---

3.20.41 RISC-V Options ¶

These command-line options are defined for RISC-V targets:

-mbranch-cost=n ¶

Set the cost of branches to roughly n instructions.

-mabi=ABI-string ¶

Specify integer and floating-point calling convention. ABI-string contains two parts: the size of integer types and the registers used for floating-point types. For example ‘-march=rv64ifd -mabi=lp64d’ means that ‘long’ and pointers are 64-bit (implicitly defining ‘int’ to be 32-bit), and that floating-point values up to 64 bits wide are passed in F registers. Contrast this with ‘-march=rv64ifd -mabi=lp64f’, which still allows the compiler to generate code that uses the F and D extensions but only allows floating-point values up to 32 bits long to be passed in registers; or ‘-march=rv64ifd -mabi=lp64’, in which no floating-point arguments are passed in registers.

The default for this argument is system dependent; if you want a specific calling convention you should specify one explicitly. The valid calling conventions are: ‘ilp32’, ‘ilp32f’, ‘ilp32d’, ‘lp64’, ‘lp64f’, and ‘lp64d’. Some calling conventions are impossible to implement on some ISAs: for example, ‘-march=rv32if -mabi=ilp32d’ is invalid because the ABI requires 64-bit values be passed in F registers, but F registers are only 32 bits wide. There are also the ‘ilp32e’ ABI that can only be used with the ‘rv32e’ architecture and the ‘lp64e’ ABI that can only be used with the ‘rv64e’. Those ABIs are not well-specified at present, and are subject to change.

-mfdiv ¶

-mno-fdiv

Do or don’t use hardware floating-point divide and square root instructions. This requires the F or D extensions for floating-point registers. The default is to use them if the specified architecture has these instructions.

-mfence-tso ¶

-mno-fence-tso

Do or don’t use the ‘fence.tso’ instruction, which is unimplemented on some processors (including those from T-Head). If the ‘fence.tso’ instruction is not available then a stronger fence is used instead.

-mdiv ¶

-mno-div

Do or don’t use hardware instructions for integer division. This requires the M extension. The default is to use them if the specified architecture has these instructions.

-misa-spec=ISA-spec-string ¶

Specify the version of the RISC-V Unprivileged (formerly User-Level) ISA specification generated code should conform to. The possibilities for ISA-spec-string are:

2.2

Produce code conforming to version 2.2.

20190608

Produce code conforming to version 20190608.

20191213

Produce code conforming to version 20191213.

The default is -misa-spec=20191213 unless GCC has been configured with --with-isa-spec= specifying a different default version.

-march=[ISA|Profile|Profile_ISA|processor-string] ¶

Generate code for given RISC-V ISA or profile or a combination of them (e.g. ‘rv64im’ ‘rvi20u64’ ‘rvi20u64_zbb’). The names of ISAs and profiles must be lower case. Examples include ‘rv64i’, ‘rv32g’, ‘rv32e’, ‘rv32imaf’, ‘rva22u64’ and ‘rva23u64’. To combine a named profile with optional RISC-V ISA extensions, give the profile first and then append the extension name(s) using an underscore as a delimiter (e.g. ‘rvi20u64_zca_zcb’ ‘rva23u64_zacas’). Additionally, a special value ‘help’ (-march=help) is accepted to list all supported extensions.

-march=unset causes the compiler to ignore any -march=… options that appear earlier on the command line, behaving as if the option was never passed. This is useful for ensuring that the architecture is taken from the -mcpu option, and an error results if no -mcpu option is given when -march=unset is used.

-march=native causes the compiler to work out the ISA of the machine it is running on. It is only available on a native GNU/Linux compiler; anywhere else the value is rejected like any other unknown ISA string. On a system whose cores are not all alike, such as one pairing fast and efficient cores, which of them the detected ISA describes is unspecified; write the ISA string out in full if it matters. If the running machine cannot be identified the option is ignored, and the architecture is taken from -mcpu or from the configured default as usual.

The syntax of the ISA string is defined as follows:

• The string must start with ‘rv32’ or ‘rv64’, followed by ‘i’, ‘e’, or ‘g’, referred to as the base ISA.
• The subsequent part of the string is a list of extension names. Extension names can be categorized as multi-letter (e.g. ‘zba’) and single-letter (e.g. ‘v’). Single-letter extensions can appear consecutively, but multi-letter extensions must be separated by underscores.
• An underscore can appear anywhere after the base ISA. It has no specific effect but is used to improve readability and can act as a separator.
• Extension names may include an optional version number, following the syntax ‘<major>p<minor>’ or ‘<major>’, (e.g. ‘m2p1’ or ‘m2’).

Supported extensions are listed below:

Extension Name	Supported Version	Description
‘g’	-	General-purpose computing base extension; ‘g’ expands to ‘i’, ‘m’, ‘a’, ‘f’, ‘d’, ‘zicsr’ and ‘zifencei’.
‘e’	2.0	Reduced base integer extension
‘i’	2.0 2.1	Base integer extension
‘m’	2.0	Integer multiplication and division extension
‘a’	2.0 2.1	Atomic extension
‘f’	2.0 2.2	Single-precision floating-point extension
‘d’	2.0 2.2	Double-precision floating-point extension
‘c’	2.0	Compressed extension
‘b’	1.0	Standard extension for bit manipulation functions
‘v’	1.0	Vector extension
‘h’	1.0	Hypervisor extension
‘zic64b’	1.0	Cache block size is 64 bytes
‘zicbom’	1.0	Cache-block management extension
‘zicbop’	1.0	Cache-block prefetch extension
‘zicboz’	1.0	Cache-block zero extension
‘ziccamoa’	1.0	Main memory supports all atomics in A
‘ziccamoc’	1.0	Main memory supports compare-and-swap atomics
‘ziccid’	1.0	Instruction/data coherence and consistency extension
‘ziccif’	1.0	Main memory supports instruction fetch with atomicity requirement
‘zicclsm’	1.0	Main memory supports misaligned loads/stores
‘ziccrse’	1.0	Main memory supports forward progress on LR/SC sequences
‘zicfilp’	1.0	Control-flow integrity landing pad extension
‘zicfiss’	1.0	Control-flow integrity shadow stack extension
‘zicntr’	2.0	Standard extension for base counters and timers
‘zicond’	1.0	Integer conditional operations extension
‘zicsr’	2.0	Control and status register access extension
‘zifencei’	2.0	Instruction-fetch fence extension
‘zihintntl’	1.0	Non-temporal locality hints extension
‘zihintpause’	2.0	Pause hint extension
‘zihpm’	2.0	Standard extension for hardware performance counters
‘zimop’	1.0	May-be-operations extension
‘zilsd’	1.0	Load/store pair instructions extension
‘zmmul’	1.0	Integer multiplication extension
‘za128rs’	1.0	Reservation set size of 128 bytes
‘za64rs’	1.0	Reservation set size of 64 bytes
‘zaamo’	1.0	Atomic memory operations extension
‘zabha’	1.0	Byte and halfword atomic memory operations extension
‘zacas’	1.0	Atomic compare-and-swap instructions extension
‘zalasr’	1.0	Atomic load-acquire and store-release extension
‘zalrsc’	1.0	Load-reserved/store-conditional subset of the A extension
‘zawrs’	1.0	Wait-on-reservation-set extension
‘zama16b’	1.0	Misaligned loads, stores, and AMOs that are fully contained within a naturally-aligned 16-byte boundary are atomic
‘zfa’	1.0	Additional floating-point extension
‘zfbfmin’	1.0	Minimal BF16 support extension
‘zfh’	1.0	Half-precision floating-point extension
‘zfhmin’	1.0	Minimal half-precision floating-point extension
‘zfinx’	1.0	Single-precision floating-point in integer registers extension
‘zdinx’	1.0	Double-precision floating-point in integer registers extension
‘zca’	1.0	Integer compressed instruction extension
‘zcb’	1.0	Simple compressed instruction extension
‘zcd’	1.0	Compressed double-precision floating point loads and stores extension
‘zce’	1.0	Compressed instruction extensions for embedded processors
‘zcf’	1.0	Compressed single-precision floating point loads and stores extension
‘zcmop’	1.0	Compressed may-be-operations extension
‘zcmp’	1.0	Compressed push pop extension
‘zcmt’	1.0	Table jump instruction extension
‘zclsd’	1.0	Compressed load/store pair instructions extension
‘zba’	1.0	Address calculation extension
‘zbb’	1.0	Basic bit manipulation extension
‘zbc’	1.0	Carry-less multiplication extension
‘zbkb’	1.0	Cryptography bit-manipulation extension
‘zbkc’	1.0	Cryptography carry-less multiply extension
‘zbkx’	1.0	Cryptography crossbar permutation extension
‘zbs’	1.0	Single-bit operation extension
‘zk’	1.0	Standard scalar cryptography extension
‘zkn’	1.0	NIST algorithm suite extension
‘zknd’	1.0	AES Decryption extension
‘zkne’	1.0	AES Encryption extension
‘zknh’	1.0	Hash function extension
‘zkr’	1.0	Entropy source extension
‘zks’	1.0	ShangMi algorithm suite extension
‘zksed’	1.0	SM4 block cipher extension
‘zksh’	1.0	SM3 hash function extension
‘zkt’	1.0	Data independent execution latency extension
‘ztso’	1.0	Total store ordering extension
‘zvbb’	1.0	Vector basic bit-manipulation extension
‘zvbc’	1.0	Vector carry-less multiplication extension
‘zve32f’	1.0	Vector extensions for embedded processors
‘zve32x’	1.0	Vector extensions for embedded processors
‘zve64d’	1.0	Vector extensions for embedded processors
‘zve64f’	1.0	Vector extensions for embedded processors
‘zve64x’	1.0	Vector extensions for embedded processors
‘zvfbfmin’	1.0	Vector BF16 converts extension
‘zvfofp8min’	0.2	Vector FP8 minimum extension
‘zvfbfwma’	1.0	Vector BF16 widening multiply/add extension
‘zvfh’	1.0	Vector half-precision floating-point extension
‘zvfhmin’	1.0	Vector minimal half-precision floating-point extension
‘zvkb’	1.0	Vector cryptography bit-manipulation extension
‘zvkg’	1.0	Vector GCM/GMAC extension
‘zvkn’	1.0	Vector NIST Algorithm Suite extension, ‘zvkn’ will expand to
‘zvknc’	1.0	Vector NIST Algorithm Suite with carry-less multiply extension, ‘zvknc’
‘zvkned’	1.0	Vector AES block cipher extension
‘zvkng’	1.0	Vector NIST Algorithm Suite with GCM extension, ‘zvkng’ will expand
‘zvknha’	1.0	Vector SHA-2 secure hash extension
‘zvknhb’	1.0	Vector SHA-2 secure hash extension
‘zvks’	1.0	Vector ShangMi algorithm suite extension, ‘zvks’ will expand
‘zvksc’	1.0	Vector ShangMi algorithm suite with carry-less multiplication extension,
‘zvksed’	1.0	Vector SM4 block cipher extension
‘zvksg’	1.0	Vector ShangMi algorithm suite with GCM extension
‘zvksh’	1.0	Vector SM3 secure hash extension
‘zvkt’	1.0	Vector data independent execution latency extension
‘zvl1024b’	1.0	Minimum vector length standard extensions
‘zvl128b’	1.0	Minimum vector length standard extensions
‘zvl16384b’	1.0	Minimum vector length standard extension
‘zvl2048b’	1.0	Minimum vector length standard extensions
‘zvl256b’	1.0	Minimum vector length standard extensions
‘zvl32768b’	1.0	Minimum vector length standard extension
‘zvl32b’	1.0	Minimum vector length standard extensions
‘zvl4096b’	1.0	Minimum vector length standard extensions
‘zvl512b’	1.0	Minimum vector length standard extensions
‘zvl64b’	1.0	Minimum vector length standard extensions
‘zvl65536b’	1.0	Minimum vector length standard extension
‘zvl8192b’	1.0	Minimum vector length standard extension
‘zhinx’	1.0	Half-precision floating-point in integer registers extension
‘zhinxmin’	1.0	Minimal half-precision floating-point in integer registers extension
‘zvabd’	0.7	Vector absolute difference extension.
‘sdtrig’	1.0	Debug triggers extension
‘sha’	1.0	The augmented hypervisor extension
‘shcounterenw’	1.0	Support writeable enables for any supported counter
‘shgatpa’	1.0	SvNNx4 mode supported for all modes supported by satp
‘shlcofideleg’	1.0	Delegating LCOFI interrupts to VS-mode
‘shtvala’	1.0	The htval register provides all needed values
‘shvstvala’	1.0	The vstval register provides all needed values
‘shvstvecd’	1.0	The vstvec register supports direct mode
‘shvsatpa’	1.0	The vsatp register supports all modes supported by satp
‘smaia’	1.0	Advanced interrupt architecture extension
‘smcdeleg’	1.0	Machine-mode counter delegation extension
‘smcntrpmf’	1.0	Cycle and instret privilege mode filtering
‘smctr’	1.0	Machine-mode control transfer records extension
‘smcsrind’	1.0	Machine-level indirect CSR access
‘smepmp’	1.0	PMP enhancements for memory access and execution prevention on machine mode
‘smmpm’	1.0	Supervisor-mode pointer masking extension
‘smnpm’	1.0	Supervisor-mode pointer masking extension
‘smrnmi’	1.0	Resumable non-maskable interrupts
‘smstateen’	1.0	State enable extension
‘smdbltrp’	1.0	Double trap extensions
‘ssaia’	1.0	Advanced interrupt architecture extension for supervisor mode
‘ssccfg’	1.0	Supervisor-mode counter configuration extension
‘ssccptr’	1.0	Main memory supports page table reads
‘sscofpmf’	1.0	Count overflow and filtering extension
‘sscounterenw’	1.0	Support writeable enables for any supported counter
‘ssctr’	1.0	Supervisor-mode control transfer records extension
‘sscsrind’	1.0	Supervisor-mode indirect CSR access
‘ssnpm’	1.0	Supervisor-mode pointer masking extension
‘sspm’	1.0	Supervisor-mode pointer masking extension
‘ssqosid’	1.0	Supervisor-mode quality-of-service ID extension
‘ssstateen’	1.0	Supervisor-mode state-enable extension
‘sstc’	1.0	Supervisor-mode timer interrupts extension
‘sstvala’	1.0	Stval provides all needed values
‘sstvecd’	1.0	Stvec supports direct mode
‘ssstrict’	1.0	Unimplemented reserved encodings raise illegal instruction exceptions and no non-conforming extensions are present
‘ssdbltrp’	1.0	Double trap extensions
‘ssu64xl’	1.0	UXLEN=64 must be supported
‘supm’	1.0	User-mode pointer masking extension
‘svinval’	1.0	Fine-grained address-translation cache invalidation extension
‘svnapot’	1.0	NAPOT translation contiguity extension
‘svpbmt’	1.0	Page-based memory types extension
‘svrsw60t59b’	1.0	Page-table reserved-for-software bits 60-59 extension
‘svvptc’	1.0	Extension for obviating memory-management instructions after marking PTEs valid
‘svadu’	1.0	Hardware updating of A/D bits extension
‘svade’	1.0	Cause exception when hardware updating of A/D bits is disabled
‘svbare’	1.0	Satp mode bare is supported
‘xcvalu’	1.0	Core-V miscellaneous ALU extension
‘xcvbi’	1.0	Core-V immediate branch extension
‘xcvelw’	1.0	Core-V event load word extension
‘xcvmac’	1.0	Core-V multiply-accumulate extension
‘xcvsimd’	1.0	Core-V SIMD extension
‘xsfcease’	1.0	SiFive CEASE instruction extension
‘xsfvcp’	1.0	SiFive VCIX vector coprocessor extension
‘xsfvfnrclipxfqf’	1.0	SiFive FP32-to-int8 ranged clip instructions
‘xsfvqmaccdod’	1.0	SiFive int8 matrix multiplication extension
‘xsfvqmaccqoq’	1.0	SiFive int8 matrix multiplication extension
‘xtheadba’	1.0	T-head address calculation extension
‘xtheadbb’	1.0	T-head basic bit-manipulation extension
‘xtheadbs’	1.0	T-head single-bit instructions extension
‘xtheadcmo’	1.0	T-head cache management operations extension
‘xtheadcondmov’	1.0	T-head conditional move extension
‘xtheadfmemidx’	1.0	T-head indexed memory operations for floating-point registers extension
‘xtheadfmv’	1.0	T-head double floating-point high-bit data transmission extension
‘xtheadint’	1.0	T-head acceleration interruption extension
‘xtheadmac’	1.0	T-head multiply-accumulate extension
‘xtheadmemidx’	1.0	T-head indexed memory operation extension
‘xtheadmempair’	1.0	T-head two-GPR memory operation extension
‘xtheadsync’	1.0	T-head multi-core synchronization extension
‘xtheadvector’	1.0	T-head vector extension
‘xventanacondops’	1.0	Ventana integer conditional operations extension
‘xmipscmov’	1.0	Mips conditional move extension
‘xmipscbop’	1.0	Mips prefetch extension
‘xandesperf’	5.0	Andes performance extension
‘xandesbfhcvt’	5.0	Andes bfloat16 conversion extension
‘xandesvbfhcvt’	5.0	Andes vector bfloat16 conversion extension
‘xandesvsintload’	5.0	Andes vector INT4 load extension
‘xandesvpackfph’	5.0	Andes vector packed FP16 extension
‘xandesvdot’	5.0	Andes vector dot product extension
‘xsmtvdot’	1.0	SpacemiT vector dot product extension
‘xsmtvdotii’	1.0	SpacemiT vector dot product II extension

When -march= is not specified, GCC uses the setting from -mcpu.

If both -march and -mcpu= are not specified, the default for this argument is system dependent; if you want a specific architecture extension, you should specify one explicitly.

When the RISC-V specifications define an extension as depending on other extensions, GCC implicitly adds the dependent extensions to the enabled extension set if they weren’t added explicitly.

‘Core Name’

-mcpu=processor-string ¶

Use architecture of and optimize the output for the given processor, specified by particular CPU name. Permissible values for this option are:

‘sifive-e20’,

‘sifive-e21’,

‘sifive-e24’,

‘sifive-e31’,

‘sifive-e34’,

‘sifive-e76’,

‘sifive-s21’,

‘sifive-s51’,

‘sifive-s54’,

‘sifive-s76’,

‘sifive-u54’,

‘sifive-u74’,

‘sifive-x280’,

‘sifive-p450’,

‘sifive-p550’,

‘sifive-p670’,

‘sifive-p870-d’,

‘thead-c906’,

‘xt-c908’,

‘xt-c908v’,

‘xt-c910’,

‘xt-c910v2’,

‘xt-c920’,

‘xt-c920v2’,

‘xt-c9501fdvt’,

‘tt-ascalon-x’,

‘tt-ascalon-xg’,

‘xiangshan-nanhu’,

‘xiangshan-kunminghu’,

‘mips-p8700’,

‘andes-n22’,

‘andes-n25’,

‘andes-a25’,

‘andes-nx25’,

‘andes-ax25’,

‘andes-a27’,

‘andes-ax27’,

‘andes-n225’,

‘andes-d23’,

‘andes-n45’,

‘andes-nx45’,

‘andes-a45’,

‘andes-ax45’,

‘spacemit-x60’,

‘spacemit-x100’,

‘spacemit-a100’.

Note that -mcpu does not override -march or -mtune.

‘Tune Name’

-mtune=processor-string ¶

Optimize the output for the given processor, specified by microarchitecture or particular CPU name. Permissible values for this option are:

‘generic’,

‘rocket’,

‘sifive-3-series’,

‘sifive-5-series’,

‘sifive-7-series’,

‘sifive-p400-series’,

‘sifive-p600-series’,

‘sifive-p550’,

‘sifive-p870-d’,

‘tt-ascalon-x’,

‘tt-ascalon-xg’,

‘thead-c906’,

‘xt-c908’,

‘xt-c908v’,

‘xt-c910’,

‘xt-c910v2’,

‘xt-c920’,

‘xt-c920v2’,

‘xt-c9501fdvt’,

‘xiangshan-nanhu’,

‘xiangshan-kunminghu’,

‘spacemit-x60’,

‘spacemit-x100’,

‘spacemit-a100’,

‘arc-v-rhx-100-series’,

‘generic-ooo’,

‘size’,

‘mips-p8700’,

‘andes-25-series’,

‘andes-23-series’,

‘andes-45-series’,

‘arc-v-rmx-100-series’,

and all valid options for -mcpu=.

When -mtune= is not specified, GCC uses the setting from -mcpu. The default is ‘generic’ if neither is specified.

-mtune=native and -mcpu=native identify the core the compiler is running on, subject to the same restrictions and the same caveat about unlike cores as -march=native above. -mcpu=native additionally selects the detected architecture. A core the compiler has no hardware identification for leaves the tuning at its default rather than causing an error.

The ‘size’ choice is not intended for use by end-users. This is used when -Os is specified. It overrides the instruction cost info provided by -mtune=, but does not override the pipeline info. This helps reduce code size while still giving good performance.

-mpreferred-stack-boundary=num ¶

Attempt to keep the stack boundary aligned to a 2 raised to num byte boundary. If -mpreferred-stack-boundary is not specified, the default is 4 (16 bytes or 128-bits).

Warning: If you use this switch, then you must build all modules with the same value, including any libraries. This includes the system libraries and startup modules.

-msmall-data-limit=n ¶

Put global and static data smaller than n bytes into a special section (on some targets).

-msave-restore ¶

-mno-save-restore

Do or don’t use smaller but slower prologue and epilogue code that uses library function calls. The default is to use fast inline prologues and epilogues.

-mmovcc ¶

-mno-movcc

Do or don’t produce branchless conditional-move code sequences even with targets that do not have specific instructions for conditional operations. If enabled, sequences of ALU operations are produced using base integer ISA instructions where profitable.

-minline-atomics ¶

-mno-inline-atomics

Do or don’t use smaller but slower subword atomic emulation code that uses libatomic function calls. The default is to use fast inline subword atomics that do not require libatomic.

-minline-strlen ¶

-mno-inline-strlen

Do or do not attempt to inline strlen calls if possible. Inlining can only be done if the string is properly aligned and instructions for accelerated processing are available. The default is to inline strlen calls.

-minline-strcmp ¶

-mno-inline-strcmp

Do or do not attempt to inline strcmp calls if possible. Inlining can only be done if the strings are properly aligned and instructions for accelerated processing are available. The default is to inline strcmp calls.

-minline-strncmp ¶

-mno-inline-strncmp

Do or do not attempt to inline strncmp calls if possible. Inlining can only be done if the strings are properly aligned and instructions for accelerated processing are available. The default is to inline strncmp calls.

-mstringop-strategy=strategy ¶

Specify a particular strategy for inlining string and memory operations. strategy may be one of ‘auto’, ‘libcall’, ‘scalar’, or ‘vector’.

-mshorten-memrefs ¶

-mno-shorten-memrefs

Do or do not attempt to make more use of compressed load/store instructions by replacing a load/store of ’base register + large offset’ with a new load/store of ’new base + small offset’. If the new base gets stored in a compressed register, then the new load/store can be compressed. Currently targets 32-bit integer load/stores only.

-mstrict-align ¶

-mno-strict-align

Do not or do generate unaligned memory accesses. The default is set depending on whether the processor we are optimizing for supports fast unaligned access or not.

-mscalar-strict-align ¶

-mno-scalar-strict-align

Do not or do generate unaligned memory accesses. The default is set depending on whether the processor we are optimizing for supports fast unaligned access or not. This is an alias for -mstrict-align.

-mvector-strict-align ¶

-mno-vector-strict-align

Do not or do generate unaligned vector memory accesses. The default is set to off unless the processor we are optimizing for explicitly supports element-misaligned vector memory access.

-mzilsd-word-align ¶

Allow Zilsd/Zclsd memory accesses to be generated when they are known to be 4-byte aligned.

-mzilsd-strict-align ¶

Require Zilsd/Zclsd memory accesses to be naturally 8-byte aligned.

The precedence among -mzilsd-word-align, -mzilsd-strict-align, -mstrict-align, -mno-strict-align, -mscalar-strict-align, and -mno-scalar-strict-align is determined by the last one specified.

-mmax-vectorization ¶

-mno-max-vectorization

Enable or disable an override to vectorizer cost model making vectorization always appear profitable. Unlike -fno-vect-cost-model or -fvect-cost-model=unlimited this option does not turn off cost comparison between different vector modes.

-mcmodel=medlow ¶

Generate code for the medium-low code model. The program and its statically defined symbols must lie within a single 2 GiB address range and must lie between absolute addresses −2 GiB and +2 GiB. Programs can be statically or dynamically linked. This is the default code model unless GCC has been configured with --with-cmodel= specifying a different default code model.

-mcmodel=medany ¶

Generate code for the medium-any code model. The program and its statically defined symbols must be within any single 2 GiB address range. Programs can be statically or dynamically linked.

The code generated by the medium-any code model is position-independent, but is not guaranteed to function correctly when linked into position-independent executables or libraries.

-mcmodel=large ¶

Generate code for a large code model, which has no restrictions on size or placement of symbols.

-mexplicit-relocs ¶

-mno-explicit-relocs

Use or do not use assembler relocation operators when dealing with symbolic addresses. The alternative is to use assembler macros instead, which may limit optimization.

-mrelax ¶

-mno-relax

Take advantage of linker relaxations to reduce the number of instructions required to materialize symbol addresses. The default is to take advantage of linker relaxations.

-mriscv-attribute ¶

-mno-riscv-attribute

Emit (do not emit) RISC-V attribute to record extra information into ELF objects. This feature requires at least binutils 2.32.

-mcsr-check ¶

-mno-csr-check

Enables or disables the CSR checking.

-momit-leaf-frame-pointer ¶

Don’t keep the frame pointer in a register for leaf functions. This avoids the instructions to save, set up and restore frame pointers and makes an extra register available in leaf functions.

-malign-data=type ¶

Control how GCC aligns variables and constants of array, structure, or union types. Supported values for type are ‘xlen’ which uses x register width as the alignment value, and ‘natural’ which uses natural alignment. ‘xlen’ is the default.

-mbig-endian ¶

Generate big-endian code. This is the default when GCC is configured for a ‘riscv64be-*-*’ or ‘riscv32be-*-*’ target. Support for RISC-V big-endian is experimental. The ABI is not yet stable and could change in incompatible ways in future releases.

-mlittle-endian ¶

Generate little-endian code. This is the default when GCC is configured for a ‘riscv64-*-*’ or ‘riscv32-*-*’ but not a ‘riscv64be-*-*’ or ‘riscv32be-*-*’ target.

-mstack-protector-guard=guard ¶

-mstack-protector-guard-reg=reg

-mstack-protector-guard-offset=offset

Generate stack protection code using canary at guard. Supported locations are ‘global’ for a global canary or ‘tls’ for per-thread canary in the TLS block.

With the latter choice the options -mstack-protector-guard-reg=reg and -mstack-protector-guard-offset=offset furthermore specify which register to use as base register for reading the canary, and from what offset from that base register. There is no default register or offset as this is entirely for use within the Linux kernel.

-mtls-dialect=desc ¶

Use TLS descriptors as the thread-local storage mechanism for dynamic accesses of TLS variables.

-mtls-dialect=trad ¶

Use traditional TLS as the thread-local storage mechanism for dynamic accesses of TLS variables. This is the default.

-munroll-only-small-loops ¶

-mno-unroll-only-small-loops

Controls conservative small loop unrolling. It is enabled by default at -O2 and above. When enabled, only loops whose RTL instruction count does not exceed a tune-specific threshold are unrolled, and the unroll factor is capped by a tune-specific parameter. This gives better utilization of the instruction decoding pipeline without enlarging bigger loops. You can disable it with -mno-unroll-only-small-loops; it is also turned off when -funroll-loops or -funroll-all-loops is explicitly enabled, so that an explicit unrolling request applies to loops of any size.

-mrvv-vector-bits=value ¶

Specify how the number of bits for an RVV vector register, as taken from the -march= option, is interpreted. The value parameter is specified as a string keyword and may be one of ‘scalable’ or ‘zvl’. The default is ‘scalable’, which tells GCC to interpret the number as a minimum, while ‘zvl’ tells GCC to use exactly the number of bits specified.

-mrvv-max-lmul=value ¶

This option allows explicit control over the maximum length multiplier (LMUL) used when generating code for the RISC-V Vector Extensions (RVV). The value parameter is specified as a string keyword and may be one of ‘m1’, ‘m2’, ‘m4’, ‘m8’, or ‘dynamic’. The default is ‘m1’ for compatibility with existing hardware that does not support the other options.

-madjust-lmul-cost ¶

-mno-adjust-lmul-cost

This option adjusts the cost model used to schedule vector instructions to multiply the latency of instructions by the RVV length multiplier, LMUL. It is disabled by default.

-mautovec-segment ¶

-mno-autovec-segment

Enable or disable generation of vector segment load/store instructions. This option is enabled by default.