'.
25843 * doc/gcc.texi: Set document language to en_US.
25844 (@copying): Wrap front cover texts in quotations, move in manual
25847 2023-03-21 Arsen Arsenović <arsen@aarsen.me>
25849 * doc/gcc.texi: Add the Indices appendix, to make texinfo
25850 generate nice indices overview page.
25852 2023-03-21 Richard Biener <rguenther@suse.de>
25854 PR tree-optimization/109170
25855 * gimple-range-op.cc (cfn_pass_through_arg1): New.
25856 (gimple_range_op_handler::maybe_builtin_call): Handle
25857 __builtin_expect via cfn_pass_through_arg1.
25859 2023-03-20 Michael Meissner <meissner@linux.ibm.com>
25862 * config/rs6000/rs6000.cc (create_complex_muldiv): Delete.
25863 (init_float128_ieee): Delete code to switch complex multiply and divide
25865 (complex_multiply_builtin_code): New helper function.
25866 (complex_divide_builtin_code): Likewise.
25867 (rs6000_mangle_decl_assembler_name): Add support for mangling the name
25868 of complex 128-bit multiply and divide built-in functions.
25870 2023-03-20 Peter Bergner <bergner@linux.ibm.com>
25873 * config/rs6000/rs6000-builtin.cc (stv_expand_builtin): Use tmode.
25875 2023-03-19 Jonny Grant <jg@jguk.org>
25877 * doc/extend.texi (Common Function Attributes) <nonnull>:
25880 2023-03-18 Peter Bergner <bergner@linux.ibm.com>
25882 PR rtl-optimization/109179
25883 * lra-constraints.cc (combine_reload_insn): Enforce TO is not a debug
25884 insn or note. Move the tests earlier to guard lra_get_insn_recog_data.
25886 2023-03-17 Jakub Jelinek <jakub@redhat.com>
25889 * function.h (push_struct_function): Add ABSTRACT_P argument defaulted
25891 * function.cc (push_struct_function): Add ABSTRACT_P argument, pass it
25892 to allocate_struct_function instead of false.
25893 * tree-inline.cc (initialize_cfun): Don't copy DECL_ARGUMENTS
25894 nor DECL_RESULT here. Pass true as ABSTRACT_P to
25895 push_struct_function. Call targetm.target_option.relayout_function
25897 (tree_function_versioning): Formatting fix.
25899 2023-03-17 Vladimir N. Makarov <vmakarov@redhat.com>
25901 * lra-constraints.cc: Include hooks.h.
25902 (combine_reload_insn): New function.
25903 (lra_constraints): Call it.
25905 2023-03-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
25906 kito-cheng <kito.cheng@sifive.com>
25908 * config/riscv/riscv-v.cc (legitimize_move): Allow undef value
25909 as legitimate value.
25910 * config/riscv/riscv-vector-builtins.cc
25911 (function_expander::use_ternop_insn): Fix bugs of ternary intrinsic.
25912 (function_expander::use_widen_ternop_insn): Ditto.
25913 * config/riscv/vector.md (@vundefined<mode>): New pattern.
25914 (pred_mul_<optab><mode>_undef_merge): Remove.
25915 (*pred_mul_<optab><mode>_undef_merge_scalar): Ditto.
25916 (*pred_mul_<optab><mode>_undef_merge_extended_scalar): Ditto.
25917 (pred_neg_mul_<optab><mode>_undef_merge): Ditto.
25918 (*pred_neg_mul_<optab><mode>_undef_merge_scalar): Ditto.
25920 2023-03-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
25923 * config/riscv/riscv.md: Fix subreg bug.
25925 2023-03-17 Jakub Jelinek <jakub@redhat.com>
25927 PR middle-end/108685
25928 * omp-expand.cc (expand_omp_for_ordered_loops): Add L0_BB argument,
25929 use its loop_father rather than BODY_BB's loop_father.
25930 (expand_omp_for_generic): Adjust expand_omp_for_ordered_loops caller.
25931 If broken_loop with ordered > collapse and at least one of those
25932 extra loops aren't guaranteed to have at least one iteration, change
25933 l0_bb's loop_father to entry_bb's loop_father. Set cont_bb's
25934 loop_father to l0_bb's loop_father rather than l1_bb's.
25936 2023-03-17 Jakub Jelinek <jakub@redhat.com>
25939 * gdbhooks.py (TreePrinter.to_string): Wrap
25940 gdb.parse_and_eval('tree_code_type') in a try block, parse
25941 and eval 'tree_code_type_tmpl<0>::tree_code_type' instead if it
25942 raises exception. Update comments for the recent tree_code_type
25945 2023-03-17 Sandra Loosemore <sandra@codesourcery.com>
25947 * doc/extend.texi (BPF Built-in Functions): Fix numerous markup
25948 issues. Add more line breaks to example so it doesn't overflow
25951 2023-03-17 Sandra Loosemore <sandra@codesourcery.com>
25953 * doc/extend.texi (Common Function Attributes) <access>: Fix bad
25954 line breaks in examples.
25955 <malloc>: Fix bad line breaks in running text, also copy-edit
25957 (Extended Asm) <Generic Operand Modifiers>: Fix @multitable width.
25958 * doc/invoke.texi (Option Summary) <Developer Options>: Fix misplaced
25960 (C++ Dialect Options) <-fcontracts>: Add line break in example.
25961 <-Wctad-maybe-unsupported>: Likewise.
25962 <-Winvalid-constexpr>: Likewise.
25963 (Warning Options) <-Wdangling-pointer>: Likewise.
25964 <-Winterference-size>: Likewise.
25965 <-Wvla-parameter>: Likewise.
25966 (Static Analyzer Options): Fix bad line breaks in running text,
25967 plus add some missing markup.
25968 (Optimize Options) <openacc-privatization>: Fix more bad line
25969 breaks in running text.
25971 2023-03-16 Uros Bizjak <ubizjak@gmail.com>
25973 * config/i386/i386-expand.cc (expand_vec_perm_pblendv):
25974 Handle 8-byte modes only with TARGET_MMX_WITH_SSE.
25975 (expand_vec_perm_2perm_pblendv): Ditto.
25977 2023-03-16 Martin Liska <mliska@suse.cz>
25979 PR middle-end/106133
25980 * gcc.cc (driver_handle_option): Use x_main_input_basename
25981 if x_dump_base_name is null.
25982 * opts.cc (common_handle_option): Likewise.
25984 2023-03-16 Richard Biener <rguenther@suse.de>
25986 PR tree-optimization/109123
25987 * gimple-ssa-warn-access.cc (pass_waccess::warn_invalid_pointer):
25988 Do not emit -Wuse-after-free late.
25989 (pass_waccess::check_call): Always check call pointer uses.
25991 2023-03-16 Richard Biener <rguenther@suse.de>
25993 PR tree-optimization/109141
25994 * tree-dfa.h (renumber_gimple_stmt_uids_in_block): New.
25995 * tree-dfa.cc (renumber_gimple_stmt_uids_in_block): Split
25997 (renumber_gimple_stmt_uids): ... here and
25998 (renumber_gimple_stmt_uids_in_blocks): ... here.
25999 * gimple-ssa-warn-access.cc (pass_waccess::use_after_inval_p):
26000 Use renumber_gimple_stmt_uids_in_block to also assign UIDs
26002 (pass_waccess::check_pointer_uses): Process all PHIs.
26004 2023-03-15 David Malcolm <dmalcolm@redhat.com>
26007 * diagnostic-format-sarif.cc (class sarif_invocation): New.
26008 (class sarif_ice_notification): New.
26009 (sarif_builder::m_invocation_obj): New field.
26010 (sarif_invocation::add_notification_for_ice): New.
26011 (sarif_invocation::prepare_to_flush): New.
26012 (sarif_ice_notification::sarif_ice_notification): New.
26013 (sarif_builder::sarif_builder): Add m_invocation_obj.
26014 (sarif_builder::end_diagnostic): Special-case DK_ICE and
26016 (sarif_builder::flush_to_file): Call prepare_to_flush on
26017 m_invocation_obj. Pass the latter to make_top_level_object.
26018 (sarif_builder::make_result_object): Move creation of "locations"
26020 (sarif_builder::make_locations_arr): ...this new function.
26021 (sarif_builder::make_top_level_object): Add "invocation_obj" param
26022 and pass it to make_run_object.
26023 (sarif_builder::make_run_object): Add "invocation_obj" param and
26025 (sarif_ice_handler): New callback.
26026 (diagnostic_output_format_init_sarif): Wire up sarif_ice_handler.
26027 * diagnostic.cc (diagnostic_initialize): Initialize new field
26029 (diagnostic_action_after_output): If it is set, make one attempt
26030 to call ice_handler_cb.
26031 * diagnostic.h (diagnostic_context::ice_handler_cb): New field.
26033 2023-03-15 Uros Bizjak <ubizjak@gmail.com>
26035 * config/i386/i386-expand.cc (expand_vec_perm_blend):
26036 Handle 8-byte modes only with TARGET_MMX_WITH_SSE. Handle V2SFmode
26037 and fix V2HImode handling.
26038 (expand_vec_perm_1): Try to emit BLEND instruction
26039 before MOVSS/MOVSD.
26040 * config/i386/mmx.md (*mmx_blendps): New insn pattern.
26042 2023-03-15 Tobias Burnus <tobias@codesourcery.com>
26044 * omp-low.cc (omp_runtime_api_call): Add omp_in_explicit_task.
26046 2023-03-15 Richard Biener <rguenther@suse.de>
26048 * gimple-ssa-warn-access.cc (pass_waccess::check_pointer_uses):
26049 Do not diagnose clobbers.
26051 2023-03-15 Richard Biener <rguenther@suse.de>
26053 PR tree-optimization/109139
26054 * tree-ssa-live.cc (remove_unused_locals): Look at the
26055 base address for unused decls on the LHS of .DEFERRED_INIT.
26057 2023-03-15 Xi Ruoyao <xry111@xry111.site>
26060 * builtins.cc (inline_string_cmp): Force the character
26061 difference into "result" pseudo-register, instead of reassign
26062 the pseudo-register.
26064 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
26066 * config.gcc: Add thead.o to RISC-V extra_objs.
26067 * config/riscv/peephole.md: Add mempair peephole passes.
26068 * config/riscv/riscv-protos.h (riscv_split_64bit_move_p): New
26070 (th_mempair_operands_p): Likewise.
26071 (th_mempair_order_operands): Likewise.
26072 (th_mempair_prepare_save_restore_operands): Likewise.
26073 (th_mempair_save_restore_regs): Likewise.
26074 (th_mempair_output_move): Likewise.
26075 * config/riscv/riscv.cc (riscv_save_reg): Move code.
26076 (riscv_restore_reg): Move code.
26077 (riscv_for_each_saved_reg): Add code to emit mempair insns.
26078 * config/riscv/t-riscv: Add thead.cc.
26079 * config/riscv/thead.md (*th_mempair_load_<GPR:mode>2):
26081 (*th_mempair_store_<GPR:mode>2): Likewise.
26082 (*th_mempair_load_extendsidi2): Likewise.
26083 (*th_mempair_load_zero_extendsidi2): Likewise.
26084 * config/riscv/thead.cc: New file.
26086 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
26088 * config/riscv/constraints.md (TARGET_XTHEADFMV ? FP_REGS : NO_REGS)
26089 New constraint "th_f_fmv".
26090 (TARGET_XTHEADFMV ? GR_REGS : NO_REGS): New constraint
26092 * config/riscv/riscv.cc (riscv_split_doubleword_move):
26093 Add split code for XTheadFmv.
26094 (riscv_secondary_memory_needed): XTheadFmv does not need
26096 * config/riscv/riscv.md: Add new UNSPEC_XTHEADFMV and
26097 UNSPEC_XTHEADFMV_HW. Add support for XTheadFmv to
26098 movdf_hardfloat_rv32.
26099 * config/riscv/thead.md (th_fmv_hw_w_x): New INSN.
26100 (th_fmv_x_w): New INSN.
26101 (th_fmv_x_hw): New INSN.
26103 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
26105 * config/riscv/riscv.md (maddhisi4): New expand.
26106 (msubhisi4): New expand.
26107 * config/riscv/thead.md (*th_mula<mode>): New pattern.
26108 (*th_mulawsi): New pattern.
26109 (*th_mulawsi2): New pattern.
26110 (*th_maddhisi4): New pattern.
26111 (*th_sextw_maddhisi4): New pattern.
26112 (*th_muls<mode>): New pattern.
26113 (*th_mulswsi): New pattern.
26114 (*th_mulswsi2): New pattern.
26115 (*th_msubhisi4): New pattern.
26116 (*th_sextw_msubhisi4): New pattern.
26118 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
26120 * config/riscv/iterators.md (TARGET_64BIT): Add GPR2 iterator.
26121 * config/riscv/riscv-protos.h (riscv_expand_conditional_move):
26123 * config/riscv/riscv.cc (riscv_rtx_costs): Add costs for
26125 (riscv_expand_conditional_move): New function.
26126 (riscv_expand_conditional_move_onesided): New function.
26127 * config/riscv/riscv.md: Add support for XTheadCondMov.
26128 * config/riscv/thead.md (*th_cond_mov<GPR:mode><GPR2:mode>): Add
26129 support for XTheadCondMov.
26130 (*th_cond_gpr_mov<GPR:mode><GPR2:mode>): Likewise.
26132 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
26134 * config/riscv/bitmanip.md (clzdi2): New expand.
26135 (clzsi2): New expand.
26136 (ctz<mode>2): New expand.
26137 (popcount<mode>2): New expand.
26138 (<bitmanip_optab>si2): Rename INSN.
26139 (*<bitmanip_optab>si2): Hide INSN name.
26140 (<bitmanip_optab>di2): Rename INSN.
26141 (*<bitmanip_optab>di2): Hide INSN name.
26142 (rotrsi3): Remove INSN.
26143 (rotr<mode>3): Add expand.
26144 (*rotrsi3): New INSN.
26145 (rotrdi3): Rename INSN.
26146 (*rotrdi3): Hide INSN name.
26147 (rotrsi3_sext): Rename INSN.
26148 (*rotrsi3_sext): Hide INSN name.
26149 (bswap<mode>2): Remove INSN.
26150 (bswapdi2): Add expand.
26151 (bswapsi2): Add expand.
26152 (*bswap<mode>2): Hide INSN name.
26153 * config/riscv/riscv.cc (riscv_rtx_costs): Add costs for sign
26155 * config/riscv/riscv.md (extv<mode>): New expand.
26156 (extzv<mode>): New expand.
26157 * config/riscv/thead.md (*th_srri<mode>3): New INSN.
26158 (*th_ext<mode>): New INSN.
26159 (*th_extu<mode>): New INSN.
26160 (*th_clz<mode>2): New INSN.
26161 (*th_rev<mode>2): New INSN.
26163 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
26165 * config/riscv/riscv.cc (riscv_rtx_costs): Add xthead:tst cost.
26166 * config/riscv/thead.md (*th_tst<mode>3): New INSN.
26168 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
26170 * config/riscv/riscv.md: Include thead.md
26171 * config/riscv/thead.md: New file.
26173 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
26175 * config/riscv/riscv-cores.def (RISCV_CORE): Add "thead-c906".
26177 2023-03-15 Christoph Müllner <christoph.muellner@vrull.eu>
26179 * common/config/riscv/riscv-common.cc: Add xthead* extensions.
26180 * config/riscv/riscv-opts.h (MASK_XTHEADBA): New.
26181 (MASK_XTHEADBB): New.
26182 (MASK_XTHEADBS): New.
26183 (MASK_XTHEADCMO): New.
26184 (MASK_XTHEADCONDMOV): New.
26185 (MASK_XTHEADFMEMIDX): New.
26186 (MASK_XTHEADFMV): New.
26187 (MASK_XTHEADINT): New.
26188 (MASK_XTHEADMAC): New.
26189 (MASK_XTHEADMEMIDX): New.
26190 (MASK_XTHEADMEMPAIR): New.
26191 (MASK_XTHEADSYNC): New.
26192 (TARGET_XTHEADBA): New.
26193 (TARGET_XTHEADBB): New.
26194 (TARGET_XTHEADBS): New.
26195 (TARGET_XTHEADCMO): New.
26196 (TARGET_XTHEADCONDMOV): New.
26197 (TARGET_XTHEADFMEMIDX): New.
26198 (TARGET_XTHEADFMV): New.
26199 (TARGET_XTHEADINT): New.
26200 (TARGET_XTHEADMAC): New.
26201 (TARGET_XTHEADMEMIDX): New.
26202 (TARGET_XTHEADMEMPAIR): new.
26203 (TARGET_XTHEADSYNC): New.
26204 * config/riscv/riscv.opt: Add riscv_xthead_subext.
26206 2023-03-15 Hu, Lin1 <lin1.hu@intel.com>
26209 * config/i386/i386-builtin.def (__builtin_ia32_vaesdec_v16qi,
26210 __builtin_ia32_vaesdeclast_v16qi,__builtin_ia32_vaesenc_v16qi,
26211 __builtin_ia32_vaesenclast_v16qi): Require OPTION_MASK_ISA_AVX512VL.
26213 2023-03-14 Jakub Jelinek <jakub@redhat.com>
26216 * config/i386/i386-expand.cc (split_double_concat): Fix splitting
26217 when lo is equal to dhi and hi is a MEM which uses dlo register.
26219 2023-03-14 Martin Jambor <mjambor@suse.cz>
26222 * ipa-cp.cc (update_profiling_info): Drop counts of orig_node to
26223 global0 instead of zeroing when it does not have as many counts as
26226 2023-03-14 Martin Jambor <mjambor@suse.cz>
26229 * ipa-cp.cc (update_specialized_profile): Drop orig_node_count to
26230 ipa count, remove assert, lenient_count_portion_handling, dump
26231 also orig_node_count.
26233 2023-03-14 Uros Bizjak <ubizjak@gmail.com>
26235 * config/i386/i386-expand.cc (expand_vec_perm_movs):
26236 Handle V2SImode for TARGET_MMX_WITH_SSE.
26237 * config/i386/mmx.md (*mmx_movss_<mode>): Rename from *mmx_movss
26238 using V2FI mode iterator to handle both V2SI and V2SF modes.
26240 2023-03-14 Sam James <sam@gentoo.org>
26242 * config/riscv/genrvv-type-indexer.cc: Avoid calloc() poisoning on musl by
26243 including <sstream> earlier.
26244 * system.h: Add INCLUDE_SSTREAM.
26246 2023-03-14 Richard Biener <rguenther@suse.de>
26248 * tree-ssa-live.cc (remove_unused_locals): Do not treat
26249 the .DEFERRED_INIT of a variable as use, instead remove
26250 that if it is the only use.
26252 2023-03-14 Eric Botcazou <ebotcazou@adacore.com>
26254 PR rtl-optimization/107762
26255 * expr.cc (emit_group_store): Revert latest change.
26257 2023-03-14 Andre Vieira <andre.simoesdiasvieira@arm.com>
26259 PR tree-optimization/109005
26260 * tree-if-conv.cc (get_bitfield_rep): Replace BLKmode check with
26261 aggregate type check.
26263 2023-03-14 Jakub Jelinek <jakub@redhat.com>
26265 PR tree-optimization/109115
26266 * tree-vect-patterns.cc (vect_recog_divmod_pattern): Don't use
26267 r.upper_bound () on r.undefined_p () range.
26269 2023-03-14 Jan Hubicka <hubicka@ucw.cz>
26271 PR tree-optimization/106896
26272 * profile-count.cc (profile_count::to_sreal_scale): Synchronize
26273 implementatoin with probability_in; avoid some asserts.
26275 2023-03-13 Max Filippov <jcmvbkbc@gmail.com>
26277 * config/xtensa/linux.h (TARGET_ASM_FILE_END): New macro.
26279 2023-03-13 Sean Bright <sean@seanbright.com>
26281 * doc/invoke.texi (Warning Options): Remove errant 'See'
26284 2023-03-13 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
26286 * config/xtensa/xtensa.h (REG_OK_STRICT, REG_OK_FOR_INDEX_P,
26287 REG_OK_FOR_BASE_P): Remove.
26289 2023-03-13 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26291 * config/riscv/vector-iterators.md (=vd,vr): Fine tune.
26292 (=vd,vd,vr,vr): Ditto.
26293 * config/riscv/vector.md: Ditto.
26295 2023-03-13 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26297 * config/riscv/riscv-vector-builtins.cc
26298 (function_expander::use_compare_insn): Add operand predicate check.
26300 2023-03-13 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26302 * config/riscv/vector.md: Fine tune RA constraints.
26304 2023-03-13 Tobias Burnus <tobias@codesourcery.com>
26306 * config/gcn/mkoffload.cc (main): Pass -save-temps on for the
26307 hsaco assemble/link.
26309 2023-03-13 Richard Biener <rguenther@suse.de>
26311 PR tree-optimization/109046
26312 * tree-ssa-forwprop.cc (pass_forwprop::execute): Combine
26313 piecewise complex loads.
26315 2023-03-12 Jakub Jelinek <jakub@redhat.com>
26317 * config/aarch64/aarch64.h (aarch64_bf16_type_node): Remove.
26318 (aarch64_bf16_ptr_type_node): Adjust comment.
26319 * config/aarch64/aarch64.cc (aarch64_gimplify_va_arg_expr): Use
26320 bfloat16_type_node rather than aarch64_bf16_type_node.
26321 (aarch64_libgcc_floating_mode_supported_p,
26322 aarch64_scalar_mode_supported_p): Also support BFmode.
26323 (aarch64_invalid_conversion, aarch64_invalid_unary_op): Remove.
26324 (aarch64_invalid_binary_op): Remove BFmode related rejections.
26325 (TARGET_INVALID_CONVERSION, TARGET_INVALID_UNARY_OP): Don't redefine.
26326 * config/aarch64/aarch64-builtins.cc (aarch64_bf16_type_node): Remove.
26327 (aarch64_int_or_fp_type): Use bfloat16_type_node rather than
26328 aarch64_bf16_type_node.
26329 (aarch64_init_simd_builtin_types): Likewise.
26330 (aarch64_init_bf16_types): Likewise. Don't create bfloat16_type_node,
26331 which is created in tree.cc already.
26332 * config/aarch64/aarch64-sve-builtins.def (svbfloat16_t): Likewise.
26334 2023-03-12 Roger Sayle <roger@nextmovesoftware.com>
26336 PR middle-end/109031
26337 * tree-chrec.cc (chrec_apply): When folding "{a, +, a} (x-1)",
26338 ensure that the type of x is as wide or wider than the type of a.
26340 2023-03-12 Tamar Christina <tamar.christina@arm.com>
26343 * config/aarch64/aarch64-simd.md (@aarch64_bitmask_udiv<mode>3): Remove.
26344 (*bitmask_shift_plus<mode>): New.
26345 * config/aarch64/aarch64-sve2.md (*bitmask_shift_plus<mode>): New.
26346 (@aarch64_bitmask_udiv<mode>3): Remove.
26347 * config/aarch64/aarch64.cc
26348 (aarch64_vectorize_can_special_div_by_constant,
26349 TARGET_VECTORIZE_CAN_SPECIAL_DIV_BY_CONST): Removed.
26350 (TARGET_VECTORIZE_PREFERRED_DIV_AS_SHIFTS_OVER_MULT,
26351 aarch64_vectorize_preferred_div_as_shifts_over_mult): New.
26353 2023-03-12 Tamar Christina <tamar.christina@arm.com>
26356 * target.def (preferred_div_as_shifts_over_mult): New.
26357 * doc/tm.texi.in: Document it.
26358 * doc/tm.texi: Regenerate.
26359 * targhooks.cc (default_preferred_div_as_shifts_over_mult): New.
26360 * targhooks.h (default_preferred_div_as_shifts_over_mult): New.
26361 * tree-vect-patterns.cc (vect_recog_divmod_pattern): Use it.
26363 2023-03-12 Tamar Christina <tamar.christina@arm.com>
26364 Richard Sandiford <richard.sandiford@arm.com>
26367 * tree-ssa-math-opts.cc (convert_mult_to_fma): Inhibit FMA in case not
26370 2023-03-12 Tamar Christina <tamar.christina@arm.com>
26371 Andrew MacLeod <amacleod@redhat.com>
26374 * gimple-range-op.h (gimple_range_op_handler): Add maybe_non_standard.
26375 * gimple-range-op.cc (gimple_range_op_handler::gimple_range_op_handler):
26377 (gimple_range_op_handler::maybe_non_standard): New.
26378 * range-op.cc (class operator_widen_plus_signed,
26379 operator_widen_plus_signed::wi_fold, class operator_widen_plus_unsigned,
26380 operator_widen_plus_unsigned::wi_fold, class operator_widen_mult_signed,
26381 operator_widen_mult_signed::wi_fold, class operator_widen_mult_unsigned,
26382 operator_widen_mult_unsigned::wi_fold,
26383 ptr_op_widen_mult_signed, ptr_op_widen_mult_unsigned,
26384 ptr_op_widen_plus_signed, ptr_op_widen_plus_unsigned): New.
26385 * range-op.h (ptr_op_widen_mult_signed, ptr_op_widen_mult_unsigned,
26386 ptr_op_widen_plus_signed, ptr_op_widen_plus_unsigned): New
26388 2023-03-12 Tamar Christina <tamar.christina@arm.com>
26391 * doc/tm.texi (TARGET_VECTORIZE_CAN_SPECIAL_DIV_BY_CONST): Remove.
26392 * doc/tm.texi.in: Likewise.
26393 * explow.cc (round_push, align_dynamic_address): Revert previous patch.
26394 * expmed.cc (expand_divmod): Likewise.
26395 * expmed.h (expand_divmod): Likewise.
26396 * expr.cc (force_operand, expand_expr_divmod): Likewise.
26397 * optabs.cc (expand_doubleword_mod, expand_doubleword_divmod): Likewise.
26398 * target.def (can_special_div_by_const): Remove.
26399 * target.h: Remove tree-core.h include
26400 * targhooks.cc (default_can_special_div_by_const): Remove.
26401 * targhooks.h (default_can_special_div_by_const): Remove.
26402 * tree-vect-generic.cc (expand_vector_operation): Remove hook.
26403 * tree-vect-patterns.cc (vect_recog_divmod_pattern): Remove hook.
26404 * tree-vect-stmts.cc (vectorizable_operation): Remove hook.
26406 2023-03-12 Sandra Loosemore <sandra@codesourcery.com>
26408 * doc/install.texi2html: Fix issue number typo in comment.
26410 2023-03-12 Gaius Mulley <gaiusmod2@gmail.com>
26412 * doc/gm2.texi (Elementary data types): Equivalence BOOLEAN with
26415 2023-03-12 Sandra Loosemore <sandra@codesourcery.com>
26417 * doc/invoke.texi (Optimize Options): Add markup to
26418 description of asan-kernel-mem-intrinsic-prefix, and clarify
26421 2023-03-11 Gerald Pfeifer <gerald@pfeifer.com>
26423 * doc/extend.texi (Named Address Spaces): Drop a redundant link
26426 2023-03-11 Jeff Law <jlaw@ventanamicro>
26429 * doc/extend.texi: Clarify Attribute Syntax a bit.
26431 2023-03-11 Sandra Loosemore <sandra@codesourcery.com>
26433 * doc/install.texi (Prerequisites): Suggest using newer versions
26435 (Final install): Clean up and modernize discussion of how to
26436 build or obtain the GCC manuals.
26437 * doc/install.texi2html: Update comment to point to the PR instead
26438 of "makeinfo 4.7 brokenness" (it's not specific to that version).
26440 2023-03-10 Jakub Jelinek <jakub@redhat.com>
26443 * optabs.cc (expand_fix): For conversions from BFmode to integral,
26444 use shifts to convert it to SFmode first and then convert SFmode
26447 2023-03-10 Andrew Pinski <apinski@marvell.com>
26449 * config/aarch64/aarch64.md: Add a new define_split
26452 2023-03-10 Richard Biener <rguenther@suse.de>
26454 * tree-ssa-structalias.cc (solve_graph): Immediately
26455 iterate self-cycles.
26457 2023-03-10 Jakub Jelinek <jakub@redhat.com>
26459 PR tree-optimization/109008
26460 * range-op-float.cc (float_widen_lhs_range): If not
26461 -frounding-math and not IBM double double format, extend lhs
26462 range just by 0.5ulp rather than 1ulp in each direction.
26464 2023-03-10 Jakub Jelinek <jakub@redhat.com>
26467 * config.gcc (x86_64-*-cygwin*): Don't add i386/t-cygwin-w64 into
26469 * config/i386/t-cygwin-w64: Remove.
26471 2023-03-10 Jakub Jelinek <jakub@redhat.com>
26474 * tree-core.h (tree_code_type, tree_code_length): For C++11 or
26475 C++14, don't declare as extern const arrays.
26476 (tree_code_type_tmpl, tree_code_length_tmpl): New types with
26477 static constexpr member arrays for C++11 or C++14.
26478 * tree.h (TREE_CODE_CLASS): For C++11 or C++14 use
26479 tree_code_type_tmpl <0>::tree_code_type instead of tree_code_type.
26480 (TREE_CODE_LENGTH): For C++11 or C++14 use
26481 tree_code_length_tmpl <0>::tree_code_length instead of
26483 * tree.cc (tree_code_type, tree_code_length): Remove.
26485 2023-03-10 Jakub Jelinek <jakub@redhat.com>
26488 * common.opt (fcanon-prefix-map): New option.
26489 * opts.cc: Include file-prefix-map.h.
26490 (flag_canon_prefix_map): New variable.
26491 (common_handle_option): Handle OPT_fcanon_prefix_map.
26492 (gen_command_line_string): Ignore OPT_fcanon_prefix_map.
26493 * file-prefix-map.h (flag_canon_prefix_map): Declare.
26494 * file-prefix-map.cc (struct file_prefix_map): Add canonicalize
26496 (add_prefix_map): Initialize canonicalize member from
26497 flag_canon_prefix_map, and if true canonicalize it using lrealpath.
26498 (remap_filename): Revert 2022-11-01 and 2022-11-07 changes,
26499 use lrealpath result only for map->canonicalize map entries.
26500 * lto-opts.cc (lto_write_options): Ignore OPT_fcanon_prefix_map.
26501 * opts-global.cc (handle_common_deferred_options): Clear
26502 flag_canon_prefix_map at the start and handle OPT_fcanon_prefix_map.
26503 * doc/invoke.texi (-fcanon-prefix-map): Document.
26504 (-ffile-prefix-map, -fdebug-prefix-map, -fprofile-prefix-map): Add
26505 see also for -fcanon-prefix-map.
26506 * doc/cppopts.texi (-fmacro-prefix-map): Likewise.
26508 2023-03-10 Jakub Jelinek <jakub@redhat.com>
26511 * cgraphunit.cc (check_global_declaration): Don't warn for unused
26512 variables which have OPT_Wunused_variable warning suppressed.
26514 2023-03-10 Jakub Jelinek <jakub@redhat.com>
26516 PR tree-optimization/109008
26517 * range-op-float.cc (float_widen_lhs_range): If lb is
26518 minimum representable finite number or ub is maximum
26519 representable finite number, instead of widening it to
26520 -inf or inf widen it to negative or positive 0x0.8p+(EMAX+1).
26521 Temporarily clear flag_finite_math_only when canonicalizing
26524 2023-03-10 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26526 * config/riscv/riscv-builtins.cc (riscv_gimple_fold_builtin): New function.
26527 * config/riscv/riscv-protos.h (riscv_gimple_fold_builtin): Ditto.
26528 (gimple_fold_builtin): Ditto.
26529 * config/riscv/riscv-vector-builtins-bases.cc (class read_vl): New class.
26530 (class vleff): Ditto.
26532 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
26533 * config/riscv/riscv-vector-builtins-functions.def (read_vl): Ditto.
26535 * config/riscv/riscv-vector-builtins-shapes.cc (struct read_vl_def): Ditto.
26536 (struct fault_load_def): Ditto.
26538 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
26539 * config/riscv/riscv-vector-builtins.cc
26540 (rvv_arg_type_info::get_tree_type): Add size_ptr.
26541 (gimple_folder::gimple_folder): New class.
26542 (gimple_folder::fold): Ditto.
26543 (gimple_fold_builtin): New function.
26544 (get_read_vl_instance): Ditto.
26545 (get_read_vl_decl): Ditto.
26546 * config/riscv/riscv-vector-builtins.def (size_ptr): Add size_ptr.
26547 * config/riscv/riscv-vector-builtins.h (class gimple_folder): New class.
26548 (get_read_vl_instance): New function.
26549 (get_read_vl_decl): Ditto.
26550 * config/riscv/riscv-vsetvl.cc (fault_first_load_p): Ditto.
26551 (read_vl_insn_p): Ditto.
26552 (available_occurrence_p): Ditto.
26553 (backward_propagate_worthwhile_p): Ditto.
26554 (gen_vsetvl_pat): Adapt for vleff support.
26555 (get_forward_read_vl_insn): New function.
26556 (get_backward_fault_first_load_insn): Ditto.
26557 (source_equal_p): Adapt for vleff support.
26558 (first_ratio_invalid_for_second_sew_p): Remove.
26559 (first_ratio_invalid_for_second_lmul_p): Ditto.
26560 (first_lmul_less_than_second_lmul_p): Ditto.
26561 (first_ratio_less_than_second_ratio_p): Ditto.
26562 (support_relaxed_compatible_p): New function.
26563 (vector_insn_info::operator>): Remove.
26564 (vector_insn_info::operator>=): Refine.
26565 (vector_insn_info::parse_insn): Adapt for vleff support.
26566 (vector_insn_info::compatible_p): Ditto.
26567 (vector_insn_info::update_fault_first_load_avl): New function.
26568 (pass_vsetvl::transfer_after): Adapt for vleff support.
26569 (pass_vsetvl::demand_fusion): Ditto.
26570 (pass_vsetvl::cleanup_insns): Ditto.
26571 * config/riscv/riscv-vsetvl.def (DEF_INCOMPATIBLE_COND): Remove
26572 redundant condtions.
26573 * config/riscv/riscv-vsetvl.h (struct demands_cond): New function.
26574 * config/riscv/riscv.cc (TARGET_GIMPLE_FOLD_BUILTIN): New target hook.
26575 * config/riscv/riscv.md: Adapt for vleff support.
26576 * config/riscv/t-riscv: Ditto.
26577 * config/riscv/vector-iterators.md: New iterator.
26578 * config/riscv/vector.md (read_vlsi): New pattern.
26579 (read_vldi_zero_extend): Ditto.
26580 (@pred_fault_load<mode>): Ditto.
26582 2023-03-10 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26584 * config/riscv/riscv-vector-builtins.cc
26585 (function_expander::use_ternop_insn): Use maybe_gen_insn instead.
26586 (function_expander::use_widen_ternop_insn): Ditto.
26587 * optabs.cc (maybe_gen_insn): Extend nops handling.
26589 2023-03-10 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26591 * config/riscv/riscv-vector-builtins-bases.cc: Split indexed load
26592 patterns according to RVV ISA.
26593 * config/riscv/vector-iterators.md: New iterators.
26594 * config/riscv/vector.md
26595 (@pred_indexed_<order>load<VNX1_QHSD:mode><VNX1_QHSDI:mode>): Remove.
26596 (@pred_indexed_<order>load<mode>_same_eew): New pattern.
26597 (@pred_indexed_<order>load<mode>_x2_greater_eew): Ditto.
26598 (@pred_indexed_<order>load<mode>_x4_greater_eew): Ditto.
26599 (@pred_indexed_<order>load<mode>_x8_greater_eew): Ditto.
26600 (@pred_indexed_<order>load<mode>_x2_smaller_eew): Ditto.
26601 (@pred_indexed_<order>load<mode>_x4_smaller_eew): Ditto.
26602 (@pred_indexed_<order>load<mode>_x8_smaller_eew): Ditto.
26603 (@pred_indexed_<order>load<VNX2_QHSD:mode><VNX2_QHSDI:mode>): Remove.
26604 (@pred_indexed_<order>load<VNX4_QHSD:mode><VNX4_QHSDI:mode>): Ditto.
26605 (@pred_indexed_<order>load<VNX8_QHSD:mode><VNX8_QHSDI:mode>): Ditto.
26606 (@pred_indexed_<order>load<VNX16_QHS:mode><VNX16_QHSI:mode>): Ditto.
26607 (@pred_indexed_<order>load<VNX32_QH:mode><VNX32_QHI:mode>): Ditto.
26608 (@pred_indexed_<order>load<VNX64_Q:mode><VNX64_Q:mode>): Ditto.
26610 2023-03-10 Michael Collison <collison@rivosinc.com>
26612 * tree-vect-loop-manip.cc (vect_do_peeling): Use
26613 result of constant_lower_bound instead of vf for the lower
26614 bound of the epilog loop trip count.
26616 2023-03-09 Tamar Christina <tamar.christina@arm.com>
26618 * passes.cc (emergency_dump_function): Finish graph generation.
26620 2023-03-09 Tamar Christina <tamar.christina@arm.com>
26622 * config/aarch64/aarch64.md (tbranch_<code><mode>3): Restrict to SHORT
26623 and bottom bit only.
26625 2023-03-09 Andrew Pinski <apinski@marvell.com>
26627 PR tree-optimization/108980
26628 * gimple-array-bounds.cc (array_bounds_checker::check_array_ref):
26629 Reorgnize the call to warning for not strict flexible arrays
26630 to be before the check of warned.
26632 2023-03-09 Jason Merrill <jason@redhat.com>
26634 * doc/extend.texi: Comment out __is_deducible docs.
26636 2023-03-09 Jason Merrill <jason@redhat.com>
26639 * doc/extend.texi (Type Traits):: Document __is_deducible.
26641 2023-03-09 Costas Argyris <costas.argyris@gmail.com>
26644 * config.host: add object for x86_64-*-mingw*.
26645 * config/i386/sym-mingw32.cc: dummy file to attach
26647 * config/i386/utf8-mingw32.rc: windres resource file.
26648 * config/i386/winnt-utf8.manifest: XML manifest to
26650 * config/i386/x-mingw32: reference to x-mingw32-utf8.
26651 * config/i386/x-mingw32-utf8: Makefile fragment to
26652 embed UTF-8 manifest.
26654 2023-03-09 Vladimir N. Makarov <vmakarov@redhat.com>
26656 * lra-constraints.cc (process_alt_operands): Use operand modes for
26657 clobbered regs instead of the biggest access mode.
26659 2023-03-09 Richard Biener <rguenther@suse.de>
26661 PR middle-end/108995
26662 * fold-const.cc (extract_muldiv_1): Avoid folding
26663 (CST * b) / CST2 when sanitizing overflow and we rely on
26664 overflow being undefined.
26666 2023-03-09 Jakub Jelinek <jakub@redhat.com>
26667 Richard Biener <rguenther@suse.de>
26669 PR tree-optimization/109008
26670 * range-op-float.cc (float_widen_lhs_range): New function.
26671 (foperator_plus::op1_range, foperator_minus::op1_range,
26672 foperator_minus::op2_range, foperator_mult::op1_range,
26673 foperator_div::op1_range, foperator_div::op2_range): Use it.
26675 2023-03-07 Jonathan Grant <jg@jguk.org>
26678 * doc/invoke.texi (Instrumentation Options): Clarify
26679 LeakSanitizer behavior.
26681 2023-03-07 Benson Muite <benson_muite@emailplus.org>
26683 * doc/install.texi (Prerequisites): Add link to gmplib.org.
26685 2023-03-07 Pan Li <pan2.li@intel.com>
26686 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26690 * config/riscv/riscv-modes.def (ADJUST_PRECISION): Adjust VNx*BI
26692 * config/riscv/riscv.cc (riscv_v_adjust_precision): New.
26693 * config/riscv/riscv.h (riscv_v_adjust_precision): New.
26694 * genmodes.cc (adj_precision): New.
26695 (ADJUST_PRECISION): New.
26696 (emit_mode_adjustments): Handle ADJUST_PRECISION.
26698 2023-03-07 Hans-Peter Nilsson <hp@axis.com>
26700 * doc/sourcebuild.texi: Document check_effective_target_tail_call.
26702 2023-03-06 Paul-Antoine Arras <pa@codesourcery.com>
26704 * config/gcn/gcn-valu.md (<expander><mode>3_exec): Add patterns for
26705 {s|u}{max|min} in QI, HI and DI modes.
26706 (<expander><mode>3): Add pattern for {s|u}{max|min} in DI mode.
26707 (cond_<fexpander><mode>): Add pattern for cond_f{max|min}.
26708 (cond_<expander><mode>): Add pattern for cond_{s|u}{max|min}.
26709 * config/gcn/gcn.cc (gcn_spill_class): Allow the exec register to be
26712 2023-03-06 Richard Biener <rguenther@suse.de>
26714 PR tree-optimization/109025
26715 * tree-vect-loop.cc (vect_is_simple_reduction): Verify
26716 the inner LC PHI use is the inner loop PHI latch definition
26717 before classifying an outer PHI as double reduction.
26719 2023-03-06 Jan Hubicka <hubicka@ucw.cz>
26722 * config/i386/x86-tune.def (X86_TUNE_USE_SCATTER_2PARTS): Enable for
26724 (X86_TUNE_USE_SCATTER_4PARTS): Likewise.
26725 (X86_TUNE_USE_SCATTER): Likewise.
26727 2023-03-06 Xi Ruoyao <xry111@xry111.site>
26730 * config/loongarch/loongarch.h (FP_RETURN): Use
26731 TARGET_*_FLOAT_ABI instead of TARGET_*_FLOAT.
26732 (UNITS_PER_FP_ARG): Likewise.
26734 2023-03-05 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26736 * config/riscv/riscv-vsetvl.cc (reg_available_p): Fix bug.
26737 (pass_vsetvl::backward_demand_fusion): Ditto.
26739 2023-03-05 Liao Shihua <shihua@iscas.ac.cn>
26740 SiYu Wu <siyu@isrc.iscas.ac.cn>
26742 * config/riscv/crypto.md (riscv_sm3p0_<mode>): Add ZKSED's and ZKSH's
26744 (riscv_sm3p1_<mode>): New.
26745 (riscv_sm4ed_<mode>): New.
26746 (riscv_sm4ks_<mode>): New.
26747 * config/riscv/riscv-builtins.cc (AVAIL): Add ZKSED's and ZKSH's AVAIL.
26748 * config/riscv/riscv-scalar-crypto.def (RISCV_BUILTIN): Add ZKSED's and
26749 ZKSH's built-in functions.
26751 2023-03-05 Liao Shihua <shihua@iscas.ac.cn>
26752 SiYu Wu <siyu@isrc.iscas.ac.cn>
26754 * config/riscv/crypto.md (riscv_sha256sig0_<mode>): Add ZKNH's instructions.
26755 (riscv_sha256sig1_<mode>): New.
26756 (riscv_sha256sum0_<mode>): New.
26757 (riscv_sha256sum1_<mode>): New.
26758 (riscv_sha512sig0h): New.
26759 (riscv_sha512sig0l): New.
26760 (riscv_sha512sig1h): New.
26761 (riscv_sha512sig1l): New.
26762 (riscv_sha512sum0r): New.
26763 (riscv_sha512sum1r): New.
26764 (riscv_sha512sig0): New.
26765 (riscv_sha512sig1): New.
26766 (riscv_sha512sum0): New.
26767 (riscv_sha512sum1): New.
26768 * config/riscv/riscv-builtins.cc (AVAIL): And ZKNH's AVAIL.
26769 * config/riscv/riscv-scalar-crypto.def (RISCV_BUILTIN): And ZKNH's
26770 built-in functions.
26771 (DIRECT_BUILTIN): Add new.
26773 2023-03-05 Liao Shihua <shihua@iscas.ac.cn>
26774 SiYu Wu <siyu@isrc.iscas.ac.cn>
26776 * config/riscv/constraints.md (D03): Add constants of bs and rnum.
26778 * config/riscv/crypto.md (riscv_aes32dsi): Add ZKND's and ZKNE's instructions.
26779 (riscv_aes32dsmi): New.
26780 (riscv_aes64ds): New.
26781 (riscv_aes64dsm): New.
26782 (riscv_aes64im): New.
26783 (riscv_aes64ks1i): New.
26784 (riscv_aes64ks2): New.
26785 (riscv_aes32esi): New.
26786 (riscv_aes32esmi): New.
26787 (riscv_aes64es): New.
26788 (riscv_aes64esm): New.
26789 * config/riscv/riscv-builtins.cc (AVAIL): Add ZKND's and ZKNE's AVAIL.
26790 * config/riscv/riscv-scalar-crypto.def (DIRECT_BUILTIN): Add ZKND's and
26791 ZKNE's built-in functions.
26793 2023-03-05 Liao Shihua <shihua@iscas.ac.cn>
26794 SiYu Wu <siyu@isrc.iscas.ac.cn>
26796 * config/riscv/bitmanip.md: Add ZBKB's instructions.
26797 * config/riscv/riscv-builtins.cc (AVAIL): Add new.
26798 * config/riscv/riscv.md: Add new type for crypto instructions.
26799 * config/riscv/crypto.md: Add Scalar Cryptography extension's machine
26801 * config/riscv/riscv-scalar-crypto.def: Add Scalar Cryptography
26802 extension's built-in function file.
26804 2023-03-05 Liao Shihua <shihua@iscas.ac.cn>
26805 SiYu Wu <siyu@isrc.iscas.ac.cn>
26807 * config/riscv/riscv-builtins.cc (RISCV_FTYPE_NAME2): New.
26808 (RISCV_FTYPE_NAME3): New.
26809 (RISCV_ATYPE_QI): New.
26810 (RISCV_ATYPE_HI): New.
26811 (RISCV_FTYPE_ATYPES2): New.
26812 (RISCV_FTYPE_ATYPES3): New.
26813 * config/riscv/riscv-ftypes.def (2): New.
26816 2023-03-05 Vineet Gupta <vineetg@rivosinc.com>
26818 * config/riscv/riscv.cc (riscv_rtx_costs): Fixed IN_RANGE() to
26821 2023-03-05 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
26822 kito-cheng <kito.cheng@sifive.com>
26824 * config/riscv/predicates.md (vector_any_register_operand): New predicate.
26825 * config/riscv/riscv-c.cc (riscv_check_builtin_call): New function.
26826 (riscv_register_pragmas): Add builtin function check call.
26827 * config/riscv/riscv-protos.h (RVV_VUNDEF): Adapt macro.
26828 (check_builtin_call): New function.
26829 * config/riscv/riscv-vector-builtins-bases.cc (class vundefined): New class.
26830 (class vreinterpret): Ditto.
26831 (class vlmul_ext): Ditto.
26832 (class vlmul_trunc): Ditto.
26833 (class vset): Ditto.
26834 (class vget): Ditto.
26836 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
26837 * config/riscv/riscv-vector-builtins-functions.def (vluxei8): Change name.
26853 (vundefined): Add new intrinsic.
26854 (vreinterpret): Ditto.
26855 (vlmul_ext): Ditto.
26856 (vlmul_trunc): Ditto.
26859 * config/riscv/riscv-vector-builtins-shapes.cc (struct return_mask_def): New class.
26860 (struct narrow_alu_def): Ditto.
26861 (struct reduc_alu_def): Ditto.
26862 (struct vundefined_def): Ditto.
26863 (struct misc_def): Ditto.
26864 (struct vset_def): Ditto.
26865 (struct vget_def): Ditto.
26867 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
26868 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_EEW8_INTERPRET_OPS): New def.
26869 (DEF_RVV_EEW16_INTERPRET_OPS): Ditto.
26870 (DEF_RVV_EEW32_INTERPRET_OPS): Ditto.
26871 (DEF_RVV_EEW64_INTERPRET_OPS): Ditto.
26872 (DEF_RVV_X2_VLMUL_EXT_OPS): Ditto.
26873 (DEF_RVV_X4_VLMUL_EXT_OPS): Ditto.
26874 (DEF_RVV_X8_VLMUL_EXT_OPS): Ditto.
26875 (DEF_RVV_X16_VLMUL_EXT_OPS): Ditto.
26876 (DEF_RVV_X32_VLMUL_EXT_OPS): Ditto.
26877 (DEF_RVV_X64_VLMUL_EXT_OPS): Ditto.
26878 (DEF_RVV_LMUL1_OPS): Ditto.
26879 (DEF_RVV_LMUL2_OPS): Ditto.
26880 (DEF_RVV_LMUL4_OPS): Ditto.
26881 (vint16mf4_t): Ditto.
26882 (vint16mf2_t): Ditto.
26883 (vint16m1_t): Ditto.
26884 (vint16m2_t): Ditto.
26885 (vint16m4_t): Ditto.
26886 (vint16m8_t): Ditto.
26887 (vint32mf2_t): Ditto.
26888 (vint32m1_t): Ditto.
26889 (vint32m2_t): Ditto.
26890 (vint32m4_t): Ditto.
26891 (vint32m8_t): Ditto.
26892 (vint64m1_t): Ditto.
26893 (vint64m2_t): Ditto.
26894 (vint64m4_t): Ditto.
26895 (vint64m8_t): Ditto.
26896 (vuint16mf4_t): Ditto.
26897 (vuint16mf2_t): Ditto.
26898 (vuint16m1_t): Ditto.
26899 (vuint16m2_t): Ditto.
26900 (vuint16m4_t): Ditto.
26901 (vuint16m8_t): Ditto.
26902 (vuint32mf2_t): Ditto.
26903 (vuint32m1_t): Ditto.
26904 (vuint32m2_t): Ditto.
26905 (vuint32m4_t): Ditto.
26906 (vuint32m8_t): Ditto.
26907 (vuint64m1_t): Ditto.
26908 (vuint64m2_t): Ditto.
26909 (vuint64m4_t): Ditto.
26910 (vuint64m8_t): Ditto.
26911 (vint8mf4_t): Ditto.
26912 (vint8mf2_t): Ditto.
26913 (vint8m1_t): Ditto.
26914 (vint8m2_t): Ditto.
26915 (vint8m4_t): Ditto.
26916 (vint8m8_t): Ditto.
26917 (vuint8mf4_t): Ditto.
26918 (vuint8mf2_t): Ditto.
26919 (vuint8m1_t): Ditto.
26920 (vuint8m2_t): Ditto.
26921 (vuint8m4_t): Ditto.
26922 (vuint8m8_t): Ditto.
26923 (vint8mf8_t): Ditto.
26924 (vuint8mf8_t): Ditto.
26925 (vfloat32mf2_t): Ditto.
26926 (vfloat32m1_t): Ditto.
26927 (vfloat32m2_t): Ditto.
26928 (vfloat32m4_t): Ditto.
26929 (vfloat64m1_t): Ditto.
26930 (vfloat64m2_t): Ditto.
26931 (vfloat64m4_t): Ditto.
26932 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_TYPE): Ditto.
26933 (DEF_RVV_EEW8_INTERPRET_OPS): Ditto.
26934 (DEF_RVV_EEW16_INTERPRET_OPS): Ditto.
26935 (DEF_RVV_EEW32_INTERPRET_OPS): Ditto.
26936 (DEF_RVV_EEW64_INTERPRET_OPS): Ditto.
26937 (DEF_RVV_X2_VLMUL_EXT_OPS): Ditto.
26938 (DEF_RVV_X4_VLMUL_EXT_OPS): Ditto.
26939 (DEF_RVV_X8_VLMUL_EXT_OPS): Ditto.
26940 (DEF_RVV_X16_VLMUL_EXT_OPS): Ditto.
26941 (DEF_RVV_X32_VLMUL_EXT_OPS): Ditto.
26942 (DEF_RVV_X64_VLMUL_EXT_OPS): Ditto.
26943 (DEF_RVV_LMUL1_OPS): Ditto.
26944 (DEF_RVV_LMUL2_OPS): Ditto.
26945 (DEF_RVV_LMUL4_OPS): Ditto.
26946 (DEF_RVV_TYPE_INDEX): Ditto.
26947 (required_extensions_p): Adapt for new intrinsic support/
26948 (get_required_extensions): New function.
26949 (check_required_extensions): Ditto.
26950 (unsigned_base_type_p): Remove.
26951 (rvv_arg_type_info::get_scalar_ptr_type): New function.
26952 (get_mode_for_bitsize): Remove.
26953 (rvv_arg_type_info::get_scalar_const_ptr_type): New function.
26954 (rvv_arg_type_info::get_base_vector_type): Ditto.
26955 (rvv_arg_type_info::get_function_type_index): Ditto.
26956 (DEF_RVV_BASE_TYPE): New def.
26957 (function_builder::apply_predication): New class.
26958 (function_expander::mask_mode): Ditto.
26959 (function_checker::function_checker): Ditto.
26960 (function_checker::report_non_ice): Ditto.
26961 (function_checker::report_out_of_range): Ditto.
26962 (function_checker::require_immediate): Ditto.
26963 (function_checker::require_immediate_range): Ditto.
26964 (function_checker::check): Ditto.
26965 (check_builtin_call): Ditto.
26966 * config/riscv/riscv-vector-builtins.def (DEF_RVV_TYPE): New def.
26967 (DEF_RVV_BASE_TYPE): Ditto.
26968 (DEF_RVV_TYPE_INDEX): Ditto.
26969 (vbool64_t): Ditto.
26970 (vbool32_t): Ditto.
26971 (vbool16_t): Ditto.
26976 (vuint8mf8_t): Ditto.
26977 (vuint8mf4_t): Ditto.
26978 (vuint8mf2_t): Ditto.
26979 (vuint8m1_t): Ditto.
26980 (vuint8m2_t): Ditto.
26981 (vint8m4_t): Ditto.
26982 (vuint8m4_t): Ditto.
26983 (vint8m8_t): Ditto.
26984 (vuint8m8_t): Ditto.
26985 (vint16mf4_t): Ditto.
26986 (vuint16mf2_t): Ditto.
26987 (vuint16m1_t): Ditto.
26988 (vuint16m2_t): Ditto.
26989 (vuint16m4_t): Ditto.
26990 (vuint16m8_t): Ditto.
26991 (vint32mf2_t): Ditto.
26992 (vuint32m1_t): Ditto.
26993 (vuint32m2_t): Ditto.
26994 (vuint32m4_t): Ditto.
26995 (vuint32m8_t): Ditto.
26996 (vuint64m1_t): Ditto.
26997 (vuint64m2_t): Ditto.
26998 (vuint64m4_t): Ditto.
26999 (vuint64m8_t): Ditto.
27000 (vfloat32mf2_t): Ditto.
27001 (vfloat32m1_t): Ditto.
27002 (vfloat32m2_t): Ditto.
27003 (vfloat32m4_t): Ditto.
27004 (vfloat32m8_t): Ditto.
27005 (vfloat64m1_t): Ditto.
27006 (vfloat64m4_t): Ditto.
27007 (vector): Move it def.
27010 (signed_vector): Ditto.
27011 (unsigned_vector): Ditto.
27012 (unsigned_scalar): Ditto.
27013 (vector_ptr): Ditto.
27014 (scalar_ptr): Ditto.
27015 (scalar_const_ptr): Ditto.
27019 (unsigned_long): Ditto.
27021 (eew8_index): Ditto.
27022 (eew16_index): Ditto.
27023 (eew32_index): Ditto.
27024 (eew64_index): Ditto.
27025 (shift_vector): Ditto.
27026 (double_trunc_vector): Ditto.
27027 (quad_trunc_vector): Ditto.
27028 (oct_trunc_vector): Ditto.
27029 (double_trunc_scalar): Ditto.
27030 (double_trunc_signed_vector): Ditto.
27031 (double_trunc_unsigned_vector): Ditto.
27032 (double_trunc_unsigned_scalar): Ditto.
27033 (double_trunc_float_vector): Ditto.
27034 (float_vector): Ditto.
27035 (lmul1_vector): Ditto.
27036 (widen_lmul1_vector): Ditto.
27037 (eew8_interpret): Ditto.
27038 (eew16_interpret): Ditto.
27039 (eew32_interpret): Ditto.
27040 (eew64_interpret): Ditto.
27041 (vlmul_ext_x2): Ditto.
27042 (vlmul_ext_x4): Ditto.
27043 (vlmul_ext_x8): Ditto.
27044 (vlmul_ext_x16): Ditto.
27045 (vlmul_ext_x32): Ditto.
27046 (vlmul_ext_x64): Ditto.
27047 * config/riscv/riscv-vector-builtins.h (DEF_RVV_BASE_TYPE): New def.
27048 (struct function_type_info): New function.
27049 (struct rvv_arg_type_info): Ditto.
27050 (class function_checker): New class.
27051 (rvv_arg_type_info::get_scalar_type): New function.
27052 (rvv_arg_type_info::get_vector_type): Ditto.
27053 (function_expander::ret_mode): New function.
27054 (function_checker::arg_mode): Ditto.
27055 (function_checker::ret_mode): Ditto.
27056 * config/riscv/t-riscv: Add generator.
27057 * config/riscv/vector-iterators.md: New iterators.
27058 * config/riscv/vector.md (vundefined<mode>): New pattern.
27059 (@vundefined<mode>): Ditto.
27060 (@vreinterpret<mode>): Ditto.
27061 (@vlmul_extx2<mode>): Ditto.
27062 (@vlmul_extx4<mode>): Ditto.
27063 (@vlmul_extx8<mode>): Ditto.
27064 (@vlmul_extx16<mode>): Ditto.
27065 (@vlmul_extx32<mode>): Ditto.
27066 (@vlmul_extx64<mode>): Ditto.
27067 (*vlmul_extx2<mode>): Ditto.
27068 (*vlmul_extx4<mode>): Ditto.
27069 (*vlmul_extx8<mode>): Ditto.
27070 (*vlmul_extx16<mode>): Ditto.
27071 (*vlmul_extx32<mode>): Ditto.
27072 (*vlmul_extx64<mode>): Ditto.
27073 * config/riscv/genrvv-type-indexer.cc: New file.
27075 2023-03-05 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
27077 * config/riscv/riscv-protos.h (enum vlen_enum): New enum.
27078 (slide1_sew64_helper): New function.
27079 * config/riscv/riscv-v.cc (compute_vlmax): Ditto.
27080 (get_unknown_min_value): Ditto.
27081 (force_vector_length_operand): Ditto.
27082 (gen_no_side_effects_vsetvl_rtx): Ditto.
27083 (get_vl_x2_rtx): Ditto.
27084 (slide1_sew64_helper): Ditto.
27085 * config/riscv/riscv-vector-builtins-bases.cc (class slideop): New class.
27086 (class vrgather): Ditto.
27087 (class vrgatherei16): Ditto.
27088 (class vcompress): Ditto.
27090 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
27091 * config/riscv/riscv-vector-builtins-functions.def (vslideup): Ditto.
27092 (vslidedown): Ditto.
27093 (vslide1up): Ditto.
27094 (vslide1down): Ditto.
27095 (vfslide1up): Ditto.
27096 (vfslide1down): Ditto.
27098 (vrgatherei16): Ditto.
27099 (vcompress): Ditto.
27100 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_EI16_OPS): New macro.
27101 (vint8mf8_t): Ditto.
27102 (vint8mf4_t): Ditto.
27103 (vint8mf2_t): Ditto.
27104 (vint8m1_t): Ditto.
27105 (vint8m2_t): Ditto.
27106 (vint8m4_t): Ditto.
27107 (vint16mf4_t): Ditto.
27108 (vint16mf2_t): Ditto.
27109 (vint16m1_t): Ditto.
27110 (vint16m2_t): Ditto.
27111 (vint16m4_t): Ditto.
27112 (vint16m8_t): Ditto.
27113 (vint32mf2_t): Ditto.
27114 (vint32m1_t): Ditto.
27115 (vint32m2_t): Ditto.
27116 (vint32m4_t): Ditto.
27117 (vint32m8_t): Ditto.
27118 (vint64m1_t): Ditto.
27119 (vint64m2_t): Ditto.
27120 (vint64m4_t): Ditto.
27121 (vint64m8_t): Ditto.
27122 (vuint8mf8_t): Ditto.
27123 (vuint8mf4_t): Ditto.
27124 (vuint8mf2_t): Ditto.
27125 (vuint8m1_t): Ditto.
27126 (vuint8m2_t): Ditto.
27127 (vuint8m4_t): Ditto.
27128 (vuint16mf4_t): Ditto.
27129 (vuint16mf2_t): Ditto.
27130 (vuint16m1_t): Ditto.
27131 (vuint16m2_t): Ditto.
27132 (vuint16m4_t): Ditto.
27133 (vuint16m8_t): Ditto.
27134 (vuint32mf2_t): Ditto.
27135 (vuint32m1_t): Ditto.
27136 (vuint32m2_t): Ditto.
27137 (vuint32m4_t): Ditto.
27138 (vuint32m8_t): Ditto.
27139 (vuint64m1_t): Ditto.
27140 (vuint64m2_t): Ditto.
27141 (vuint64m4_t): Ditto.
27142 (vuint64m8_t): Ditto.
27143 (vfloat32mf2_t): Ditto.
27144 (vfloat32m1_t): Ditto.
27145 (vfloat32m2_t): Ditto.
27146 (vfloat32m4_t): Ditto.
27147 (vfloat32m8_t): Ditto.
27148 (vfloat64m1_t): Ditto.
27149 (vfloat64m2_t): Ditto.
27150 (vfloat64m4_t): Ditto.
27151 (vfloat64m8_t): Ditto.
27152 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_EI16_OPS): Ditto.
27153 * config/riscv/riscv.md: Adjust RVV instruction types.
27154 * config/riscv/vector-iterators.md (down): New iterator.
27155 (=vd,vr): New attribute.
27156 (UNSPEC_VSLIDE1UP): New unspec.
27157 * config/riscv/vector.md (@pred_slide<ud><mode>): New pattern.
27158 (*pred_slide<ud><mode>): Ditto.
27159 (*pred_slide<ud><mode>_extended): Ditto.
27160 (@pred_gather<mode>): Ditto.
27161 (@pred_gather<mode>_scalar): Ditto.
27162 (@pred_gatherei16<mode>): Ditto.
27163 (@pred_compress<mode>): Ditto.
27165 2023-03-05 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
27167 * config/riscv/riscv-vector-builtins.cc: Remove void_type_node.
27169 2023-03-05 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
27171 * config/riscv/constraints.md (Wb1): New constraint.
27172 * config/riscv/predicates.md
27173 (vector_least_significant_set_mask_operand): New predicate.
27174 (vector_broadcast_mask_operand): Ditto.
27175 * config/riscv/riscv-protos.h (enum vlmul_type): Adjust.
27176 (gen_scalar_move_mask): New function.
27177 * config/riscv/riscv-v.cc (gen_scalar_move_mask): Ditto.
27178 * config/riscv/riscv-vector-builtins-bases.cc (class vmv): New class.
27179 (class vmv_s): Ditto.
27181 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
27182 * config/riscv/riscv-vector-builtins-functions.def (vmv_x): Ditto.
27186 * config/riscv/riscv-vector-builtins-shapes.cc (struct scalar_move_def): Ditto.
27188 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
27189 * config/riscv/riscv-vector-builtins.cc (function_expander::mask_mode): Ditto.
27190 (function_expander::use_exact_insn): New function.
27191 (function_expander::use_contiguous_load_insn): New function.
27192 (function_expander::use_contiguous_store_insn): New function.
27193 (function_expander::use_ternop_insn): New function.
27194 (function_expander::use_widen_ternop_insn): New function.
27195 (function_expander::use_scalar_move_insn): New function.
27196 * config/riscv/riscv-vector-builtins.def (s): New operand suffix.
27197 * config/riscv/riscv-vector-builtins.h
27198 (function_expander::add_scalar_move_mask_operand): New class.
27199 * config/riscv/riscv-vsetvl.cc (ignore_vlmul_insn_p): New function.
27200 (scalar_move_insn_p): Ditto.
27201 (has_vsetvl_killed_avl_p): Ditto.
27202 (anticipatable_occurrence_p): Ditto.
27203 (insert_vsetvl): Ditto.
27204 (get_vl_vtype_info): Ditto.
27205 (calculate_sew): Ditto.
27206 (calculate_vlmul): Ditto.
27207 (incompatible_avl_p): Ditto.
27208 (different_sew_p): Ditto.
27209 (different_lmul_p): Ditto.
27210 (different_ratio_p): Ditto.
27211 (different_tail_policy_p): Ditto.
27212 (different_mask_policy_p): Ditto.
27213 (possible_zero_avl_p): Ditto.
27214 (first_ratio_invalid_for_second_sew_p): Ditto.
27215 (first_ratio_invalid_for_second_lmul_p): Ditto.
27216 (second_ratio_invalid_for_first_sew_p): Ditto.
27217 (second_ratio_invalid_for_first_lmul_p): Ditto.
27218 (second_sew_less_than_first_sew_p): Ditto.
27219 (first_sew_less_than_second_sew_p): Ditto.
27220 (compare_lmul): Ditto.
27221 (second_lmul_less_than_first_lmul_p): Ditto.
27222 (first_lmul_less_than_second_lmul_p): Ditto.
27223 (first_ratio_less_than_second_ratio_p): Ditto.
27224 (second_ratio_less_than_first_ratio_p): Ditto.
27225 (DEF_INCOMPATIBLE_COND): Ditto.
27226 (greatest_sew): Ditto.
27227 (first_sew): Ditto.
27228 (second_sew): Ditto.
27229 (first_vlmul): Ditto.
27230 (second_vlmul): Ditto.
27231 (first_ratio): Ditto.
27232 (second_ratio): Ditto.
27233 (vlmul_for_first_sew_second_ratio): Ditto.
27234 (ratio_for_second_sew_first_vlmul): Ditto.
27235 (DEF_SEW_LMUL_FUSE_RULE): Ditto.
27236 (always_unavailable): Ditto.
27237 (avl_unavailable_p): Ditto.
27238 (sew_unavailable_p): Ditto.
27239 (lmul_unavailable_p): Ditto.
27240 (ge_sew_unavailable_p): Ditto.
27241 (ge_sew_lmul_unavailable_p): Ditto.
27242 (ge_sew_ratio_unavailable_p): Ditto.
27243 (DEF_UNAVAILABLE_COND): Ditto.
27244 (same_sew_lmul_demand_p): Ditto.
27245 (propagate_avl_across_demands_p): Ditto.
27246 (reg_available_p): Ditto.
27247 (avl_info::has_non_zero_avl): Ditto.
27248 (vl_vtype_info::has_non_zero_avl): Ditto.
27249 (vector_insn_info::operator>=): Refactor.
27250 (vector_insn_info::parse_insn): Adjust for scalar move.
27251 (vector_insn_info::demand_vl_vtype): Remove.
27252 (vector_insn_info::compatible_p): New function.
27253 (vector_insn_info::compatible_avl_p): Ditto.
27254 (vector_insn_info::compatible_vtype_p): Ditto.
27255 (vector_insn_info::available_p): Ditto.
27256 (vector_insn_info::merge): Ditto.
27257 (vector_insn_info::fuse_avl): Ditto.
27258 (vector_insn_info::fuse_sew_lmul): Ditto.
27259 (vector_insn_info::fuse_tail_policy): Ditto.
27260 (vector_insn_info::fuse_mask_policy): Ditto.
27261 (vector_insn_info::dump): Ditto.
27262 (vector_infos_manager::release): Ditto.
27263 (pass_vsetvl::compute_local_backward_infos): Adjust for scalar move support.
27264 (pass_vsetvl::get_backward_fusion_type): Adjust for scalar move support.
27265 (pass_vsetvl::hard_empty_block_p): Ditto.
27266 (pass_vsetvl::backward_demand_fusion): Ditto.
27267 (pass_vsetvl::forward_demand_fusion): Ditto.
27268 (pass_vsetvl::refine_vsetvls): Ditto.
27269 (pass_vsetvl::cleanup_vsetvls): Ditto.
27270 (pass_vsetvl::commit_vsetvls): Ditto.
27271 (pass_vsetvl::propagate_avl): Ditto.
27272 * config/riscv/riscv-vsetvl.h (enum demand_status): New class.
27273 (struct demands_pair): Ditto.
27274 (struct demands_cond): Ditto.
27275 (struct demands_fuse_rule): Ditto.
27276 * config/riscv/vector-iterators.md: New iterator.
27277 * config/riscv/vector.md (@pred_broadcast<mode>): New pattern.
27278 (*pred_broadcast<mode>): Ditto.
27279 (*pred_broadcast<mode>_extended_scalar): Ditto.
27280 (@pred_extract_first<mode>): Ditto.
27281 (*pred_extract_first<mode>): Ditto.
27282 (@pred_extract_first_trunc<mode>): Ditto.
27283 * config/riscv/riscv-vsetvl.def: New file.
27285 2023-03-05 Lin Sinan <sinan.lin@linux.alibaba.com>
27287 * config/riscv/bitmanip.md: allow 0 constant in max/min
27290 2023-03-05 Lin Sinan <sinan.lin@linux.alibaba.com>
27292 * config/riscv/bitmanip.md: Fix wrong index in the check.
27294 2023-03-04 Jakub Jelinek <jakub@redhat.com>
27296 PR middle-end/109006
27297 * vec.cc (test_auto_alias): Adjust comment for removal of
27299 * read-rtl-function.cc (function_reader::parse_block): Likewise.
27300 * gdbhooks.py: Likewise.
27302 2023-03-04 Jakub Jelinek <jakub@redhat.com>
27304 PR testsuite/108973
27305 * selftest-diagnostic.cc
27306 (test_diagnostic_context::test_diagnostic_context): Set
27307 caret_max_width to 80.
27309 2023-03-03 Alexandre Oliva <oliva@adacore.com>
27311 * gimple-ssa-warn-access.cc
27312 (pass_waccess::check_dangling_stores): Skip non-stores.
27314 2023-03-03 Alexandre Oliva <oliva@adacore.com>
27316 * config/arm/vfp.md (*thumb2_movsi_vfp): Drop blank after tab
27317 after vmsr and vmrs, and lower the case of P0.
27319 2023-03-03 Jonathan Wakely <jwakely@redhat.com>
27321 PR middle-end/109006
27322 * gdbhooks.py (VecPrinter): Handle vec<T> as well as vec<T>*.
27324 2023-03-03 Jonathan Wakely <jwakely@redhat.com>
27326 PR middle-end/109006
27327 * gdbhooks.py (VecPrinter): Adjust for new vec layout.
27329 2023-03-03 Jakub Jelinek <jakub@redhat.com>
27332 * gimple-ssa-warn-access.cc (pass_waccess::maybe_check_access_sizes):
27333 Return immediately if OPT_Wnonnull or OPT_Wstringop_overflow_ is
27334 suppressed on stmt. For [static %E] warning, print access_nelts
27335 rather than access_size. Fix up comment wording.
27337 2023-03-03 Robin Dapp <rdapp@linux.ibm.com>
27339 * config/s390/driver-native.cc (s390_host_detect_local_cpu): Use
27340 arch14 instead of z16.
27342 2023-03-03 Anthony Green <green@moxielogic.com>
27344 * config/moxie/moxie.cc (TARGET_LRA_P): Remove.
27346 2023-03-03 Anthony Green <green@moxielogic.com>
27348 * config/moxie/constraints.md (A, B, W): Change
27349 define_constraint to define_memory_constraint.
27351 2023-03-03 Xi Ruoyao <xry111@xry111.site>
27353 * toplev.cc (process_options): Fix the spelling of
27354 "-fstack-clash-protection".
27356 2023-03-03 Richard Biener <rguenther@suse.de>
27358 PR tree-optimization/109002
27359 * tree-ssa-pre.cc (compute_partial_antic_aux): Properly
27360 PHI-translate ANTIC_IN.
27362 2023-03-03 Jakub Jelinek <jakub@redhat.com>
27364 PR tree-optimization/108988
27365 * gimple-fold.cc (gimple_fold_builtin_fputs): Fold len to
27366 size_type_node before passing it as argument to fwrite. Formatting
27369 2023-03-03 Richard Biener <rguenther@suse.de>
27372 * config/i386/i386.opt (--param x86-stv-max-visits): New param.
27373 * doc/invoke.texi (--param x86-stv-max-visits): Document it.
27374 * config/i386/i386-features.h (scalar_chain::max_visits): New.
27375 (scalar_chain::build): Add bitmap parameter, return boolean.
27376 (scalar_chain::add_insn): Likewise.
27377 (scalar_chain::analyze_register_chain): Likewise.
27378 * config/i386/i386-features.cc (scalar_chain::scalar_chain):
27379 Initialize max_visits.
27380 (scalar_chain::analyze_register_chain): When we exhaust
27381 max_visits, abort. Also abort when running into any
27383 (scalar_chain::add_insn): Propagate abort.
27384 (scalar_chain::build): Likewise. When aborting amend
27385 the set of disallowed insn with the insns set.
27386 (convert_scalars_to_vector): Adjust. Do not convert aborted
27389 2023-03-03 Richard Biener <rguenther@suse.de>
27392 * dwarf2out.cc (dwarf2out_late_global_decl): Do not
27393 generate a DIE for a function scope static.
27395 2023-03-03 Alexandre Oliva <oliva@adacore.com>
27397 * config/vx-common.h (WINT_TYPE): Alias to "wchar_t".
27399 2023-03-02 Jakub Jelinek <jakub@redhat.com>
27402 * target.h (emit_support_tinfos_callback): New typedef.
27403 * targhooks.h (default_emit_support_tinfos): Declare.
27404 * targhooks.cc (default_emit_support_tinfos): New function.
27405 * target.def (emit_support_tinfos): New target hook.
27406 * doc/tm.texi.in (emit_support_tinfos): Document it.
27407 * doc/tm.texi: Regenerated.
27408 * config/i386/i386.cc (ix86_emit_support_tinfos): New function.
27409 (TARGET_EMIT_SUPPORT_TINFOS): Redefine.
27411 2023-03-02 Vladimir N. Makarov <vmakarov@redhat.com>
27413 * ira-costs.cc: Include print-rtl.h.
27414 (record_reg_classes, scan_one_insn): Add code to print debug info.
27415 (record_operand_costs): Find and use smaller cost for hard reg
27418 2023-03-02 Kwok Cheung Yeung <kcy@codesourcery.com>
27419 Paul-Antoine Arras <pa@codesourcery.com>
27421 * builtins.cc (mathfn_built_in_explicit): New.
27422 * config/gcn/gcn.cc: Include case-cfn-macros.h.
27423 (mathfn_built_in_explicit): Add prototype.
27424 (gcn_vectorize_builtin_vectorized_function): New.
27425 (gcn_libc_has_function): New.
27426 (TARGET_LIBC_HAS_FUNCTION): Define.
27427 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Define.
27429 2023-03-02 Richard Sandiford <richard.sandiford@arm.com>
27431 PR tree-optimization/108979
27432 * tree-vect-stmts.cc (vectorizable_operation): Don't mask
27433 operations on invariants.
27435 2023-03-02 Robin Dapp <rdapp@linux.ibm.com>
27437 * config/s390/predicates.md (vll_bias_operand): Add -1 bias.
27438 * config/s390/s390.cc (s390_option_override_internal): Make
27439 partial vector usage the default from z13 on.
27440 * config/s390/vector.md (len_load_v16qi): Add.
27441 (len_store_v16qi): Add.
27443 2023-03-02 Andre Vieira <andre.simoesdiasvieira@arm.com>
27445 * simplify-rtx.cc (simplify_context::simplify_subreg): Use byte instead
27446 of constant 0 offset.
27448 2023-03-02 Robert Suchanek <robert.suchanek@imgtec.com>
27450 * config/mips/mips.cc (mips_set_text_contents_type): Use HOST_WIDE_INT
27452 * config/mips/mips-protos.h (mips_set_text_contents_type): Likewise.
27454 2023-03-02 Junxian Zhu <zhujunxian@oss.cipunited.com>
27456 * config.gcc: add -with-{no-}msa build option.
27457 * config/mips/mips.h: Likewise.
27458 * doc/install.texi: Likewise.
27460 2023-03-02 Richard Sandiford <richard.sandiford@arm.com>
27462 PR tree-optimization/108603
27463 * explow.cc (convert_memory_address_addr_space_1): Only wrap
27464 the result of a recursive call in a CONST if no instructions
27467 2023-03-02 Richard Sandiford <richard.sandiford@arm.com>
27469 PR tree-optimization/108430
27470 * tree-vect-stmts.cc (vectorizable_condition): Fix handling
27471 of inverted condition.
27473 2023-03-02 Jakub Jelinek <jakub@redhat.com>
27476 * fold-const.cc (native_interpret_expr) <case REAL_CST>: Before memcmp
27477 comparison copy the bytes from ptr to a temporary buffer and clearing
27478 padding bits in there.
27480 2023-03-01 Tobias Burnus <tobias@codesourcery.com>
27482 PR middle-end/108545
27483 * gimplify.cc (struct tree_operand_hash_no_se): New.
27484 (omp_index_mapping_groups_1, omp_index_mapping_groups,
27485 omp_reindex_mapping_groups, omp_mapped_by_containing_struct,
27486 omp_tsort_mapping_groups_1, omp_tsort_mapping_groups,
27487 oacc_resolve_clause_dependencies, omp_build_struct_sibling_lists,
27488 gimplify_scan_omp_clauses): Use tree_operand_hash_no_se instead
27489 of tree_operand_hash.
27491 2023-03-01 LIU Hao <lh_mouse@126.com>
27494 * config/i386/host-mingw32.cc (mingw32_gt_pch_get_address):
27495 Remove the size limit `pch_VA_max_size`
27497 2023-03-01 Tobias Burnus <tobias@codesourcery.com>
27499 PR middle-end/108546
27500 * omp-low.cc (lower_omp_target): Remove optional handling
27501 on the receiver side, i.e. inside target (data), for
27504 2023-03-01 Jakub Jelinek <jakub@redhat.com>
27507 * cfgexpand.cc (expand_debug_expr): Handle WIDEN_{PLUS,MINUS}_EXPR
27508 and VEC_WIDEN_{PLUS,MINUS}_{HI,LO}_EXPR.
27510 2023-03-01 Richard Biener <rguenther@suse.de>
27512 PR tree-optimization/108970
27513 * tree-vect-loop-manip.cc (slpeel_can_duplicate_loop_p):
27514 Check we can copy the BBs.
27515 (slpeel_tree_duplicate_loop_to_edge_cfg): Avoid redundant
27517 (vect_do_peeling): Streamline error handling.
27519 2023-03-01 Richard Biener <rguenther@suse.de>
27521 PR tree-optimization/108950
27522 * tree-vect-patterns.cc (vect_recog_widen_sum_pattern):
27523 Check oprnd0 is defined in the loop.
27524 * tree-vect-loop.cc (vectorizable_reduction): Record all
27525 operands vector types, compute that of invariants and
27526 properly update their SLP nodes.
27528 2023-03-01 Kewen Lin <linkw@linux.ibm.com>
27531 * config/rs6000/rs6000.cc (rs6000_option_override_internal): Allow
27532 implicit powerpc64 setting to be unset if 64 bit is enabled implicitly.
27534 2023-02-28 Qing Zhao <qing.zhao@oracle.com>
27536 PR middle-end/107411
27537 PR middle-end/107411
27538 * gimplify.cc (gimple_add_init_for_auto_var): Use sprintf to replace
27540 * tree-ssa-uninit.cc (warn_uninit): Handle the case when the
27541 LHS varaible of a .DEFERRED_INIT call doesn't have a DECL_NAME.
27543 2023-02-28 Jakub Jelinek <jakub@redhat.com>
27545 PR sanitizer/108894
27546 * ubsan.cc (ubsan_expand_bounds_ifn): Emit index >= bound
27547 comparison rather than index > bound.
27548 * gimple-fold.cc (gimple_fold_call): Use tree_int_cst_lt
27549 rather than tree_int_cst_le for IFN_UBSAN_BOUND comparison.
27550 * doc/invoke.texi (-fsanitize=bounds): Document that whether
27551 flexible array member-like arrays are instrumented or not depends
27552 on -fstrict-flex-arrays* options of strict_flex_array attributes.
27553 (-fsanitize=bounds-strict): Document that flexible array members
27554 are not instrumented.
27556 2023-02-27 Uroš Bizjak <ubizjak@gmail.com>
27560 * config/i386/i386.md (fmodxf3): Enable for flag_finite_math_only only.
27561 (fmod<mode>3): Ditto.
27562 (fpremxf4_i387): Ditto.
27563 (reminderxf3): Ditto.
27564 (reminder<mode>3): Ditto.
27565 (fprem1xf4_i387): Ditto.
27567 2023-02-27 Roger Sayle <roger@nextmovesoftware.com>
27569 * simplify-rtx.cc (simplify_unary_operation_1) <case FFS>: Avoid
27570 generating FFS with mismatched operand and result modes, by using
27571 an explicit SIGN_EXTEND/ZERO_EXTEND.
27572 <case POPCOUNT>: Likewise, for POPCOUNT of ZERO_EXTEND.
27573 <case PARITY>: Likewise, for PARITY of {ZERO,SIGN}_EXTEND.
27575 2023-02-27 Patrick Palka <ppalka@redhat.com>
27577 * hash-table.h (gt_pch_nx(hash_table<D>)): Remove static.
27578 * lra-int.h (lra_change_class): Likewise.
27579 * recog.h (which_op_alt): Likewise.
27580 * sel-sched-ir.h (sel_bb_empty_or_nop_p): Declare inline
27583 2023-02-27 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
27585 * config/xtensa/xtensa-protos.h (xtensa_match_CLAMPS_imms_p):
27587 * config/xtensa/xtensa.cc (xtensa_match_CLAMPS_imms_p):
27589 * config/xtensa/xtensa.h (TARGET_CLAMPS): New macro definition.
27590 * config/xtensa/xtensa.md (*xtensa_clamps): New insn pattern.
27592 2023-02-27 Max Filippov <jcmvbkbc@gmail.com>
27594 * config/xtensa/xtensa-dynconfig.cc (xtensa_get_config_v2)
27595 (xtensa_get_config_v3): New functions.
27597 2023-02-27 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27599 * config/aarch64/aarch64-simd.md (aarch64_abs<mode>): Fix typo in comment.
27601 2023-02-27 Lulu Cheng <chenglulu@loongson.cn>
27603 * config/host-linux.cc (TRY_EMPTY_VM_SPACE): Modify the value of
27604 the macro to 0x1000000000.
27606 2023-02-25 Gaius Mulley <gaiusmod2@gmail.com>
27609 * doc/gm2.texi (-fm2-pathname): New option documented.
27610 (-fm2-pathnameI): New option documented.
27611 (-fm2-prefix=): New option documented.
27612 (-fruntime-modules=): Update default module list.
27614 2023-02-25 Max Filippov <jcmvbkbc@gmail.com>
27617 * config/xtensa/xtensa-protos.h
27618 (xtensa_prepare_expand_call): Rename to xtensa_expand_call.
27619 * config/xtensa/xtensa.cc (xtensa_prepare_expand_call): Rename
27620 to xtensa_expand_call.
27621 (xtensa_expand_call): Emit the call and add a clobber expression
27622 for the static chain to it in case of windowed ABI.
27623 * config/xtensa/xtensa.md (call, call_value, sibcall)
27624 (sibcall_value): Call xtensa_expand_call and complete expansion
27625 right after that call.
27627 2023-02-24 Richard Biener <rguenther@suse.de>
27629 * vec.h (vec<T, A, vl_embed>::m_vecdata): Remove.
27630 (vec<T, A, vl_embed>::m_vecpfx): Align as T to avoid
27631 changing alignment of vec<T, A, vl_embed> and simplifying
27633 (vec<T, A, vl_embed>::address): Compute as this + 1.
27634 (vec<T, A, vl_embed>::embedded_size): Use sizeof the
27635 vector instead of the offset of the m_vecdata member.
27636 (auto_vec<T, N>::m_data): Turn storage into
27637 uninitialized unsigned char.
27638 (auto_vec<T, N>::auto_vec): Allow allocation of one
27639 stack member. Initialize m_vec in a special way to
27640 avoid later stringop overflow diagnostics.
27641 * vec.cc (test_auto_alias): New.
27642 (vec_cc_tests): Call it.
27644 2023-02-24 Richard Biener <rguenther@suse.de>
27646 * vec.h (vec<T, A, vl_embed>::lower_bound): Adjust to
27647 take a const reference to the object, use address to
27649 (vec<T, A, vl_embed>::contains): Use address to access data.
27650 (vec<T, A, vl_embed>::operator[]): Use address instead of
27651 m_vecdata to access data.
27652 (vec<T, A, vl_embed>::iterate): Likewise.
27653 (vec<T, A, vl_embed>::copy): Likewise.
27654 (vec<T, A, vl_embed>::quick_push): Likewise.
27655 (vec<T, A, vl_embed>::pop): Likewise.
27656 (vec<T, A, vl_embed>::quick_insert): Likewise.
27657 (vec<T, A, vl_embed>::ordered_remove): Likewise.
27658 (vec<T, A, vl_embed>::unordered_remove): Likewise.
27659 (vec<T, A, vl_embed>::block_remove): Likewise.
27660 (vec<T, A, vl_heap>::address): Likewise.
27662 2023-02-24 Martin Liska <mliska@suse.cz>
27664 PR sanitizer/108834
27665 * asan.cc (asan_add_global): Use proper TU name for normal
27666 global variables (and aux_base_name for the artificial one).
27668 2023-02-24 Jakub Jelinek <jakub@redhat.com>
27670 * config/i386/i386-builtin.def: Update description of BDESC
27671 and BDESC_FIRST in file comment to include mask2.
27673 2023-02-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27675 * config/aarch64/aarch64-cores.def (FLAGS): Update comment.
27677 2023-02-24 Jakub Jelinek <jakub@redhat.com>
27679 PR middle-end/108854
27680 * cgraphclones.cc (duplicate_thunk_for_node): If no parameter
27681 changes are needed, copy at least DECL_ARGUMENTS PARM_DECL
27682 nodes and adjust their DECL_CONTEXT.
27684 2023-02-24 Jakub Jelinek <jakub@redhat.com>
27687 * config/i386/i386-builtin.def (__builtin_ia32_cvtne2ps2bf16_v16bf,
27688 __builtin_ia32_cvtne2ps2bf16_v16bf_mask,
27689 __builtin_ia32_cvtne2ps2bf16_v16bf_maskz,
27690 __builtin_ia32_cvtne2ps2bf16_v8bf,
27691 __builtin_ia32_cvtne2ps2bf16_v8bf_mask,
27692 __builtin_ia32_cvtne2ps2bf16_v8bf_maskz,
27693 __builtin_ia32_cvtneps2bf16_v8sf_mask,
27694 __builtin_ia32_cvtneps2bf16_v8sf_maskz,
27695 __builtin_ia32_cvtneps2bf16_v4sf_mask,
27696 __builtin_ia32_cvtneps2bf16_v4sf_maskz,
27697 __builtin_ia32_dpbf16ps_v8sf, __builtin_ia32_dpbf16ps_v8sf_mask,
27698 __builtin_ia32_dpbf16ps_v8sf_maskz, __builtin_ia32_dpbf16ps_v4sf,
27699 __builtin_ia32_dpbf16ps_v4sf_mask,
27700 __builtin_ia32_dpbf16ps_v4sf_maskz): Require also
27701 OPTION_MASK_ISA_AVX512VL.
27703 2023-02-24 Sebastian Huber <sebastian.huber@embedded-brains.de>
27705 * config/riscv/t-rtems: Keep only -mcmodel=medany 64-bit multilibs.
27706 Add non-compact 32-bit multilibs.
27708 2023-02-24 Junxian Zhu <zhujunxian@oss.cipunited.com>
27710 * config/mips/mips.md (*clo<mode>2): New pattern.
27712 2023-02-24 Prachi Godbole <prachi.godbole@imgtec.com>
27714 * config/mips/mips.h (machine_function): New variable
27715 use_hazard_barrier_return_p.
27716 * config/mips/mips.md (UNSPEC_JRHB): New unspec.
27717 (mips_hb_return_internal): New insn pattern.
27718 * config/mips/mips.cc (mips_attribute_table): Add attribute
27719 use_hazard_barrier_return.
27720 (mips_use_hazard_barrier_return_p): New static function.
27721 (mips_function_attr_inlinable_p): Likewise.
27722 (mips_compute_frame_info): Set use_hazard_barrier_return_p.
27723 Emit error for unsupported architecture choice.
27724 (mips_function_ok_for_sibcall, mips_can_use_return_insn):
27725 Return false for use_hazard_barrier_return.
27726 (mips_expand_epilogue): Emit hazard barrier return.
27727 * doc/extend.texi: Document use_hazard_barrier_return.
27729 2023-02-23 Max Filippov <jcmvbkbc@gmail.com>
27731 * config/xtensa/xtensa-dynconfig.cc (config.h, system.h)
27732 (coretypes.h, diagnostic.h, intl.h): Use "..." instead of <...>
27733 for the gcc-internal headers.
27735 2023-02-23 Max Filippov <jcmvbkbc@gmail.com>
27737 * config/xtensa/t-xtensa (xtensa-dynconfig.o): Use $(COMPILE)
27738 and $(POSTCOMPILE) instead of manual dependency listing.
27739 * config/xtensa/xtensa-dynconfig.c: Rename to ...
27740 * config/xtensa/xtensa-dynconfig.cc: ... this.
27742 2023-02-23 Arsen Arsenović <arsen@aarsen.me>
27744 * doc/cfg.texi: Reorder index entries around @items.
27745 * doc/cpp.texi: Ditto.
27746 * doc/cppenv.texi: Ditto.
27747 * doc/cppopts.texi: Ditto.
27748 * doc/generic.texi: Ditto.
27749 * doc/install.texi: Ditto.
27750 * doc/extend.texi: Ditto.
27751 * doc/invoke.texi: Ditto.
27752 * doc/md.texi: Ditto.
27753 * doc/rtl.texi: Ditto.
27754 * doc/tm.texi.in: Ditto.
27755 * doc/trouble.texi: Ditto.
27756 * doc/tm.texi: Regenerate.
27758 2023-02-23 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
27760 * config/xtensa/xtensa.md: New peephole2 pattern that eliminates
27761 the occurrence of general-purpose register used only once and for
27762 transferring intermediate value.
27764 2023-02-23 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
27766 * config/xtensa/xtensa.cc (machine_function): Add new member
27767 'eliminated_callee_saved_bmp'.
27768 (xtensa_can_eliminate_callee_saved_reg_p): New function to
27769 determine whether the register can be eliminated or not.
27770 (xtensa_expand_prologue): Add invoking the above function and
27771 elimination the use of callee-saved register by using its stack
27772 slot through the stack pointer (or the frame pointer if needed)
27774 (xtensa_expand_prologue): Modify to not emit register restoration
27775 insn from its stack slot if the register is already eliminated.
27777 2023-02-23 Jakub Jelinek <jakub@redhat.com>
27779 PR translation/108890
27780 * config/xtensa/xtensa-dynconfig.c (xtensa_load_config): Drop _()s
27781 around fatal_error format strings.
27783 2023-02-23 Richard Biener <rguenther@suse.de>
27785 * tree-ssa-structalias.cc (handle_lhs_call): Do not
27786 re-create rhsc, only truncate it.
27788 2023-02-23 Jakub Jelinek <jakub@redhat.com>
27790 PR middle-end/106258
27791 * ipa-prop.cc (try_make_edge_direct_virtual_call): Handle
27792 BUILT_IN_UNREACHABLE_TRAP like BUILT_IN_UNREACHABLE.
27794 2023-02-23 Richard Biener <rguenther@suse.de>
27796 * tree-if-conv.cc (tree_if_conversion): Properly manage
27797 memory of refs and the contained data references.
27799 2023-02-23 Richard Biener <rguenther@suse.de>
27801 PR tree-optimization/108888
27802 * tree-if-conv.cc (if_convertible_stmt_p): Set PLF_2 on
27803 calls to predicate.
27804 (predicate_statements): Only predicate calls with PLF_2.
27806 2023-02-23 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
27808 * config/xtensa/xtensa.md
27809 (zero_cost_loop_start, zero_cost_loop_end, loop_end):
27810 Add missing "SI:" to PLUS RTXes.
27812 2023-02-23 Max Filippov <jcmvbkbc@gmail.com>
27815 * config/xtensa/xtensa.cc (xtensa_expand_epilogue):
27816 Emit (use (reg:SI A0_REG)) at the end in the sibling call
27817 (i.e. the same place as (return) in the normal call).
27819 2023-02-23 Max Filippov <jcmvbkbc@gmail.com>
27822 2023-02-21 Max Filippov <jcmvbkbc@gmail.com>
27825 * config/xtensa/xtensa.cc (xtensa_expand_epilogue): Drop emit_use
27827 * config/xtensa/xtensa.md (sibcall, sibcall_internal)
27828 (sibcall_value, sibcall_value_internal): Add 'use' expression
27831 2023-02-23 Arsen Arsenović <arsen@aarsen.me>
27833 * doc/cppdiropts.texi: Reorder @opindex commands to precede
27834 @items they relate to.
27835 * doc/cppopts.texi: Ditto.
27836 * doc/cppwarnopts.texi: Ditto.
27837 * doc/invoke.texi: Ditto.
27838 * doc/lto.texi: Ditto.
27840 2023-02-22 Andrew Stubbs <ams@codesourcery.com>
27842 * internal-fn.cc (expand_MASK_CALL): New.
27843 * internal-fn.def (MASK_CALL): New.
27844 * internal-fn.h (expand_MASK_CALL): New prototype.
27845 * omp-simd-clone.cc (simd_clone_adjust_argument_types): Set vector_type
27846 for mask arguments also.
27847 * tree-if-conv.cc: Include cgraph.h.
27848 (if_convertible_stmt_p): Do if conversions for calls to SIMD calls.
27849 (predicate_statements): Convert functions to IFN_MASK_CALL.
27850 * tree-vect-loop.cc (vect_get_datarefs_in_loop): Recognise
27851 IFN_MASK_CALL as a SIMD function call.
27852 * tree-vect-stmts.cc (vectorizable_simd_clone_call): Handle
27853 IFN_MASK_CALL as an inbranch SIMD function call.
27854 Generate the mask vector arguments.
27856 2023-02-22 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
27858 * config/riscv/riscv-vector-builtins-bases.cc (class reducop): New class.
27859 (class widen_reducop): Ditto.
27860 (class freducop): Ditto.
27861 (class widen_freducop): Ditto.
27863 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
27864 * config/riscv/riscv-vector-builtins-functions.def (vredsum): Add reduction support.
27873 (vwredsumu): Ditto.
27874 (vfredusum): Ditto.
27875 (vfredosum): Ditto.
27878 (vfwredosum): Ditto.
27879 (vfwredusum): Ditto.
27880 * config/riscv/riscv-vector-builtins-shapes.cc (struct reduc_alu_def): Ditto.
27882 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
27883 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_WI_OPS): New macro.
27884 (DEF_RVV_WU_OPS): Ditto.
27885 (DEF_RVV_WF_OPS): Ditto.
27886 (vint8mf8_t): Ditto.
27887 (vint8mf4_t): Ditto.
27888 (vint8mf2_t): Ditto.
27889 (vint8m1_t): Ditto.
27890 (vint8m2_t): Ditto.
27891 (vint8m4_t): Ditto.
27892 (vint8m8_t): Ditto.
27893 (vint16mf4_t): Ditto.
27894 (vint16mf2_t): Ditto.
27895 (vint16m1_t): Ditto.
27896 (vint16m2_t): Ditto.
27897 (vint16m4_t): Ditto.
27898 (vint16m8_t): Ditto.
27899 (vint32mf2_t): Ditto.
27900 (vint32m1_t): Ditto.
27901 (vint32m2_t): Ditto.
27902 (vint32m4_t): Ditto.
27903 (vint32m8_t): Ditto.
27904 (vuint8mf8_t): Ditto.
27905 (vuint8mf4_t): Ditto.
27906 (vuint8mf2_t): Ditto.
27907 (vuint8m1_t): Ditto.
27908 (vuint8m2_t): Ditto.
27909 (vuint8m4_t): Ditto.
27910 (vuint8m8_t): Ditto.
27911 (vuint16mf4_t): Ditto.
27912 (vuint16mf2_t): Ditto.
27913 (vuint16m1_t): Ditto.
27914 (vuint16m2_t): Ditto.
27915 (vuint16m4_t): Ditto.
27916 (vuint16m8_t): Ditto.
27917 (vuint32mf2_t): Ditto.
27918 (vuint32m1_t): Ditto.
27919 (vuint32m2_t): Ditto.
27920 (vuint32m4_t): Ditto.
27921 (vuint32m8_t): Ditto.
27922 (vfloat32mf2_t): Ditto.
27923 (vfloat32m1_t): Ditto.
27924 (vfloat32m2_t): Ditto.
27925 (vfloat32m4_t): Ditto.
27926 (vfloat32m8_t): Ditto.
27927 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_WI_OPS): Ditto.
27928 (DEF_RVV_WU_OPS): Ditto.
27929 (DEF_RVV_WF_OPS): Ditto.
27930 (required_extensions_p): Add reduction support.
27931 (rvv_arg_type_info::get_base_vector_type): Ditto.
27932 (rvv_arg_type_info::get_tree_type): Ditto.
27933 * config/riscv/riscv-vector-builtins.h (enum rvv_base_type): Ditto.
27934 * config/riscv/riscv.md: Ditto.
27935 * config/riscv/vector-iterators.md (minu): Ditto.
27936 * config/riscv/vector.md (@pred_reduc_<reduc><mode><vlmul1>): New patern.
27937 (@pred_reduc_<reduc><mode><vlmul1_zve32>): Ditto.
27938 (@pred_widen_reduc_plus<v_su><mode><vwlmul1>): Ditto.
27939 (@pred_widen_reduc_plus<v_su><mode><vwlmul1_zve32>):Ditto.
27940 (@pred_reduc_plus<order><mode><vlmul1>): Ditto.
27941 (@pred_reduc_plus<order><mode><vlmul1_zve32>): Ditto.
27942 (@pred_widen_reduc_plus<order><mode><vwlmul1>): Ditto.
27944 2023-02-22 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
27946 * config/riscv/iterators.md: New iterator.
27947 * config/riscv/riscv-vector-builtins-bases.cc (class widen_binop): New class.
27948 (enum ternop_type): New enum.
27949 (class vmacc): New class.
27950 (class imac): Ditto.
27951 (class vnmsac): Ditto.
27952 (enum widen_ternop_type): New enum.
27953 (class vmadd): Ditto.
27954 (class vnmsub): Ditto.
27955 (class iwmac): Ditto.
27956 (class vwmacc): Ditto.
27957 (class vwmaccu): Ditto.
27958 (class vwmaccsu): Ditto.
27959 (class vwmaccus): Ditto.
27960 (class reverse_binop): Ditto.
27961 (class vfmacc): Ditto.
27962 (class vfnmsac): Ditto.
27963 (class vfmadd): Ditto.
27964 (class vfnmsub): Ditto.
27965 (class vfnmacc): Ditto.
27966 (class vfmsac): Ditto.
27967 (class vfnmadd): Ditto.
27968 (class vfmsub): Ditto.
27969 (class vfwmacc): Ditto.
27970 (class vfwnmacc): Ditto.
27971 (class vfwmsac): Ditto.
27972 (class vfwnmsac): Ditto.
27973 (class float_misc): Ditto.
27974 (class fcmp): Ditto.
27975 (class vfclass): Ditto.
27976 (class vfcvt_x): Ditto.
27977 (class vfcvt_rtz_x): Ditto.
27978 (class vfcvt_f): Ditto.
27979 (class vfwcvt_x): Ditto.
27980 (class vfwcvt_rtz_x): Ditto.
27981 (class vfwcvt_f): Ditto.
27982 (class vfncvt_x): Ditto.
27983 (class vfncvt_rtz_x): Ditto.
27984 (class vfncvt_f): Ditto.
27985 (class vfncvt_rod_f): Ditto.
27987 * config/riscv/riscv-vector-builtins-bases.h:
27988 * config/riscv/riscv-vector-builtins-functions.def (vzext): Ditto.
28032 (vfcvt_rtz_x): Ditto.
28033 (vfcvt_rtz_xu): Ditto.
28036 (vfwcvt_xu): Ditto.
28037 (vfwcvt_rtz_x): Ditto.
28038 (vfwcvt_rtz_xu): Ditto.
28041 (vfncvt_xu): Ditto.
28042 (vfncvt_rtz_x): Ditto.
28043 (vfncvt_rtz_xu): Ditto.
28045 (vfncvt_rod_f): Ditto.
28046 * config/riscv/riscv-vector-builtins-shapes.cc (struct alu_def): Ditto.
28047 (struct move_def): Ditto.
28048 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_WEXTF_OPS): New macro.
28049 (DEF_RVV_CONVERT_I_OPS): Ditto.
28050 (DEF_RVV_CONVERT_U_OPS): Ditto.
28051 (DEF_RVV_WCONVERT_I_OPS): Ditto.
28052 (DEF_RVV_WCONVERT_U_OPS): Ditto.
28053 (DEF_RVV_WCONVERT_F_OPS): Ditto.
28054 (vfloat64m1_t): Ditto.
28055 (vfloat64m2_t): Ditto.
28056 (vfloat64m4_t): Ditto.
28057 (vfloat64m8_t): Ditto.
28058 (vint32mf2_t): Ditto.
28059 (vint32m1_t): Ditto.
28060 (vint32m2_t): Ditto.
28061 (vint32m4_t): Ditto.
28062 (vint32m8_t): Ditto.
28063 (vint64m1_t): Ditto.
28064 (vint64m2_t): Ditto.
28065 (vint64m4_t): Ditto.
28066 (vint64m8_t): Ditto.
28067 (vuint32mf2_t): Ditto.
28068 (vuint32m1_t): Ditto.
28069 (vuint32m2_t): Ditto.
28070 (vuint32m4_t): Ditto.
28071 (vuint32m8_t): Ditto.
28072 (vuint64m1_t): Ditto.
28073 (vuint64m2_t): Ditto.
28074 (vuint64m4_t): Ditto.
28075 (vuint64m8_t): Ditto.
28076 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_CONVERT_I_OPS): Ditto.
28077 (DEF_RVV_CONVERT_U_OPS): Ditto.
28078 (DEF_RVV_WCONVERT_I_OPS): Ditto.
28079 (DEF_RVV_WCONVERT_U_OPS): Ditto.
28080 (DEF_RVV_WCONVERT_F_OPS): Ditto.
28081 (DEF_RVV_F_OPS): Ditto.
28082 (DEF_RVV_WEXTF_OPS): Ditto.
28083 (required_extensions_p): Adjust for floating-point support.
28084 (check_required_extensions): Ditto.
28085 (unsigned_base_type_p): Ditto.
28086 (get_mode_for_bitsize): Ditto.
28087 (rvv_arg_type_info::get_base_vector_type): Ditto.
28088 (rvv_arg_type_info::get_tree_type): Ditto.
28089 * config/riscv/riscv-vector-builtins.def (v_f): New define.
28092 (xu_v): New define.
28094 (xu_w): New define.
28095 * config/riscv/riscv-vector-builtins.h (enum rvv_base_type): New enum.
28096 (function_expander::arg_mode): New function.
28097 * config/riscv/vector-iterators.md (sof): New iterator.
28103 (fixuns_trunc): Ditto.
28105 * config/riscv/vector.md (@pred_broadcast<mode>): New pattern.
28106 (@pred_<optab><mode>): Ditto.
28107 (@pred_<optab><mode>_scalar): Ditto.
28108 (@pred_<optab><mode>_reverse_scalar): Ditto.
28109 (@pred_<copysign><mode>): Ditto.
28110 (@pred_<copysign><mode>_scalar): Ditto.
28111 (@pred_mul_<optab><mode>): Ditto.
28112 (pred_mul_<optab><mode>_undef_merge): Ditto.
28113 (*pred_<madd_nmsub><mode>): Ditto.
28114 (*pred_<macc_nmsac><mode>): Ditto.
28115 (*pred_mul_<optab><mode>): Ditto.
28116 (@pred_mul_<optab><mode>_scalar): Ditto.
28117 (*pred_mul_<optab><mode>_undef_merge_scalar): Ditto.
28118 (*pred_<madd_nmsub><mode>_scalar): Ditto.
28119 (*pred_<macc_nmsac><mode>_scalar): Ditto.
28120 (*pred_mul_<optab><mode>_scalar): Ditto.
28121 (@pred_neg_mul_<optab><mode>): Ditto.
28122 (pred_neg_mul_<optab><mode>_undef_merge): Ditto.
28123 (*pred_<nmadd_msub><mode>): Ditto.
28124 (*pred_<nmacc_msac><mode>): Ditto.
28125 (*pred_neg_mul_<optab><mode>): Ditto.
28126 (@pred_neg_mul_<optab><mode>_scalar): Ditto.
28127 (*pred_neg_mul_<optab><mode>_undef_merge_scalar): Ditto.
28128 (*pred_<nmadd_msub><mode>_scalar): Ditto.
28129 (*pred_<nmacc_msac><mode>_scalar): Ditto.
28130 (*pred_neg_mul_<optab><mode>_scalar): Ditto.
28131 (@pred_<misc_op><mode>): Ditto.
28132 (@pred_class<mode>): Ditto.
28133 (@pred_dual_widen_<optab><mode>): Ditto.
28134 (@pred_dual_widen_<optab><mode>_scalar): Ditto.
28135 (@pred_single_widen_<plus_minus:optab><mode>): Ditto.
28136 (@pred_single_widen_<plus_minus:optab><mode>_scalar): Ditto.
28137 (@pred_widen_mul_<optab><mode>): Ditto.
28138 (@pred_widen_mul_<optab><mode>_scalar): Ditto.
28139 (@pred_widen_neg_mul_<optab><mode>): Ditto.
28140 (@pred_widen_neg_mul_<optab><mode>_scalar): Ditto.
28141 (@pred_cmp<mode>): Ditto.
28142 (*pred_cmp<mode>): Ditto.
28143 (*pred_cmp<mode>_narrow): Ditto.
28144 (@pred_cmp<mode>_scalar): Ditto.
28145 (*pred_cmp<mode>_scalar): Ditto.
28146 (*pred_cmp<mode>_scalar_narrow): Ditto.
28147 (@pred_eqne<mode>_scalar): Ditto.
28148 (*pred_eqne<mode>_scalar): Ditto.
28149 (*pred_eqne<mode>_scalar_narrow): Ditto.
28150 (@pred_merge<mode>_scalar): Ditto.
28151 (@pred_fcvt_x<v_su>_f<mode>): Ditto.
28152 (@pred_<fix_cvt><mode>): Ditto.
28153 (@pred_<float_cvt><mode>): Ditto.
28154 (@pred_widen_fcvt_x<v_su>_f<mode>): Ditto.
28155 (@pred_widen_<fix_cvt><mode>): Ditto.
28156 (@pred_widen_<float_cvt><mode>): Ditto.
28157 (@pred_extend<mode>): Ditto.
28158 (@pred_narrow_fcvt_x<v_su>_f<mode>): Ditto.
28159 (@pred_narrow_<fix_cvt><mode>): Ditto.
28160 (@pred_narrow_<float_cvt><mode>): Ditto.
28161 (@pred_trunc<mode>): Ditto.
28162 (@pred_rod_trunc<mode>): Ditto.
28164 2023-02-22 Jakub Jelinek <jakub@redhat.com>
28166 PR middle-end/106258
28167 * cgraph.cc (cgraph_edge::redirect_call_stmt_to_callee,
28168 cgraph_update_edges_for_call_stmt_node, cgraph_node::verify_node):
28169 Handle BUILT_IN_UNREACHABLE_TRAP like BUILT_IN_UNREACHABLE.
28170 * cgraphclones.cc (cgraph_node::create_clone): Likewise.
28172 2023-02-22 Thomas Schwinge <thomas@codesourcery.com>
28174 * common.opt (-Wcomplain-wrong-lang): New.
28175 * doc/invoke.texi (-Wno-complain-wrong-lang): Document it.
28176 * opts-common.cc (prune_options): Handle it.
28177 * opts-global.cc (complain_wrong_lang): Use it.
28179 2023-02-21 David Malcolm <dmalcolm@redhat.com>
28182 * doc/invoke.texi: Document -fno-analyzer-suppress-followups.
28184 2023-02-21 Max Filippov <jcmvbkbc@gmail.com>
28187 * config/xtensa/xtensa.cc (xtensa_expand_epilogue): Drop emit_use
28189 * config/xtensa/xtensa.md (sibcall, sibcall_internal)
28190 (sibcall_value, sibcall_value_internal): Add 'use' expression
28193 2023-02-21 Richard Biener <rguenther@suse.de>
28195 PR tree-optimization/108691
28196 * tree-ssa-dce.cc (eliminate_unnecessary_stmts): Remove
28197 assert about calls_setjmp not becoming true when it was false.
28199 2023-02-21 Richard Biener <rguenther@suse.de>
28201 PR tree-optimization/108793
28202 * tree-ssa-loop-niter.cc (number_of_iterations_until_wrap):
28203 Use convert operands to niter_type when computing num.
28205 2023-02-21 Richard Biener <rguenther@suse.de>
28208 2023-02-13 Richard Biener <rguenther@suse.de>
28210 PR tree-optimization/108691
28211 * tree-cfg.cc (notice_special_calls): When the CFG is built
28212 honor gimple_call_ctrl_altering_p.
28213 * cfgexpand.cc (expand_call_stmt): Clear cfun->calls_setjmp
28214 temporarily if the call is not control-altering.
28215 * calls.cc (emit_call_1): Do not add REG_SETJMP if
28216 cfun->calls_setjmp is not set. Do not alter cfun->calls_setjmp.
28218 2023-02-21 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
28220 * config/xtensa/xtensa.cc (xtensa_call_save_reg): Change to return
28221 true if register A0 (return address register) when -Og is specified.
28223 2023-02-20 Uroš Bizjak <ubizjak@gmail.com>
28225 * config/i386/predicates.md
28226 (general_x64constmem_operand): New predicate.
28227 * config/i386/i386.md (*cmpqi_ext<mode>_1):
28228 Use nonimm_x64constmem_operand.
28229 (*cmpqi_ext<mode>_3): Use general_x64constmem_operand.
28230 (*addqi_ext<mode>_1): Ditto.
28231 (*testqi_ext<mode>_1): Ditto.
28232 (*andqi_ext<mode>_1): Ditto.
28233 (*andqi_ext<mode>_1_cc): Ditto.
28234 (*<any_or:code>qi_ext<mode>_1): Ditto.
28235 (*xorqi_ext<mode>_1_cc): Ditto.
28237 2023-02-20 Jakub Jelinek <jakub2redhat.com>
28240 * config/rs6000/rs6000.md (umaddditi4): Swap gen_maddlddi4 with
28241 gen_umadddi4_highpart{,_le}.
28243 2023-02-20 Kito Cheng <kito.cheng@sifive.com>
28245 * config/riscv/riscv.md (prefetch): Use r instead of p for the
28247 (riscv_prefetchi_<mode>): Ditto.
28249 2023-02-20 Richard Biener <rguenther@suse.de>
28251 PR tree-optimization/108816
28252 * tree-vect-loop-manip.cc (vect_loop_versioning): Adjust
28253 versioning condition split prerequesite, assert required
28256 2023-02-20 Richard Biener <rguenther@suse.de>
28258 PR tree-optimization/108825
28259 * tree-ssa-loop-manip.cc (verify_loop_closed_ssa): For
28260 loop-local verfication only verify there's no pending SSA
28263 2023-02-20 Richard Biener <rguenther@suse.de>
28265 PR tree-optimization/108819
28266 * tree-ssa-loop-niter.cc (number_of_iterations_cltz): Check
28267 we have an SSA name as iv_2 as expected.
28269 2023-02-18 Jakub Jelinek <jakub@redhat.com>
28271 PR tree-optimization/108819
28272 * tree-ssa-reassoc.cc (update_ops): Fold new stmt in place.
28274 2023-02-18 Jakub Jelinek <jakub@redhat.com>
28277 * config/i386/i386-protos.h (ix86_replace_reg_with_reg): Declare.
28278 * config/i386/i386-expand.cc (ix86_replace_reg_with_reg): New
28280 * config/i386/i386.md: Replace replace_rtx calls in all peephole2s
28281 with ix86_replace_reg_with_reg.
28283 2023-02-18 Gerald Pfeifer <gerald@pfeifer.com>
28285 * doc/invoke.texi (AVR Options): Update link to AVR-LibC.
28287 2023-02-18 Xi Ruoyao <xry111@xry111.site>
28289 * config.gcc (triplet_abi): Set its value based on $with_abi,
28290 instead of $target.
28291 (la_canonical_triplet): Set it after $triplet_abi is set
28293 * config/loongarch/t-linux (MULTILIB_OSDIRNAMES): Make the
28294 multiarch tuple for lp64d "loongarch64-linux-gnu" (without
28297 2023-02-18 Andrew Pinski <apinski@marvell.com>
28299 * match.pd: Remove #if GIMPLE around the
28302 2023-02-18 Andrew Pinski <apinski@marvell.com>
28304 * value-query.h (get_range_query): Return the global ranges
28305 for a nullptr func.
28307 2023-02-17 Siddhesh Poyarekar <siddhesh@gotplt.org>
28309 * doc/invoke.texi (@item -Wall): Fix typo in
28312 2023-02-17 Uroš Bizjak <ubizjak@gmail.com>
28315 * config/i386/predicates.md
28316 (nonimm_x64constmem_operand): New predicate.
28317 * config/i386/i386.md (*addqi_ext<mode>_0): New insn pattern.
28318 (*subqi_ext<mode>_0): Ditto.
28319 (*andqi_ext<mode>_0): Ditto.
28320 (*<any_or:code>qi_ext<mode>_0): Ditto.
28322 2023-02-17 Uroš Bizjak <ubizjak@gmail.com>
28325 * simplify-rtx.cc (simplify_context::simplify_subreg): Use
28326 int_outermode instead of GET_MODE (tem) to prevent
28327 VOIDmode from entering simplify_gen_subreg.
28329 2023-02-17 Richard Biener <rguenther@suse.de>
28331 PR tree-optimization/108821
28332 * tree-ssa-loop-im.cc (sm_seq_valid_bb): We can also not
28333 move volatile accesses.
28335 2023-02-17 Richard Biener <rguenther@suse.de>
28337 * tree-ssa.cc (ssa_undefined_value_p): Assert we are not
28338 called on virtual operands.
28339 * tree-ssa-sccvn.cc (vn_phi_lookup): Guard
28340 ssa_undefined_value_p calls.
28341 (vn_phi_insert): Likewise.
28342 (set_ssa_val_to): Likewise.
28343 (visit_phi): Avoid extra work with equivalences for
28344 virtual operand PHIs.
28346 2023-02-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
28348 * config/riscv/riscv-vector-builtins-bases.cc (class mask_logic): New
28350 (class mask_nlogic): Ditto.
28351 (class mask_notlogic): Ditto.
28352 (class vmmv): Ditto.
28353 (class vmclr): Ditto.
28354 (class vmset): Ditto.
28355 (class vmnot): Ditto.
28356 (class vcpop): Ditto.
28357 (class vfirst): Ditto.
28358 (class mask_misc): Ditto.
28359 (class viota): Ditto.
28360 (class vid): Ditto.
28362 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
28363 * config/riscv/riscv-vector-builtins-functions.def (vmand): Ditto.
28382 * config/riscv/riscv-vector-builtins-shapes.cc (struct alu_def): Ditto.
28383 (struct mask_alu_def): Ditto.
28385 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
28386 * config/riscv/riscv-vector-builtins.cc: Ditto.
28387 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::cleanup_insns): Fix bug
28388 for dest it scalar RVV intrinsics.
28389 * config/riscv/vector-iterators.md (sof): New iterator.
28390 * config/riscv/vector.md (@pred_<optab>n<mode>): New pattern.
28391 (@pred_<optab>not<mode>): New pattern.
28392 (@pred_popcount<VB:mode><P:mode>): New pattern.
28393 (@pred_ffs<VB:mode><P:mode>): New pattern.
28394 (@pred_<misc_op><mode>): New pattern.
28395 (@pred_iota<mode>): New pattern.
28396 (@pred_series<mode>): New pattern.
28398 2023-02-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
28400 * config/riscv/riscv-vector-builtins-functions.def (vadc): Rename.
28404 * config/riscv/riscv-vector-builtins.cc: Ditto.
28406 2023-02-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
28407 kito-cheng <kito.cheng@sifive.com>
28409 * config/riscv/riscv-protos.h (sew64_scalar_helper): New function.
28410 * config/riscv/riscv-v.cc (has_vi_variant_p): Adjust.
28411 (sew64_scalar_helper): New function.
28412 * config/riscv/vector.md: Normalization.
28414 2023-02-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
28416 * config/riscv/riscv-vector-builtins-functions.def (vsetvlmax): Rearrange.
28478 2023-02-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
28480 * config/riscv/vector.md (@pred_<optab><mode>): Rearrange.
28481 (@pred_<optab><mode>_scalar): Ditto.
28482 (*pred_<optab><mode>_scalar): Ditto.
28483 (*pred_<optab><mode>_extended_scalar): Ditto.
28485 2023-02-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
28487 * config/riscv/riscv-protos.h (riscv_run_selftests): Remove 'extern'.
28488 (init_builtins): Ditto.
28489 (mangle_builtin_type): Ditto.
28490 (verify_type_context): Ditto.
28491 (handle_pragma_vector): Ditto.
28492 (builtin_decl): Ditto.
28493 (expand_builtin): Ditto.
28494 (const_vec_all_same_in_range_p): Ditto.
28495 (legitimize_move): Ditto.
28496 (emit_vlmax_op): Ditto.
28497 (emit_nonvlmax_op): Ditto.
28498 (get_vlmul): Ditto.
28499 (get_ratio): Ditto.
28502 (get_avl_type): Ditto.
28503 (calculate_ratio): Ditto.
28504 (enum vlmul_type): Ditto.
28506 (neg_simm5_p): Ditto.
28507 (has_vi_variant_p): Ditto.
28509 2023-02-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
28511 * config/riscv/riscv-protos.h (simm32_p): Remove.
28512 * config/riscv/riscv-v.cc (simm32_p): Ditto.
28513 * config/riscv/vector.md: Use immediate_operand
28514 instead of riscv_vector::simm32_p.
28516 2023-02-16 Gerald Pfeifer <gerald@pfeifer.com>
28518 * doc/invoke.texi (Optimize Options): Reword the explanation
28519 getting minimal, maximal and default values of a parameter.
28521 2023-02-16 Patrick Palka <ppalka@redhat.com>
28523 * addresses.h: Mechanically drop 'static' from 'static inline'
28524 functions via s/^static inline/inline/g.
28525 * asan.h: Likewise.
28526 * attribs.h: Likewise.
28527 * basic-block.h: Likewise.
28528 * bitmap.h: Likewise.
28529 * cfghooks.h: Likewise.
28530 * cfgloop.h: Likewise.
28531 * cgraph.h: Likewise.
28532 * cselib.h: Likewise.
28533 * data-streamer.h: Likewise.
28534 * debug.h: Likewise.
28536 * diagnostic.h: Likewise.
28537 * dominance.h: Likewise.
28538 * dumpfile.h: Likewise.
28539 * emit-rtl.h: Likewise.
28540 * except.h: Likewise.
28541 * expmed.h: Likewise.
28542 * expr.h: Likewise.
28543 * fixed-value.h: Likewise.
28544 * gengtype.h: Likewise.
28545 * gimple-expr.h: Likewise.
28546 * gimple-iterator.h: Likewise.
28547 * gimple-predict.h: Likewise.
28548 * gimple-range-fold.h: Likewise.
28549 * gimple-ssa.h: Likewise.
28550 * gimple.h: Likewise.
28551 * graphite.h: Likewise.
28552 * hard-reg-set.h: Likewise.
28553 * hash-map.h: Likewise.
28554 * hash-set.h: Likewise.
28555 * hash-table.h: Likewise.
28556 * hwint.h: Likewise.
28557 * input.h: Likewise.
28558 * insn-addr.h: Likewise.
28559 * internal-fn.h: Likewise.
28560 * ipa-fnsummary.h: Likewise.
28561 * ipa-icf-gimple.h: Likewise.
28562 * ipa-inline.h: Likewise.
28563 * ipa-modref.h: Likewise.
28564 * ipa-prop.h: Likewise.
28565 * ira-int.h: Likewise.
28567 * lra-int.h: Likewise.
28569 * lto-streamer.h: Likewise.
28570 * memmodel.h: Likewise.
28571 * omp-general.h: Likewise.
28572 * optabs-query.h: Likewise.
28573 * optabs.h: Likewise.
28574 * plugin.h: Likewise.
28575 * pretty-print.h: Likewise.
28576 * range.h: Likewise.
28577 * read-md.h: Likewise.
28578 * recog.h: Likewise.
28579 * regs.h: Likewise.
28580 * rtl-iter.h: Likewise.
28582 * sbitmap.h: Likewise.
28583 * sched-int.h: Likewise.
28584 * sel-sched-ir.h: Likewise.
28585 * sese.h: Likewise.
28586 * sparseset.h: Likewise.
28587 * ssa-iterators.h: Likewise.
28588 * system.h: Likewise.
28589 * target-globals.h: Likewise.
28590 * target.h: Likewise.
28591 * timevar.h: Likewise.
28592 * tree-chrec.h: Likewise.
28593 * tree-data-ref.h: Likewise.
28594 * tree-iterator.h: Likewise.
28595 * tree-outof-ssa.h: Likewise.
28596 * tree-phinodes.h: Likewise.
28597 * tree-scalar-evolution.h: Likewise.
28598 * tree-sra.h: Likewise.
28599 * tree-ssa-alias.h: Likewise.
28600 * tree-ssa-live.h: Likewise.
28601 * tree-ssa-loop-manip.h: Likewise.
28602 * tree-ssa-loop.h: Likewise.
28603 * tree-ssa-operands.h: Likewise.
28604 * tree-ssa-propagate.h: Likewise.
28605 * tree-ssa-sccvn.h: Likewise.
28606 * tree-ssa.h: Likewise.
28607 * tree-ssanames.h: Likewise.
28608 * tree-streamer.h: Likewise.
28609 * tree-switch-conversion.h: Likewise.
28610 * tree-vectorizer.h: Likewise.
28611 * tree.h: Likewise.
28612 * wide-int.h: Likewise.
28614 2023-02-16 Jakub Jelinek <jakub@redhat.com>
28616 PR tree-optimization/108657
28617 * tree-ssa-dse.cc (initialize_ao_ref_for_dse): If lhs of stmt
28618 exists and is not a SSA_NAME, call ao_ref_init even if the stmt
28619 is a call to internal or builtin function.
28621 2023-02-16 Jonathan Wakely <jwakely@redhat.com>
28623 * doc/invoke.texi (C++ Dialect Options): Suggest adding a
28624 using-declaration to unhide functions.
28626 2023-02-16 Jakub Jelinek <jakub@redhat.com>
28628 PR tree-optimization/108783
28629 * tree-ssa-reassoc.cc (eliminate_redundant_comparison): If lcode
28630 is equal to TREE_CODE (t), op1 to newop1 and op2 to newop2, set
28631 t to curr->op. Otherwise, punt if either newop1 or newop2 are
28632 SSA_NAME_OCCURS_IN_ABNORMAL_PHI SSA_NAMEs.
28634 2023-02-16 Richard Biener <rguenther@suse.de>
28636 PR tree-optimization/108791
28637 * tree-ssa-forwprop.cc (optimize_vector_load): Build
28638 the ADDR_EXPR of a TARGET_MEM_REF using a more meaningful
28641 2023-02-15 Eric Botcazou <ebotcazou@adacore.com>
28644 * config/i386/i386.cc (ix86_compute_frame_layout): Disable the
28645 effects of -fstack-clash-protection for TARGET_STACK_PROBE.
28646 (ix86_expand_prologue): Likewise.
28648 2023-02-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
28650 * config/bpf/bpf.cc (bpf_option_override): Fix doubled space.
28652 2023-02-15 Uroš Bizjak <ubizjak@gmail.com>
28654 * config/i386/i386.md (*cmpqi_ext<mode>_1): Use
28655 int248_register_operand predicate in zero_extract sub-RTX.
28656 (*cmpqi_ext<mode>_2): Ditto.
28657 (*cmpqi_ext<mode>_3): Ditto.
28658 (*cmpqi_ext<mode>_4): Ditto.
28659 (*extzvqi_mem_rex64): Ditto.
28661 (*insvqi_1_mem_rex64): Ditto.
28662 (@insv<mode>_1): Ditto.
28663 (*insvqi_1): Ditto.
28664 (*insvqi_2): Ditto.
28665 (*insvqi_3): Ditto.
28666 (*extendqi<SWI24:mode>_ext_1): Ditto.
28667 (*addqi_ext<mode>_1): Ditto.
28668 (*addqi_ext<mode>_2): Ditto.
28669 (*subqi_ext<mode>_2): Ditto.
28670 (*testqi_ext<mode>_1): Ditto.
28671 (*testqi_ext<mode>_2): Ditto.
28672 (*andqi_ext<mode>_1): Ditto.
28673 (*andqi_ext<mode>_1_cc): Ditto.
28674 (*andqi_ext<mode>_2): Ditto.
28675 (*<any_or:code>qi_ext<mode>_1): Ditto.
28676 (*<any_or:code>qi_ext<mode>_2): Ditto.
28677 (*xorqi_ext<mode>_1_cc): Ditto.
28678 (*negqi_ext<mode>_2): Ditto.
28679 (*ashlqi_ext<mode>_2): Ditto.
28680 (*<any_shiftrt:insn>qi_ext<mode>_2): Ditto.
28682 2023-02-15 Uroš Bizjak <ubizjak@gmail.com>
28684 * config/i386/predicates.md (int248_register_operand):
28685 Rename from extr_register_operand.
28686 * config/i386/i386.md (*extv<mode>): Update for renamed predicate.
28687 (*extzx<mode>): Ditto.
28688 (*ashl<dwi>3_doubleword_mask): Use int248_register_operand predicate.
28689 (*ashl<mode>3_mask): Ditto.
28690 (*<any_shiftrt:insn><mode>3_mask): Ditto.
28691 (*<any_shiftrt:insn><dwi>3_doubleword_mask): Ditto.
28692 (*<any_rotate:insn><mode>3_mask): Ditto.
28693 (*<btsc><mode>_mask): Ditto.
28694 (*btr<mode>_mask): Ditto.
28695 (*jcc_bt<mode>_mask_1): Ditto.
28697 2023-02-15 Richard Biener <rguenther@suse.de>
28699 PR middle-end/26854
28700 * df-core.cc (df_worklist_propagate_forward): Put later
28701 blocks on worklist and only earlier blocks on pending.
28702 (df_worklist_propagate_backward): Likewise.
28703 (df_worklist_dataflow_doublequeue): Change the iteration
28704 to process new blocks in the same iteration if that
28705 maintains the iteration order.
28707 2023-02-15 Marek Polacek <polacek@redhat.com>
28709 PR middle-end/106080
28710 * gimple-ssa-warn-access.cc (is_auto_decl): Remove. Use auto_var_p
28713 2023-02-15 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
28715 * config/riscv/predicates.md: Refine codes.
28716 * config/riscv/riscv-protos.h (RVV_VUNDEF): New macro.
28717 * config/riscv/riscv-v.cc: Refine codes.
28718 * config/riscv/riscv-vector-builtins-bases.cc (enum ternop_type): New
28720 (class imac): New class.
28721 (enum widen_ternop_type): New enum.
28722 (class iwmac): New class.
28724 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
28725 * config/riscv/riscv-vector-builtins-functions.def (vmacc): Ditto.
28733 * config/riscv/riscv-vector-builtins.cc
28734 (function_builder::apply_predication): Adjust for multiply-add support.
28735 (function_expander::add_vundef_operand): Refine codes.
28736 (function_expander::use_ternop_insn): New function.
28737 (function_expander::use_widen_ternop_insn): Ditto.
28738 * config/riscv/riscv-vector-builtins.h: New function.
28739 * config/riscv/vector.md (@pred_mul_<optab><mode>): New pattern.
28740 (pred_mul_<optab><mode>_undef_merge): Ditto.
28741 (*pred_<madd_nmsub><mode>): Ditto.
28742 (*pred_<macc_nmsac><mode>): Ditto.
28743 (*pred_mul_<optab><mode>): Ditto.
28744 (@pred_mul_<optab><mode>_scalar): Ditto.
28745 (*pred_mul_<optab><mode>_undef_merge_scalar): Ditto.
28746 (*pred_<madd_nmsub><mode>_scalar): Ditto.
28747 (*pred_<macc_nmsac><mode>_scalar): Ditto.
28748 (*pred_mul_<optab><mode>_scalar): Ditto.
28749 (*pred_mul_<optab><mode>_undef_merge_extended_scalar): Ditto.
28750 (*pred_<madd_nmsub><mode>_extended_scalar): Ditto.
28751 (*pred_<macc_nmsac><mode>_extended_scalar): Ditto.
28752 (*pred_mul_<optab><mode>_extended_scalar): Ditto.
28753 (@pred_widen_mul_plus<su><mode>): Ditto.
28754 (@pred_widen_mul_plus<su><mode>_scalar): Ditto.
28755 (@pred_widen_mul_plussu<mode>): Ditto.
28756 (@pred_widen_mul_plussu<mode>_scalar): Ditto.
28757 (@pred_widen_mul_plusus<mode>_scalar): Ditto.
28759 2023-02-15 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
28761 * config/riscv/predicates.md (vector_mask_operand): Refine the codes.
28762 (vector_all_trues_mask_operand): New predicate.
28763 (vector_undef_operand): New predicate.
28764 (ltge_operator): New predicate.
28765 (comparison_except_ltge_operator): New predicate.
28766 (comparison_except_eqge_operator): New predicate.
28767 (ge_operator): New predicate.
28768 * config/riscv/riscv-v.cc (has_vi_variant_p): Add compare support.
28769 * config/riscv/riscv-vector-builtins-bases.cc (class icmp): New class.
28771 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
28772 * config/riscv/riscv-vector-builtins-functions.def (vmseq): Ditto.
28782 * config/riscv/riscv-vector-builtins-shapes.cc
28783 (struct return_mask_def): Adjust for compare support.
28784 * config/riscv/riscv-vector-builtins.cc
28785 (function_expander::use_compare_insn): New function.
28786 * config/riscv/riscv-vector-builtins.h
28787 (function_expander::add_integer_operand): Ditto.
28788 * config/riscv/riscv.cc (riscv_print_operand): Add compare support.
28789 * config/riscv/riscv.md: Add vector min/max attributes.
28790 * config/riscv/vector-iterators.md (xnor): New iterator.
28791 * config/riscv/vector.md (@pred_cmp<mode>): New pattern.
28792 (*pred_cmp<mode>): Ditto.
28793 (*pred_cmp<mode>_narrow): Ditto.
28794 (@pred_ltge<mode>): Ditto.
28795 (*pred_ltge<mode>): Ditto.
28796 (*pred_ltge<mode>_narrow): Ditto.
28797 (@pred_cmp<mode>_scalar): Ditto.
28798 (*pred_cmp<mode>_scalar): Ditto.
28799 (*pred_cmp<mode>_scalar_narrow): Ditto.
28800 (@pred_eqne<mode>_scalar): Ditto.
28801 (*pred_eqne<mode>_scalar): Ditto.
28802 (*pred_eqne<mode>_scalar_narrow): Ditto.
28803 (*pred_cmp<mode>_extended_scalar): Ditto.
28804 (*pred_cmp<mode>_extended_scalar_narrow): Ditto.
28805 (*pred_eqne<mode>_extended_scalar): Ditto.
28806 (*pred_eqne<mode>_extended_scalar_narrow): Ditto.
28807 (@pred_ge<mode>_scalar): Ditto.
28808 (@pred_<optab><mode>): Ditto.
28809 (@pred_n<optab><mode>): Ditto.
28810 (@pred_<optab>n<mode>): Ditto.
28811 (@pred_not<mode>): Ditto.
28813 2023-02-15 Martin Jambor <mjambor@suse.cz>
28816 * ipa-sra.cc (push_param_adjustments_for_index): Do not omit
28817 creation of non-scalar replacements even if IPA-CP knows their
28820 2023-02-15 Jakub Jelinek <jakub@redhat.com>
28824 * config/rs6000/rs6000.md (<u>maddditi4): Change into umaddditi4 only
28825 expander, change operand 3 to be TImode, emit maddlddi4 and
28826 umadddi4_highpart{,_le} with its low half and finally add the high
28827 half to the result.
28829 2023-02-15 Martin Liska <mliska@suse.cz>
28831 * doc/invoke.texi: Document --param=asan-kernel-mem-intrinsic-prefix.
28833 2023-02-15 Richard Biener <rguenther@suse.de>
28835 * sanopt.cc (sanitize_asan_mark_unpoison): Use bitmap
28836 for with_poison and alias worklist to it.
28837 (sanitize_asan_mark_poison): Likewise.
28839 2023-02-15 Richard Biener <rguenther@suse.de>
28842 * config/i386/i386-features.cc (scalar_chain::add_to_queue):
28843 Combine bitmap test and set.
28844 (scalar_chain::add_insn): Likewise.
28845 (scalar_chain::analyze_register_chain): Remove redundant
28846 attempt to add to queue and instead strengthen assert.
28847 Sink common attempts to mark the def dual-mode.
28848 (scalar_chain::add_to_queue): Remove redundant insn bitmap
28851 2023-02-15 Richard Biener <rguenther@suse.de>
28854 * config/i386/i386-features.cc (convert_scalars_to_vector):
28855 Switch candidates bitmaps to tree view before building the chains.
28857 2023-02-15 Hans-Peter Nilsson <hp@axis.com>
28859 * reload1.cc (gen_reload): Correct rtx parameter for fatal_insn
28860 "failure trying to reload" call.
28862 2023-02-15 Hans-Peter Nilsson <hp@axis.com>
28864 * gdbinit.in (phrs): New command.
28865 * sel-sched-dump.cc (debug_hard_reg_set): Remove debug-function.
28866 * ira-color.cc (debug_hard_reg_set): New, calling print_hard_reg_set.
28868 2023-02-14 David Faust <david.faust@oracle.com>
28871 * config/bpf/constraints.md (q): New memory constraint.
28872 * config/bpf/bpf.md (zero_extendhidi2): Use it here.
28873 (zero_extendqidi2): Likewise.
28874 (zero_extendsidi2): Likewise.
28875 (*mov<MM:mode>): Likewise.
28877 2023-02-14 Andrew Pinski <apinski@marvell.com>
28879 PR tree-optimization/108355
28880 PR tree-optimization/96921
28881 * match.pd: Add pattern for "1 - bool_val".
28883 2023-02-14 Richard Biener <rguenther@suse.de>
28885 * tree-ssa-sccvn.cc (vn_phi_compute_hash): Key skipping
28886 basic block index hashing on the availability of ->cclhs.
28887 (vn_phi_eq): Avoid re-doing sanity checks for CSE but
28888 rely on ->cclhs availability.
28889 (vn_phi_lookup): Set ->cclhs only when we are eventually
28890 going to CSE the PHI.
28891 (vn_phi_insert): Likewise.
28893 2023-02-14 Eric Botcazou <ebotcazou@adacore.com>
28895 * gimplify.cc (gimplify_save_expr): Add missing guard.
28897 2023-02-14 Richard Biener <rguenther@suse.de>
28899 PR tree-optimization/108782
28900 * tree-vect-loop.cc (vect_phi_first_order_recurrence_p):
28901 Make sure we're not vectorizing an inner loop.
28903 2023-02-14 Jakub Jelinek <jakub@redhat.com>
28905 PR sanitizer/108777
28906 * params.opt (-param=asan-kernel-mem-intrinsic-prefix=): New param.
28907 * asan.h (asan_memfn_rtl): Declare.
28908 * asan.cc (asan_memfn_rtls): New variable.
28909 (asan_memfn_rtl): New function.
28910 * builtins.cc (expand_builtin): If
28911 param_asan_kernel_mem_intrinsic_prefix and function is
28912 kernel-{,hw}address sanitized, emit calls to
28913 __{,hw}asan_{memcpy,memmove,memset} rather than
28914 {memcpy,memmove,memset}. Use sanitize_flags_p (SANITIZE_ADDRESS)
28915 instead of flag_sanitize & SANITIZE_ADDRESS to check if
28916 asan_intercepted_p functions shouldn't be expanded inline.
28918 2023-02-14 Richard Sandiford <richard.sandiford@arm.com>
28920 PR tree-optimization/96373
28921 * tree-vect-stmts.cc (vectorizable_operation): Predicate trapping
28922 operations on the loop mask. Reject partial vectors if this isn't
28925 2023-02-13 Richard Sandiford <richard.sandiford@arm.com>
28927 PR rtl-optimization/108681
28928 * lra-spills.cc (lra_final_code_change): Extend subreg replacement
28929 code to handle bare uses and clobbers.
28931 2023-02-13 Vladimir N. Makarov <vmakarov@redhat.com>
28933 * ira.cc (ira_update_equiv_info_by_shuffle_insn): Clear equiv
28934 caller_save_p flag when clearing defined_p flag.
28935 (setup_reg_equiv): Ditto.
28936 * lra-constraints.cc (lra_constraints): Ditto.
28938 2023-02-13 Uroš Bizjak <ubizjak@gmail.com>
28941 * config/i386/predicates.md (extr_register_operand):
28942 New special predicate.
28943 * config/i386/i386.md (*extv<mode>): Use extr_register_operand
28944 as operand 1 predicate.
28945 (*exzv<mode>): Ditto.
28946 (*extendqi<SWI24:mode>_ext_1): New insn pattern.
28948 2023-02-13 Richard Biener <rguenther@suse.de>
28950 PR tree-optimization/28614
28951 * tree-ssa-sccvn.cc (can_track_predicate_on_edge): Avoid
28952 walking all edges in most cases.
28953 (vn_nary_op_insert_pieces_predicated): Avoid repeated
28954 calls to can_track_predicate_on_edge unless checking is
28956 (process_bb): Instead call it once here for each edge
28957 we register possibly multiple predicates on.
28959 2023-02-13 Richard Biener <rguenther@suse.de>
28961 PR tree-optimization/108691
28962 * tree-cfg.cc (notice_special_calls): When the CFG is built
28963 honor gimple_call_ctrl_altering_p.
28964 * cfgexpand.cc (expand_call_stmt): Clear cfun->calls_setjmp
28965 temporarily if the call is not control-altering.
28966 * calls.cc (emit_call_1): Do not add REG_SETJMP if
28967 cfun->calls_setjmp is not set. Do not alter cfun->calls_setjmp.
28969 2023-02-13 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
28972 * config/s390/s390.cc (s390_bb_fallthru_entry_likely): Remove.
28973 (struct s390_sched_state): Initialise to zero.
28974 (s390_sched_variable_issue): For better debuggability also emit
28976 (s390_sched_init): Unconditionally reset scheduler state.
28978 2023-02-13 Richard Sandiford <richard.sandiford@arm.com>
28980 * ifcvt.h (noce_if_info::cond_inverted): New field.
28981 * ifcvt.cc (cond_move_convert_if_block): Swap the then and else
28982 values when cond_inverted is true.
28983 (noce_find_if_block): Allow the condition to be inverted when
28984 handling conditional moves.
28986 2023-02-13 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
28988 * config/s390/predicates.md (execute_operation): Use
28989 constrain_operands instead of extract_constrain_insn in order to
28990 determine wheter there exists a valid alternative.
28992 2023-02-13 Claudiu Zissulescu <claziss@gmail.com>
28994 * common/config/arc/arc-common.cc (arc_option_optimization_table):
28995 Remove millicode from list.
28997 2023-02-13 Martin Liska <mliska@suse.cz>
28999 * doc/invoke.texi: Document ira-simple-lra-insn-threshold.
29001 2023-02-13 Richard Biener <rguenther@suse.de>
29003 PR tree-optimization/106722
29004 * tree-ssa-dce.cc (mark_last_stmt_necessary): Return
29005 whether we marked a stmt.
29006 (mark_control_dependent_edges_necessary): When
29007 mark_last_stmt_necessary didn't mark any stmt make sure
29008 to mark its control dependent edges.
29009 (propagate_necessity): Likewise.
29011 2023-02-13 Kito Cheng <kito.cheng@sifive.com>
29013 * config/riscv/riscv.h (RISCV_DWARF_VLENB): New.
29014 (DWARF_FRAME_REGISTERS): New.
29015 (DWARF_REG_TO_UNWIND_COLUMN): New.
29017 2023-02-12 Gerald Pfeifer <gerald@pfeifer.com>
29019 * doc/sourcebuild.texi: Remove (broken) direct reference to
29020 "The GNU configure and build system".
29022 2023-02-12 Jin Ma <jinma@linux.alibaba.com>
29024 * config/riscv/riscv.cc (riscv_adjust_libcall_cfi_prologue): Change
29025 gen_add3_insn to gen_rtx_SET.
29026 (riscv_adjust_libcall_cfi_epilogue): Likewise.
29028 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29030 * config/riscv/riscv-vector-builtins-bases.cc (class sat_op): New class.
29031 (class vnclip): Ditto.
29033 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
29034 * config/riscv/riscv-vector-builtins-functions.def (vaadd): Ditto.
29043 * config/riscv/vector-iterators.md (su): Add instruction.
29046 * config/riscv/vector.md (@pred_<sat_op><mode>): New pattern.
29047 (@pred_<sat_op><mode>_scalar): Ditto.
29048 (*pred_<sat_op><mode>_scalar): Ditto.
29049 (*pred_<sat_op><mode>_extended_scalar): Ditto.
29050 (@pred_narrow_clip<v_su><mode>): Ditto.
29051 (@pred_narrow_clip<v_su><mode>_scalar): Ditto.
29053 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29055 * config/riscv/constraints.md (Wbr): Remove unused constraint.
29056 * config/riscv/predicates.md: Fix move operand predicate.
29057 * config/riscv/riscv-vector-builtins-bases.cc (class vnshift): New class.
29058 (class vncvt_x): Ditto.
29059 (class vmerge): Ditto.
29060 (class vmv_v): Ditto.
29062 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
29063 * config/riscv/riscv-vector-builtins-functions.def (vsra): Ditto.
29070 * config/riscv/riscv-vector-builtins-shapes.cc (struct narrow_alu_def): Ditto.
29071 (struct move_def): Ditto.
29073 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
29074 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_WEXTI_OPS): New variable.
29075 (DEF_RVV_WEXTU_OPS): Ditto
29076 * config/riscv/riscv-vector-builtins.def (x_x_w): Fix type for suffix.
29081 * config/riscv/riscv.cc (riscv_print_operand): Refine ASM printting rule.
29082 * config/riscv/vector-iterators.md (nmsac):New iterator.
29083 (nmsub): New iterator.
29084 * config/riscv/vector.md (@pred_merge<mode>): New pattern.
29085 (@pred_merge<mode>_scalar): New pattern.
29086 (*pred_merge<mode>_scalar): New pattern.
29087 (*pred_merge<mode>_extended_scalar): New pattern.
29088 (@pred_narrow_<optab><mode>): New pattern.
29089 (@pred_narrow_<optab><mode>_scalar): New pattern.
29090 (@pred_trunc<mode>): New pattern.
29092 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29094 * config/riscv/riscv-vector-builtins-bases.cc (class vmadc): New class.
29095 (class vmsbc): Ditto.
29096 (BASE): Define new class.
29097 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
29098 * config/riscv/riscv-vector-builtins-functions.def (vmadc): New define.
29100 * config/riscv/riscv-vector-builtins-shapes.cc (struct return_mask_def):
29103 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
29104 * config/riscv/riscv-vector-builtins.cc
29105 (function_expander::use_exact_insn): Adjust for new support
29106 * config/riscv/riscv-vector-builtins.h
29107 (function_base::has_merge_operand_p): New function.
29108 * config/riscv/vector-iterators.md: New iterator.
29109 * config/riscv/vector.md (@pred_madc<mode>): New pattern.
29110 (@pred_msbc<mode>): Ditto.
29111 (@pred_madc<mode>_scalar): Ditto.
29112 (@pred_msbc<mode>_scalar): Ditto.
29113 (*pred_madc<mode>_scalar): Ditto.
29114 (*pred_madc<mode>_extended_scalar): Ditto.
29115 (*pred_msbc<mode>_scalar): Ditto.
29116 (*pred_msbc<mode>_extended_scalar): Ditto.
29117 (@pred_madc<mode>_overflow): Ditto.
29118 (@pred_msbc<mode>_overflow): Ditto.
29119 (@pred_madc<mode>_overflow_scalar): Ditto.
29120 (@pred_msbc<mode>_overflow_scalar): Ditto.
29121 (*pred_madc<mode>_overflow_scalar): Ditto.
29122 (*pred_madc<mode>_overflow_extended_scalar): Ditto.
29123 (*pred_msbc<mode>_overflow_scalar): Ditto.
29124 (*pred_msbc<mode>_overflow_extended_scalar): Ditto.
29126 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29128 * config/riscv/riscv-protos.h (simm5_p): Add vadc/vsbc support.
29129 * config/riscv/riscv-v.cc (simm32_p): Ditto.
29130 * config/riscv/riscv-vector-builtins-bases.cc (class vadc): New class.
29131 (class vsbc): Ditto.
29133 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
29134 * config/riscv/riscv-vector-builtins-functions.def (vadc): Ditto.
29136 * config/riscv/riscv-vector-builtins-shapes.cc
29137 (struct no_mask_policy_def): Ditto.
29139 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
29140 * config/riscv/riscv-vector-builtins.cc
29141 (rvv_arg_type_info::get_base_vector_type): Add vadc/vsbc support.
29142 (rvv_arg_type_info::get_tree_type): Ditto.
29143 (function_expander::use_exact_insn): Ditto.
29144 * config/riscv/riscv-vector-builtins.h (enum rvv_base_type): Ditto.
29145 (function_base::use_mask_predication_p): New function.
29146 * config/riscv/vector-iterators.md: New iterator.
29147 * config/riscv/vector.md (@pred_adc<mode>): New pattern.
29148 (@pred_sbc<mode>): Ditto.
29149 (@pred_adc<mode>_scalar): Ditto.
29150 (@pred_sbc<mode>_scalar): Ditto.
29151 (*pred_adc<mode>_scalar): Ditto.
29152 (*pred_adc<mode>_extended_scalar): Ditto.
29153 (*pred_sbc<mode>_scalar): Ditto.
29154 (*pred_sbc<mode>_extended_scalar): Ditto.
29156 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29158 * config/riscv/vector.md: use "zero" reg.
29160 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29162 * config/riscv/riscv-vector-builtins-bases.cc (class widen_binop): New
29164 (class vwmulsu): Ditto.
29165 (class vwcvt): Ditto.
29166 (BASE): Add integer widening support.
29167 * config/riscv/riscv-vector-builtins-bases.h: Ditto
29168 * config/riscv/riscv-vector-builtins-functions.def (vwadd): New class.
29169 (vwsub): New class.
29170 (vwmul): New class.
29171 (vwmulu): New class.
29172 (vwmulsu): New class.
29173 (vwaddu): New class.
29174 (vwsubu): New class.
29175 (vwcvt_x): New class.
29176 (vwcvtu_x): New class.
29177 * config/riscv/riscv-vector-builtins-shapes.cc (struct alu_def): New
29179 (struct widen_alu_def): New class.
29180 (SHAPE): New class.
29181 * config/riscv/riscv-vector-builtins-shapes.h: New class.
29182 * config/riscv/riscv-vector-builtins.cc
29183 (rvv_arg_type_info::get_base_vector_type): Add integer widening support.
29184 (rvv_arg_type_info::get_tree_type): Ditto.
29185 * config/riscv/riscv-vector-builtins.def (x_x_v): Change into "x_v"
29187 * config/riscv/riscv-vector-builtins.h (enum rvv_base_type): Add integer
29189 * config/riscv/riscv-vsetvl.cc (change_insn): Fix reg_equal use bug.
29190 * config/riscv/riscv.h (X0_REGNUM): New constant.
29191 * config/riscv/vector-iterators.md: New iterators.
29192 * config/riscv/vector.md
29193 (@pred_dual_widen_<any_widen_binop:optab><any_extend:su><mode>): New
29195 (@pred_dual_widen_<any_widen_binop:optab><any_extend:su><mode>_scalar):
29197 (@pred_single_widen_<plus_minus:optab><any_extend:su><mode>): Ditto.
29198 (@pred_single_widen_<plus_minus:optab><any_extend:su><mode>_scalar):
29200 (@pred_widen_mulsu<mode>): Ditto.
29201 (@pred_widen_mulsu<mode>_scalar): Ditto.
29202 (@pred_<optab><mode>): Ditto.
29204 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29205 kito-cheng <kito.cheng@sifive.com>
29207 * common/config/riscv/riscv-common.cc: Add flag for 'V' extension.
29208 * config/riscv/riscv-vector-builtins-bases.cc (class vmulh): New class.
29210 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
29211 * config/riscv/riscv-vector-builtins-functions.def (vmulh): Add vmulh
29215 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_FULL_V_I_OPS):
29217 (DEF_RVV_FULL_V_U_OPS): Ditto.
29218 (vint8mf8_t): Ditto.
29219 (vint8mf4_t): Ditto.
29220 (vint8mf2_t): Ditto.
29221 (vint8m1_t): Ditto.
29222 (vint8m2_t): Ditto.
29223 (vint8m4_t): Ditto.
29224 (vint8m8_t): Ditto.
29225 (vint16mf4_t): Ditto.
29226 (vint16mf2_t): Ditto.
29227 (vint16m1_t): Ditto.
29228 (vint16m2_t): Ditto.
29229 (vint16m4_t): Ditto.
29230 (vint16m8_t): Ditto.
29231 (vint32mf2_t): Ditto.
29232 (vint32m1_t): Ditto.
29233 (vint32m2_t): Ditto.
29234 (vint32m4_t): Ditto.
29235 (vint32m8_t): Ditto.
29236 (vint64m1_t): Ditto.
29237 (vint64m2_t): Ditto.
29238 (vint64m4_t): Ditto.
29239 (vint64m8_t): Ditto.
29240 (vuint8mf8_t): Ditto.
29241 (vuint8mf4_t): Ditto.
29242 (vuint8mf2_t): Ditto.
29243 (vuint8m1_t): Ditto.
29244 (vuint8m2_t): Ditto.
29245 (vuint8m4_t): Ditto.
29246 (vuint8m8_t): Ditto.
29247 (vuint16mf4_t): Ditto.
29248 (vuint16mf2_t): Ditto.
29249 (vuint16m1_t): Ditto.
29250 (vuint16m2_t): Ditto.
29251 (vuint16m4_t): Ditto.
29252 (vuint16m8_t): Ditto.
29253 (vuint32mf2_t): Ditto.
29254 (vuint32m1_t): Ditto.
29255 (vuint32m2_t): Ditto.
29256 (vuint32m4_t): Ditto.
29257 (vuint32m8_t): Ditto.
29258 (vuint64m1_t): Ditto.
29259 (vuint64m2_t): Ditto.
29260 (vuint64m4_t): Ditto.
29261 (vuint64m8_t): Ditto.
29262 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_FULL_V_I_OPS): Ditto.
29263 (DEF_RVV_FULL_V_U_OPS): Ditto.
29264 (check_required_extensions): Add vmulh support.
29265 (rvv_arg_type_info::get_tree_type): Ditto.
29266 * config/riscv/riscv-vector-builtins.h (RVV_REQUIRE_FULL_V): Ditto.
29267 (enum rvv_base_type): Ditto.
29268 * config/riscv/riscv.opt: Add 'V' extension flag.
29269 * config/riscv/vector-iterators.md (su): New iterator.
29270 * config/riscv/vector.md (@pred_mulh<v_su><mode>): New pattern.
29271 (@pred_mulh<v_su><mode>_scalar): Ditto.
29272 (*pred_mulh<v_su><mode>_scalar): Ditto.
29273 (*pred_mulh<v_su><mode>_extended_scalar): Ditto.
29275 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29277 * config/riscv/iterators.md: Add sign_extend/zero_extend.
29278 * config/riscv/riscv-vector-builtins-bases.cc (class ext): New class.
29280 * config/riscv/riscv-vector-builtins-bases.h: Add vsext/vzext support.
29281 * config/riscv/riscv-vector-builtins-functions.def (vsext): New macro
29284 * config/riscv/riscv-vector-builtins-shapes.cc (struct alu_def): Adjust
29285 for vsext/vzext support.
29286 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_WEXTI_OPS): New
29288 (DEF_RVV_QEXTI_OPS): Ditto.
29289 (DEF_RVV_OEXTI_OPS): Ditto.
29290 (DEF_RVV_WEXTU_OPS): Ditto.
29291 (DEF_RVV_QEXTU_OPS): Ditto.
29292 (DEF_RVV_OEXTU_OPS): Ditto.
29293 (vint16mf4_t): Ditto.
29294 (vint16mf2_t): Ditto.
29295 (vint16m1_t): Ditto.
29296 (vint16m2_t): Ditto.
29297 (vint16m4_t): Ditto.
29298 (vint16m8_t): Ditto.
29299 (vint32mf2_t): Ditto.
29300 (vint32m1_t): Ditto.
29301 (vint32m2_t): Ditto.
29302 (vint32m4_t): Ditto.
29303 (vint32m8_t): Ditto.
29304 (vint64m1_t): Ditto.
29305 (vint64m2_t): Ditto.
29306 (vint64m4_t): Ditto.
29307 (vint64m8_t): Ditto.
29308 (vuint16mf4_t): Ditto.
29309 (vuint16mf2_t): Ditto.
29310 (vuint16m1_t): Ditto.
29311 (vuint16m2_t): Ditto.
29312 (vuint16m4_t): Ditto.
29313 (vuint16m8_t): Ditto.
29314 (vuint32mf2_t): Ditto.
29315 (vuint32m1_t): Ditto.
29316 (vuint32m2_t): Ditto.
29317 (vuint32m4_t): Ditto.
29318 (vuint32m8_t): Ditto.
29319 (vuint64m1_t): Ditto.
29320 (vuint64m2_t): Ditto.
29321 (vuint64m4_t): Ditto.
29322 (vuint64m8_t): Ditto.
29323 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_WEXTI_OPS): Ditto.
29324 (DEF_RVV_QEXTI_OPS): Ditto.
29325 (DEF_RVV_OEXTI_OPS): Ditto.
29326 (DEF_RVV_WEXTU_OPS): Ditto.
29327 (DEF_RVV_QEXTU_OPS): Ditto.
29328 (DEF_RVV_OEXTU_OPS): Ditto.
29329 (rvv_arg_type_info::get_base_vector_type): Add sign_exted/zero_extend
29331 (rvv_arg_type_info::get_tree_type): Ditto.
29332 * config/riscv/riscv-vector-builtins.h (enum rvv_base_type): Ditto.
29333 * config/riscv/vector-iterators.md (z): New attribute.
29334 * config/riscv/vector.md (@pred_<optab><mode>_vf2): New pattern.
29335 (@pred_<optab><mode>_vf4): Ditto.
29336 (@pred_<optab><mode>_vf8): Ditto.
29338 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29340 * config/riscv/iterators.md: Add saturating Addition && Subtraction.
29341 * config/riscv/riscv-v.cc (has_vi_variant_p): Ditto.
29342 * config/riscv/riscv-vector-builtins-bases.cc (BASE): Ditto.
29343 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
29344 * config/riscv/riscv-vector-builtins-functions.def (vsadd): New def.
29348 * config/riscv/vector-iterators.md (sll.vi): Adjust for Saturating
29353 * config/riscv/vector.md (@pred_<optab><mode>): New pattern.
29354 (@pred_<optab><mode>_scalar): New pattern.
29355 (*pred_<optab><mode>_scalar): New pattern.
29356 (*pred_<optab><mode>_extended_scalar): New pattern.
29358 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29360 * config/riscv/iterators.md: Add neg and not.
29361 * config/riscv/riscv-vector-builtins-bases.cc (class unop): New class.
29363 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
29364 * config/riscv/riscv-vector-builtins-functions.def (vadd): Rename binop
29385 * config/riscv/riscv-vector-builtins-shapes.cc (struct binop_def): Ditto.
29386 (struct alu_def): Ditto.
29388 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
29389 * config/riscv/riscv-vector-builtins.cc: Support unary C/C/++.
29390 * config/riscv/vector-iterators.md: New iterator.
29391 * config/riscv/vector.md (@pred_<optab><mode>): New pattern
29393 2023-02-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29395 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::compute_probabilities): Skip exit block.
29397 2023-02-11 Jakub Jelinek <jakub@redhat.com>
29400 * ipa-cp.cc (ipa_agg_value_from_jfunc): Return NULL_TREE also if
29401 item->offset bit position is too large to be representable as
29402 unsigned int byte position.
29404 2023-02-11 Gerald Pfeifer <gerald@pfeifer.com>
29406 * doc/extend.texi (Other Builtins): Adjust link to WG14 N965.
29408 2023-02-10 Vladimir N. Makarov <vmakarov@redhat.com>
29410 * ira.cc (update_equiv_regs): Set up ira_reg_equiv for
29411 valid_combine only when ira_use_lra_p is true.
29413 2023-02-10 Vladimir N. Makarov <vmakarov@redhat.com>
29415 * params.opt (ira-simple-lra-insn-threshold): Add new param.
29416 * ira.cc (ira): Use the param to switch on simple LRA.
29418 2023-02-10 Andrew MacLeod <amacleod@redhat.com>
29420 PR tree-optimization/108687
29421 * gimple-range-cache.cc (ranger_cache::range_on_edge): Revert
29422 back to RFD_NONE mode for calculations.
29423 (ranger_cache::propagate_cache): Call the internal edge range API
29424 with RFD_READ_ONLY instead of changing the external routine.
29426 2023-02-10 Andrew MacLeod <amacleod@redhat.com>
29428 PR tree-optimization/108520
29429 * gimple-range-infer.cc (check_assume_func): Invoke
29430 gimple_range_global directly instead using global_range_query.
29431 * value-query.cc (get_range_global): Add function context and
29432 avoid calling nonnull_arg_p if not cfun.
29433 (gimple_range_global): Add function context pointer.
29434 * value-query.h (imple_range_global): Add function context.
29436 2023-02-10 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29438 * config/riscv/constraints.md (Wdm): Adjust constraint.
29439 (Wbr): New constraint.
29440 * config/riscv/predicates.md (reg_or_int_operand): New predicate.
29441 * config/riscv/riscv-protos.h (emit_pred_op): Remove function.
29442 (emit_vlmax_op): New function.
29443 (emit_nonvlmax_op): Ditto.
29445 (neg_simm5_p): Ditto.
29446 (has_vi_variant_p): Ditto.
29447 * config/riscv/riscv-v.cc (emit_pred_op): Adjust function.
29448 (emit_vlmax_op): New function.
29449 (emit_nonvlmax_op): Ditto.
29450 (expand_const_vector): Adjust function.
29451 (legitimize_move): Ditto.
29452 (simm32_p): New function.
29454 (neg_simm5_p): Ditto.
29455 (has_vi_variant_p): Ditto.
29456 * config/riscv/riscv-vector-builtins-bases.cc (class vrsub): New class.
29458 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
29459 * config/riscv/riscv-vector-builtins-functions.def (vmin): Remove
29462 (vminu): Remove signed cases.
29464 (vdiv): Remove unsigned cases.
29466 (vdivu): Remove signed cases.
29470 (vrsub): New class.
29475 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_U_OPS): New macro.
29476 * config/riscv/riscv.h: change VL/VTYPE as fixed reg.
29477 * config/riscv/vector-iterators.md: New iterators.
29478 * config/riscv/vector.md (@pred_broadcast<mode>): Adjust pattern for vx
29480 (@pred_<optab><mode>_scalar): New pattern.
29481 (@pred_sub<mode>_reverse_scalar): Ditto.
29482 (*pred_<optab><mode>_scalar): Ditto.
29483 (*pred_<optab><mode>_extended_scalar): Ditto.
29484 (*pred_sub<mode>_reverse_scalar): Ditto.
29485 (*pred_sub<mode>_extended_reverse_scalar): Ditto.
29487 2023-02-10 Richard Biener <rguenther@suse.de>
29489 PR tree-optimization/108724
29490 * tree-vect-stmts.cc (vectorizable_operation): Avoid
29491 using word_mode vectors when vector lowering will
29492 decompose them to elementwise operations.
29494 2023-02-10 Jakub Jelinek <jakub@redhat.com>
29497 2023-02-09 Martin Liska <mliska@suse.cz>
29500 * doc/extend.texi: Document that the function
29501 does not work correctly for old VIA processors.
29503 2023-02-10 Andrew Pinski <apinski@marvell.com>
29504 Andrew Macleod <amacleod@redhat.com>
29506 PR tree-optimization/108684
29507 * tree-ssa-dce.cc (simple_dce_from_worklist):
29508 Check all ssa names and not just non-vdef ones
29509 before accepting the inline-asm.
29510 Call unlink_stmt_vdef on the statement before
29513 2023-02-09 Vladimir N. Makarov <vmakarov@redhat.com>
29515 * ira.h (struct ira_reg_equiv_s): Add new field caller_save_p.
29516 * ira.cc (validate_equiv_mem): Check memref address variance.
29517 (no_equiv): Clear caller_save_p flag.
29518 (update_equiv_regs): Define caller save equivalence for
29520 (setup_reg_equiv): Clear defined_p flag for caller save equivalence.
29521 * lra-constraints.cc (lra_copy_reg_equiv): Add new arg
29522 call_save_p. Use caller save equivalence depending on the arg.
29523 (split_reg): Adjust the call.
29525 2023-02-09 Jakub Jelinek <jakub@redhat.com>
29528 * common/config/i386/cpuinfo.h (get_zhaoxin_cpu): Formatting fixes.
29529 (cpu_indicator_init): Call get_available_features for all CPUs with
29530 max_level >= 1, rather than just Intel, AMD or Zhaoxin. Formatting
29533 2023-02-09 Jakub Jelinek <jakub@redhat.com>
29535 PR tree-optimization/108688
29536 * match.pd (bit_field_ref [bit_insert]): Simplify BIT_FIELD_REF
29537 of BIT_INSERT_EXPR extracting exactly all inserted bits even
29538 when without mode precision. Formatting fixes.
29540 2023-02-09 Andrew Pinski <apinski@marvell.com>
29542 PR tree-optimization/108688
29543 * match.pd (bit_field_ref [bit_insert]): Avoid generating
29544 BIT_FIELD_REFs of non-mode-precision integral operands.
29546 2023-02-09 Martin Liska <mliska@suse.cz>
29549 * doc/extend.texi: Document that the function
29550 does not work correctly for old VIA processors.
29552 2023-02-09 Andreas Schwab <schwab@suse.de>
29554 * lto-wrapper.cc (merge_and_complain): Handle
29555 -funwind-tables and -fasynchronous-unwind-tables.
29556 (append_compiler_options): Likewise.
29558 2023-02-09 Richard Biener <rguenther@suse.de>
29560 PR tree-optimization/26854
29561 * tree-into-ssa.cc (update_ssa): Turn blocks_to_update to tree
29562 view around insert_updated_phi_nodes_for.
29563 * tree-ssa-alias.cc (maybe_skip_until): Allocate visited bitmap
29565 (walk_aliased_vdefs_1): Likewise.
29567 2023-02-08 Gerald Pfeifer <gerald@pfeifer.com>
29569 * doc/include/gpl_v3.texi: Change fsf.org to www.fsf.org.
29571 2023-02-08 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
29574 * config.gcc (tm_mlib_file): Define new variable.
29576 2023-02-08 Jakub Jelinek <jakub@redhat.com>
29578 PR tree-optimization/108692
29579 * tree-vect-patterns.cc (vect_widened_op_tree): If rhs_code is
29580 widened_code which is different from code, don't call
29581 vect_look_through_possible_promotion but instead just check op is
29582 SSA_NAME with integral type for which vect_is_simple_use is true
29583 and call set_op on this_unprom.
29585 2023-02-08 Andrea Corallo <andrea.corallo@arm.com>
29587 * config/aarch64/aarch64-protos.h (aarch_ra_sign_key): Remove
29589 * config/aarch64/aarch64.cc (aarch_ra_sign_key): Remove
29591 * config/aarch64/aarch64.opt (aarch64_ra_sign_key): Rename
29592 to 'aarch_ra_sign_key'.
29593 * config/arm/aarch-common.cc (aarch_ra_sign_key): Remove
29595 * config/arm/arm-protos.h (aarch_ra_sign_key): Likewise.
29596 * config/arm/arm.cc (enum aarch_key_type): Remove definition.
29597 * config/arm/arm.opt: Define.
29599 2023-02-08 Richard Sandiford <richard.sandiford@arm.com>
29601 PR tree-optimization/108316
29602 * tree-vect-stmts.cc (get_load_store_type): When using
29603 internal functions for gather/scatter, make sure that the type
29604 of the offset argument is consistent with the offset vector type.
29606 2023-02-08 Vladimir N. Makarov <vmakarov@redhat.com>
29609 2023-02-07 Vladimir N. Makarov <vmakarov@redhat.com>
29611 * ira.h (struct ira_reg_equiv_s): Add new field caller_save_p.
29612 * ira.cc (validate_equiv_mem): Check memref address variance.
29613 (update_equiv_regs): Define caller save equivalence for
29615 (setup_reg_equiv): Clear defined_p flag for caller save equivalence.
29616 * lra-constraints.cc (lra_copy_reg_equiv): Add new arg
29617 call_save_p. Use caller save equivalence depending on the arg.
29618 (split_reg): Adjust the call.
29620 2023-02-08 Jakub Jelinek <jakub@redhat.com>
29622 * tree.def (SAD_EXPR): Remove outdated comment about missing
29625 2023-02-07 Marek Polacek <polacek@redhat.com>
29627 * doc/invoke.texi: Update -fchar8_t documentation.
29629 2023-02-07 Vladimir N. Makarov <vmakarov@redhat.com>
29631 * ira.h (struct ira_reg_equiv_s): Add new field caller_save_p.
29632 * ira.cc (validate_equiv_mem): Check memref address variance.
29633 (update_equiv_regs): Define caller save equivalence for
29635 (setup_reg_equiv): Clear defined_p flag for caller save equivalence.
29636 * lra-constraints.cc (lra_copy_reg_equiv): Add new arg
29637 call_save_p. Use caller save equivalence depending on the arg.
29638 (split_reg): Adjust the call.
29640 2023-02-07 Richard Biener <rguenther@suse.de>
29642 PR tree-optimization/26854
29643 * gimple-fold.cc (has_use_on_stmt): Look at stmt operands
29644 instead of immediate uses.
29646 2023-02-07 Jakub Jelinek <jakub@redhat.com>
29648 PR tree-optimization/106923
29649 * ipa-split.cc (execute_split_functions): Don't split returns_twice
29652 2023-02-07 Jakub Jelinek <jakub@redhat.com>
29654 PR tree-optimization/106433
29655 * cgraph.cc (set_const_flag_1): Recurse on simd clones too.
29656 (cgraph_node::set_pure_flag): Call set_pure_flag_1 on simd clones too.
29658 2023-02-07 Jan Hubicka <jh@suse.cz>
29660 * config/i386/x86-tune.def (X86_TUNE_AVX256_OPTIMAL): Turn off
29663 2023-02-06 Andrew Stubbs <ams@codesourcery.com>
29665 * config/gcn/mkoffload.cc (gcn_stack_size): New global variable.
29666 (process_asm): Create a constructor for GCN_STACK_SIZE.
29667 (main): Parse the -mstack-size option.
29669 2023-02-06 Alex Coplan <alex.coplan@arm.com>
29672 * config/aarch64/aarch64-simd.md (aarch64_bfmlal<bt>_lane<q>v4sf):
29673 Use correct constraint for operand 3.
29675 2023-02-06 Martin Jambor <mjambor@suse.cz>
29677 * ipa-sra.cc (adjust_parameter_descriptions): Fix a typo in a dump.
29679 2023-02-06 Xi Ruoyao <xry111@xry111.site>
29681 * config/loongarch/loongarch.md (bytepick_w_ashift_amount):
29682 New define_int_iterator.
29683 (bytepick_d_ashift_amount): Likewise.
29684 (bytepick_imm): New define_int_attr.
29685 (bytepick_w_lshiftrt_amount): Likewise.
29686 (bytepick_d_lshiftrt_amount): Likewise.
29687 (bytepick_w_<bytepick_imm>): New define_insn template.
29688 (bytepick_w_<bytepick_imm>_extend): Likewise.
29689 (bytepick_d_<bytepick_imm>): Likewise.
29690 (bytepick_w): Remove unused define_insn.
29691 (bytepick_d): Likewise.
29692 (UNSPEC_BYTEPICK_W): Remove unused unspec.
29693 (UNSPEC_BYTEPICK_D): Likewise.
29694 * config/loongarch/predicates.md (const_0_to_3_operand):
29695 Remove unused define_predicate.
29696 (const_0_to_7_operand): Likewise.
29698 2023-02-06 Jakub Jelinek <jakub@redhat.com>
29700 PR tree-optimization/108655
29701 * ubsan.cc (sanitize_unreachable_fn): For -funreachable-traps
29702 or -fsanitize=unreachable -fsanitize-trap=unreachable return
29703 BUILT_IN_UNREACHABLE_TRAP decl rather than BUILT_IN_TRAP.
29705 2023-02-05 Gerald Pfeifer <gerald@pfeifer.com>
29707 * doc/install.texi (Specific): Remove PW32.
29709 2023-02-03 Jakub Jelinek <jakub@redhat.com>
29711 PR tree-optimization/108647
29712 * range-op.cc (operator_equal::op1_range,
29713 operator_not_equal::op1_range): Don't test op2 bound
29714 equality if op2.undefined_p (), instead set_varying.
29715 (operator_lt::op1_range, operator_le::op1_range,
29716 operator_gt::op1_range, operator_ge::op1_range): Return false if
29717 op2.undefined_p ().
29718 (operator_lt::op2_range, operator_le::op2_range,
29719 operator_gt::op2_range, operator_ge::op2_range): Return false if
29720 op1.undefined_p ().
29722 2023-02-03 Aldy Hernandez <aldyh@redhat.com>
29724 PR tree-optimization/108639
29725 * value-range.cc (irange::legacy_equal_p): Compare nonzero bits as
29727 (irange::operator==): Same.
29729 2023-02-03 Aldy Hernandez <aldyh@redhat.com>
29731 PR tree-optimization/108647
29732 * range-op-float.cc (foperator_lt::op1_range): Handle undefined ranges.
29733 (foperator_lt::op2_range): Same.
29734 (foperator_le::op1_range): Same.
29735 (foperator_le::op2_range): Same.
29736 (foperator_gt::op1_range): Same.
29737 (foperator_gt::op2_range): Same.
29738 (foperator_ge::op1_range): Same.
29739 (foperator_ge::op2_range): Same.
29740 (foperator_unordered_lt::op1_range): Same.
29741 (foperator_unordered_lt::op2_range): Same.
29742 (foperator_unordered_le::op1_range): Same.
29743 (foperator_unordered_le::op2_range): Same.
29744 (foperator_unordered_gt::op1_range): Same.
29745 (foperator_unordered_gt::op2_range): Same.
29746 (foperator_unordered_ge::op1_range): Same.
29747 (foperator_unordered_ge::op2_range): Same.
29749 2023-02-03 Andrew MacLeod <amacleod@redhat.com>
29751 PR tree-optimization/107570
29752 * tree-vrp.cc (remove_and_update_globals): Reset SCEV.
29754 2023-02-03 Gaius Mulley <gaiusmod2@gmail.com>
29756 * doc/gm2.texi (Internals): Remove from menu.
29757 (Using): Comment out ifnohtml conditional.
29758 (Documentation): Use gcc url.
29759 (License): Node simplified.
29760 (Copying): New node. Include gpl_v3_without_node.
29761 (Contributing): Node simplified.
29762 (Internals): Commented out.
29763 (Libraries): Node simplified.
29766 (Functions): Ditto.
29768 2023-02-03 Christophe Lyon <christophe.lyon@arm.com>
29770 * config/arm/mve.md (mve_vabavq_p_<supf><mode>): Add length
29772 (mve_vqshluq_m_n_s<mode>): Likewise.
29773 (mve_vshlq_m_<supf><mode>): Likewise.
29774 (mve_vsriq_m_n_<supf><mode>): Likewise.
29775 (mve_vsubq_m_<supf><mode>): Likewise.
29777 2023-02-03 Martin Jambor <mjambor@suse.cz>
29780 * ipa-sra.cc (push_param_adjustments_for_index): Remove a size check
29781 when comparing to an IPA-CP value.
29782 (dump_list_of_param_indices): New function.
29783 (adjust_parameter_descriptions): Check for mismatching IPA-CP values.
29784 Dump removed candidates using dump_list_of_param_indices.
29785 * ipa-param-manipulation.cc
29786 (ipa_param_body_adjustments::modify_expression): Add assert checking
29787 sizes of a VIEW_CONVERT_EXPR will match.
29788 (ipa_param_body_adjustments::modify_assignment): Likewise.
29790 2023-02-03 Monk Chiang <monk.chiang@sifive.com>
29792 * config/riscv/riscv.h: Remove VL_REGS, VTYPE_REGS class.
29793 * config/riscv/riscv.cc: Ditto.
29795 2023-02-03 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29797 * config/riscv/vector-iterators.md (sll.vi): Fix constraint bug.
29801 * config/riscv/vector.md: Ditto.
29803 2023-02-03 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
29805 * config/riscv/predicates.md (pmode_reg_or_uimm5_operand): New predicate.
29806 * config/riscv/riscv-vector-builtins-bases.cc: New class.
29807 * config/riscv/riscv-vector-builtins-functions.def (vsll): Ditto.
29810 * config/riscv/riscv-vector-builtins.cc: Ditto.
29811 * config/riscv/vector.md (@pred_<optab><mode>_scalar): New pattern.
29813 2023-02-02 Iain Sandoe <iain@sandoe.co.uk>
29815 * toplev.cc (toplev::main): Only print the version information header
29816 from toplevel main().
29818 2023-02-02 Paul-Antoine Arras <pa@codesourcery.com>
29820 * config/gcn/gcn-valu.md (cond_<expander><mode>): Add
29821 cond_{ashl|ashr|lshr}
29823 2023-02-02 Richard Sandiford <richard.sandiford@arm.com>
29825 PR rtl-optimization/108086
29826 * rtl-ssa/insns.h (insn_info): Make m_num_defs a full unsigned int.
29827 Adjust size-related commentary accordingly.
29829 2023-02-02 Richard Sandiford <richard.sandiford@arm.com>
29831 PR rtl-optimization/108508
29832 * rtl-ssa/accesses.cc (function_info::split_clobber_group): When
29833 the splay tree search gives the first clobber in the second group,
29834 make sure that the root of the first clobber group is updated
29835 correctly. Enter the new clobber group into the definition splay
29838 2023-02-02 Jin Ma <jinma@linux.alibaba.com>
29840 * common/config/riscv/riscv-common.cc (riscv_compute_multilib):
29841 Fix finding best match score.
29843 2023-02-02 Jakub Jelinek <jakub@redhat.com>
29846 PR rtl-optimization/108463
29848 * cselib.cc (cselib_current_insn): Move declaration earlier.
29849 (cselib_hasher::equal): For debug only locs, temporarily override
29850 cselib_current_insn to their l->setting_insn for the
29851 rtx_equal_for_cselib_1 call, so that unsuccessful comparisons don't
29852 promote some debug locs.
29853 * sched-deps.cc (sched_analyze_2) <case MEM>: For MEMs in DEBUG_INSNs
29854 when using cselib call cselib_lookup_from_insn on the address but
29855 don't substitute it.
29857 2023-02-02 Richard Biener <rguenther@suse.de>
29859 PR middle-end/108625
29860 * genmatch.cc (expr::gen_transform): Also disallow resimplification
29861 from pushing to lseq with force_leaf.
29862 (dt_simplify::gen_1): Likewise.
29864 2023-02-02 Andrew Stubbs <ams@codesourcery.com>
29866 * config/gcn/gcn-run.cc: Include libgomp-gcn.h.
29867 (struct kernargs): Replace the common content with kernargs_abi.
29868 (struct heap): Delete.
29869 (main): Read GCN_STACK_SIZE envvar.
29870 Allocate space for the device stacks.
29871 Write the new kernargs fields.
29872 * config/gcn/gcn.cc (gcn_option_override): Remove stack_size_opt.
29873 (default_requested_args): Remove PRIVATE_SEGMENT_BUFFER_ARG and
29874 PRIVATE_SEGMENT_WAVE_OFFSET_ARG.
29875 (gcn_addr_space_convert): Mask the QUEUE_PTR_ARG content.
29876 (gcn_expand_prologue): Move the TARGET_PACKED_WORK_ITEMS to the top.
29877 Set up the stacks from the values in the kernargs, not private.
29878 (gcn_expand_builtin_1): Match the stack configuration in the prologue.
29879 (gcn_hsa_declare_function_name): Turn off the private segment.
29880 (gcn_conditional_register_usage): Ensure QUEUE_PTR is fixed.
29881 * config/gcn/gcn.h (FIXED_REGISTERS): Fix the QUEUE_PTR register.
29882 * config/gcn/gcn.opt (mstack-size): Change the description.
29884 2023-02-02 Andre Vieira <andre.simoesdiasvieira@arm.com>
29887 * config/arm/arm.h (VALID_MVE_PRED_MODE): Add V2QI.
29888 * config/arm/arm.cc (thumb2_legitimate_address_p): Use HImode for
29889 addressing MVE predicate modes.
29890 (mve_bool_vec_to_const): Change to represent correct MVE predicate
29892 (arm_hard_regno_mode_ok): Use VALID_MVE_PRED_MODE instead of checking
29894 (arm_vector_mode_supported_p): Likewise.
29895 (arm_mode_to_pred_mode): Add V2QI.
29896 * config/arm/arm-builtins.cc (UNOP_PRED_UNONE_QUALIFIERS): New
29898 (UNOP_PRED_PRED_QUALIFIERS): New qualifier
29899 (BINOP_PRED_UNONE_PRED_QUALIFIERS): New qualifier.
29900 (v2qi_UP): New macro.
29901 (v4bi_UP): New macro.
29902 (v8bi_UP): New macro.
29903 (v16bi_UP): New macro.
29904 (arm_expand_builtin_args): Make it able to expand the new predicate
29906 * config/arm/arm-modes.def (V2QI): New mode.
29907 * config/arm/arm-simd-builtin-types.def (Pred1x16_t, Pred2x8_t
29908 Pred4x4_t): Remove unused predicate builtin types.
29909 * config/arm/arm_mve.h (__arm_vctp16q, __arm_vctp32q, __arm_vctp64q,
29910 __arm_vctp8q, __arm_vpnot, __arm_vctp8q_m, __arm_vctp64q_m,
29911 __arm_vctp32q_m, __arm_vctp16q_m): Use predicate modes.
29912 * config/arm/arm_mve_builtins.def (vctp16q, vctp32q, vctp64q, vctp8q,
29913 vpnot, vctp8q_m, vctp16q_m, vctp32q_m, vctp64q_m): Likewise.
29914 * config/arm/constraints.md (DB): Check for VALID_MVE_PRED_MODE instead
29915 of MODE_VECTOR_BOOL.
29916 * config/arm/iterators.md (MVE_7, MVE_7_HI): Add V2QI
29917 (MVE_VPRED): Likewise.
29918 (MVE_vpred): Add V2QI and map upper case predicate modes to lower case.
29919 (MVE_vctp): New mode attribute.
29923 * config/arm/mve.md (mve_vctp<mode1>qhi): Rename this...
29924 (mve_vctp<MVE_vctp>q<MVE_vpred>): ... to this. And use new mode
29926 (mve_vpnothi): Rename this...
29927 (mve_vpnotv16bi): ... to this.
29928 (mve_vctp<mode1>q_mhi): Rename this...
29929 (mve_vctp<MVE_vctp>q_m<MVE_vpred>):... to this.
29930 (mve_vldrdq_gather_base_z_<supf>v2di,
29931 mve_vldrdq_gather_offset_z_<supf>v2di,
29932 mve_vldrdq_gather_shifted_offset_z_<supf>v2di,
29933 mve_vstrdq_scatter_base_p_<supf>v2di,
29934 mve_vstrdq_scatter_offset_p_<supf>v2di,
29935 mve_vstrdq_scatter_offset_p_<supf>v2di_insn,
29936 mve_vstrdq_scatter_shifted_offset_p_<supf>v2di,
29937 mve_vstrdq_scatter_shifted_offset_p_<supf>v2di_insn,
29938 mve_vstrdq_scatter_base_wb_p_<supf>v2di,
29939 mve_vldrdq_gather_base_wb_z_<supf>v2di,
29940 mve_vldrdq_gather_base_nowb_z_<supf>v2di,
29941 mve_vldrdq_gather_base_wb_z_<supf>v2di_insn): Use V2QI insead of HI for
29943 * config/arm/unspecs.md (VCTP8Q, VCTP16Q, VCTP32Q, VCTP64Q): Replace
29945 (VCTP): ... with this.
29946 (VCTP8Q_M, VCTP16Q_M, VCTP32Q_M, VCTP64Q_M): Replace these...
29947 (VCTP_M): ... with this.
29948 * config/arm/vfp.md (*thumb2_movhi_vfp, *thumb2_movhi_fp16): Use
29949 VALID_MVE_PRED_MODE instead of checking for MODE_VECTOR_BOOL class.
29951 2023-02-02 Andre Vieira <andre.simoesdiasvieira@arm.com>
29954 * config/arm/arm.cc (arm_hard_regno_mode_ok): Use new MACRO.
29955 (arm_modes_tieable_p): Make MVE predicate modes tieable.
29956 * config/arm/arm.h (VALID_MVE_PRED_MODE): New define.
29957 * simplify-rtx.cc (simplify_context::simplify_subreg): Teach
29958 simplify_subreg to simplify subregs where the outermode is not scalar.
29960 2023-02-02 Andre Vieira <andre.simoesdiasvieira@arm.com>
29963 * config/arm/arm-builtins.cc (arm_simd_builtin_type): Rewrite to use
29964 new qualifiers parameter and use unsigned short type for MVE predicate.
29965 (arm_init_builtin): Call arm_simd_builtin_type with qualifiers
29967 (arm_init_crypto_builtins): Likewise.
29969 2023-02-02 Jakub Jelinek <jakub@redhat.com>
29972 * builtins.def (BUILT_IN_UNREACHABLE_TRAP): New builtin.
29973 * internal-fn.def (TRAP): Remove.
29974 * internal-fn.cc (expand_TRAP): Remove.
29975 * tree.cc (build_common_builtin_nodes): Define
29976 BUILT_IN_UNREACHABLE_TRAP if not yet defined.
29977 (builtin_decl_unreachable): Use BUILT_IN_UNREACHABLE_TRAP
29978 instead of BUILT_IN_TRAP.
29979 * gimple.cc (gimple_build_builtin_unreachable): Remove
29980 emitting internal function for BUILT_IN_TRAP.
29981 * asan.cc (maybe_instrument_call): Handle BUILT_IN_UNREACHABLE_TRAP.
29982 * cgraph.cc (cgraph_edge::verify_corresponds_to_fndecl): Handle
29983 BUILT_IN_UNREACHABLE_TRAP instead of BUILT_IN_TRAP.
29984 * ipa-devirt.cc (possible_polymorphic_call_target_p): Handle
29985 BUILT_IN_UNREACHABLE_TRAP.
29986 * builtins.cc (expand_builtin, is_inexpensive_builtin): Likewise.
29987 * tree-cfg.cc (verify_gimple_call,
29988 pass_warn_function_return::execute): Likewise.
29989 * attribs.cc (decl_attributes): Don't report exclusions on
29990 BUILT_IN_UNREACHABLE_TRAP either.
29992 2023-02-02 liuhongt <hongtao.liu@intel.com>
29994 PR tree-optimization/108601
29995 * tree-vectorizer.h (vect_can_peel_nonlinear_iv_p): Removed.
29996 * tree-vect-loop.cc
29997 (vectorizable_nonlinear_induction): Remove
29998 vect_can_peel_nonlinear_iv_p.
29999 (vect_can_peel_nonlinear_iv_p): Don't peel
30000 nonlinear iv(mult or shift) for epilog when vf is not
30001 constant and moved the defination to ..
30002 * tree-vect-loop-manip.cc (vect_can_peel_nonlinear_iv_p):
30005 2023-02-02 Jakub Jelinek <jakub@redhat.com>
30007 PR middle-end/108435
30008 * tree-nested.cc (convert_nonlocal_omp_clauses)
30009 <case OMP_CLAUSE_LASTPRIVATE>: If info->new_local_var_chain and *seq
30010 is not a GIMPLE_BIND, wrap the sequence into a new GIMPLE_BIND
30011 before calling declare_vars.
30012 (convert_nonlocal_omp_clauses) <case OMP_CLAUSE_LINEAR>: Merge
30013 with the OMP_CLAUSE_LASTPRIVATE handling except for whether
30014 seq is initialized to &OMP_CLAUSE_LASTPRIVATE_GIMPLE_SEQ (clause)
30015 or &OMP_CLAUSE_LINEAR_GIMPLE_SEQ (clause).
30017 2023-02-01 Tamar Christina <tamar.christina@arm.com>
30019 * common/config/aarch64/aarch64-common.cc
30020 (struct aarch64_option_extension): Add native_detect and document struct
30022 (all_extensions): Set new field native_detect.
30023 * config/aarch64/aarch64.cc (struct aarch64_option_extension): Delete
30026 2023-02-01 Martin Liska <mliska@suse.cz>
30028 * ipa-devirt.cc (odr_types_equivalent_p): Respect *warned
30031 2023-02-01 Andrew MacLeod <amacleod@redhat.com>
30033 PR tree-optimization/108356
30034 * gimple-range-cache.cc (ranger_cache::range_on_edge): Always
30035 do a search of the DOM tree for a range.
30037 2023-02-01 Martin Liska <mliska@suse.cz>
30040 * cgraphunit.cc (walk_polymorphic_call_targets): Insert
30041 ony non-null values.
30042 * ipa.cc (walk_polymorphic_call_targets): Likewise.
30044 2023-02-01 Martin Liska <mliska@suse.cz>
30047 * gcc.cc (LINK_COMPRESS_DEBUG_SPEC): Report error only for
30050 2023-02-01 Jakub Jelinek <jakub@redhat.com>
30053 * ree.cc (combine_reaching_defs): Don't return false for paradoxical
30054 subregs in DEBUG_INSNs.
30056 2023-02-01 Richard Sandiford <richard.sandiford@arm.com>
30058 * compare-elim.cc (find_flags_uses_in_insn): Guard use of SET_SRC.
30060 2023-02-01 Andreas Krebbel <krebbel@linux.ibm.com>
30062 * config/s390/s390.cc (s390_restore_gpr_p): New function.
30063 (s390_preserve_gpr_arg_in_range_p): New function.
30064 (s390_preserve_gpr_arg_p): New function.
30065 (s390_preserve_fpr_arg_p): New function.
30066 (s390_register_info_stdarg_fpr): Rename to ...
30067 (s390_register_info_arg_fpr): ... this. Add -mpreserve-args handling.
30068 (s390_register_info_stdarg_gpr): Rename to ...
30069 (s390_register_info_arg_gpr): ... this. Add -mpreserve-args handling.
30070 (s390_register_info): Use the renamed functions above.
30071 (s390_optimize_register_info): Likewise.
30072 (save_fpr): Generate CFI for -mpreserve-args.
30073 (save_gprs): Generate CFI for -mpreserve-args. Drop return value.
30074 (s390_emit_prologue): Adjust to changed calling convention of save_gprs.
30075 (s390_optimize_prologue): Likewise.
30076 * config/s390/s390.opt: New option -mpreserve-args
30078 2023-02-01 Andreas Krebbel <krebbel@linux.ibm.com>
30080 * config/s390/s390.cc (save_gprs): Use gen_frame_mem.
30081 (restore_gprs): Likewise.
30082 (s390_emit_stack_tie): Make the stack_tie to be dependent on the
30083 frame pointer if a frame-pointer is used.
30084 (s390_emit_prologue): Emit stack_tie when frame-pointer is needed.
30085 * config/s390/s390.md (stack_tie): Add a register operand and
30087 (@stack_tie<mode>): ... this.
30089 2023-02-01 Andreas Krebbel <krebbel@linux.ibm.com>
30091 * dwarf2cfi.cc (dwarf2out_frame_debug_cfa_restore): Add
30092 EMIT_CFI parameter.
30093 (dwarf2out_frame_debug): Add case for REG_CFA_NORESTORE.
30094 * reg-notes.def (REG_CFA_NOTE): New reg note definition.
30096 2023-02-01 Richard Biener <rguenther@suse.de>
30098 PR middle-end/108500
30099 * dominance.cc (assign_dfs_numbers): Replace recursive DFS
30100 with tree traversal algorithm.
30102 2023-02-01 Jason Merrill <jason@redhat.com>
30104 * doc/invoke.texi: Document -Wno-changes-meaning.
30106 2023-02-01 David Malcolm <dmalcolm@redhat.com>
30108 * doc/invoke.texi (Static Analyzer Options): Add notes about
30109 limitations of -fanalyzer.
30111 2023-01-31 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30113 * config/riscv/constraints.md (vj): New.
30115 * config/riscv/iterators.md: Add more opcode.
30116 * config/riscv/predicates.md (vector_arith_operand): New.
30117 (vector_neg_arith_operand): New.
30118 (vector_shift_operand): New.
30119 * config/riscv/riscv-vector-builtins-bases.cc (class binop): New.
30120 * config/riscv/riscv-vector-builtins-bases.h: (vadd): New.
30137 * config/riscv/riscv-vector-builtins-functions.def (vadd): New.
30154 * config/riscv/riscv-vector-builtins-shapes.cc (struct binop_def): New.
30155 * config/riscv/riscv-vector-builtins-shapes.h (binop): New.
30156 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_I_OPS): New.
30157 (DEF_RVV_U_OPS): New.
30158 (rvv_arg_type_info::get_base_vector_type): Handle
30159 RVV_BASE_shift_vector.
30160 (rvv_arg_type_info::get_tree_type): Ditto.
30161 * config/riscv/riscv-vector-builtins.h (enum rvv_base_type): Add
30162 RVV_BASE_shift_vector.
30163 * config/riscv/riscv.cc (riscv_print_operand): Handle 'V'.
30164 * config/riscv/vector-iterators.md: Handle more opcode.
30165 * config/riscv/vector.md (@pred_<optab><mode>): New.
30167 2023-01-31 Philipp Tomsich <philipp.tomsich@vrull.eu>
30170 * config/aarch64/aarch64.cc (aarch_macro_fusion_pair_p): Check
30173 2023-01-31 Richard Sandiford <richard.sandiford@arm.com>
30175 PR tree-optimization/108608
30176 * tree-vect-loop.cc (vect_transform_reduction): Handle single
30177 def-use cycles that involve function calls rather than tree codes.
30179 2023-01-31 Andrew MacLeod <amacleod@redhat.com>
30181 PR tree-optimization/108385
30182 * gimple-range-gori.cc (gori_compute::compute_operand_range):
30183 Allow VARYING computations to continue if there is a relation.
30184 * range-op.cc (pointer_plus_operator::op2_range): New.
30186 2023-01-31 Andrew MacLeod <amacleod@redhat.com>
30188 PR tree-optimization/108359
30189 * range-op.cc (range_operator::wi_fold_in_parts_equiv): New.
30190 (range_operator::fold_range): If op1 is equivalent to op2 then
30191 invoke new fold_in_parts_equiv to operate on sub-components.
30192 * range-op.h (wi_fold_in_parts_equiv): New prototype.
30194 2023-01-31 Andrew MacLeod <amacleod@redhat.com>
30196 * gimple-range-gori.cc (gori_compute::compute_operand_range): Do
30197 not abort calculations if there is a valid relation available.
30198 (gori_compute::refine_using_relation): Pass correct relation trio.
30199 (gori_compute::compute_operand1_range): Create trio and use it.
30200 (gori_compute::compute_operand2_range): Ditto.
30201 * range-op.cc (operator_plus::op1_range): Use correct trio member.
30202 (operator_minus::op1_range): Use correct trio member.
30203 * value-relation.cc (value_relation::create_trio): New.
30204 * value-relation.h (value_relation::create_trio): New prototype.
30206 2023-01-31 Jakub Jelinek <jakub@redhat.com>
30209 * config/i386/i386-expand.cc
30210 (ix86_convert_const_wide_int_to_broadcast): Return nullptr if
30211 CONST_WIDE_INT_NUNITS (op) times HOST_BITS_PER_WIDE_INT isn't
30212 equal to bitsize of mode.
30214 2023-01-31 Jakub Jelinek <jakub@redhat.com>
30216 PR rtl-optimization/108596
30217 * bb-reorder.cc (fix_up_fall_thru_edges): Handle the case where cur_bb
30218 ends with asm goto and has a crossing fallthrough edge to the same bb
30219 that contains at least one of its labels by restoring EDGE_CROSSING
30220 flag even on possible edge from cur_bb to new_bb successor.
30222 2023-01-31 Jakub Jelinek <jakub@redhat.com>
30225 * config/i386/avx512erintrin.h (_mm512_exp2a23_round_pd,
30226 _mm512_exp2a23_round_ps, _mm512_rcp28_round_pd, _mm512_rcp28_round_ps,
30227 _mm512_rsqrt28_round_pd, _mm512_rsqrt28_round_ps): Use
30228 _mm512_undefined_pd () or _mm512_undefined_ps () instead of using
30229 uninitialized automatic variable __W.
30231 2023-01-31 Gerald Pfeifer <gerald@pfeifer.com>
30233 * doc/include/fdl.texi: Change fsf.org to www.fsf.org.
30235 2023-01-30 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30237 * config/riscv/riscv-protos.h (get_vector_mode): New function.
30238 * config/riscv/riscv-v.cc (get_vector_mode): Ditto.
30239 * config/riscv/riscv-vector-builtins-bases.cc (enum lst_type): New enum.
30240 (class loadstore): Adjust for indexed loads/stores support.
30242 * config/riscv/riscv-vector-builtins-bases.h: New function declare.
30243 * config/riscv/riscv-vector-builtins-functions.def (vluxei8): Ditto.
30259 * config/riscv/riscv-vector-builtins-shapes.cc
30260 (struct indexed_loadstore_def): New class.
30262 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
30263 * config/riscv/riscv-vector-builtins.cc (required_extensions_p): Adjust
30264 for indexed loads/stores support.
30265 (check_required_extensions): Ditto.
30266 (rvv_arg_type_info::get_base_vector_type): New function.
30267 (rvv_arg_type_info::get_tree_type): Ditto.
30268 (function_builder::add_unique_function): Adjust for indexed loads/stores
30270 (function_expander::use_exact_insn): New function.
30271 * config/riscv/riscv-vector-builtins.h (enum rvv_base_type): Adjust for
30272 indexed loads/stores support.
30273 (struct rvv_arg_type_info): Ditto.
30274 (function_expander::index_mode): New function.
30275 (function_base::apply_tail_policy_p): Ditto.
30276 (function_base::apply_mask_policy_p): Ditto.
30277 * config/riscv/vector-iterators.md (unspec): New unspec.
30278 * config/riscv/vector.md (unspec): Ditto.
30279 (@pred_indexed_<order>load<VNX1_QHSD:mode><VNX1_QHSDI:mode>): New
30281 (@pred_indexed_<order>store<VNX1_QHSD:mode><VNX1_QHSDI:mode>): Ditto.
30282 (@pred_indexed_<order>load<VNX2_QHSD:mode><VNX2_QHSDI:mode>): Ditto.
30283 (@pred_indexed_<order>store<VNX2_QHSD:mode><VNX2_QHSDI:mode>): Ditto.
30284 (@pred_indexed_<order>load<VNX4_QHSD:mode><VNX4_QHSDI:mode>): Ditto.
30285 (@pred_indexed_<order>store<VNX4_QHSD:mode><VNX4_QHSDI:mode>): Ditto.
30286 (@pred_indexed_<order>load<VNX8_QHSD:mode><VNX8_QHSDI:mode>): Ditto.
30287 (@pred_indexed_<order>store<VNX8_QHSD:mode><VNX8_QHSDI:mode>): Ditto.
30288 (@pred_indexed_<order>load<VNX16_QHS:mode><VNX16_QHSI:mode>): Ditto.
30289 (@pred_indexed_<order>store<VNX16_QHS:mode><VNX16_QHSI:mode>): Ditto.
30290 (@pred_indexed_<order>load<VNX32_QH:mode><VNX32_QHI:mode>): Ditto.
30291 (@pred_indexed_<order>store<VNX32_QH:mode><VNX32_QHI:mode>): Ditto.
30292 (@pred_indexed_<order>load<VNX64_Q:mode><VNX64_Q:mode>): Ditto.
30293 (@pred_indexed_<order>store<VNX64_Q:mode><VNX64_Q:mode>): Ditto.
30295 2023-01-30 Flavio Cruz <flaviocruz@gmail.com>
30297 * config.gcc: Recognize x86_64-*-gnu* targets and include
30299 * config/i386/gnu64.h: Define configuration for new target
30300 including ld.so location.
30302 2023-01-30 Philipp Tomsich <philipp.tomsich@vrull.eu>
30304 * config/aarch64/aarch64-cores.def (AARCH64_CORE): Update
30305 ampere1a to include SM4.
30307 2023-01-30 Andrew Pinski <apinski@marvell.com>
30309 PR tree-optimization/108582
30310 * tree-ssa-phiopt.cc (match_simplify_replacement): Add check
30311 for middlebb to have no phi nodes.
30313 2023-01-30 Richard Biener <rguenther@suse.de>
30315 PR tree-optimization/108574
30316 * tree-ssa-sccvn.cc (visit_phi): Instead of swapping
30317 sameval and def, ignore the equivalence if there's the
30318 danger of oscillating between two values.
30320 2023-01-30 Andreas Schwab <schwab@suse.de>
30322 * common/config/riscv/riscv-common.cc
30323 (riscv_option_optimization_table)
30324 [TARGET_DEFAULT_ASYNC_UNWIND_TABLES]: Enable
30325 -fasynchronous-unwind-tables and -funwind-tables.
30326 * config.gcc (riscv*-*-linux*): Define
30327 TARGET_DEFAULT_ASYNC_UNWIND_TABLES.
30329 2023-01-30 YunQiang Su <yunqiang.su@cipunited.com>
30331 * Makefile.in (CROSS_SYSTEM_HEADER_DIR): set according the
30332 value of includedir.
30334 2023-01-30 Richard Biener <rguenther@suse.de>
30337 * cgraph.cc (possibly_call_in_translation_unit_p): Relax
30340 2023-01-30 liuhongt <hongtao.liu@intel.com>
30342 * config/i386/i386.opt: Change AVX512FP16 to AVX512-FP16.
30343 * doc/invoke.texi: Ditto.
30345 2023-01-29 Jan Hubicka <hubicka@ucw.cz>
30347 * ipa-utils.cc: Include calls.h, cfgloop.h and cfganal.h
30348 (stmt_may_terminate_function_p): If assuming return or EH
30349 volatile asm is safe.
30350 (find_always_executed_bbs): Fix handling of terminating BBS and
30351 infinite loops; add debug output.
30352 * tree-ssa-alias.cc (stmt_kills_ref_p): Fix debug output
30354 2023-01-28 Philipp Tomsich <philipp.tomsich@vrull.eu>
30356 * config/aarch64/aarch64.cc (aarch64_uxt_size): fix an
30357 off-by-one in checking the permissible shift-amount.
30359 2023-01-28 Gerald Pfeifer <gerald@pfeifer.com>
30361 * doc/extend.texi (Named Address Spaces): Update link to the
30364 2023-01-28 Gerald Pfeifer <gerald@pfeifer.com>
30366 * doc/standards.texi (Standards): Fix markup.
30368 2023-01-28 Gerald Pfeifer <gerald@pfeifer.com>
30370 * doc/standards.texi (Standards): Update link to Objective-C book.
30372 2023-01-28 Gerald Pfeifer <gerald@pfeifer.com>
30374 * doc/invoke.texi (Instrumentation Options): Update reference to
30377 2023-01-28 Gerald Pfeifer <gerald@pfeifer.com>
30379 * doc/standards.texi: Update Go1 link.
30381 2023-01-28 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30383 * config/riscv/predicates.md (pmode_reg_or_0_operand): New predicate.
30384 * config/riscv/riscv-vector-builtins-bases.cc (class loadstore):
30387 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
30388 * config/riscv/riscv-vector-builtins-functions.def (vlse): New class.
30390 * config/riscv/riscv-vector-builtins.cc
30391 (function_expander::use_contiguous_load_insn): Support vlse/vsse.
30392 * config/riscv/vector.md (@pred_strided_load<mode>): New md pattern.
30393 (@pred_strided_store<mode>): Ditto.
30395 2023-01-28 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30397 * config/riscv/vector.md (tail_policy_op_idx): Remove.
30398 (mask_policy_op_idx): Remove.
30399 (avl_type_op_idx): Remove.
30401 2023-01-27 Richard Sandiford <richard.sandiford@arm.com>
30403 PR tree-optimization/96373
30404 * tree.h (sign_mask_for): Declare.
30405 * tree.cc (sign_mask_for): New function.
30406 (signed_or_unsigned_type_for): For vector types, try to use the
30407 related_int_vector_mode.
30408 * genmatch.cc (commutative_op): Handle conditional internal functions.
30409 * match.pd: Fold an IFN_COND_MUL+copysign into an IFN_COND_XOR+and.
30411 2023-01-27 Richard Sandiford <richard.sandiford@arm.com>
30413 * tree-vectorizer.cc (vector_costs::compare_inside_loop_cost):
30414 Use the likely minimum VF when bounding the denominators to
30415 the estimated number of iterations.
30417 2023-01-27 Richard Biener <rguenther@suse.de>
30420 * doc/invoke.texi (-shared): Clarify effect on -ffast-math
30421 and -Ofast FP environment side-effects.
30423 2023-01-27 Richard Biener <rguenther@suse.de>
30426 * config/mips/gnu-user.h (GNU_USER_TARGET_MATHFILE_SPEC):
30427 Don't add crtfastmath.o for -shared.
30429 2023-01-27 Richard Biener <rguenther@suse.de>
30432 * config/ia64/linux.h (ENDFILE_SPEC): Don't add crtfastmath.o
30435 2023-01-27 Richard Biener <rguenther@suse.de>
30438 * config/alpha/linux.h (ENDFILE_SPEC): Don't add
30439 crtfastmath.o for -shared.
30441 2023-01-27 Andrew MacLeod <amacleod@redhat.com>
30443 PR tree-optimization/108306
30444 * range-op.cc (operator_lshift::fold_range): Return [0, 0] not
30445 varying for shifts that are always out of void range.
30446 (operator_rshift::fold_range): Return [0, 0] not
30447 varying for shifts that are always out of void range.
30449 2023-01-27 Andrew MacLeod <amacleod@redhat.com>
30451 PR tree-optimization/108447
30452 * gimple-range-fold.cc (old_using_range::relation_fold_and_or):
30453 Do not attempt to fold HONOR_NAN types.
30455 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30457 * config/riscv/riscv-vector-builtins-shapes.cc (struct loadstore_def):
30458 Remove _m suffix for "vop_m" C++ overloaded API name.
30460 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30462 * config/riscv/riscv-vector-builtins-bases.cc (BASE): Add vlm/vsm support.
30463 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
30464 * config/riscv/riscv-vector-builtins-functions.def (vlm): New define.
30466 * config/riscv/riscv-vector-builtins-shapes.cc (struct loadstore_def): Add vlm/vsm support.
30467 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_B_OPS): Ditto.
30468 (vbool64_t): Ditto.
30469 (vbool32_t): Ditto.
30470 (vbool16_t): Ditto.
30475 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_B_OPS): Ditto.
30476 (rvv_arg_type_info::get_tree_type): Ditto.
30477 (function_expander::use_contiguous_load_insn): Ditto.
30478 * config/riscv/vector.md (@pred_store<mode>): Ditto.
30480 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30482 * config/riscv/riscv-vsetvl.cc (vsetvl_insn_p): Add condition to avoid ICE.
30483 (vsetvl_discard_result_insn_p): New function.
30484 (reg_killed_by_bb_p): rename to find_reg_killed_by.
30485 (find_reg_killed_by): New name.
30486 (get_vl): allow it to be called by more functions.
30487 (has_vsetvl_killed_avl_p): Add condition.
30488 (get_avl): allow it to be called by more functions.
30489 (insn_should_be_added_p): New function.
30490 (get_all_nonphi_defs): Refine function.
30491 (get_all_sets): Ditto.
30492 (get_same_bb_set): New function.
30493 (any_insn_in_bb_p): Ditto.
30494 (any_set_in_bb_p): Ditto.
30495 (get_vl_vtype_info): Add VLMAX forward optimization.
30496 (source_equal_p): Fix issues.
30497 (extract_single_source): Refine.
30498 (avl_info::multiple_source_equal_p): New function.
30499 (avl_info::operator==): Adjust for final version.
30500 (vl_vtype_info::operator==): Ditto.
30501 (vl_vtype_info::same_avl_p): Ditto.
30502 (vector_insn_info::parse_insn): Ditto.
30503 (vector_insn_info::available_p): New function.
30504 (vector_insn_info::merge): Adjust for final version.
30505 (vector_insn_info::dump): Add hard_empty.
30506 (pass_vsetvl::hard_empty_block_p): New function.
30507 (pass_vsetvl::backward_demand_fusion): Adjust for final version.
30508 (pass_vsetvl::forward_demand_fusion): Ditto.
30509 (pass_vsetvl::demand_fusion): Ditto.
30510 (pass_vsetvl::cleanup_illegal_dirty_blocks): New function.
30511 (pass_vsetvl::compute_local_properties): Adjust for final version.
30512 (pass_vsetvl::can_refine_vsetvl_p): Ditto.
30513 (pass_vsetvl::refine_vsetvls): Ditto.
30514 (pass_vsetvl::commit_vsetvls): Ditto.
30515 (pass_vsetvl::propagate_avl): New function.
30516 (pass_vsetvl::lazy_vsetvl): Adjust for new version.
30517 * config/riscv/riscv-vsetvl.h (enum def_type): New enum.
30519 2023-01-27 Jakub Jelinek <jakub@redhat.com>
30522 * doc/extend.texi: Fix up return type of __builtin_va_arg_pack_len
30523 from size_t to int.
30525 2023-01-27 Jakub Jelinek <jakub@redhat.com>
30528 * cgraph.cc (cgraph_edge::verify_corresponds_to_fndecl): Allow
30529 redirection of calls to __builtin_trap in addition to redirection
30530 to __builtin_unreachable.
30532 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30534 * config/riscv/riscv-vsetvl.cc (before_p): Fix bug.
30536 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30538 * config/riscv/riscv-vsetvl.cc (gen_vsetvl_pat): Refine function args.
30539 (emit_vsetvl_insn): Ditto.
30541 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30543 * config/riscv/vector.md: Fix constraints.
30545 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30547 * config/riscv/vector-iterators.md: Add TARGET_MIN_VLEN > 32 predicates.
30549 2023-01-27 Patrick Palka <ppalka@redhat.com>
30550 Jakub Jelinek <jakub@redhat.com>
30552 * tree-core.h (tree_code_type, tree_code_length): For
30553 C++17 and later, add inline keyword, otherwise don't define
30554 the arrays, but declare extern arrays.
30555 * tree.cc (tree_code_type, tree_code_length): Define these
30556 arrays for C++14 and older.
30558 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30560 * config/riscv/riscv-vsetvl.h: Change it into public.
30562 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30564 * config/riscv/riscv-passes.def (INSERT_PASS_BEFORE): Reorder VSETVL
30567 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30569 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::execute): Always call split_all_insns.
30571 2023-01-27 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30573 * config/riscv/vector.md: Fix incorrect attributes.
30575 2023-01-27 Richard Biener <rguenther@suse.de>
30578 * config/loongarch/gnu-user.h (GNU_USER_TARGET_MATHFILE_SPEC):
30579 Don't add crtfastmath.o for -shared.
30581 2023-01-27 Alexandre Oliva <oliva@gnu.org>
30583 * doc/options.texi (option, RejectNegative): Mention that
30584 -g-started options are also implicitly negatable.
30586 2023-01-26 Kito Cheng <kito.cheng@sifive.com>
30588 * config/riscv/riscv-vector-builtins.cc (register_builtin_types):
30589 Use get_typenode_from_name to get fixed-width integer type
30591 * config/riscv/riscv-vector-builtins.def: Update define with
30592 fixed-width integer type nodes.
30594 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30596 * config/riscv/riscv-vsetvl.cc (same_bb_and_before_p): Remove it.
30597 (real_insn_and_same_bb_p): New function.
30598 (same_bb_and_after_or_equal_p): Remove it.
30599 (before_p): New function.
30600 (reg_killed_by_bb_p): Ditto.
30601 (has_vsetvl_killed_avl_p): Ditto.
30602 (get_vl): Move location so that we can call it.
30603 (anticipatable_occurrence_p): Fix issue of AVL=REG support.
30604 (available_occurrence_p): Ditto.
30605 (dominate_probability_p): Remove it.
30606 (can_backward_propagate_p): Remove it.
30607 (get_all_nonphi_defs): New function.
30608 (get_all_predecessors): Ditto.
30609 (any_insn_in_bb_p): Ditto.
30610 (insert_vsetvl): Adjust AVL REG.
30611 (source_equal_p): New function.
30612 (extract_single_source): Ditto.
30613 (avl_info::single_source_equal_p): Ditto.
30614 (avl_info::operator==): Adjust for AVL=REG.
30615 (vl_vtype_info::same_avl_p): Ditto.
30616 (vector_insn_info::set_demand_info): Remove it.
30617 (vector_insn_info::compatible_p): Adjust for AVL=REG.
30618 (vector_insn_info::compatible_avl_p): New function.
30619 (vector_insn_info::merge): Adjust AVL=REG.
30620 (vector_insn_info::dump): Ditto.
30621 (pass_vsetvl::merge_successors): Remove it.
30622 (enum fusion_type): New enum.
30623 (pass_vsetvl::get_backward_fusion_type): New function.
30624 (pass_vsetvl::backward_demand_fusion): Adjust for AVL=REG.
30625 (pass_vsetvl::forward_demand_fusion): Ditto.
30626 (pass_vsetvl::demand_fusion): Ditto.
30627 (pass_vsetvl::prune_expressions): Ditto.
30628 (pass_vsetvl::compute_local_properties): Ditto.
30629 (pass_vsetvl::cleanup_vsetvls): Ditto.
30630 (pass_vsetvl::commit_vsetvls): Ditto.
30631 (pass_vsetvl::init): Ditto.
30632 * config/riscv/riscv-vsetvl.h (enum fusion_type): New enum.
30633 (enum merge_type): New enum.
30635 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30637 * config/riscv/riscv-vsetvl.cc
30638 (vector_infos_manager::vector_infos_manager): Add probability.
30639 (vector_infos_manager::dump): Ditto.
30640 (pass_vsetvl::compute_probabilities): Ditto.
30641 * config/riscv/riscv-vsetvl.h (struct vector_block_info): Ditto.
30643 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30645 * config/riscv/riscv-vsetvl.cc (vector_insn_info::operator==): Remove dirty_pat.
30646 (vector_insn_info::merge): Ditto.
30647 (vector_insn_info::dump): Ditto.
30648 (pass_vsetvl::merge_successors): Ditto.
30649 (pass_vsetvl::backward_demand_fusion): Ditto.
30650 (pass_vsetvl::forward_demand_fusion): Ditto.
30651 (pass_vsetvl::commit_vsetvls): Ditto.
30652 * config/riscv/riscv-vsetvl.h: Ditto.
30654 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30656 * config/riscv/riscv-vsetvl.cc (add_label_notes): Rename insn to
30659 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30661 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::backward_demand_fusion): Refine codes.
30663 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30665 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::forward_demand_fusion):
30666 Add pre-check for redundant flow.
30668 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30670 * config/riscv/riscv-vsetvl.cc (vector_infos_manager::create_bitmap_vectors): New function.
30671 (vector_infos_manager::free_bitmap_vectors): Ditto.
30672 (pass_vsetvl::pre_vsetvl): Adjust codes.
30673 * config/riscv/riscv-vsetvl.h: New function declaration.
30675 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30677 * config/riscv/riscv-vsetvl.cc (can_backward_propagate_p): Fix for null iter_bb.
30678 (vector_insn_info::set_demand_info): New function.
30679 (pass_vsetvl::emit_local_forward_vsetvls): Adjust for refinement of Phase 3.
30680 (pass_vsetvl::merge_successors): Ditto.
30681 (pass_vsetvl::compute_global_backward_infos): Ditto.
30682 (pass_vsetvl::backward_demand_fusion): Ditto.
30683 (pass_vsetvl::forward_demand_fusion): Ditto.
30684 (pass_vsetvl::demand_fusion): New function.
30685 (pass_vsetvl::lazy_vsetvl): Adjust for refinement of phase 3.
30686 * config/riscv/riscv-vsetvl.h: New function declaration.
30688 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30690 * config/riscv/riscv-vsetvl.cc (vector_insn_info::operator>=): Fix available condition.
30692 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30694 * config/riscv/riscv-vsetvl.cc (change_vsetvl_insn): New function.
30695 (pass_vsetvl::compute_global_backward_infos): Simplify codes.
30697 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30699 * config/riscv/riscv-vsetvl.cc (loop_basic_block_p): Adjust function.
30700 (backward_propagate_worthwhile_p): Fix non-worthwhile.
30702 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30704 * config/riscv/riscv-vsetvl.cc (change_insn): Adjust in_group in validate_change.
30706 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30708 * config/riscv/riscv-vsetvl.cc (vector_infos_manager::all_same_avl_p): New function.
30709 (pass_vsetvl::can_refine_vsetvl_p): Add AVL check.
30710 (pass_vsetvl::commit_vsetvls): Ditto.
30711 * config/riscv/riscv-vsetvl.h: New function declaration.
30713 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30715 * config/riscv/vector.md:
30717 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30719 * config/riscv/riscv-vector-builtins-bases.cc (class loadstore): use
30720 pred_store for vse.
30721 * config/riscv/riscv-vector-builtins.cc
30722 (function_expander::add_mem_operand): Refine function.
30723 (function_expander::use_contiguous_load_insn): Adjust new
30725 (function_expander::use_contiguous_store_insn): Ditto.
30726 * config/riscv/riscv-vector-builtins.h: Refine function.
30727 * config/riscv/vector.md (@pred_store<mode>): New pattern.
30729 2023-01-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
30731 * config/riscv/riscv-vector-builtins.cc: Change to scalar pointer.
30733 2023-01-26 Marek Polacek <polacek@redhat.com>
30735 PR middle-end/108543
30736 * opts.cc (parse_sanitizer_options): Don't always clear SANITIZE_ADDRESS
30737 if it was previously set.
30739 2023-01-26 Jakub Jelinek <jakub@redhat.com>
30741 PR tree-optimization/108540
30742 * range-op-float.cc (foperator_equal::fold_range): If both op1 and op2
30743 are singletons, use range_true even if op1 != op2
30744 when one range is [-0.0, -0.0] and another [0.0, 0.0]. Similarly,
30745 even if intersection of the ranges is empty and one has
30746 zero low bound and another zero high bound, use range_true_and_false
30747 rather than range_false.
30748 (foperator_not_equal::fold_range): If both op1 and op2
30749 are singletons, use range_false even if op1 != op2
30750 when one range is [-0.0, -0.0] and another [0.0, 0.0]. Similarly,
30751 even if intersection of the ranges is empty and one has
30752 zero low bound and another zero high bound, use range_true_and_false
30753 rather than range_true.
30755 2023-01-26 Jakub Jelinek <jakub@redhat.com>
30757 * value-relation.cc (kind_string): Add const.
30758 (rr_negate_table, rr_swap_table, rr_intersect_table,
30759 rr_union_table, rr_transitive_table): Add static const, change
30760 element type from relation_kind to unsigned char.
30761 (relation_negate, relation_swap, relation_intersect, relation_union,
30762 relation_transitive): Cast rr_*_table element to relation_kind.
30763 (relation_to_code): Add static const.
30764 (relation_tests): Assert VREL_LAST is smaller than UCHAR_MAX.
30766 2023-01-26 Richard Biener <rguenther@suse.de>
30768 PR tree-optimization/108547
30769 * gimple-predicate-analysis.cc (value_sat_pred_p):
30772 2023-01-26 Siddhesh Poyarekar <siddhesh@gotplt.org>
30774 PR tree-optimization/108522
30775 * tree-object-size.cc (compute_object_offset): Make EXPR
30776 argument non-const. Call component_ref_field_offset.
30778 2023-01-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
30780 * config/aarch64/aarch64-option-extensions.def (cssc): Specify
30781 FEATURE_STRING field.
30783 2023-01-26 Gerald Pfeifer <gerald@pfeifer.com>
30785 * doc/sourcebuild.texi: Refer to projects as GCC and GDB.
30787 2023-01-25 Iain Sandoe <iain@sandoe.co.uk>
30791 * gcc.cc: Provide default specs for Modula-2 so that when the
30792 language is not built-in better diagnostics are emitted for
30793 attempts to use .mod or .m2i file extensions.
30795 2023-01-25 Andrea Corallo <andrea.corallo@arm.com>
30797 * config/arm/mve.md (mve_vqnegq_s<mode>): Fix spacing.
30799 2023-01-25 Andrea Corallo <andrea.corallo@arm.com>
30801 * config/arm/mve.md (mve_vqabsq_s<mode>): Fix spacing.
30803 2023-01-25 Andrea Corallo <andrea.corallo@arm.com>
30805 * config/arm/mve.md (mve_vnegq_f<mode>, mve_vnegq_s<mode>):
30808 2023-01-25 Andrea Corallo <andrea.corallo@arm.com>
30810 * config/arm/mve.md (@mve_vclzq_s<mode>): Fix spacing.
30812 2023-01-25 Andrea Corallo <andrea.corallo@arm.com>
30814 * config/arm/mve.md (mve_vclsq_s<mode>): Fix spacing.
30816 2023-01-25 Richard Biener <rguenther@suse.de>
30818 PR tree-optimization/108523
30819 * tree-ssa-sccvn.cc (visit_phi): Avoid using the exclusive
30820 backedge value for the result when using predication to
30823 2023-01-25 Richard Biener <rguenther@suse.de>
30825 * doc/lto.texi (Command line options): Reword and update reference
30826 to removed lto_read_all_file_options.
30828 2023-01-25 Richard Sandiford <richard.sandiford@arm.com>
30830 * config/aarch64/aarch64.md (umax<mode>3): Separate the CNT and CSSC
30833 2023-01-25 Gerald Pfeifer <gerald@pfeifer.com>
30835 * doc/contrib.texi: Add Jose E. Marchesi.
30837 2023-01-25 Jakub Jelinek <jakub@redhat.com>
30839 PR tree-optimization/108498
30840 * gimple-ssa-store-merging.cc (class store_operand_info):
30841 End coment with full stop rather than comma.
30842 (split_group): Likewise.
30843 (merged_store_group::apply_stores): Clear string_concatenation if
30844 start or end aren't on a byte boundary.
30846 2023-01-25 Siddhesh Poyarekar <siddhesh@gotplt.org>
30847 Jakub Jelinek <jakub@redhat.com>
30849 PR tree-optimization/108522
30850 * tree-object-size.cc (compute_object_offset): Use
30851 TREE_OPERAND(ref, 2) for COMPONENT_REF when available.
30853 2023-01-24 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
30855 * config/xtensa/xtensa.md:
30856 Fix exit from loops detecting references before overwriting in the
30859 2023-01-24 Vladimir N. Makarov <vmakarov@redhat.com>
30861 * lra-constraints.cc (get_hard_regno): Remove final_p arg. Always
30862 do elimination but only for hard register.
30863 (operands_match_p, uses_hard_regs_p, process_alt_operands): Adjust
30864 calls of get_hard_regno.
30866 2023-01-24 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
30868 * config/s390/s390-d.cc (s390_d_target_versions): Fix detection
30871 2023-01-24 Andre Vieira <andre.simoesdiasvieira@arm.com>
30874 * config/arm/mve.md (mve_vstrbq_p_<supf><mode>, mve_vstrhq_p_fv8hf,
30875 mve_vstrhq_p_<supf><mode>, mve_vstrwq_p_<supf>v4si): Add memory operand
30878 2023-01-24 Xianmiao Qu <cooper.qu@linux.alibaba.com>
30880 * config.gcc(csky-*-linux*): Define CSKY_ENABLE_MULTILIB
30881 and only include 'csky/t-csky-linux' when enable multilib.
30882 * config/csky/csky-linux-elf.h(SYSROOT_SUFFIX_SPEC): Don't
30883 define it when disable multilib.
30885 2023-01-24 Richard Biener <rguenther@suse.de>
30887 PR tree-optimization/108500
30888 * dominance.h (calculate_dominance_info): Add parameter
30889 to indicate fast-query compute, defaulted to true.
30890 * dominance.cc (calculate_dominance_info): Honor
30891 fast-query compute parameter.
30892 * tree-cfgcleanup.cc (cleanup_tree_cfg_noloop): Do
30893 not compute the dominator fast-query DFS numbers.
30895 2023-01-24 Eric Biggers <ebiggers@google.com>
30898 * optc-save-gen.awk: Fix copy-and-paste error.
30900 2023-01-24 Jakub Jelinek <jakub@redhat.com>
30903 * cgraphbuild.cc: Include gimplify.h.
30904 (record_reference): Replace VAR_DECLs with DECL_HAS_VALUE_EXPR_P with
30905 their corresponding DECL_VALUE_EXPR expressions after unsharing.
30907 2023-01-24 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
30910 * config.gcc (tm_file): Move the variable out of loop.
30912 2023-01-24 Lulu Cheng <chenglulu@loongson.cn>
30913 Yang Yujie <yangyujie@loongson.cn>
30916 * config/loongarch/loongarch.cc (loongarch_classify_address):
30917 Add precessint for CONST_INT.
30918 (loongarch_print_operand_reloc): Operand modifier 'c' is supported.
30919 (loongarch_print_operand): Increase the processing of '%c'.
30920 * doc/extend.texi: Adds documents for LoongArch operand modifiers.
30921 And port the public operand modifiers information to this document.
30923 2023-01-23 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
30925 * doc/invoke.texi (-mbranch-protection): Update documentation.
30927 2023-01-23 Richard Biener <rguenther@suse.de>
30930 * config/sparc/freebsd.h (ENDFILE_SPEC): Don't add crtfastmath.o
30932 * config/sparc/linux.h (ENDFILE_SPEC): Likewise.
30933 * config/sparc/linux64.h (ENDFILE_SPEC): Likewise.
30934 * config/sparc/sp-elf.h (ENDFILE_SPEC): Likewise.
30935 * config/sparc/sp64-elf.h (ENDFILE_SPEC): Likewise.
30937 2023-01-23 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
30939 * config/arm/aout.h (ra_auth_code): Add entry in enum.
30940 * config/arm/arm.cc (emit_multi_reg_push): Add RA_AUTH_CODE register
30941 to dwarf frame expression.
30942 (arm_emit_multi_reg_pop): Restore RA_AUTH_CODE register.
30943 (arm_expand_prologue): Update frame related information and reg notes
30944 for pac/pacbit insn.
30945 (arm_regno_class): Check for pac pseudo reigster.
30946 (arm_dbx_register_number): Assign ra_auth_code register number in dwarf.
30947 (arm_init_machine_status): Set pacspval_needed to zero.
30948 (arm_debugger_regno): Check for PAC register.
30949 (arm_unwind_emit_sequence): Print .save directive with ra_auth_code
30951 (arm_unwind_emit_set): Add entry for IP_REGNUM in switch case.
30952 (arm_unwind_emit): Update REG_CFA_REGISTER case._
30953 * config/arm/arm.h (FIRST_PSEUDO_REGISTER): Modify.
30954 (DWARF_PAC_REGNUM): Define.
30955 (IS_PAC_REGNUM): Likewise.
30956 (enum reg_class): Add PAC_REG entry.
30957 (machine_function): Add pacbti_needed state to structure.
30958 * config/arm/arm.md (RA_AUTH_CODE): Define.
30960 2023-01-23 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
30962 * config.gcc ($tm_file): Update variable.
30963 * config/arm/arm-mlib.h: Create new header file.
30964 * config/arm/t-rmprofile (MULTI_ARCH_DIRS_RM): Rename mbranch-protection
30965 multilib arch directory.
30966 (MULTILIB_REUSE): Add multilib reuse rules.
30967 (MULTILIB_MATCHES): Add multilib match rules.
30969 2023-01-23 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
30971 * config/arm/arm-cpus.in (cortex-m85): Define new CPU.
30972 * config/arm/arm-tables.opt: Regenerate.
30973 * config/arm/arm-tune.md: Likewise.
30974 * doc/invoke.texi (Arm Options): Document -mcpu=cortex-m85.
30975 * (-mfix-cmse-cve-2021-35465): Likewise.
30977 2023-01-23 Richard Biener <rguenther@suse.de>
30979 PR tree-optimization/108482
30980 * tree-vect-generic.cc (expand_vector_operations): Fold remaining
30981 .LOOP_DIST_ALIAS calls.
30983 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
30985 * config.gcc (arm*-*-*): Add 'aarch-bti-insert.o' object.
30986 * config/arm/arm-protos.h: Update.
30987 * config/arm/aarch-common-protos.h: Declare
30988 'aarch_bti_arch_check'.
30989 * config/arm/arm.cc (aarch_bti_enabled) Update.
30990 (aarch_bti_j_insn_p, aarch_pac_insn_p, aarch_gen_bti_c)
30991 (aarch_gen_bti_j, aarch_bti_arch_check): New functions.
30992 * config/arm/arm.md (bti_nop): New insn.
30993 * config/arm/t-arm (PASSES_EXTRA): Add 'arm-passes.def'.
30994 (aarch-bti-insert.o): New target.
30995 * config/arm/unspecs.md (VUNSPEC_BTI_NOP): New unspec.
30996 * config/arm/aarch-bti-insert.cc (rest_of_insert_bti): Verify arch
30998 (gate): Make use of 'aarch_bti_arch_check'.
30999 * config/arm/arm-passes.def: New file.
31000 * config/aarch64/aarch64.cc (aarch_bti_arch_check): New function.
31002 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31004 * config.gcc (aarch64*-*-*): Rename 'aarch64-bti-insert.o' into
31005 'aarch-bti-insert.o'.
31006 * config/aarch64/aarch64-protos.h: Remove 'aarch64_bti_enabled'
31008 * config/aarch64/aarch64.cc (aarch_bti_enabled): Rename.
31009 (aarch_bti_j_insn_p, aarch_pac_insn_p): New functions.
31010 (aarch64_output_mi_thunk)
31011 (aarch64_print_patchable_function_entry)
31012 (aarch64_file_end_indicate_exec_stack): Update renamed function
31013 calls to renamed functions.
31014 * config/aarch64/aarch64-c.cc (aarch64_update_cpp_builtins): Likewise.
31015 * config/aarch64/t-aarch64 (aarch-bti-insert.o): Update
31017 * config/aarch64/aarch64-bti-insert.cc: Delete.
31018 * config/arm/aarch-bti-insert.cc: New file including and
31019 generalizing code from aarch64-bti-insert.cc.
31020 * config/arm/aarch-common-protos.h: Update.
31022 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31024 * config/arm/arm.h (arm_arch8m_main): Declare it.
31025 * config/arm/arm-protos.h (arm_current_function_pac_enabled_p):
31027 * config/arm/arm.cc (arm_arch8m_main): Define it.
31028 (arm_option_reconfigure_globals): Set arm_arch8m_main.
31029 (arm_compute_frame_layout, arm_expand_prologue)
31030 (thumb2_expand_return, arm_expand_epilogue)
31031 (arm_conditional_register_usage): Update for pac codegen.
31032 (arm_current_function_pac_enabled_p): New function.
31033 (aarch_bti_enabled) New function.
31034 (use_return_insn): Return zero when pac is enabled.
31035 * config/arm/arm.md (pac_ip_lr_sp, pacbti_ip_lr_sp, aut_ip_lr_sp):
31037 * config/arm/unspecs.md (UNSPEC_PAC_NOP)
31038 (VUNSPEC_PACBTI_NOP, VUNSPEC_AUT_NOP): Add unspecs.
31040 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31042 * config/arm/t-rmprofile: Add multilib rules for march +pacbti and
31043 mbranch-protection.
31045 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31046 Tejas Belagod <tbelagod@arm.com>
31048 * config/arm/arm.cc (arm_file_start): Emit EABI attributes for
31049 Tag_PAC_extension, Tag_BTI_extension, TAG_BTI_use, TAG_PACRET_use.
31051 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31052 Tejas Belagod <tbelagod@arm.com>
31053 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
31055 * ginclude/unwind-arm-common.h (_Unwind_VRS_RegClass): Introduce
31056 new pseudo register class _UVRSC_PAC.
31058 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31059 Tejas Belagod <tbelagod@arm.com>
31061 * config/arm/arm-c.cc (arm_cpu_builtins): Define
31062 __ARM_FEATURE_BTI_DEFAULT, __ARM_FEATURE_PAC_DEFAULT,
31063 __ARM_FEATURE_PAUTH and __ARM_FEATURE_BTI.
31065 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31066 Tejas Belagod <tbelagod@arm.com>
31068 * doc/sourcebuild.texi: Document arm_pacbti_hw.
31070 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31071 Tejas Belagod <tbelagod@arm.com>
31072 Richard Earnshaw <Richard.Earnshaw@arm.com>
31074 * config/arm/arm.cc (arm_configure_build_target): Parse and validate
31075 -mbranch-protection option and initialize appropriate data structures.
31076 * config/arm/arm.opt (-mbranch-protection): New option.
31077 * doc/invoke.texi (Arm Options): Document it.
31079 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31080 Tejas Belagod <tbelagod@arm.com>
31082 * config/arm/arm.h (TARGET_HAVE_PACBTI): New macro.
31083 * config/arm/arm-cpus.in (pacbti): New feature.
31084 * doc/invoke.texi (Arm Options): Document it.
31086 2023-01-23 Andrea Corallo <andrea.corallo@arm.com>
31087 Tejas Belagod <tbelagod@arm.com>
31089 * common/config/aarch64/aarch64-common.cc: Include aarch-common.h.
31090 (all_architectures): Fix comment.
31091 (aarch64_parse_extension): Rename return type, enum value names.
31092 * config/aarch64/aarch64-c.cc (aarch64_update_cpp_builtins): Rename
31093 factored out aarch_ra_sign_scope and aarch_ra_sign_key variables.
31094 Also rename corresponding enum values.
31095 * config/aarch64/aarch64-opts.h (aarch64_function_type): Factor
31096 out aarch64_function_type and move it to common code as
31097 aarch_function_type in aarch-common.h.
31098 * config/aarch64/aarch64-protos.h: Include common types header,
31099 move out types aarch64_parse_opt_result and aarch64_key_type to
31101 * config/aarch64/aarch64.cc: Move mbranch-protection parsing types
31102 and functions out into aarch-common.h and aarch-common.cc. Fix up
31103 all the name changes resulting from the move.
31104 * config/aarch64/aarch64.md: Fix up aarch64_ra_sign_key type name change
31106 * config/aarch64/aarch64.opt: Include aarch-common.h to import
31107 type move. Fix up name changes from factoring out common code and
31109 * config/arm/aarch-common-protos.h: Export factored out routines to both
31111 * config/arm/aarch-common.cc: Include newly factored out types.
31112 Move all mbranch-protection code and data structures from
31114 * config/arm/aarch-common.h: New header that declares types shared
31115 between aarch32 and aarch64 backends.
31116 * config/arm/arm-protos.h: Declare types and variables that are
31117 made common to aarch64 and aarch32 backends - aarch_ra_sign_key,
31118 aarch_ra_sign_scope and aarch_enable_bti.
31119 * config/arm/arm.opt (config/arm/aarch-common.h): Include header.
31120 (aarch_ra_sign_scope, aarch_enable_bti): Declare variable.
31121 * config/arm/arm.cc: Add missing includes.
31123 2023-01-23 Tobias Burnus <tobias@codesourcery.com>
31125 * doc/install.texi (amdgcn, nvptx): Require newlib 4.3.0.
31127 2023-01-23 Richard Biener <rguenther@suse.de>
31129 PR tree-optimization/108449
31130 * cgraphunit.cc (check_global_declaration): Do not turn
31131 undefined statics into externs.
31133 2023-01-22 Dimitar Dimitrov <dimitar@dinux.eu>
31135 * config/pru/pru.h (CLZ_DEFINED_VALUE_AT_ZERO): Fix value for QI
31136 and HI input modes.
31137 * config/pru/pru.md (clz): Fix generated code for QI and HI
31140 2023-01-22 Cupertino Miranda <cupertino.miranda@oracle.com>
31142 * config/v850/v850.cc (v850_select_section): Put const volatile
31143 objects into read-only sections.
31145 2023-01-20 Tejas Belagod <tejas.belagod@arm.com>
31147 * config/aarch64/arm_neon.h (vmull_p64, vmull_high_p64, vaeseq_u8,
31148 vaesdq_u8, vaesmcq_u8, vaesimcq_u8): Gate under "nothing+aes".
31149 (vsha1*_u32, vsha256*_u32): Gate under "nothing+sha2".
31151 2023-01-20 Jakub Jelinek <jakub@redhat.com>
31153 PR tree-optimization/108457
31154 * tree-ssa-loop-niter.cc (build_cltz_expr): Use
31155 SCALAR_INT_TYPE_MODE (utype) directly as C[LT]Z_DEFINED_VALUE_AT_ZERO
31156 argument instead of a temporary. Formatting fixes.
31158 2023-01-19 Jakub Jelinek <jakub@redhat.com>
31160 PR tree-optimization/108447
31161 * value-relation.cc (rr_union_table): Fix VREL_UNDEFINED row order.
31162 (relation_tests): Add self-tests for relation_{intersect,union}
31164 * selftest.h (relation_tests): Declare.
31165 * function-tests.cc (test_ranges): Call it.
31167 2023-01-19 H.J. Lu <hjl.tools@gmail.com>
31170 * config/i386/i386-expand.cc (ix86_expand_builtin): Check
31171 invalid third argument to __builtin_ia32_prefetch.
31173 2023-01-19 Jakub Jelinek <jakub@redhat.com>
31175 PR middle-end/108459
31176 * omp-expand.cc (expand_omp_for_init_counts): Use fold_build1 rather
31177 than fold_unary for NEGATE_EXPR.
31179 2023-01-19 Christophe Lyon <christophe.lyon@arm.com>
31182 * config/aarch64/aarch64.cc (aarch64_layout_arg): Improve
31183 comment. Move assert about alignment a bit later.
31185 2023-01-19 Jakub Jelinek <jakub@redhat.com>
31187 PR tree-optimization/108440
31188 * tree-ssa-forwprop.cc: Include gimple-range.h.
31189 (simplify_rotate): For the forms with T2 wider than T and shift counts of
31190 Y and B - Y add & (B - 1) masking for the rotate count if Y could be equal
31191 to B. For the forms with T2 wider than T and shift counts of
31192 Y and (-Y) & (B - 1), don't punt if range could be [B, B2], but only if
31193 range doesn't guarantee Y < B or Y = N * B. If range doesn't guarantee
31194 Y < B, also add & (B - 1) masking for the rotate count. Use lazily created
31195 pass specific ranger instead of get_global_range_query.
31196 (pass_forwprop::execute): Disable that ranger at the end of pass if it has
31199 2023-01-19 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
31201 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_set<mode>): Use
31202 exact_log2 (INTVAL (operands[2])) >= 0 as condition for gating
31204 (aarch64_simd_vec_copy_lane<mode>): Likewise.
31205 (aarch64_simd_vec_copy_lane_<vswap_width_name><mode>): Likewise.
31207 2023-01-19 Alexandre Oliva <oliva@adacore.com>
31210 * sched-deps.cc (sched_analyze_2): Skip cselib address lookup
31211 within debug insns.
31213 2023-01-18 Martin Jambor <mjambor@suse.cz>
31216 * cgraph.cc (cgraph_node::remove): Check whether nodes up the
31217 lcone_of chain also do not need the body.
31219 2023-01-18 Richard Biener <rguenther@suse.de>
31222 2022-12-16 Richard Biener <rguenther@suse.de>
31224 PR middle-end/108086
31225 * tree-inline.cc (remap_ssa_name): Do not unshare the
31226 result from the decl_map.
31228 2023-01-18 Murray Steele <murray.steele@arm.com>
31231 * config/arm/arm_mve.h (__arm_vst1q_p_u8): Use prefixed intrinsic
31233 (__arm_vst1q_p_s8): Likewise.
31234 (__arm_vld1q_z_u8): Likewise.
31235 (__arm_vld1q_z_s8): Likewise.
31236 (__arm_vst1q_p_u16): Likewise.
31237 (__arm_vst1q_p_s16): Likewise.
31238 (__arm_vld1q_z_u16): Likewise.
31239 (__arm_vld1q_z_s16): Likewise.
31240 (__arm_vst1q_p_u32): Likewise.
31241 (__arm_vst1q_p_s32): Likewise.
31242 (__arm_vld1q_z_u32): Likewise.
31243 (__arm_vld1q_z_s32): Likewise.
31244 (__arm_vld1q_z_f16): Likewise.
31245 (__arm_vst1q_p_f16): Likewise.
31246 (__arm_vld1q_z_f32): Likewise.
31247 (__arm_vst1q_p_f32): Likewise.
31249 2023-01-18 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
31251 * config/xtensa/xtensa.md (xorsi3_internal):
31252 Rename from the original of "xorsi3".
31253 (xorsi3): New expansion pattern that emits addition rather than
31254 bitwise-XOR when the second source is a constant of -2147483648
31257 2023-01-18 Kewen Lin <linkw@linux.ibm.com>
31258 Andrew Pinski <apinski@marvell.com>
31261 * config/rs6000/rs6000-overload.def (VEC_VSUBCUQ): Fix typo
31262 vec_vsubcuqP with vec_vsubcuq.
31264 2023-01-18 Kewen Lin <linkw@linux.ibm.com>
31267 * config/rs6000/rs6000.cc (rs6000_opaque_type_invalid_use_p): Add the
31268 support for invalid uses of MMA opaque type in function arguments.
31270 2023-01-18 liuhongt <hongtao.liu@intel.com>
31273 * config/i386/cygwin.h (ENDFILE_SPEC): Link crtfastmath.o
31274 whenever -mdaz-ftz is specified. Don't link crtfastmath.o when
31275 -share or -mno-daz-ftz is specified.
31276 * config/i386/darwin.h (ENDFILE_SPEC): Ditto.
31277 * config/i386/mingw32.h (ENDFILE_SPEC): Ditto.
31279 2023-01-17 Jose E. Marchesi <jose.marchesi@oracle.com>
31281 * config/bpf/bpf.cc (bpf_option_override): Disable
31284 2023-01-17 Jakub Jelinek <jakub@redhat.com>
31286 PR tree-optimization/106523
31287 * tree-ssa-forwprop.cc (simplify_rotate): For the
31288 patterns with (-Y) & (B - 1) in one operand's shift
31289 count and Y in another, if T2 has wider precision than T,
31290 punt if Y could have a value in [B, B2 - 1] range.
31292 2023-01-16 H.J. Lu <hjl.tools@gmail.com>
31295 * config/i386/i386.cc (x86_output_mi_thunk): Disable
31296 -mforce-indirect-call for PIC in 32-bit mode.
31298 2023-01-16 Jan Hubicka <hubicka@ucw.cz>
31301 * ipa-modref.cc (modref_access_analysis::analyze): Use
31302 find_always_executed_bbs.
31303 * ipa-sra.cc (process_scan_results): Likewise.
31304 * ipa-utils.cc (stmt_may_terminate_function_p): New function.
31305 (find_always_executed_bbs): New function.
31306 * ipa-utils.h (stmt_may_terminate_function_p): Declare.
31307 (find_always_executed_bbs): Declare.
31309 2023-01-16 Jan Hubicka <jh@suse.cz>
31311 * config/i386/i386.cc (ix86_vectorize_builtin_scatter): Guard scatter
31312 by TARGET_USE_SCATTER.
31313 * config/i386/i386.h (TARGET_USE_SCATTER_2PARTS,
31314 TARGET_USE_SCATTER_4PARTS, TARGET_USE_SCATTER): New macros.
31315 * config/i386/x86-tune.def (TARGET_USE_SCATTER_2PARTS,
31316 TARGET_USE_SCATTER_4PARTS, TARGET_USE_SCATTER): New tunes.
31317 (X86_TUNE_AVOID_256FMA_CHAINS, X86_TUNE_AVOID_512FMA_CHAINS): Disable
31318 for znver4. (X86_TUNE_USE_GATHER): Disable for zen4.
31320 2023-01-16 Richard Biener <rguenther@suse.de>
31323 * config/sol2.h (ENDFILE_SPEC): Don't add crtfastmath.o for -shared.
31325 2023-01-16 Stam Markianos-Wright <stam.markianos-wright@arm.com>
31329 * config/arm/arm_mve.h (__ARM_mve_coerce2): Split types.
31330 (__ARM_mve_coerce3): Likewise.
31332 2023-01-16 Andrew Carlotti <andrew.carlotti@arm.com>
31334 * tree-ssa-loop-niter.cc (build_popcount_expr): Add IFN support.
31336 2023-01-16 Andrew Carlotti <andrew.carlotti@arm.com>
31338 * tree-ssa-loop-niter.cc (number_of_iterations_cltz): New.
31339 (number_of_iterations_bitcount): Add call to the above.
31340 (number_of_iterations_exit_assumptions): Add EQ_EXPR case for
31341 c[lt]z idiom recognition.
31343 2023-01-16 Andrew Carlotti <andrew.carlotti@arm.com>
31345 * doc/sourcebuild.texi: Add missing target attributes.
31347 2023-01-16 Andrew Carlotti <andrew.carlotti@arm.com>
31349 PR tree-optimization/94793
31350 * tree-scalar-evolution.cc (expression_expensive_p): Add checks
31352 * tree-ssa-loop-niter.cc (build_cltz_expr): New.
31353 (number_of_iterations_cltz_complement): New.
31354 (number_of_iterations_bitcount): Add call to the above.
31356 2023-01-16 Jonathan Wakely <jwakely@redhat.com>
31358 * doc/extend.texi (Common Function Attributes): Fix grammar.
31360 2023-01-16 Jakub Jelinek <jakub@redhat.com>
31363 * config/riscv/riscv-vsetvl.h: Add space in between Copyright and (C).
31364 * config/riscv/riscv-vsetvl.cc: Likewise.
31366 2023-01-16 Jakub Jelinek <jakub@redhat.com>
31369 * config/i386/xmmintrin.h (_mm_undefined_ps): Temporarily
31370 disable -Winit-self using pragma GCC diagnostic ignored.
31371 * config/i386/emmintrin.h (_mm_undefined_pd, _mm_undefined_si128):
31373 * config/i386/avxintrin.h (_mm256_undefined_pd, _mm256_undefined_ps,
31374 _mm256_undefined_si256): Likewise.
31375 * config/i386/avx512fintrin.h (_mm512_undefined_pd,
31376 _mm512_undefined_ps, _mm512_undefined_epi32): Likewise.
31377 * config/i386/avx512fp16intrin.h (_mm_undefined_ph,
31378 _mm256_undefined_ph, _mm512_undefined_ph): Likewise.
31380 2023-01-16 Kewen Lin <linkw@linux.ibm.com>
31383 * config/rs6000/rs6000.cc (rs6000_opaque_type_invalid_use_p): Add the
31384 support for invalid uses in inline asm, factor out the checking and
31385 erroring to lambda function check_and_error_invalid_use.
31387 2023-01-15 Aldy Hernandez <aldyh@redhat.com>
31389 PR tree-optimization/107608
31390 * range-op-float.cc (range_operator_float::fold_range): Avoid
31391 folding into INF when flag_trapping_math.
31392 * value-range.h (frange::known_isinf): Return false for possible NANs.
31394 2023-01-15 Xianmiao Qu <cooper.qu@linux.alibaba.com>
31396 * config.gcc (csky-*-*): Support --with-float=softfp.
31398 2023-01-14 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
31400 * config/xtensa/xtensa-protos.h (order_regs_for_local_alloc):
31401 Rename to xtensa_adjust_reg_alloc_order.
31402 * config/xtensa/xtensa.cc (xtensa_adjust_reg_alloc_order):
31403 Ditto. And also remove code to reorder register numbers for
31404 leaf functions, rename the tables, and adjust the allocation
31405 order for the call0 ABI to use register A0 more.
31406 (xtensa_leaf_regs): Remove.
31407 * config/xtensa/xtensa.h (REG_ALLOC_ORDER): Cosmetics.
31408 (order_regs_for_local_alloc): Rename as the above.
31409 (LEAF_REGISTERS, LEAF_REG_REMAP, leaf_function): Remove.
31411 2023-01-14 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
31413 * config/aarch64/aarch64-sve.md (aarch64_vec_duplicate_vq<mode>_le):
31414 Change to define_insn_and_split to fold ldr+dup to ld1rq.
31415 * config/aarch64/predicates.md (aarch64_sve_dup_ld1rq_operand): New.
31417 2023-01-14 Alexandre Oliva <oliva@adacore.com>
31419 * hash-table.h (is_deleted): Precheck !is_empty.
31420 (mark_deleted): Postcheck !is_empty.
31421 (copy constructor): Test is_empty before is_deleted.
31423 2023-01-14 Alexandre Oliva <oliva@adacore.com>
31426 * config/arm/arm.md (movmisaligndi): Prefer aligned SImode
31429 2023-01-13 Eric Botcazou <ebotcazou@adacore.com>
31431 PR rtl-optimization/108274
31432 * function.cc (thread_prologue_and_epilogue_insns): Also update the
31433 DF information for calls in a few more cases.
31435 2023-01-13 John David Anglin <danglin@gcc.gnu.org>
31437 * config/pa/pa-linux.h (TARGET_SYNC_LIBCALL): Delete define.
31438 * config/pa/pa.cc (pa_init_libfuncs): Use MAX_SYNC_LIBFUNC_SIZE
31440 * config/pa/pa.h (TARGET_SYNC_LIBCALLS): Use flag_sync_libcalls.
31441 (MAX_SYNC_LIBFUNC_SIZE): Define.
31442 (TARGET_CPU_CPP_BUILTINS): Define __SOFTFP__ when soft float is
31444 * config/pa/pa.md (atomic_storeqi): Emit __atomic_exchange_1
31445 libcall when sync libcalls are disabled.
31446 (atomic_storehi, atomic_storesi, atomic_storedi): Likewise.
31447 (atomic_loaddi): Emit __atomic_load_8 libcall when sync libcalls
31448 are disabled on 32-bit target.
31449 * config/pa/pa.opt (matomic-libcalls): New option.
31450 * doc/invoke.texi (HPPA Options): Update.
31452 2023-01-13 Alexander Monakov <amonakov@ispras.ru>
31454 PR rtl-optimization/108117
31455 PR rtl-optimization/108132
31456 * sched-deps.cc (deps_analyze_insn): Do not schedule across
31457 calls before reload.
31459 2023-01-13 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
31461 * common/config/arm/arm-common.cc (arm_canon_arch_option_1): Ignore cde
31462 options for -mlibarch.
31463 * config/arm/arm-cpus.in (begin cpu cortex-m55): Add cde options.
31464 * doc/invoke.texi (CDE): Document options for Cortex-M55 CPU.
31466 2023-01-13 Qing Zhao <qing.zhao@oracle.com>
31468 * attribs.cc (strict_flex_array_level_of): Move this function to ...
31469 * attribs.h (strict_flex_array_level_of): Remove the declaration.
31470 * gimple-array-bounds.cc (array_bounds_checker::check_array_ref):
31471 replace the referece to strict_flex_array_level_of with
31472 DECL_NOT_FLEXARRAY.
31473 * tree.cc (component_ref_size): Likewise.
31475 2023-01-13 Richard Biener <rguenther@suse.de>
31478 * config/arm/linux-eabi.h (ENDFILE_SPEC): Don't add
31479 crtfastmath.o for -shared.
31480 * config/arm/unknown-elf.h (STARTFILE_SPEC): Likewise.
31482 2023-01-13 Richard Biener <rguenther@suse.de>
31485 * config/aarch64/aarch64-elf-raw.h (ENDFILE_SPEC): Don't add
31486 crtfastmath.o for -shared.
31487 * config/aarch64/aarch64-freebsd.h (GNU_USER_TARGET_MATHFILE_SPEC):
31489 * config/aarch64/aarch64-linux.h (GNU_USER_TARGET_MATHFILE_SPEC):
31492 2023-01-13 Richard Sandiford <richard.sandiford@arm.com>
31494 * config/aarch64/aarch64.cc (aarch64_dwarf_frame_reg_mode): New
31496 (TARGET_DWARF_FRAME_REG_MODE): Define.
31498 2023-01-13 Richard Biener <rguenther@suse.de>
31501 * config/aarch64/aarch64.cc (aarch64_gimple_fold_builtin): Don't
31502 update EH info on the fly.
31504 2023-01-13 Richard Biener <rguenther@suse.de>
31506 PR tree-optimization/108387
31507 * tree-ssa-sccvn.cc (visit_nary_op): Check for SSA_NAME
31508 value before inserting expression into the tables.
31510 2023-01-12 Andrew Pinski <apinski@marvell.com>
31511 Roger Sayle <roger@nextmovesoftware.com>
31513 PR tree-optimization/92342
31514 * match.pd ((m1 CMP m2) * d -> (m1 CMP m2) ? d : 0):
31515 Use tcc_comparison and :c for the multiply.
31516 (b & -(a CMP c) -> (a CMP c)?b:0): New pattern.
31518 2023-01-12 Christophe Lyon <christophe.lyon@arm.com>
31519 Richard Sandiford <richard.sandiford@arm.com>
31522 * config/aarch64/aarch64.cc (aarch64_function_arg_alignment):
31523 Check DECL_PACKED for bitfield.
31524 (aarch64_layout_arg): Warn when parameter passing ABI changes.
31525 (aarch64_function_arg_boundary): Do not warn here.
31526 (aarch64_gimplify_va_arg_expr): Warn when parameter passing ABI
31529 2023-01-12 Christophe Lyon <christophe.lyon@arm.com>
31530 Richard Sandiford <richard.sandiford@arm.com>
31532 * config/aarch64/aarch64.cc (aarch64_function_arg_alignment): Fix
31534 (aarch64_layout_arg): Factorize warning conditions.
31535 (aarch64_function_arg_boundary): Fix typo.
31536 * function.cc (currently_expanding_function_start): New variable.
31537 (expand_function_start): Handle
31538 currently_expanding_function_start.
31539 * function.h (currently_expanding_function_start): Declare.
31541 2023-01-12 Richard Biener <rguenther@suse.de>
31543 PR tree-optimization/99412
31544 * tree-ssa-reassoc.cc (is_phi_for_stmt): Remove.
31545 (swap_ops_for_binary_stmt): Remove reduction handling.
31546 (rewrite_expr_tree_parallel): Adjust.
31547 (reassociate_bb): Likewise.
31548 * tree-parloops.cc (build_new_reduction): Handle MINUS_EXPR.
31550 2023-01-12 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
31552 * config/xtensa/xtensa.md (ctzsi2, ffssi2):
31553 Rearrange the emitting codes.
31555 2023-01-12 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
31557 * config/xtensa/xtensa.md (*btrue):
31558 Correct value of the attribute "length" that depends on
31559 TARGET_DENSITY and operands, and add '?' character to the register
31560 constraint of the compared operand.
31562 2023-01-12 Alexandre Oliva <oliva@adacore.com>
31564 * hash-table.h (expand): Check elements and deleted counts.
31565 (verify): Likewise.
31567 2023-01-11 Roger Sayle <roger@nextmovesoftware.com>
31569 PR tree-optimization/71343
31570 * tree-ssa-sccvn.cc (visit_nary_op) <case LSHIFT_EXPR>: Make
31571 the value number of the expression X << C the same as the value
31572 number for the multiplication X * (1<<C).
31574 2023-01-11 David Faust <david.faust@oracle.com>
31577 * config/bpf/bpf.cc (bpf_print_operand): Correct handling for
31578 floating point modes.
31580 2023-01-11 Eric Botcazou <ebotcazou@adacore.com>
31582 PR tree-optimization/108199
31583 * tree-sra.cc (sra_modify_expr): Deal with reverse storage order
31584 for bit-field references.
31586 2023-01-11 Kewen Lin <linkw@linux.ibm.com>
31588 * config/rs6000/rs6000.cc (rs6000_option_override_internal): Make
31589 OPTION_MASK_P10_FUSION implicit setting honour Power10 tuning setting.
31590 * config/rs6000/rs6000-cpus.def (ISA_3_1_MASKS_SERVER): Remove
31591 OPTION_MASK_P10_FUSION.
31593 2023-01-11 Richard Biener <rguenther@suse.de>
31595 PR tree-optimization/107767
31596 * tree-cfgcleanup.cc (phi_alternatives_equal): Export.
31597 * tree-cfgcleanup.h (phi_alternatives_equal): Declare.
31598 * tree-switch-conversion.cc (switch_conversion::collect):
31599 Count unique non-default targets accounting for later
31600 merging opportunities.
31602 2023-01-11 Martin Liska <mliska@suse.cz>
31604 PR middle-end/107976
31605 * params.opt: Limit JT params.
31606 * stmt.cc (emit_case_dispatch_table): Use auto_vec.
31608 2023-01-11 Richard Biener <rguenther@suse.de>
31610 PR tree-optimization/108352
31611 * tree-ssa-threadbackward.cc
31612 (back_threader_profitability::profitable_path_p): Adjust
31613 heuristic that allows non-multi-way branch threads creating
31615 * doc/invoke.texi (--param fsm-scale-path-blocks): Remove.
31616 (--param fsm-scale-path-stmts): Adjust.
31617 * params.opt (--param=fsm-scale-path-blocks=): Remove.
31618 (-param=fsm-scale-path-stmts=): Adjust description.
31620 2023-01-11 Richard Biener <rguenther@suse.de>
31622 PR tree-optimization/108353
31623 * tree-ssa-propagate.cc (cfg_blocks_back, ssa_edge_worklist_back):
31625 (add_ssa_edge): Simplify.
31626 (add_control_edge): Likewise.
31627 (ssa_prop_init): Likewise.
31628 (ssa_prop_fini): Likewise.
31629 (ssa_propagation_engine::ssa_propagate): Likewise.
31631 2023-01-11 Andreas Krebbel <krebbel@linux.ibm.com>
31633 * config/s390/s390.md (*not<mode>): New pattern.
31635 2023-01-11 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
31637 * config/xtensa/xtensa.cc (xtensa_insn_cost):
31638 Let insn cost for size be obtained by applying COSTS_N_INSNS()
31639 to instruction length and then dividing by 3.
31641 2023-01-10 Richard Biener <rguenther@suse.de>
31643 PR tree-optimization/106293
31644 * tree-ssa-dse.cc (dse_classify_store): Use a worklist to
31645 process degenerate PHI defs.
31647 2023-01-10 Roger Sayle <roger@nextmovesoftware.com>
31649 PR rtl-optimization/106421
31650 * cprop.cc (bypass_block): Check that DEST is local to this
31651 function (non-NULL) before calling find_edge.
31653 2023-01-10 Martin Jambor <mjambor@suse.cz>
31656 * ipa-param-manipulation.h (ipa_param_body_adjustments): New members
31657 sort_replacements, lookup_first_base_replacement and
31658 m_sorted_replacements_p.
31659 * ipa-param-manipulation.cc: Define INCLUDE_ALGORITHM.
31660 (ipa_param_body_adjustments::register_replacement): Set
31661 m_sorted_replacements_p to false.
31662 (compare_param_body_replacement): New function.
31663 (ipa_param_body_adjustments::sort_replacements): Likewise.
31664 (ipa_param_body_adjustments::common_initialization): Call
31666 (ipa_param_body_adjustments::ipa_param_body_adjustments): Initialize
31667 m_sorted_replacements_p.
31668 (ipa_param_body_adjustments::lookup_replacement_1): Rework to use
31670 (ipa_param_body_adjustments::lookup_first_base_replacement): New
31672 (ipa_param_body_adjustments::modify_call_stmt): Use
31673 lookup_first_base_replacement.
31674 * omp-simd-clone.cc (ipa_simd_modify_function_body): Call
31675 adjustments->sort_replacements.
31677 2023-01-10 Richard Biener <rguenther@suse.de>
31679 PR tree-optimization/108314
31680 * tree-vect-stmts.cc (vectorizable_condition): Do not
31681 perform BIT_NOT_EXPR optimization for EXTRACT_LAST_REDUCTION.
31683 2023-01-10 Xianmiao Qu <cooper.qu@linux.alibaba.com>
31685 * config/csky/csky-linux-elf.h (SYSROOT_SUFFIX_SPEC): New.
31687 2023-01-10 Xianmiao Qu <cooper.qu@linux.alibaba.com>
31689 * config/csky/csky.h (MULTILIB_DEFAULTS): Fix float abi option.
31691 2023-01-10 Xianmiao Qu <cooper.qu@linux.alibaba.com>
31693 * config/csky/csky.cc (csky_cpu_cpp_builtins): Add builtin
31694 defines for soft float abi.
31696 2023-01-10 Xianmiao Qu <cooper.qu@linux.alibaba.com>
31698 * config/csky/csky.md (smart_bseti): Change condition to CSKY_ISA_FEATURE (E1).
31699 (smart_bclri): Likewise.
31700 (fast_bseti): Change condition to CSKY_ISA_FEATURE (E2).
31701 (fast_bclri): Likewise.
31702 (fast_cmpnesi_i): Likewise.
31703 (*fast_cmpltsi_i): Likewise.
31704 (*fast_cmpgeusi_i): Likewise.
31706 2023-01-10 Xianmiao Qu <cooper.qu@linux.alibaba.com>
31708 * config/csky/csky_insn_fpuv3.md (l<frm_pattern><fixsuop><mode>si2): Test
31709 flag_fp_int_builtin_inexact || !flag_trapping_math.
31710 (<frm_pattern><mode>2): Likewise.
31712 2023-01-10 Andreas Krebbel <krebbel@linux.ibm.com>
31714 * config/s390/s390.cc (s390_register_info): Check call_used_regs
31715 instead of hard-coding the register numbers for call saved
31717 (s390_optimize_register_info): Likewise.
31719 2023-01-09 Eric Botcazou <ebotcazou@adacore.com>
31721 * doc/gm2.texi (Overview): Fix @node markers.
31722 (Using): Likewise. Remove subsections that were moved to Overview
31723 from the menu and move others around.
31725 2023-01-09 Richard Biener <rguenther@suse.de>
31727 PR middle-end/108209
31728 * genmatch.cc (commutative_op): Fix return value for
31729 user-id with non-commutative first replacement.
31731 2023-01-09 Jakub Jelinek <jakub@redhat.com>
31734 * calls.cc (expand_call): For calls with
31735 TYPE_NO_NAMED_ARGS_STDARG_P (funtype) use zero for n_named_args.
31738 2023-01-09 Richard Biener <rguenther@suse.de>
31740 PR middle-end/69482
31741 * cfgexpand.cc (discover_nonconstant_array_refs_r): Volatile
31742 qualified accesses also force objects to memory.
31744 2023-01-09 Martin Liska <mliska@suse.cz>
31747 * lto-cgraph.cc (compute_ltrans_boundary): Do not insert
31748 NULL (deleleted value) to a hash_set.
31750 2023-01-08 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
31752 * config/xtensa/xtensa.md (*splice_bits):
31753 New insn_and_split pattern.
31755 2023-01-07 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
31757 * config/xtensa/xtensa.cc
31758 (xtensa_split_imm_two_addends, xtensa_emit_add_imm):
31759 New helper functions.
31760 (xtensa_set_return_address, xtensa_output_mi_thunk):
31761 Change to use the helper function.
31762 (xtensa_emit_adjust_stack_ptr): Ditto.
31763 And also change to try reusing the content of scratch register
31764 A9 if the register is not modified in the function body.
31766 2023-01-07 LIU Hao <lh_mouse@126.com>
31768 PR middle-end/108300
31769 * config/xtensa/xtensa-dynconfig.c: Define `WIN32_LEAN_AND_MEAN`
31770 before <windows.h>.
31771 * diagnostic-color.cc: Likewise.
31772 * plugin.cc: Likewise.
31773 * prefix.cc: Likewise.
31775 2023-01-06 Joseph Myers <joseph@codesourcery.com>
31777 * doc/extend.texi (__builtin_tgmath): Do not restate standard rule
31778 for handling real integer types.
31780 2023-01-06 Tamar Christina <tamar.christina@arm.com>
31783 2022-12-12 Tamar Christina <tamar.christina@arm.com>
31785 * config/aarch64/aarch64-simd.md (*aarch64_simd_movv2hf): New.
31786 (mov<mode>, movmisalign<mode>, aarch64_dup_lane<mode>,
31787 aarch64_store_lane0<mode>, aarch64_simd_vec_set<mode>,
31788 @aarch64_simd_vec_copy_lane<mode>, vec_set<mode>,
31789 reduc_<optab>_scal_<mode>, reduc_<fmaxmin>_scal_<mode>,
31790 aarch64_reduc_<optab>_internal<mode>, aarch64_get_lane<mode>,
31791 vec_init<mode><Vel>, vec_extract<mode><Vel>): Support V2HF.
31792 (aarch64_simd_dupv2hf): New.
31793 * config/aarch64/aarch64.cc (aarch64_classify_vector_mode):
31795 * config/aarch64/iterators.md (VHSDF_P): New.
31796 (V2F, VMOVE, nunits, Vtype, Vmtype, Vetype, stype, VEL,
31797 Vel, q, vp): Add V2HF.
31798 * config/arm/types.md (neon_fp_reduc_add_h): New.
31800 2023-01-06 Martin Liska <mliska@suse.cz>
31802 PR middle-end/107966
31803 * doc/options.texi: Fix Var documentation in internal manual.
31805 2023-01-05 Roger Sayle <roger@nextmovesoftware.com>
31808 2023-01-03 Roger Sayle <roger@nextmovesoftware.com>
31810 * config/i386/i386-expand.cc (ix86_expand_int_movcc): Rewrite
31811 RTL expansion to allow condition (mask) to be shared/reused,
31812 by avoiding overwriting pseudos and adding REG_EQUAL notes.
31814 2023-01-05 Iain Sandoe <iain@sandoe.co.uk>
31816 * common.opt: Add -static-libgm2.
31817 * config/darwin.h (LINK_SPEC): Handle static-libgm2.
31818 * doc/gm2.texi: Document static-libgm2.
31819 * gcc.cc (driver_handle_option): Allow static-libgm2.
31821 2023-01-05 Tejas Joshi <TejasSanjay.Joshi@amd.com>
31823 * common/config/i386/i386-common.cc (processor_alias_table):
31824 Use CPU_ZNVER4 for znver4.
31825 * config/i386/i386.md: Add znver4.md.
31826 * config/i386/znver4.md: New.
31828 2023-01-04 Jakub Jelinek <jakub@redhat.com>
31830 PR tree-optimization/108253
31831 * tree-vrp.cc (maybe_set_nonzero_bits): Handle var with pointer
31834 2023-01-04 Jakub Jelinek <jakub@redhat.com>
31836 PR middle-end/108237
31837 * generic-match-head.cc: Include tree-pass.h.
31838 (canonicalize_math_p, optimize_vectors_before_lowering_p): Define
31839 to false if cfun and cfun->curr_properties has PROP_gimple_opt_math
31840 resp. PROP_gimple_lvec property set.
31842 2023-01-04 Jakub Jelinek <jakub@redhat.com>
31844 PR sanitizer/108256
31845 * convert.cc (do_narrow): Punt for MULT_EXPR if original
31846 type doesn't wrap around and -fsanitize=signed-integer-overflow
31848 * fold-const.cc (fold_unary_loc) <CASE_CONVERT>: Likewise.
31850 2023-01-04 Hu, Lin1 <lin1.hu@intel.com>
31852 * common/config/i386/cpuinfo.h (get_intel_cpu): Handle Emeraldrapids.
31853 * common/config/i386/i386-common.cc: Add Emeraldrapids.
31855 2023-01-04 Hu, Lin1 <lin1.hu@intel.com>
31857 * common/config/i386/cpuinfo.h (get_intel_cpu): Remove case 0xb5
31860 2023-01-03 Sandra Loosemore <sandra@codesourcery.com>
31862 * cgraph.h (struct cgraph_node): Add gc_candidate bit, modify
31863 default constructor to initialize it.
31864 * cgraphunit.cc (expand_all_functions): Save gc_candidate functions
31865 for last and iterate to handle recursive calls. Delete leftover
31866 candidates at the end.
31867 * omp-simd-clone.cc (simd_clone_create): Set gc_candidate bit
31869 * tree-vect-stmts.cc (vectorizable_simd_clone_call): Clear
31870 gc_candidate bit when a clone is used.
31872 2023-01-03 Florian Weimer <fweimer@redhat.com>
31875 2023-01-02 Florian Weimer <fweimer@redhat.com>
31877 * dwarf2cfi.cc (init_return_column_size): Remove.
31878 (init_one_dwarf_reg_size): Adjust.
31879 (generate_dwarf_reg_sizes): New function. Extracted
31880 from expand_builtin_init_dwarf_reg_sizes.
31881 (expand_builtin_init_dwarf_reg_sizes): Call
31882 generate_dwarf_reg_sizes.
31883 * target.def (init_dwarf_reg_sizes_extra): Adjust
31885 * config/msp430/msp430.cc
31886 (msp430_init_dwarf_reg_sizes_extra): Adjust.
31887 * config/rs6000/rs6000.cc
31888 (rs6000_init_dwarf_reg_sizes_extra): Likewise.
31889 * doc/tm.texi: Update.
31891 2023-01-03 Florian Weimer <fweimer@redhat.com>
31894 2023-01-02 Florian Weimer <fweimer@redhat.com>
31896 * debug.h (dwarf_reg_sizes_constant): Declare.
31897 * dwarf2cfi.cc (dwarf_reg_sizes_constant): New function.
31899 2023-01-03 Siddhesh Poyarekar <siddhesh@gotplt.org>
31901 PR tree-optimization/105043
31902 * doc/extend.texi (Object Size Checking): Split out into two
31903 subsections and mention _FORTIFY_SOURCE.
31905 2023-01-03 Roger Sayle <roger@nextmovesoftware.com>
31907 * config/i386/i386-expand.cc (ix86_expand_int_movcc): Rewrite
31908 RTL expansion to allow condition (mask) to be shared/reused,
31909 by avoiding overwriting pseudos and adding REG_EQUAL notes.
31911 2023-01-03 Roger Sayle <roger@nextmovesoftware.com>
31914 * config/i386/i386-features.cc
31915 (general_scalar_chain::compute_convert_gain) <case PLUS>: Consider
31916 the gain/cost of converting a MEM operand.
31918 2023-01-03 Jakub Jelinek <jakub@redhat.com>
31920 PR middle-end/108264
31921 * expr.cc (store_expr): For stores into SUBREG_PROMOTED_* targets
31922 from source which doesn't have scalar integral mode first convert
31925 2023-01-03 Jakub Jelinek <jakub@redhat.com>
31927 PR rtl-optimization/108263
31928 * cfgrtl.cc (fixup_reorder_chain): Avoid trying to redirect
31931 2023-01-02 Alexander Monakov <amonakov@ispras.ru>
31934 * config/i386/lujiazui.md (lujiazui_div): New automaton.
31935 (lua_div): New unit.
31936 (lua_idiv_qi): Correct unit in the reservation.
31937 (lua_idiv_qi_load): Ditto.
31938 (lua_idiv_hi): Ditto.
31939 (lua_idiv_hi_load): Ditto.
31940 (lua_idiv_si): Ditto.
31941 (lua_idiv_si_load): Ditto.
31942 (lua_idiv_di): Ditto.
31943 (lua_idiv_di_load): Ditto.
31944 (lua_fdiv_SF): Ditto.
31945 (lua_fdiv_SF_load): Ditto.
31946 (lua_fdiv_DF): Ditto.
31947 (lua_fdiv_DF_load): Ditto.
31948 (lua_fdiv_XF): Ditto.
31949 (lua_fdiv_XF_load): Ditto.
31950 (lua_ssediv_SF): Ditto.
31951 (lua_ssediv_load_SF): Ditto.
31952 (lua_ssediv_V4SF): Ditto.
31953 (lua_ssediv_load_V4SF): Ditto.
31954 (lua_ssediv_V8SF): Ditto.
31955 (lua_ssediv_load_V8SF): Ditto.
31956 (lua_ssediv_SD): Ditto.
31957 (lua_ssediv_load_SD): Ditto.
31958 (lua_ssediv_V2DF): Ditto.
31959 (lua_ssediv_load_V2DF): Ditto.
31960 (lua_ssediv_V4DF): Ditto.
31961 (lua_ssediv_load_V4DF): Ditto.
31963 2023-01-02 Florian Weimer <fweimer@redhat.com>
31965 * debug.h (dwarf_reg_sizes_constant): Declare.
31966 * dwarf2cfi.cc (dwarf_reg_sizes_constant): New function.
31968 2023-01-02 Florian Weimer <fweimer@redhat.com>
31970 * dwarf2cfi.cc (init_return_column_size): Remove.
31971 (init_one_dwarf_reg_size): Adjust.
31972 (generate_dwarf_reg_sizes): New function. Extracted
31973 from expand_builtin_init_dwarf_reg_sizes.
31974 (expand_builtin_init_dwarf_reg_sizes): Call
31975 generate_dwarf_reg_sizes.
31976 * target.def (init_dwarf_reg_sizes_extra): Adjust
31978 * config/msp430/msp430.cc
31979 (msp430_init_dwarf_reg_sizes_extra): Adjust.
31980 * config/rs6000/rs6000.cc
31981 (rs6000_init_dwarf_reg_sizes_extra): Likewise.
31982 * doc/tm.texi: Update.
31984 2023-01-02 Jakub Jelinek <jakub@redhat.com>
31986 * gcc.cc (process_command): Update copyright notice dates.
31987 * gcov-dump.cc (print_version): Ditto.
31988 * gcov.cc (print_version): Ditto.
31989 * gcov-tool.cc (print_version): Ditto.
31990 * gengtype.cc (create_file): Ditto.
31991 * doc/cpp.texi: Bump @copying's copyright year.
31992 * doc/cppinternals.texi: Ditto.
31993 * doc/gcc.texi: Ditto.
31994 * doc/gccint.texi: Ditto.
31995 * doc/gcov.texi: Ditto.
31996 * doc/install.texi: Ditto.
31997 * doc/invoke.texi: Ditto.
31999 2023-01-01 Roger Sayle <roger@nextmovesoftware.com>
32000 Uroš Bizjak <ubizjak@gmail.com>
32002 * config/i386/i386.md (extendditi2): New define_insn.
32003 (define_split): Use DWIH mode iterator to treat new extendditi2
32004 identically to existing extendsidi2_1.
32005 (define_peephole2): Likewise.
32006 (define_peephole2): Likewise.
32007 (define_Split): Likewise.
32010 Copyright (C) 2023 Free Software Foundation, Inc.
32012 Copying and distribution of this file, with or without modification,
32013 are permitted in any medium without royalty provided the copyright
32014 notice and this notice are preserved.