1; Machine description for AArch64 architecture. 2; Copyright (C) 2009-2020 Free Software Foundation, Inc. 3; Contributed by ARM Ltd. 4; 5; This file is part of GCC. 6; 7; GCC is free software; you can redistribute it and/or modify it 8; under the terms of the GNU General Public License as published by 9; the Free Software Foundation; either version 3, or (at your option) 10; any later version. 11; 12; GCC is distributed in the hope that it will be useful, but 13; WITHOUT ANY WARRANTY; without even the implied warranty of 14; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU 15; General Public License for more details. 16; 17; You should have received a copy of the GNU General Public License 18; along with GCC; see the file COPYING3. If not see 19; <http://www.gnu.org/licenses/>. 20 21HeaderInclude 22config/aarch64/aarch64-opts.h 23 24TargetVariable 25enum aarch64_processor explicit_tune_core = aarch64_none 26 27TargetVariable 28enum aarch64_arch explicit_arch = aarch64_no_arch 29 30TargetSave 31const char *x_aarch64_override_tune_string 32 33TargetVariable 34uint64_t aarch64_isa_flags = 0 35 36TargetVariable 37unsigned aarch64_enable_bti = 2 38 39; The TLS dialect names to use with -mtls-dialect. 40 41Enum 42Name(tls_type) Type(enum aarch64_tls_type) 43The possible TLS dialects: 44 45EnumValue 46Enum(tls_type) String(trad) Value(TLS_TRADITIONAL) 47 48EnumValue 49Enum(tls_type) String(desc) Value(TLS_DESCRIPTORS) 50 51; The code model option names for -mcmodel. 52 53Enum 54Name(cmodel) Type(enum aarch64_code_model) 55The code model option names for -mcmodel: 56 57EnumValue 58Enum(cmodel) String(tiny) Value(AARCH64_CMODEL_TINY) 59 60EnumValue 61Enum(cmodel) String(small) Value(AARCH64_CMODEL_SMALL) 62 63EnumValue 64Enum(cmodel) String(large) Value(AARCH64_CMODEL_LARGE) 65 66mbig-endian 67Target Report RejectNegative Mask(BIG_END) 68Assume target CPU is configured as big endian. 69 70mgeneral-regs-only 71Target Report RejectNegative Mask(GENERAL_REGS_ONLY) Save 72Generate code which uses only the general registers. 73 74mharden-sls= 75Target RejectNegative Joined Var(aarch64_harden_sls_string) 76Generate code to mitigate against straight line speculation. 77 78mfix-cortex-a53-835769 79Target Report Var(aarch64_fix_a53_err835769) Init(2) Save 80Workaround for ARM Cortex-A53 Erratum number 835769. 81 82mfix-cortex-a53-843419 83Target Report Var(aarch64_fix_a53_err843419) Init(2) Save 84Workaround for ARM Cortex-A53 Erratum number 843419. 85 86mlittle-endian 87Target Report RejectNegative InverseMask(BIG_END) 88Assume target CPU is configured as little endian. 89 90mcmodel= 91Target RejectNegative Joined Enum(cmodel) Var(aarch64_cmodel_var) Init(AARCH64_CMODEL_SMALL) Save 92Specify the code model. 93 94mstrict-align 95Target Report Mask(STRICT_ALIGN) Save 96Don't assume that unaligned accesses are handled by the system. 97 98momit-leaf-frame-pointer 99Target Report Var(flag_omit_leaf_frame_pointer) Init(2) Save 100Omit the frame pointer in leaf functions. 101 102mtls-dialect= 103Target RejectNegative Joined Enum(tls_type) Var(aarch64_tls_dialect) Init(TLS_DESCRIPTORS) Save 104Specify TLS dialect. 105 106mtls-size= 107Target RejectNegative Joined Var(aarch64_tls_size) Enum(aarch64_tls_size) 108Specifies bit size of immediate TLS offsets. Valid values are 12, 24, 32, 48. 109 110Enum 111Name(aarch64_tls_size) Type(int) 112 113EnumValue 114Enum(aarch64_tls_size) String(12) Value(12) 115 116EnumValue 117Enum(aarch64_tls_size) String(24) Value(24) 118 119EnumValue 120Enum(aarch64_tls_size) String(32) Value(32) 121 122EnumValue 123Enum(aarch64_tls_size) String(48) Value(48) 124 125march= 126Target RejectNegative Negative(march=) ToLower Joined Var(aarch64_arch_string) 127Use features of architecture ARCH. 128 129mcpu= 130Target RejectNegative Negative(mcpu=) ToLower Joined Var(aarch64_cpu_string) 131Use features of and optimize for CPU. 132 133mtune= 134Target RejectNegative Negative(mtune=) ToLower Joined Var(aarch64_tune_string) 135Optimize for CPU. 136 137mabi= 138Target RejectNegative Joined Enum(aarch64_abi) Var(aarch64_abi) Init(AARCH64_ABI_DEFAULT) 139Generate code that conforms to the specified ABI. 140 141moverride= 142Target RejectNegative ToLower Joined Var(aarch64_override_tune_string) 143-moverride=<string> Power users only! Override CPU optimization parameters. 144 145Enum 146Name(aarch64_abi) Type(int) 147Known AArch64 ABIs (for use with the -mabi= option): 148 149EnumValue 150Enum(aarch64_abi) String(ilp32) Value(AARCH64_ABI_ILP32) 151 152EnumValue 153Enum(aarch64_abi) String(lp64) Value(AARCH64_ABI_LP64) 154 155mpc-relative-literal-loads 156Target Report Save Var(pcrelative_literal_loads) Init(2) Save 157PC relative literal loads. 158 159mbranch-protection= 160Target RejectNegative Joined Var(aarch64_branch_protection_string) Save 161Use branch-protection features. 162 163msign-return-address= 164Target WarnRemoved RejectNegative Joined Enum(aarch64_ra_sign_scope_t) Var(aarch64_ra_sign_scope) Init(AARCH64_FUNCTION_NONE) Save 165Select return address signing scope. 166 167Enum 168Name(aarch64_ra_sign_scope_t) Type(enum aarch64_function_type) 169Supported AArch64 return address signing scope (for use with -msign-return-address= option): 170 171EnumValue 172Enum(aarch64_ra_sign_scope_t) String(none) Value(AARCH64_FUNCTION_NONE) 173 174EnumValue 175Enum(aarch64_ra_sign_scope_t) String(non-leaf) Value(AARCH64_FUNCTION_NON_LEAF) 176 177EnumValue 178Enum(aarch64_ra_sign_scope_t) String(all) Value(AARCH64_FUNCTION_ALL) 179 180mlow-precision-recip-sqrt 181Target Var(flag_mrecip_low_precision_sqrt) Optimization 182Enable the reciprocal square root approximation. Enabling this reduces 183precision of reciprocal square root results to about 16 bits for 184single precision and to 32 bits for double precision. 185 186mlow-precision-sqrt 187Target Var(flag_mlow_precision_sqrt) Optimization 188Enable the square root approximation. Enabling this reduces 189precision of square root results to about 16 bits for 190single precision and to 32 bits for double precision. 191If enabled, it implies -mlow-precision-recip-sqrt. 192 193mlow-precision-div 194Target Var(flag_mlow_precision_div) Optimization 195Enable the division approximation. Enabling this reduces 196precision of division results to about 16 bits for 197single precision and to 32 bits for double precision. 198 199Enum 200Name(sve_vector_bits) Type(enum aarch64_sve_vector_bits_enum) 201The possible SVE vector lengths: 202 203EnumValue 204Enum(sve_vector_bits) String(scalable) Value(SVE_SCALABLE) 205 206EnumValue 207Enum(sve_vector_bits) String(128) Value(SVE_128) 208 209EnumValue 210Enum(sve_vector_bits) String(256) Value(SVE_256) 211 212EnumValue 213Enum(sve_vector_bits) String(512) Value(SVE_512) 214 215EnumValue 216Enum(sve_vector_bits) String(1024) Value(SVE_1024) 217 218EnumValue 219Enum(sve_vector_bits) String(2048) Value(SVE_2048) 220 221msve-vector-bits= 222Target RejectNegative Joined Enum(sve_vector_bits) Var(aarch64_sve_vector_bits) Init(SVE_SCALABLE) 223-msve-vector-bits=<number> Set the number of bits in an SVE vector register. 224 225mverbose-cost-dump 226Target Undocumented Var(flag_aarch64_verbose_cost) 227Enables verbose cost model dumping in the debug dump files. 228 229mtrack-speculation 230Target Var(aarch64_track_speculation) 231Generate code to track when the CPU might be speculating incorrectly. 232 233mstack-protector-guard= 234Target RejectNegative Joined Enum(stack_protector_guard) Var(aarch64_stack_protector_guard) Init(SSP_GLOBAL) 235Use given stack-protector guard. 236 237Enum 238Name(stack_protector_guard) Type(enum stack_protector_guard) 239Valid arguments to -mstack-protector-guard=: 240 241EnumValue 242Enum(stack_protector_guard) String(sysreg) Value(SSP_SYSREG) 243 244EnumValue 245Enum(stack_protector_guard) String(global) Value(SSP_GLOBAL) 246 247mstack-protector-guard-reg= 248Target Joined RejectNegative String Var(aarch64_stack_protector_guard_reg_str) 249Use the system register specified on the command line as the stack protector 250guard register. This option is for use with fstack-protector-strong and 251not for use in user-land code. 252 253mstack-protector-guard-offset= 254Target Joined RejectNegative String Var(aarch64_stack_protector_guard_offset_str) 255Use an immediate to offset from the stack protector guard register, sp_el0. 256This option is for use with fstack-protector-strong and not for use in 257user-land code. 258 259TargetVariable 260long aarch64_stack_protector_guard_offset = 0 261 262moutline-atomics 263Target Report Var(aarch64_flag_outline_atomics) Init(2) Save 264Generate local calls to out-of-line atomic operations. 265 266-param=aarch64-sve-compare-costs= 267Target Joined UInteger Var(aarch64_sve_compare_costs) Init(1) IntegerRange(0, 1) Param 268When vectorizing for SVE, consider using unpacked vectors for smaller elements and use the cost model to pick the cheapest approach. Also use the cost model to choose between SVE and Advanced SIMD vectorization. 269 270-param=aarch64-float-recp-precision= 271Target Joined UInteger Var(aarch64_float_recp_precision) Init(1) IntegerRange(1, 5) Param 272The number of Newton iterations for calculating the reciprocal for float type. The precision of division is proportional to this param when division approximation is enabled. The default value is 1. 273 274-param=aarch64-double-recp-precision= 275Target Joined UInteger Var(aarch64_double_recp_precision) Init(2) IntegerRange(1, 5) Param 276The number of Newton iterations for calculating the reciprocal for double type. The precision of division is proportional to this param when division approximation is enabled. The default value is 2. 277 278-param=aarch64-autovec-preference= 279Target Joined UInteger Var(aarch64_autovec_preference) Init(0) IntegerRange(0, 4) Param 280