Go to the source code of this file.
Data Structures | |
| class | extraction_insn |
Macros | |
| #define | find_widening_optab_handler(A, B, C) find_widening_optab_handler_and_mode (A, B, C, NULL) |
Enumerations | |
| enum | extraction_pattern { EP_insv , EP_extv , EP_extzv } |
Macro Definition Documentation
◆ find_widening_optab_handler
| #define find_widening_optab_handler | ( | A, | |
| B, | |||
| C ) find_widening_optab_handler_and_mode (A, B, C, NULL) |
Find a widening optab even if it doesn't widen as much as we want.
Referenced by can_widen_mult_without_libcall(), expand_binop(), expand_expr_real_2(), and expand_widen_pattern_expr().
Enumeration Type Documentation
◆ extraction_pattern
Function Documentation
◆ can_atomic_exchange_p()
Return true if an atomic exchange can be performed.
References can_compare_and_swap_p(), direct_optab_handler(), and ggc_alloc().
Referenced by expand_omp_atomic_store().
◆ can_atomic_load_p()
| bool can_atomic_load_p | ( | machine_mode | mode | ) |
Return true if an atomic load can be performed without falling back to a compare-and-swap.
References BITS_PER_WORD, direct_optab_handler(), GET_MODE_PRECISION(), ggc_alloc(), and known_le.
Referenced by expand_atomic_compare_and_swap(), expand_atomic_exchange(), expand_atomic_fetch_op(), expand_atomic_store(), expand_omp_atomic_cas(), expand_omp_atomic_fetch_op(), expand_omp_atomic_pipeline(), and fold_builtin_atomic_always_lock_free().
◆ can_compare_and_swap_p()
Return true if there is a compare_and_swap pattern.
References direct_optab_handler(), ggc_alloc(), optab_handler(), and optab_libfunc().
Referenced by can_atomic_exchange_p(), expand_atomic_fetch_op(), expand_omp_atomic_cas(), expand_omp_atomic_fetch_op(), expand_omp_atomic_pipeline(), fold_builtin_atomic_always_lock_free(), maybe_emit_compare_and_swap_exchange_loop(), and maybe_emit_sync_lock_test_and_set().
◆ can_conditionally_move_p()
| bool can_conditionally_move_p | ( | machine_mode | mode | ) |
Return nonzero if a conditional move of mode MODE is supported. This function is for combine so it can tell whether an insn that looks like a conditional move is actually supported by the hardware. If we guess wrong we lose a bit on optimization, but that's it.
??? sparc64 supports conditionally moving integers values based on fp comparisons, and vice versa. How do we handle them?
References direct_optab_handler(), and ggc_alloc().
Referenced by bb_ok_for_noce_convert_multiple_sets(), expand_cond_expr_using_cmove(), expand_expr_real_2(), noce_try_cmove_arith(), and simplify_set().
◆ can_extend_p()
| enum insn_code can_extend_p | ( | machine_mode | to_mode, |
| machine_mode | from_mode, | ||
| int | unsignedp ) |
Return the insn code used to extend FROM_MODE to TO_MODE. UNSIGNEDP specifies zero-extension instead of sign-extension. If no such operation exists, CODE_FOR_nothing will be returned.
References convert_optab_handler(), ggc_alloc(), and targetm.
Referenced by assign_parm_setup_reg(), compress_float_constant(), convert_mode_scalar(), gen_extend_insn(), init_expr_target(), and supportable_convert_operation().
◆ can_fix_p()
| enum insn_code can_fix_p | ( | machine_mode | fixmode, |
| machine_mode | fltmode, | ||
| int | unsignedp, | ||
| bool * | truncp_ptr ) |
Return the insn code to convert floating-point mode FLTMODE to fixed-point mode FIXMODE, or CODE_FOR_nothing if no such instruction exists. UNSIGNEDP specifies whether FIXMODE is unsigned. On a successful return, set *TRUNCP_PTR to true if it is necessary to output an explicit FTRUNC before the instruction.
References convert_optab_handler(), ggc_alloc(), and optab_handler().
Referenced by expand_fix(), and supportable_convert_operation().
◆ can_float_p()
| enum insn_code can_float_p | ( | machine_mode | fltmode, |
| machine_mode | fixmode, | ||
| int | unsignedp ) |
Return the insn code to convert fixed-point mode FIXMODE to floating-point mode FLTMODE, or CODE_FOR_nothing if no such instruction exists. UNSIGNEDP specifies whether FIXMODE is unsigned.
References convert_optab_handler(), and ggc_alloc().
Referenced by expand_float(), simplify_using_ranges::simplify_float_conversion_using_ranges(), and supportable_convert_operation().
◆ can_mult_highpart_p()
| int can_mult_highpart_p | ( | machine_mode | mode, |
| bool | uns_p ) |
Return non-zero if a highpart multiply is supported or can be synthesized. For the benefit of expand_mult_highpart, the return value is 1 for direct, 2 for integral widening, 3 for even/odd widening, 4 for hi/lo widening.
References BITS_PER_WORD, can_vec_perm_const_p(), convert_optab_handler(), GET_MODE_BITSIZE(), GET_MODE_CLASS, GET_MODE_NUNITS(), GET_MODE_WIDER_MODE(), ggc_alloc(), i, and optab_handler().
Referenced by default_preferred_div_as_shifts_over_mult(), expand_mul_overflow(), expand_mult_highpart(), expand_vector_divmod(), get_compute_type(), match_arith_overflow(), vect_recog_divmod_pattern(), and vectorizable_operation().
◆ can_vec_extract()
| bool can_vec_extract | ( | machine_mode | mode, |
| machine_mode | extr_mode ) |
Whether we can extract part of the vector mode MODE as (scalar or vector) mode EXTR_MODE.
References convert_optab_handler(), GET_MODE_BITSIZE(), GET_MODE_SIZE(), ggc_alloc(), int_mode_for_size(), related_vector_mode(), and VECTOR_MODE_P.
Referenced by vect_find_reusable_accumulator().
◆ can_vec_perm_const_p()
| bool can_vec_perm_const_p | ( | machine_mode | mode, |
| machine_mode | op_mode, | ||
| const vec_perm_indices & | sel, | ||
| bool | allow_variable_p ) |
Return true if the target directly supports VEC_PERM_EXPRs on vectors of mode OP_MODE and result vector of mode MODE using the selector SEL. ALLOW_VARIABLE_P is true if it is acceptable to force the selector into a register and use a variable permute (if the target supports that). Note that additional permutations representing whole-vector shifts may also be handled via the vec_shr or vec_shl optab, but only where the second input vector is entirely constant zeroes; this case is not dealt with here.
References direct_optab_handler(), GET_MODE_UNIT_SIZE, ggc_alloc(), NULL, NULL_RTX, qimode_for_vec_perm(), selector_fits_mode_p(), targetm, and VECTOR_MODE_P.
Referenced by can_duplicate_and_interleave_p(), can_mult_highpart_p(), have_whole_vector_shift(), interleave_supported_p(), lower_vec_perm(), perm_mask_for_reverse(), scan_store_can_perm_p(), simplify_vector_constructor(), vect_gen_perm_mask_checked(), vect_grouped_load_supported(), vect_grouped_store_supported(), vect_recog_rotate_pattern(), vect_shift_permute_load_chain(), vect_transform_slp_perm_load_1(), vectorizable_bswap(), vectorizable_nonlinear_induction(), vectorizable_recurr(), and vectorizable_slp_permutation_1().
◆ can_vec_perm_var_p()
| bool can_vec_perm_var_p | ( | machine_mode | mode | ) |
Return true if VEC_PERM_EXPRs with variable selector operands can be expanded using SIMD extensions of the CPU. MODE is the mode of the vectors being permuted.
References direct_optab_handler(), GET_MODE_MASK, GET_MODE_NUNITS(), GET_MODE_UNIT_SIZE, ggc_alloc(), maybe_gt, optab_handler(), qimode_for_vec_perm(), and VECTOR_MODE_P.
Referenced by lower_vec_perm().
◆ convert_optab_handler() [1/2]
|
inline |
Return the insn used to perform conversion OP from mode FROM_MODE to mode TO_MODE; return CODE_FOR_nothing if the target does not have such an insn.
References convert_optab_p(), gcc_assert, and ggc_alloc().
Referenced by get_vcond_eq_icode(), get_vcond_icode(), get_vcond_mask_icode(), get_vec_cmp_eq_icode(), and get_vec_cmp_icode().
◆ convert_optab_handler() [2/2]
| enum insn_code convert_optab_handler | ( | convert_optab | optab, |
| machine_mode | to_mode, | ||
| machine_mode | from_mode, | ||
| optimization_type | opt_type ) |
Return the insn used to perform conversion OP from mode FROM_MODE to mode TO_MODE; return CODE_FOR_nothing if the target does not have such an insn, or if it is unsuitable for optimization type OPT_TYPE.
References convert_optab_handler(), ggc_alloc(), and targetm.
Referenced by can_extend_p(), can_fix_p(), can_float_p(), can_mult_highpart_p(), can_vec_extract(), can_vec_extract_var_idx_p(), convert_mode_scalar(), convert_move(), convert_optab_handler(), convert_optab_supported_p(), expand_binop(), expand_convert_optab_fn(), expand_fixed_convert(), expand_gather_load_optab_fn(), expand_partial_load_optab_fn(), expand_partial_store_optab_fn(), expand_scatter_store_optab_fn(), expand_sfix_optab(), expand_vec_cond_mask_optab_fn(), expand_vec_cond_optab_fn(), expand_vector_broadcast(), expand_while_optab_fn(), expmed_mult_highpart_optab(), extract_bit_field_1(), find_widening_optab_handler_and_mode(), get_multi_vector_move(), internal_gather_scatter_fn_supported_p(), internal_len_load_store_bias(), multi_vector_optab_supported_p(), store_constructor(), supportable_convert_operation(), supports_vec_convert_optab_p(), target_supports_len_load_store_p(), target_supports_mask_load_store_p(), vect_create_partial_epilog(), vect_lanes_optab_supported_p(), vector_vector_composition_type(), and vectorizable_store().
◆ convert_optab_p()
IR-agnostic target query functions relating to optabs Copyright (C) 2001-2024 Free Software Foundation, Inc. This file is part of GCC. GCC is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 3, or (at your option) any later version. GCC is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with GCC; see the file COPYING3. If not see <http://www.gnu.org/licenses/>.
Return true if OP is a conversion optab.
References ggc_alloc(), and unknown_optab.
Referenced by convert_optab_handler(), and expand_binop().
◆ direct_optab_handler() [1/2]
| enum insn_code direct_optab_handler | ( | convert_optab | optab, |
| machine_mode | mode, | ||
| optimization_type | opt_type ) |
Return the insn used to implement mode MODE of OP; return CODE_FOR_nothing if the target does not have such an insn, or if it is unsuitable for optimization type OPT_TYPE.
References direct_optab_handler(), ggc_alloc(), and targetm.
Referenced by can_atomic_exchange_p(), can_atomic_load_p(), can_compare_and_swap_p(), can_conditionally_move_p(), can_vec_perm_const_p(), can_vec_perm_var_p(), default_secondary_reload(), direct_optab_handler(), direct_optab_supported_p(), emit_block_cmp_via_cmpmem(), emit_block_move_via_pattern(), emit_conditional_move(), emit_conditional_move_1(), emit_conditional_neg_or_complement(), expand_atomic_compare_and_swap(), expand_atomic_load(), expand_atomic_store(), expand_builtin_feclear_feraise_except(), expand_builtin_fegetround(), expand_builtin_set_thread_pointer(), expand_builtin_strcmp(), expand_builtin_strncmp(), expand_builtin_thread_pointer(), expand_direct_optab_fn(), expand_ifn_atomic_bit_test_and(), expand_ifn_atomic_op_fetch_cmp_0(), expand_partial_load_optab_fn(), expand_partial_store_optab_fn(), expand_RAWMEMCHR(), expand_vec_perm_const(), expand_vec_perm_var(), expand_vec_series_expr(), get_optab_extraction_insn(), if_convert(), internal_check_ptrs_fn_supported_p(), internal_len_load_store_bias(), maybe_emit_atomic_exchange(), maybe_emit_op(), optimize_atomic_compare_exchange_p(), prepare_cmp_insn(), set_storage_via_setmem(), target_supports_len_load_store_p(), loop_distribution::transform_reduction_loop(), validate_test_and_branch(), and vectorizable_early_exit().
◆ direct_optab_handler() [2/2]
|
inline |
Return the insn used to implement mode MODE of OP, or CODE_FOR_nothing if the target does not have such an insn.
References optab_handler().
◆ find_widening_optab_handler_and_mode() [1/2]
|
inline |
Version of find_widening_optab_handler_and_mode that operates on specific mode types.
References find_widening_optab_handler_and_mode(), and ggc_alloc().
◆ find_widening_optab_handler_and_mode() [2/2]
| enum insn_code find_widening_optab_handler_and_mode | ( | optab | op, |
| machine_mode | to_mode, | ||
| machine_mode | from_mode, | ||
| machine_mode * | found_mode ) |
Find a widening optab even if it doesn't widen as much as we want. E.g. if from_mode is HImode, and to_mode is DImode, and there is no direct HI->SI insn, then return SI->DI, if that exists.
References convert_optab_handler(), FOR_EACH_MODE, gcc_checking_assert, GET_MODE_CLASS, GET_MODE_PRECISION(), GET_MODE_WIDER_MODE(), ggc_alloc(), and known_lt.
Referenced by convert_mult_to_widen(), convert_plusminus_to_widen(), and find_widening_optab_handler_and_mode().
◆ get_best_mem_extraction_insn()
| bool get_best_mem_extraction_insn | ( | extraction_insn * | insn, |
| enum extraction_pattern | pattern, | ||
| HOST_WIDE_INT | bitsize, | ||
| HOST_WIDE_INT | bitnum, | ||
| machine_mode | field_mode ) |
Return true if an instruction exists to access a field of BITSIZE bits starting BITNUM bits into a memory structure. Describe the "best" such instruction in *INSN if so. PATTERN describes the type of insertion or extraction we want to perform and FIELDMODE is the natural mode of the extracted field. The instructions considered here only access bytes that overlap the bitfield; they do not touch any surrounding bytes.
References ET_unaligned_mem, get_best_extraction_insn(), and ggc_alloc().
Referenced by extract_integral_bit_field(), and store_integral_bit_field().
◆ get_best_reg_extraction_insn()
| bool get_best_reg_extraction_insn | ( | extraction_insn * | insn, |
| enum extraction_pattern | pattern, | ||
| unsigned HOST_WIDE_INT | struct_bits, | ||
| machine_mode | field_mode ) |
Return true if an instruction exists to access a field of mode FIELDMODE in a register structure that has STRUCT_BITS significant bits. Describe the "best" such instruction in *INSN if so. PATTERN describes the type of insertion or extraction we want to perform. For an insertion, the number of significant structure bits includes all bits of the target. For an extraction, it need only include the most significant bit of the field. Larger widths are acceptable in both cases.
References ET_reg, get_best_extraction_insn(), and ggc_alloc().
Referenced by adjust_bit_field_mem_for_reg(), extract_integral_bit_field(), make_extraction(), and store_integral_bit_field().
◆ get_vcond_eq_icode()
|
inline |
Return insn code for a conditional operator with a comparison in mode CMODE (only EQ/NE), resulting in a value of mode VMODE.
References convert_optab_handler(), and ggc_alloc().
Referenced by vcond_eq_icode_p().
◆ get_vcond_icode()
Return insn code for a conditional operator with a comparison in mode CMODE, unsigned if UNS is true, resulting in a value of mode VMODE.
References convert_optab_handler(), and ggc_alloc().
Referenced by can_vcond_compare_p(), and gimple_expand_vec_cond_expr().
◆ get_vcond_mask_icode()
|
inline |
Return insn code for a conditional operator with a mask mode MMODE resulting in a value of mode VMODE.
References convert_optab_handler(), and ggc_alloc().
Referenced by expand_vec_cond_expr_p(), and gimple_expand_vec_cond_expr().
◆ get_vec_cmp_eq_icode()
|
inline |
Return insn code for a comparison operator with VMODE resultin MASK_MODE (only for EQ/NE).
References convert_optab_handler(), and ggc_alloc().
Referenced by expand_vec_cmp_expr(), and vec_cmp_eq_icode_p().
◆ get_vec_cmp_icode()
Return insn code for a comparison operator with VMODE resultin MASK_MODE, unsigned if UNS is true.
References convert_optab_handler(), and ggc_alloc().
Referenced by can_vec_cmp_compare_p(), and expand_vec_cmp_expr().
◆ lshift_cheap_p()
Determine whether "1 << x" is relatively cheap in word_mode.
References const1_rtx, COSTS_N_INSNS, gen_raw_REG(), ggc_alloc(), optab_handler(), set_src_cost(), and word_mode.
Referenced by optimize_range_tests().
◆ optab_handler()
Return the insn used to implement mode MODE of OP, or CODE_FOR_nothing if the target does not have such an insn.
References gcc_assert, and ggc_alloc().
Referenced by add_rshift(), assign_parm_adjust_stack_rtl(), assign_parm_setup_reg(), assign_parm_setup_stack(), build_constant_desc(), builtin_memset_gen_str(), by_pieces_mode_supported_p(), by_pieces_ninsns(), can_compare_and_swap_p(), can_compare_p(), can_fix_p(), can_mult_highpart_p(), can_store_by_pieces(), can_vec_perm_var_p(), can_vec_set_var_idx_p(), check_scan_store(), convert_expand_mult_copysign(), convert_mult_to_highpart(), default_builtin_support_vector_misalignment(), direct_optab_handler(), directly_supported_p(), emit_cmp_and_jump_insn_1(), emit_conditional_add(), emit_move_ccmode(), emit_move_complex(), emit_move_insn(), emit_move_insn_1(), emit_move_via_integer(), emit_store_flag_1(), emit_store_flag_int(), emit_storent_insn(), expand_abs_nojump(), expand_addsub_overflow(), expand_assignment(), expand_atomic_compare_and_swap(), expand_binop(), expand_builtin_cexpi(), expand_builtin_issignaling(), expand_builtin_mathfn_3(), expand_builtin_mathfn_ternary(), expand_builtin_signbit(), expand_builtin_sincos(), expand_builtin_strlen(), expand_ccmp_expr(), expand_clrsb_using_clz(), expand_copysign(), expand_copysign_absneg(), expand_ctz(), expand_divmod(), expand_DIVMOD(), expand_doubleword_clz_ctz_ffs(), expand_doubleword_mod(), expand_expr_real_1(), expand_expr_real_2(), expand_ffs(), expand_fix(), expand_misaligned_mem_ref(), expand_mul_overflow(), expand_mult(), expand_mult_highpart(), expand_neg_overflow(), expand_omp_for_init_vars(), expand_one_cmpl_abs_nojump(), expand_parity(), expand_shift_1(), expand_smod_pow2(), expand_SPACESHIP(), expand_ternary_op(), expand_twoval_binop(), expand_twoval_unop(), expand_UADDC(), expand_unop(), expand_unop_direct(), expand_vec_perm_const(), expand_vec_set_optab_fn(), expand_vector_broadcast(), expand_vector_conversion(), expand_vector_divmod(), expand_vector_operations_1(), expand_vector_ubsan_overflow(), expand_widen_pattern_expr(), expmed_mult_highpart_optab(), expression_expensive_p(), find_if_header(), find_reloads_address_1(), gen_add2_insn(), gen_add3_insn(), gen_addptr3_insn(), gen_cond_trap(), gen_reload(), gen_sub2_insn(), gen_sub3_insn(), get_compute_type(), gimple_expand_builtin_cabs(), gimple_expand_builtin_pow(), gimple_fold_builtin_memory_op(), have_add2_insn(), have_addptr3_insn(), have_insn_for(), have_sub2_insn(), have_whole_vector_shift(), hoist_adjacent_loads(), interclass_mathfn_icode(), lower_vec_perm(), lshift_cheap_p(), match_arith_overflow(), match_uaddc_usubc(), maybe_emit_op(), maybe_emit_sync_lock_test_and_set(), nontemporal_store_p(), optimize_atomic_bit_test_and(), optimize_atomic_compare_exchange_p(), optimize_atomic_op_fetch_cmp_0(), optimize_spaceship(), optimize_vector_constructor(), prepare_cmp_insn(), move_by_pieces_d::prepare_mode(), store_by_pieces_d::prepare_mode(), compare_by_pieces_d::prepare_mode(), reverse_rotate_by_imm_p(), scan_store_can_perm_p(), simplify_vector_constructor(), store_bit_field_1(), store_constructor(), store_integral_bit_field(), supportable_half_widening_operation(), supportable_narrowing_operation(), supportable_widening_operation(), target_has_vecop_for_code(), target_supports_divmod_p(), target_supports_op_p(), type_for_widest_vector_mode(), undistribute_bitref_for_vector(), vect_maybe_permute_loop_masks(), vect_recog_divmod_pattern(), vect_recog_mult_pattern(), vect_recog_rotate_pattern(), vect_supportable_direct_optab_p(), vect_supportable_dr_alignment(), vect_supportable_shift(), vectorizable_comparison_1(), vectorizable_condition(), vectorizable_load(), vectorizable_operation(), vectorizable_shift(), vectorizable_store(), widen_bswap(), and widen_leading().
◆ qimode_for_vec_perm()
| opt_machine_mode qimode_for_vec_perm | ( | machine_mode | mode | ) |
If a target doesn't implement a permute on a vector with multibyte elements, we can try to do the same permute on byte elements. If this makes sense for vector mode MODE then return the appropriate byte vector mode.
References GET_MODE_INNER, GET_MODE_SIZE(), ggc_alloc(), and related_vector_mode().
Referenced by can_vec_perm_const_p(), can_vec_perm_var_p(), expand_vec_perm_const(), and expand_vec_perm_var().
◆ selector_fits_mode_p()
| bool selector_fits_mode_p | ( | machine_mode | mode, |
| const vec_perm_indices & | sel ) |
Return true if selector SEL can be represented in the integer equivalent of vector mode MODE.
References GET_MODE_INNER, GET_MODE_MASK, ggc_alloc(), and HOST_WIDE_INT_M1U.
Referenced by can_vec_perm_const_p(), and expand_vec_perm_const().
◆ supports_vec_gather_load_p()
| bool supports_vec_gather_load_p | ( | machine_mode | mode | ) |
If MODE is not VOIDmode, return true if vec_gather_load is available for that mode. If MODE is VOIDmode, return true if gather_load is available for at least one vector mode.
References ggc_alloc(), supports_vec_convert_optab_p(), and this_fn_optabs.
Referenced by vect_check_gather_scatter().
◆ supports_vec_scatter_store_p()
| bool supports_vec_scatter_store_p | ( | machine_mode | mode | ) |
If MODE is not VOIDmode, return true if vec_scatter_store is available for that mode. If MODE is VOIDmode, return true if scatter_store is available for at least one vector mode.
References ggc_alloc(), supports_vec_convert_optab_p(), and this_fn_optabs.
Referenced by vect_check_gather_scatter().
◆ trapv_binoptab_p()
Return true if BINOPTAB is for a trapping-on-overflow operation.
References ggc_alloc().
Referenced by expand_binop(), and gen_int_libfunc().
◆ trapv_unoptab_p()
Return true if UNOPTAB is for a trapping-on-overflow operation.
References ggc_alloc().
Referenced by expand_unop(), and gen_int_libfunc().
