1 2024-01-09 Jeff Law <jlaw@ventanamicro.com>
3 * config/epiphany/constraints.md (Car): Allow -1024..1023, no more,
6 2024-01-09 Richard Sandiford <richard.sandiford@arm.com>
8 * config/mn10300/mn10300.md (subdi3_degenerate): Add isa attribute.
10 2024-01-09 Tamar Christina <tamar.christina@arm.com>
12 * tree-vect-loop.cc (vectorizable_live_operation_1): Drop unused
14 (vectorizable_live_operation): Likewise.
16 2024-01-09 Tamar Christina <tamar.christina@arm.com>
18 PR tree-optimization/113199
19 * tree-vect-loop.cc (vectorizable_live_operation_1): Use
22 2024-01-09 Jakub Jelinek <jakub@redhat.com>
25 * config.gcc (aarch64*-*-*): Add aarch64-builtins.h to target_gtfiles.
26 * config/aarch64/aarch64-builtins.cc (aarch64_simd_types): Add extern
27 GTY(()) declaration before the definition, drop GTY(()) drom the
30 2024-01-09 Richard Biener <rguenther@suse.de>
32 PR tree-optimization/113026
33 * tree-vect-loop-manip.cc (vect_do_peeling): Remove
34 redundant and wrong niter bound setting. Move niter
35 bound adjustment down.
37 2024-01-09 Tamar Christina <tamar.christina@arm.com>
40 * tree-vect-loop-manip.cc (vect_can_peel_nonlinear_iv_p):
41 Reject non-linear inductions that aren't supported.
43 2024-01-09 Roger Sayle <roger@nextmovesoftware.com>
45 * config/arc/arc.cc (arc_shift_alg): New enumerated type for
46 left shift implementation strategies.
47 (arc_shift_info): Type for each entry of the shift strategy table.
48 (arc_shift_context_idx): Return a integer value for each code
49 generation context, used as an index
50 (arc_ashl_alg): Table indexed by context and shifted bit count.
51 (arc_split_ashl): Use the arc_ashl_alg table to select SImode
52 left shift implementation.
53 (arc_rtx_costs) <case ASHIFT>: Use the arc_ashl_alg table to
54 provide accurate costs, when optimizing for speed or size.
56 2024-01-09 Juzhe-Zhong <juzhe.zhong@rivai.ai>
58 * config/riscv/riscv-vector-costs.cc (loop_invariant_op_p): Fix loop invariant check.
60 2024-01-09 Julian Brown <julian@codesourcery.com>
62 * gimplify.cc (gimplify_expr): Ensure OMP_ARRAY_SECTION has been
63 processed out before gimplification.
64 * tree-pretty-print.cc (dump_generic_node): Support OMP_ARRAY_SECTION.
65 * tree.def (OMP_ARRAY_SECTION): New tree code.
67 2024-01-09 Jakub Jelinek <jakub@redhat.com>
69 PR tree-optimization/113210
70 * tree-vect-loop.cc (vect_get_loop_niters): If non-INTEGER_CST
71 value in *number_of_iterationsm1 PLUS_EXPR 1 is folded into
72 INTEGER_CST, recompute *number_of_iterationsm1 as the INTEGER_CST
75 2024-01-09 Eric Botcazou <ebotcazou@adacore.com>
77 PR rtl-optimization/113140
78 * reorg.cc (fill_slots_from_thread): If we are to branch after the
79 last instruction of the function, create an end label.
81 2024-01-09 Roger Sayle <roger@nextmovesoftware.com>
82 Hongtao Liu <hongtao.liu@intel.com>
85 * config/i386/i386-expand.cc
86 (ix86_convert_const_wide_int_to_broadcast): Allow call to
87 ix86_expand_vector_init_duplicate to fail, and return NULL_RTX.
88 (ix86_broadcast_from_constant): Revert recent change; Return a
89 suitable MEMREF independently of mode/target combinations.
90 (ix86_expand_vector_move): Allow ix86_expand_vector_init_duplicate
91 to decide whether expansion is possible/preferrable. Only try
92 forcing DImode constants to memory (and trying again) if calling
93 ix86_expand_vector_init_duplicate fails with an DImode immediate
95 (ix86_expand_vector_init_duplicate) <case E_V2DImode>: Try using
96 V4SImode for suitable immediate constants.
97 <case E_V4DImode>: Try using V8SImode for suitable constants.
98 <case E_V4HImode>: Fail for CONST_INT_P, i.e. use constant pool.
99 <case E_V2HImode>: Likewise.
100 <case E_V8HImode>: For CONST_INT_P try using V4SImode via widen.
101 <case E_V16QImode>: For CONT_INT_P try using V8HImode via widen.
102 <label widen>: Handle CONT_INTs via simplify_binary_operation.
103 Allow recursive calls to ix86_expand_vector_init_duplicate to fail.
104 <case E_V16HImode>: For CONST_INT_P try V8SImode via widen.
105 <case E_V32QImode>: For CONST_INT_P try V16HImode via widen.
106 (ix86_expand_vector_init): Move try using a broadcast for all_same
107 with ix86_expand_vector_init_duplicate before using constant pool.
109 2024-01-09 Chung-Ju Wu <jasonwucj@gmail.com>
111 * doc/invoke.texi (Arm Options): Document Cortex-M52 options.
113 2024-01-09 Chung-Ju Wu <jasonwucj@gmail.com>
115 * config/arm/arm-cpus.in (cortex-m52): New cpu.
116 * config/arm/arm-tables.opt: Regenerate.
117 * config/arm/arm-tune.md: Regenerate.
119 2024-01-09 Jiahao Xu <xujiahao@loongson.cn>
121 * config/loongarch/lasx.md (vec_initv32qiv16qi): Rename to ..
122 (vec_init<mode><lasxhalf>): .. this, and extend to mode.
123 (@vec_concatz<mode>): New insn pattern.
124 * config/loongarch/loongarch.cc (loongarch_expand_vector_group_init):
125 Handle VALS containing two vectors.
127 2024-01-09 Juzhe-Zhong <juzhe.zhong@rivai.ai>
129 * config/riscv/riscv-vector-builtins-functions.def (vleff): Move comments.
132 2024-01-09 Feng Wang <wangfeng@eswincomputing.com>
134 * config/riscv/riscv-vector-builtins-bases.cc (class vandn):
135 Add new function_base for crypto vector.
136 (class bitmanip): Ditto.
137 (class b_reverse):Ditto.
138 (class vwsll): Ditto.
139 (class clmul): Ditto.
140 (class vg_nhab): Ditto.
141 (class crypto_vv):Ditto.
142 (class crypto_vi):Ditto.
143 (class vaeskf2_vsm3c):Ditto.
144 (class vsm3me): Ditto.
145 (BASE): Add BASE declaration for crypto vector.
146 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
147 * config/riscv/riscv-vector-builtins-functions.def (REQUIRED_EXTENSIONS):
148 Add crypto vector intrinsic definition.
176 * config/riscv/riscv-vector-builtins-shapes.cc (struct crypto_vv_def):
177 Add new function_shape for crypto vector.
178 (struct crypto_vi_def): Ditto.
179 (struct crypto_vv_no_op_type_def): Ditto.
180 (SHAPE): Add SHAPE declaration of crypto vector.
181 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
182 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_CRYPTO_SEW32_OPS):
183 Add new data type for crypto vector.
184 (DEF_RVV_CRYPTO_SEW64_OPS): Ditto.
185 (vuint32mf2_t): Ditto.
186 (vuint32m1_t): Ditto.
187 (vuint32m2_t): Ditto.
188 (vuint32m4_t): Ditto.
189 (vuint32m8_t): Ditto.
190 (vuint64m1_t): Ditto.
191 (vuint64m2_t): Ditto.
192 (vuint64m4_t): Ditto.
193 (vuint64m8_t): Ditto.
194 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_CRYPTO_SEW32_OPS):
195 Add new data struct for crypto vector.
196 (DEF_RVV_CRYPTO_SEW64_OPS): Ditto.
197 (registered_function::overloaded_hash): Processing size_t uimm for C overloaded func.
198 * config/riscv/riscv-vector-builtins.def (vi): Add vi OP_TYPE.
200 2024-01-08 Ilya Leoshkevich <iii@linux.ibm.com>
203 * varasm.cc (assemble_function_label_raw): Do not call
204 asan_function_start () without the current function.
206 2024-01-08 Cupertino Miranda <cupertino.miranda@oracle.com>
209 * btfout.cc (btf_collect_datasec): Skip creating BTF info for
210 extern and kernel_helper attributed function decls.
212 2024-01-08 Cupertino Miranda <cupertino.miranda@oracle.com>
214 * btfout.cc (output_btf_strs): Changed.
216 2024-01-08 Tobias Burnus <tobias@codesourcery.com>
218 * config/gcn/mkoffload.cc (main): Handle gfx1100
219 when setting the default XNACK.
221 2024-01-08 Tobias Burnus <tobias@codesourcery.com>
223 * config.gcc (amdgcn-*-amdhsa): Accept --with-arch=gfx1100.
224 * config/gcn/gcn-hsa.h (NO_XNACK): Add gfx1100:
225 (ASM_SPEC): Handle gfx1100.
226 * config/gcn/gcn-opts.h (enum processor_type): Add PROCESSOR_GFX1100.
227 (enum gcn_isa): Add ISA_RDNA3.
228 (TARGET_GFX1100, TARGET_RDNA2_PLUS, TARGET_RDNA3): Define.
229 * config/gcn/gcn-valu.md: Change TARGET_RDNA2 to TARGET_RDNA2_PLUS.
230 * config/gcn/gcn.cc (gcn_option_override,
231 gcn_omp_device_kind_arch_isa, output_file_start): Handle gfx1100.
232 (gcn_global_address_p, gcn_addr_space_legitimate_address_p): Change
233 TARGET_RDNA2 to TARGET_RDNA2_PLUS.
234 (gcn_hsa_declare_function_name): Don't use '.amdhsa_reserve_flat_scratch'
236 * config/gcn/gcn.h (ASSEMBLER_DIALECT): Likewise.
237 (TARGET_CPU_CPP_BUILTINS): Define __RDNA3__, __gfx1030__ and
239 * config/gcn/gcn.md: Change TARGET_RDNA2 to TARGET_RDNA2_PLUS.
240 * config/gcn/gcn.opt (Enum gpu_type): Add gfx1100.
241 * config/gcn/mkoffload.cc (EF_AMDGPU_MACH_AMDGCN_GFX1100): Define.
242 (isa_has_combined_avgprs, main): Handle gfx1100.
243 * config/gcn/t-omp-device (isa): Add gfx1100.
245 2024-01-08 Richard Biener <rguenther@suse.de>
247 * doc/invoke.texi (-mmovbe): Clarify.
249 2024-01-08 Richard Biener <rguenther@suse.de>
251 PR tree-optimization/113026
252 * tree-vect-loop.cc (vect_need_peeling_or_partial_vectors_p):
253 Avoid an epilog in more cases.
254 * tree-vect-loop-manip.cc (vect_do_peeling): Adjust the
255 epilogues niter upper bounds and estimates.
257 2024-01-08 Jakub Jelinek <jakub@redhat.com>
259 PR tree-optimization/113228
260 * gimplify.cc (recalculate_side_effects): Do nothing for SSA_NAMEs.
262 2024-01-08 Jakub Jelinek <jakub@redhat.com>
264 PR tree-optimization/113120
265 * gimple-lower-bitint.cc (gimple_lower_bitint): Fix handling of very
266 large _BitInt zero INTEGER_CST PHI argument.
268 2024-01-08 Jakub Jelinek <jakub@redhat.com>
270 PR tree-optimization/113119
271 * gimple-lower-bitint.cc (optimizable_arith_overflow): Punt if
272 both REALPART_EXPR and cast from IMAGPART_EXPR appear, but cast
273 is before REALPART_EXPR.
275 2024-01-08 Georg-Johann Lay <avr@gjlay.de>
278 * config/avr/avr.cc (avr_handle_addr_attribute): Also print valid
279 range when diagnosing attribute "io" and "io_low" are out of range.
280 (avr_eval_addr_attrib): Don't ICE on empty address at that place.
281 (avr_insert_attributes): Reject if attribute "address", "io" or "io_low"
282 in contexts other than static storage.
283 (avr_asm_output_aligned_decl_common): Move output of decls with
284 attribute "address", "io", and "io_low" to...
285 (avr_output_addr_attrib): ...this new function.
286 (avr_asm_asm_output_aligned_bss): Remove output for decls with
287 attribute "address", "io", and "io_low".
288 (avr_encode_section_info): Rectify handling of decls with attribute
289 "address", "io", and "io_low".
291 2024-01-08 Andrew Stubbs <ams@codesourcery.com>
293 * config/gcn/mkoffload.cc (TEST_XNACK_UNSET): New.
294 (elf_flags): Remove XNACK from the default value.
295 (main): Set a default XNACK according to the arch.
297 2024-01-08 Andrew Stubbs <ams@codesourcery.com>
299 * config/gcn/mkoffload.cc (isa_has_combined_avgprs): Delete.
300 (process_asm): Don't count avgprs.
302 2024-01-08 Hongyu Wang <hongyu.wang@intel.com>
304 * config/i386/i386.opt: Add supported sub-features.
305 * doc/extend.texi: Add description for target attribute.
307 2024-01-08 Feng Wang <wangfeng@eswincomputing.com>
309 * config/riscv/vector.md: Modify avl_type operand index of zvbc ins.
311 2024-01-07 Roger Sayle <roger@nextmovesoftware.com>
312 Uros Bizjak <ubizjak@gmail.com>
315 * config/i386/i386-features.cc (compute_convert_gain): Include
316 the overhead of explicit load and store (movd) instructions when
317 converting non-store scalar operations with memory destinations.
318 Various indentation whitespace fixes.
320 2024-01-07 Tamar Christina <tamar.christina@arm.com>
322 * config/arm/neon.md (cbranch<mode>4): New.
324 2024-01-07 Juzhe-Zhong <juzhe.zhong@rivai.ai>
326 * config/riscv/riscv-vsetvl.cc: replace std::max by MAX.
328 2024-01-06 Jiahao Xu <xujiahao@loongson.cn>
330 * config/loongarch/lasx.md: Set the unused bits in operand[3] to 0.
332 2024-01-06 Juzhe-Zhong <juzhe.zhong@rivai.ai>
335 * config/riscv/riscv-vsetvl.cc (pre_vsetvl::fuse_local_vsetvl_info):
338 2024-01-06 Juzhe-Zhong <juzhe.zhong@rivai.ai>
340 * config/riscv/riscv-vector-costs.cc (loop_invariant_op_p): New function.
341 (variable_vectorized_p): Teach loop invariant.
342 (has_unexpected_spills_p): Ditto.
344 2024-01-06 Juzhe-Zhong <juzhe.zhong@rivai.ai>
346 * config/riscv/riscv-protos.h (whole_reg_to_reg_move_p): New function.
347 * config/riscv/riscv-v.cc (whole_reg_to_reg_move_p): Ditto.
348 * config/riscv/vector.md: Allow non-vlmax with len = NUNITS simplification.
350 2024-01-05 Richard Sandiford <richard.sandiford@arm.com>
353 * doc/invoke.texi (aarch64-sve-compare-costs): Replace with...
354 (aarch64-vect-compare-costs): ...this.
355 * config/aarch64/aarch64.opt (-param=aarch64-sve-compare-costs=):
357 (-param=aarch64-vect-compare-costs=): ...this new param.
358 * config/aarch64/aarch64.cc (aarch64_override_options_internal):
359 Don't disable it when vectorizing for Advanced SIMD only.
360 (aarch64_autovectorize_vector_modes): Apply VECT_COMPARE_COSTS
361 whenever aarch64_vect_compare_costs is true.
363 2024-01-05 Lulu Cheng <chenglulu@loongson.cn>
365 * config/loongarch/lasx.md (lasx_mxld_<lasxfmt_f>):
366 Modify the method of determining the memory offset of [x]vld/[x]vst.
367 (lasx_mxst_<lasxfmt_f>): Likewise.
368 * config/loongarch/loongarch.cc (loongarch_valid_offset_p): Delete.
369 (loongarch_address_insns): Likewise.
370 * config/loongarch/lsx.md (lsx_ld_<lsxfmt_f>): Likewise.
371 (lsx_st_<lsxfmt_f>): Likewise.
372 * config/loongarch/predicates.md (aq10b_operand): Likewise.
373 (aq10h_operand): Likewise.
374 (aq10w_operand): Likewise.
375 (aq10d_operand): Likewise.
377 2024-01-05 Alex Coplan <alex.coplan@arm.com>
380 * config/aarch64/aarch64-ldp-fusion.cc
381 (ldp_bb_info::try_fuse_pair): If the second access can throw,
382 narrow the move range to exactly that insn.
384 2024-01-05 Ilya Leoshkevich <iii@linux.ibm.com>
386 * asan.cc (asan_function_start): Drop switch_to_section ().
387 (asan_emit_stack_protection): Set .LASANPC alignment.
388 * config/i386/i386.cc: Use assemble_function_label_raw ()
389 instead of ASM_OUTPUT_LABEL ().
390 * config/s390/s390.cc (s390_asm_output_function_label):
392 * defaults.h (ASM_OUTPUT_FUNCTION_LABEL): Likewise.
393 * final.cc (final_start_function_1): Drop
394 asan_function_start ().
395 * output.h (assemble_function_label_raw): New function.
396 * varasm.cc (assemble_function_label_raw): Likewise.
398 2024-01-05 Ilya Leoshkevich <iii@linux.ibm.com>
400 * config/aarch64/aarch64.cc (aarch64_declare_function_name):
401 Use ASM_OUTPUT_FUNCTION_LABEL ().
402 * config/alpha/alpha.cc (alpha_start_function): Likewise.
403 * config/arm/aout.h (ASM_DECLARE_FUNCTION_NAME): Likewise.
404 * config/arm/arm.cc (arm_asm_declare_function_name): Likewise.
405 * config/bfin/bfin.h (ASM_DECLARE_FUNCTION_NAME): Likewise.
406 * config/c6x/c6x.h (ASM_DECLARE_FUNCTION_NAME): Likewise.
407 * config/gcn/gcn.cc (gcn_hsa_declare_function_name): Likewise.
408 * config/h8300/h8300.h (ASM_DECLARE_FUNCTION_NAME): Likewise.
409 * config/ia64/ia64.cc (ia64_start_function): Likewise.
410 * config/mcore/mcore-elf.h (ASM_DECLARE_FUNCTION_NAME):
412 * config/microblaze/microblaze.cc (microblaze_function_prologue):
414 * config/mips/mips.cc (mips_start_unique_function): Return the
416 (mips_start_function_definition): Use
417 ASM_OUTPUT_FUNCTION_LABEL ().
418 (mips_finish_stub): Pass the tree to
419 mips_start_function_definition ().
420 (mips16_build_function_stub): Likewise.
421 (mips16_build_call_stub): Likewise.
422 (mips_output_function_prologue): Likewise.
423 * config/pa/pa.cc (pa_output_function_label): Use
424 ASM_OUTPUT_FUNCTION_LABEL ().
425 * config/riscv/riscv.cc (riscv_declare_function_name): Likewise.
426 * config/rs6000/rs6000.cc (rs6000_elf_declare_function_name):
428 (rs6000_xcoff_declare_function_name): Likewise.
430 2024-01-05 Jakub Jelinek <jakub@redhat.com>
432 PR tree-optimization/113201
433 * tree-scalar-evolution.cc (final_value_replacement_loop): Don't call
434 replace_uses_by on SSA_NAME_OCCURS_IN_ABNORMAL_PHI rslt.
436 2024-01-05 Jakub Jelinek <jakub@redhat.com>
438 PR tree-optimization/90693
439 * tree-ssa-math-opts.cc (match_single_bit_test): If
440 tree_expr_nonzero_p (arg), remember it in the second argument to
441 IFN_POPCOUNT or lower it as arg & (arg - 1) == 0 rather than
442 arg ^ (arg - 1) > arg - 1.
443 * internal-fn.cc (expand_POPCOUNT): If second argument to
444 IFN_POPCOUNT suggests arg is non-zero, try to expand it as
445 arg & (arg - 1) == 0 rather than arg ^ (arg - 1) > arg - 1.
447 2024-01-05 Kito Cheng <kito.cheng@sifive.com>
449 * config/riscv/riscv-v.cc (expand_load_store):
451 (expand_cond_len_op): Ditto.
452 (expand_gather_scatter): Ditto.
453 (expand_lanes_load_store): Ditto.
454 (expand_fold_extract_last): Ditto.
456 2024-01-05 Pan Li <pan2.li@intel.com>
459 2024-01-05 Feng Wang <wangfeng@eswincomputing.com>
461 * config/riscv/riscv-vector-builtins-bases.cc (class vandn):
462 Add new function_base for crypto vector.
463 (class bitmanip): Ditto.
464 (class b_reverse):Ditto.
465 (class vwsll): Ditto.
466 (class clmul): Ditto.
467 (class vg_nhab): Ditto.
468 (class crypto_vv):Ditto.
469 (class crypto_vi):Ditto.
470 (class vaeskf2_vsm3c):Ditto.
471 (class vsm3me): Ditto.
472 (BASE): Add BASE declaration for crypto vector.
473 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
474 * config/riscv/riscv-vector-builtins-functions.def (REQUIRED_EXTENSIONS):
475 Add crypto vector intrinsic definition.
503 * config/riscv/riscv-vector-builtins-shapes.cc (struct crypto_vv_def):
504 Add new function_shape for crypto vector.
505 (struct crypto_vi_def): Ditto.
506 (struct crypto_vv_no_op_type_def): Ditto.
507 (SHAPE): Add SHAPE declaration of crypto vector.
508 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
509 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_CRYPTO_SEW32_OPS):
510 Add new data type for crypto vector.
511 (DEF_RVV_CRYPTO_SEW64_OPS): Ditto.
512 (vuint32mf2_t): Ditto.
513 (vuint32m1_t): Ditto.
514 (vuint32m2_t): Ditto.
515 (vuint32m4_t): Ditto.
516 (vuint32m8_t): Ditto.
517 (vuint64m1_t): Ditto.
518 (vuint64m2_t): Ditto.
519 (vuint64m4_t): Ditto.
520 (vuint64m8_t): Ditto.
521 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_CRYPTO_SEW32_OPS):
522 Add new data struct for crypto vector.
523 (DEF_RVV_CRYPTO_SEW64_OPS): Ditto.
524 (registered_function::overloaded_hash): Processing size_t uimm for C overloaded func.
525 * config/riscv/riscv-vector-builtins.def (vi): Add vi OP_TYPE.
527 2024-01-05 Feng Wang <wangfeng@eswincomputing.com>
529 * config/riscv/riscv-vector-builtins-bases.cc (class vandn):
530 Add new function_base for crypto vector.
531 (class bitmanip): Ditto.
532 (class b_reverse):Ditto.
533 (class vwsll): Ditto.
534 (class clmul): Ditto.
535 (class vg_nhab): Ditto.
536 (class crypto_vv):Ditto.
537 (class crypto_vi):Ditto.
538 (class vaeskf2_vsm3c):Ditto.
539 (class vsm3me): Ditto.
540 (BASE): Add BASE declaration for crypto vector.
541 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
542 * config/riscv/riscv-vector-builtins-functions.def (REQUIRED_EXTENSIONS):
543 Add crypto vector intrinsic definition.
571 * config/riscv/riscv-vector-builtins-shapes.cc (struct crypto_vv_def):
572 Add new function_shape for crypto vector.
573 (struct crypto_vi_def): Ditto.
574 (struct crypto_vv_no_op_type_def): Ditto.
575 (SHAPE): Add SHAPE declaration of crypto vector.
576 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
577 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_CRYPTO_SEW32_OPS):
578 Add new data type for crypto vector.
579 (DEF_RVV_CRYPTO_SEW64_OPS): Ditto.
580 (vuint32mf2_t): Ditto.
581 (vuint32m1_t): Ditto.
582 (vuint32m2_t): Ditto.
583 (vuint32m4_t): Ditto.
584 (vuint32m8_t): Ditto.
585 (vuint64m1_t): Ditto.
586 (vuint64m2_t): Ditto.
587 (vuint64m4_t): Ditto.
588 (vuint64m8_t): Ditto.
589 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_CRYPTO_SEW32_OPS):
590 Add new data struct for crypto vector.
591 (DEF_RVV_CRYPTO_SEW64_OPS): Ditto.
592 (registered_function::overloaded_hash): Processing size_t uimm for C overloaded func.
593 * config/riscv/riscv-vector-builtins.def (vi): Add vi OP_TYPE.
595 2024-01-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
597 * config/riscv/riscv-vector-costs.cc (variable_vectorized_p): Teach vi variant.
599 2024-01-04 Andrew Pinski <quic_apinski@quicinc.com>
601 PR tree-optimization/113186
602 * gimple-match-head.cc (gimple_bitwise_inverted_equal_p):
603 Match `^` with the `==` for 1bit integral types.
604 * match.pd (maybe_cmp): Allow for bit_xor for 1bit
607 2024-01-04 David Malcolm <dmalcolm@redhat.com>
609 * toplev.cc (general_init): Pass lang_mask to urlifier.
611 2024-01-04 David Malcolm <dmalcolm@redhat.com>
613 * diagnostic.h (diagnostic_make_option_url_cb): Add lang_mask
615 (diagnostic_context::make_option_url): Update for lang_mask param.
616 * gcc-urlifier.cc: Include "opts.h" and "options.h".
617 (gcc_urlifier::gcc_urlifier): Add lang_mask param.
618 (gcc_urlifier::m_lang_mask): New field.
619 (doc_urls): Make static.
620 (gcc_urlifier::get_url_for_quoted_text): Use label_text.
621 (gcc_urlifier::get_url_suffix_for_quoted_text): Use label_text.
622 Look for an option by name before trying a binary search in
624 (gcc_urlifier::get_url_suffix_for_quoted_text): Use label_text.
625 (gcc_urlifier::get_url_suffix_for_option): New.
626 (make_gcc_urlifier): Add lang_mask param.
627 (selftest::gcc_urlifier_cc_tests): Update for above changes.
628 Verify that a URL is found for "-fpack-struct".
629 * gcc-urlifier.def: Drop options "--version" and "-fpack-struct".
630 * gcc-urlifier.h (make_gcc_urlifier): Add lang_mask param.
631 * gcc.cc (driver::global_initializations): Pass 0 for lang_mask
632 to make_gcc_urlifier.
633 * opts-diagnostic.h (get_option_url): Add lang_mask param.
634 * opts.cc (get_option_html_page): Remove special-casing for
636 (get_option_url_suffix): New.
637 (get_option_url): Reimplement.
638 (selftest::test_get_option_html_page): Rename to...
639 (selftest::test_get_option_url_suffix): ...this and update for
641 (selftest::opts_cc_tests): Update for renaming.
642 * opts.h: Include "rich-location.h".
643 (get_option_url_suffix): New decl.
645 2024-01-04 David Malcolm <dmalcolm@redhat.com>
647 * Makefile.in (ALL_OPT_URL_FILES): New.
648 (GCC_OBJS): Add options-urls.o.
650 (OBJS-libcommon): Likewise.
651 (s-options): Depend on $(ALL_OPT_URL_FILES), and add this to
652 inputs to opt-gather.awk.
653 (options-urls.cc): New Makefile target.
654 * opt-functions.awk (url_suffix): New function.
655 (lang_url_suffix): New function.
656 * options-urls-cc-gen.awk: New file.
657 * opts.h (get_opt_url_suffix): New decl.
659 2024-01-04 David Malcolm <dmalcolm@redhat.com>
661 * params.opt.urls: New file, autogenerated by
662 regenerate-opt-urls.py.
664 2024-01-04 David Malcolm <dmalcolm@redhat.com>
666 * common.opt.urls: New file, autogenerated by
667 regenerate-opt-urls.py.
668 * config/aarch64/aarch64.opt.urls: Likewise.
669 * config/alpha/alpha.opt.urls: Likewise.
670 * config/alpha/elf.opt.urls: Likewise.
671 * config/arc/arc-tables.opt.urls: Likewise.
672 * config/arc/arc.opt.urls: Likewise.
673 * config/arm/arm-tables.opt.urls: Likewise.
674 * config/arm/arm.opt.urls: Likewise.
675 * config/arm/vxworks.opt.urls: Likewise.
676 * config/avr/avr.opt.urls: Likewise.
677 * config/bpf/bpf.opt.urls: Likewise.
678 * config/c6x/c6x-tables.opt.urls: Likewise.
679 * config/c6x/c6x.opt.urls: Likewise.
680 * config/cris/cris.opt.urls: Likewise.
681 * config/cris/elf.opt.urls: Likewise.
682 * config/csky/csky.opt.urls: Likewise.
683 * config/csky/csky_tables.opt.urls: Likewise.
684 * config/darwin.opt.urls: Likewise.
685 * config/dragonfly.opt.urls: Likewise.
686 * config/epiphany/epiphany.opt.urls: Likewise.
687 * config/fr30/fr30.opt.urls: Likewise.
688 * config/freebsd.opt.urls: Likewise.
689 * config/frv/frv.opt.urls: Likewise.
690 * config/ft32/ft32.opt.urls: Likewise.
691 * config/fused-madd.opt.urls: Likewise.
692 * config/g.opt.urls: Likewise.
693 * config/gcn/gcn.opt.urls: Likewise.
694 * config/gnu-user.opt.urls: Likewise.
695 * config/h8300/h8300.opt.urls: Likewise.
696 * config/hpux11.opt.urls: Likewise.
697 * config/i386/cygming.opt.urls: Likewise.
698 * config/i386/cygwin.opt.urls: Likewise.
699 * config/i386/djgpp.opt.urls: Likewise.
700 * config/i386/i386.opt.urls: Likewise.
701 * config/i386/mingw-w64.opt.urls: Likewise.
702 * config/i386/mingw.opt.urls: Likewise.
703 * config/i386/nto.opt.urls: Likewise.
704 * config/ia64/ia64.opt.urls: Likewise.
705 * config/ia64/ilp32.opt.urls: Likewise.
706 * config/ia64/vms.opt.urls: Likewise.
707 * config/iq2000/iq2000.opt.urls: Likewise.
708 * config/linux-android.opt.urls: Likewise.
709 * config/linux.opt.urls: Likewise.
710 * config/lm32/lm32.opt.urls: Likewise.
711 * config/loongarch/loongarch.opt.urls: Likewise.
712 * config/lynx.opt.urls: Likewise.
713 * config/m32c/m32c.opt.urls: Likewise.
714 * config/m32r/m32r.opt.urls: Likewise.
715 * config/m68k/ieee.opt.urls: Likewise.
716 * config/m68k/m68k-tables.opt.urls: Likewise.
717 * config/m68k/m68k.opt.urls: Likewise.
718 * config/m68k/uclinux.opt.urls: Likewise.
719 * config/mcore/mcore.opt.urls: Likewise.
720 * config/microblaze/microblaze.opt.urls: Likewise.
721 * config/mips/mips-tables.opt.urls: Likewise.
722 * config/mips/mips.opt.urls: Likewise.
723 * config/mips/sde.opt.urls: Likewise.
724 * config/mmix/mmix.opt.urls: Likewise.
725 * config/mn10300/mn10300.opt.urls: Likewise.
726 * config/moxie/moxie.opt.urls: Likewise.
727 * config/msp430/msp430.opt.urls: Likewise.
728 * config/nds32/nds32-elf.opt.urls: Likewise.
729 * config/nds32/nds32-linux.opt.urls: Likewise.
730 * config/nds32/nds32.opt.urls: Likewise.
731 * config/netbsd-elf.opt.urls: Likewise.
732 * config/netbsd.opt.urls: Likewise.
733 * config/nios2/elf.opt.urls: Likewise.
734 * config/nios2/nios2.opt.urls: Likewise.
735 * config/nvptx/nvptx-gen.opt.urls: Likewise.
736 * config/nvptx/nvptx.opt.urls: Likewise.
737 * config/openbsd.opt.urls: Likewise.
738 * config/or1k/elf.opt.urls: Likewise.
739 * config/or1k/or1k.opt.urls: Likewise.
740 * config/pa/pa-hpux.opt.urls: Likewise.
741 * config/pa/pa-hpux1010.opt.urls: Likewise.
742 * config/pa/pa-hpux1111.opt.urls: Likewise.
743 * config/pa/pa-hpux1131.opt.urls: Likewise.
744 * config/pa/pa.opt.urls: Likewise.
745 * config/pa/pa64-hpux.opt.urls: Likewise.
746 * config/pdp11/pdp11.opt.urls: Likewise.
747 * config/pru/pru.opt.urls: Likewise.
748 * config/riscv/riscv.opt.urls: Likewise.
749 * config/rl78/rl78.opt.urls: Likewise.
750 * config/rpath.opt.urls: Likewise.
751 * config/rs6000/476.opt.urls: Likewise.
752 * config/rs6000/aix64.opt.urls: Likewise.
753 * config/rs6000/darwin.opt.urls: Likewise.
754 * config/rs6000/linux64.opt.urls: Likewise.
755 * config/rs6000/rs6000-tables.opt.urls: Likewise.
756 * config/rs6000/rs6000.opt.urls: Likewise.
757 * config/rs6000/sysv4.opt.urls: Likewise.
758 * config/rtems.opt.urls: Likewise.
759 * config/rx/elf.opt.urls: Likewise.
760 * config/rx/rx.opt.urls: Likewise.
761 * config/s390/s390.opt.urls: Likewise.
762 * config/s390/tpf.opt.urls: Likewise.
763 * config/sh/sh.opt.urls: Likewise.
764 * config/sh/superh.opt.urls: Likewise.
765 * config/sol2.opt.urls: Likewise.
766 * config/sparc/long-double-switch.opt.urls: Likewise.
767 * config/sparc/sparc.opt.urls: Likewise.
768 * config/stormy16/stormy16.opt.urls: Likewise.
769 * config/v850/v850.opt.urls: Likewise.
770 * config/vax/elf.opt.urls: Likewise.
771 * config/vax/vax.opt.urls: Likewise.
772 * config/visium/visium.opt.urls: Likewise.
773 * config/vms/vms.opt.urls: Likewise.
774 * config/vxworks-smp.opt.urls: Likewise.
775 * config/vxworks.opt.urls: Likewise.
776 * config/xtensa/elf.opt.urls: Likewise.
777 * config/xtensa/uclinux.opt.urls: Likewise.
778 * config/xtensa/xtensa.opt.urls: Likewise.
779 * config/bfin/bfin.opt.urls: New file.
781 2024-01-04 David Malcolm <dmalcolm@redhat.com>
783 * Makefile.in (OPT_URLS_HTML_DEPS): New.
784 (regenerate-opt-urls): New target.
785 (regenerate-opt-urls-unit-test): New target.
786 * doc/options.texi (Option properties): Add UrlSuffix and
787 description of regenerate-opt-urls.py. Add LangUrlSuffix_*.
788 * doc/sourcebuild.texi (Anatomy of a Language Front End): Add
789 reference to regenerate-opt-urls.py's PER_LANGUAGE_OPTION_INDEXES
790 and Makefile.in's OPT_URLS_HTML_DEPS.
791 (Anatomy of a Target Back End): Add
792 reference to regenerate-opt-urls.py's TARGET_SPECIFIC_PAGES.
793 * regenerate-opt-urls.py: New file.
795 2024-01-04 David Malcolm <dmalcolm@redhat.com>
797 * diagnostic-format-sarif.cc
798 (sarif_builder::make_logical_location_object): Convert to...
799 (make_sarif_logical_location_object): ...this.
800 (sarif_builder::set_any_logical_locs_arr): Update for above
802 (sarif_builder::make_thread_flow_location_object): Call
803 maybe_add_sarif_properties on each diagnostic_event.
804 * diagnostic-format-sarif.h (class logical_location): New forward
806 (make_sarif_logical_location_object): New decl.
807 * diagnostic-path.h (class sarif_object): New forward decl.
808 (diagnostic_event::maybe_add_sarif_properties): New vfunc.
810 2024-01-04 Kuan-Lin Chen <rufus@andestech.com>
811 Patrick Lin <patrick@andestech.com>
812 Rufus Chen <rufus@andestech.com>
813 Monk Chiang <monk.chiang@sifive.com>
815 * config/riscv/riscv.cc (riscv_legitimize_move): Expand movfh
816 with Nan-boxing value.
817 * config/riscv/riscv.md (*movhf_softfloat_unspec): New pattern.
819 2024-01-04 Roger Sayle <roger@nextmovesoftware.com>
820 Jeff Law <jlaw@ventanamicro.com>
822 PR rtl-optimization/104914
823 * expr.cc (expand_assignment): When target is SUBREG_PROMOTED_VAR_P
824 a sign or zero extension is only required if the modified field
825 overlaps the SUBREG's most significant bit. On MODE_REP_EXTENDED
826 targets, don't refer to the temporarily incorrectly extended value
827 using a SUBREG, but instead generate an explicit TRUNCATE rtx.
829 2024-01-04 Pan Li <pan2.li@intel.com>
832 2024-01-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
834 * config/riscv/riscv-vector-costs.cc (variable_vectorized_p): Teach vi variant.
836 2024-01-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
838 * config/riscv/riscv-vector-costs.cc (variable_vectorized_p): Teach vi variant.
840 2024-01-04 Kito Cheng <kito.cheng@sifive.com>
842 * config/riscv/riscv.cc (riscv_for_each_saved_reg): Adjust the
845 2024-01-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
847 * config/riscv/riscv-vector-costs.cc (variable_vectorized_p): New function.
848 (compute_nregs_for_mode): Refine LMUL.
849 (max_number_of_live_regs): Ditto.
850 (compute_estimated_lmul): Ditto.
851 (has_unexpected_spills_p): Ditto.
853 2024-01-04 Li Wei <liwei@loongson.cn>
855 * config/loongarch/loongarch.cc (loongarch_is_odd_extraction):
856 Remove useless forward declaration.
857 (loongarch_is_even_extraction): Remove useless forward declaration.
858 (loongarch_try_expand_lsx_vshuf_const): Removed.
859 (loongarch_expand_vec_perm_const_1): Merged.
860 (loongarch_is_double_duplicate): Removed.
861 (loongarch_is_center_extraction): Ditto.
862 (loongarch_is_reversing_permutation): Ditto.
863 (loongarch_is_di_misalign_extract): Ditto.
864 (loongarch_is_si_misalign_extract): Ditto.
865 (loongarch_is_lasx_lowpart_extract): Ditto.
866 (loongarch_is_op_reverse_perm): Ditto.
867 (loongarch_is_single_op_perm): Ditto.
868 (loongarch_is_divisible_perm): Ditto.
869 (loongarch_is_triple_stride_extract): Ditto.
870 (loongarch_expand_vec_perm_const_2): Merged.
871 (loongarch_expand_vec_perm_const): New.
872 (loongarch_vectorize_vec_perm_const): Adjust.
874 2024-01-04 Sandra Loosemore <sandra@codesourcery.com>
876 * omp-general.cc: Fix comment typos and misplaced/confusing
877 comments. Delete redundant include of omp-general.h.
879 2024-01-04 YunQiang Su <syq@gcc.gnu.org>
881 PR rtl-optimization/104914
882 * config/mips/mips.md (insqisi_extended): New patterns.
883 (inshisi_extended): Ditto.
885 2024-01-04 YunQiang Su <syq@gcc.gnu.org>
887 * config/mips/mips.cc (mips_insn_cost): New function.
889 2024-01-04 YunQiang Su <syq@gcc.gnu.org>
891 * config/mips/mips.md (perf_ratio): New attribute.
893 2024-01-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
897 * config/riscv/riscv-vsetvl.cc (invalid_opt_bb_p): New function.
898 (pre_vsetvl::compute_lcm_local_properties): Disable earliest fusion on
899 blocks belong to infinite loop.
900 (pre_vsetvl::emit_vsetvl): Remove fake edges.
901 * config/riscv/t-riscv: Add a new include file.
903 2024-01-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
905 * config/riscv/vector.md: Fix indent.
907 2024-01-03 Kwok Cheung Yeung <kcy@codesourcery.com>
909 * tree-core.h (enum omp_clause_code): Move OMP_CLAUSE_INDIRECT to before
910 OMP_CLAUSE__SIMDUID_.
911 * tree.cc (omp_clause_num_ops): Update position of entry for
912 OMP_CLAUSE_INDIRECT to correspond with omp_clause_code.
913 (omp_clause_code_name): Likewise.
915 2024-01-03 Kwok Cheung Yeung <kcy@codesourcery.com>
917 * config/nvptx/nvptx.cc (nvptx_record_offload_symbol): Restucture
918 printing of FUNC_MAP/IND_FUNC_MAP labels.
920 2024-01-03 Jakub Jelinek <jakub@redhat.com>
922 * gcc.cc (process_command): Update copyright notice dates.
923 * gcov-dump.cc (print_version): Ditto.
924 * gcov.cc (print_version): Ditto.
925 * gcov-tool.cc (print_version): Ditto.
926 * gengtype.cc (create_file): Ditto.
927 * doc/cpp.texi: Bump @copying's copyright year.
928 * doc/cppinternals.texi: Ditto.
929 * doc/gcc.texi: Ditto.
930 * doc/gccint.texi: Ditto.
931 * doc/gcov.texi: Ditto.
932 * doc/install.texi: Ditto.
933 * doc/invoke.texi: Ditto.
935 2024-01-03 Xi Ruoyao <xry111@xry111.site>
937 * config/loongarch/simd.md (fmax<mode>3): New define_insn.
938 (fmin<mode>3): Likewise.
939 (reduc_fmax_scal_<mode>3): New define_expand.
940 (reduc_fmin_scal_<mode>3): Likewise.
942 2024-01-03 Juzhe-Zhong <juzhe.zhong@rivai.ai>
945 * config/riscv/riscv-vector-costs.cc (compute_nregs_for_mode): Add rgroup info.
946 (max_number_of_live_regs): Ditto.
947 (has_unexpected_spills_p): Ditto.
949 2024-01-02 Jun Sha (Joshua) <cooper.joshua@linux.alibaba.com>
950 Jin Ma <jinma@linux.alibaba.com>
951 Xianmiao Qu <cooper.qu@linux.alibaba.com>
952 Christoph Müllner <christoph.muellner@vrull.eu>
954 * config/riscv/vector.md:
955 Use vector_length_operand for vsetvl patterns.
957 2024-01-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
959 * config/riscv/riscv-v.cc (is_vlmax_len_p): Remove satisfies_constraint_K.
960 (expand_cond_len_op): Add simplification of dummy len and dummy mask.
962 2024-01-02 Di Zhao <dizhao@os.amperecomputing.com>
964 * config/aarch64/aarch64-tuning-flags.def
965 (AARCH64_EXTRA_TUNING_OPTION): New tuning option
966 AARCH64_EXTRA_TUNE_FULLY_PIPELINED_FMA.
967 * config/aarch64/aarch64.cc
968 (aarch64_override_options_internal): Set
969 param_fully_pipelined_fma according to tuning option.
970 * config/aarch64/tuning_models/ampere1.h: Add
971 AARCH64_EXTRA_TUNE_FULLY_PIPELINED_FMA to tune_flags.
972 * config/aarch64/tuning_models/ampere1a.h: Likewise.
973 * config/aarch64/tuning_models/ampere1b.h: Likewise.
975 2024-01-02 Feng Wang <wangfeng@eswincomputing.com>
977 * config/riscv/vector-crypto.md: Modify copyright year.
979 2024-01-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
981 * config/riscv/riscv-vector-costs.cc: Move STMT_VINFO_TYPE (...) to local.
983 2024-01-02 Lulu Cheng <chenglulu@loongson.cn>
985 * config.in: Regenerate.
986 * config/loongarch/loongarch-opts.h (HAVE_AS_TLS_LE_RELAXATION): Define.
987 * config/loongarch/loongarch.cc (loongarch_legitimize_tls_address):
988 Added TLS Le Relax support.
989 (loongarch_print_operand_reloc): Add the output string of TLS Le Relax.
990 * config/loongarch/loongarch.md (@add_tls_le_relax<mode>): New template.
991 * configure: Regenerate.
992 * configure.ac: Check if binutils supports TLS le relax.
994 2024-01-02 Feng Wang <wangfeng@eswincomputing.com>
996 * config/riscv/iterators.md: Add rotate insn name.
997 * config/riscv/riscv.md: Add new insns name for crypto vector.
998 * config/riscv/vector-iterators.md: Add new iterators for crypto vector.
999 * config/riscv/vector.md: Add the corresponding attr for crypto vector.
1000 * config/riscv/vector-crypto.md: New file.The machine descriptions for crypto vector.
1002 2024-01-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1005 * config/riscv/riscv-vector-costs.cc (compute_nregs_for_mode): Fix
1006 pointer type liveness count.
1008 Copyright (C) 2024 Free Software Foundation, Inc.
1010 Copying and distribution of this file, with or without modification,
1011 are permitted in any medium without royalty provided the copyright
1012 notice and this notice are preserved.