gcc: FRV Options
1
1 3.18.14 FRV Options
1 -------------------
1
1 '-mgpr-32'
1
1 Only use the first 32 general-purpose registers.
1
1 '-mgpr-64'
1
1 Use all 64 general-purpose registers.
1
1 '-mfpr-32'
1
1 Use only the first 32 floating-point registers.
1
1 '-mfpr-64'
1
1 Use all 64 floating-point registers.
1
1 '-mhard-float'
1
1 Use hardware instructions for floating-point operations.
1
1 '-msoft-float'
1
1 Use library routines for floating-point operations.
1
1 '-malloc-cc'
1
1 Dynamically allocate condition code registers.
1
1 '-mfixed-cc'
1
1 Do not try to dynamically allocate condition code registers, only
1 use 'icc0' and 'fcc0'.
1
1 '-mdword'
1
1 Change ABI to use double word insns.
1
1 '-mno-dword'
1
1 Do not use double word instructions.
1
1 '-mdouble'
1
1 Use floating-point double instructions.
1
1 '-mno-double'
1
1 Do not use floating-point double instructions.
1
1 '-mmedia'
1
1 Use media instructions.
1
1 '-mno-media'
1
1 Do not use media instructions.
1
1 '-mmuladd'
1
1 Use multiply and add/subtract instructions.
1
1 '-mno-muladd'
1
1 Do not use multiply and add/subtract instructions.
1
1 '-mfdpic'
1
1 Select the FDPIC ABI, which uses function descriptors to represent
1 pointers to functions. Without any PIC/PIE-related options, it
1 implies '-fPIE'. With '-fpic' or '-fpie', it assumes GOT entries
1 and small data are within a 12-bit range from the GOT base address;
1 with '-fPIC' or '-fPIE', GOT offsets are computed with 32 bits.
1 With a 'bfin-elf' target, this option implies '-msim'.
1
1 '-minline-plt'
1
1 Enable inlining of PLT entries in function calls to functions that
1 are not known to bind locally. It has no effect without '-mfdpic'.
1 It's enabled by default if optimizing for speed and compiling for
1 shared libraries (i.e., '-fPIC' or '-fpic'), or when an
1 optimization option such as '-O3' or above is present in the
1 command line.
1
1 '-mTLS'
1
1 Assume a large TLS segment when generating thread-local code.
1
1 '-mtls'
1
1 Do not assume a large TLS segment when generating thread-local
1 code.
1
1 '-mgprel-ro'
1
1 Enable the use of 'GPREL' relocations in the FDPIC ABI for data
1 that is known to be in read-only sections. It's enabled by
1 default, except for '-fpic' or '-fpie': even though it may help
1 make the global offset table smaller, it trades 1 instruction for
1 4. With '-fPIC' or '-fPIE', it trades 3 instructions for 4, one of
1 which may be shared by multiple symbols, and it avoids the need for
1 a GOT entry for the referenced symbol, so it's more likely to be a
1 win. If it is not, '-mno-gprel-ro' can be used to disable it.
1
1 '-multilib-library-pic'
1
1 Link with the (library, not FD) pic libraries. It's implied by
1 '-mlibrary-pic', as well as by '-fPIC' and '-fpic' without
1 '-mfdpic'. You should never have to use it explicitly.
1
1 '-mlinked-fp'
1
1 Follow the EABI requirement of always creating a frame pointer
1 whenever a stack frame is allocated. This option is enabled by
1 default and can be disabled with '-mno-linked-fp'.
1
1 '-mlong-calls'
1
1 Use indirect addressing to call functions outside the current
1 compilation unit. This allows the functions to be placed anywhere
1 within the 32-bit address space.
1
1 '-malign-labels'
1
1 Try to align labels to an 8-byte boundary by inserting NOPs into
1 the previous packet. This option only has an effect when VLIW
1 packing is enabled. It doesn't create new packets; it merely adds
1 NOPs to existing ones.
1
1 '-mlibrary-pic'
1
1 Generate position-independent EABI code.
1
1 '-macc-4'
1
1 Use only the first four media accumulator registers.
1
1 '-macc-8'
1
1 Use all eight media accumulator registers.
1
1 '-mpack'
1
1 Pack VLIW instructions.
1
1 '-mno-pack'
1
1 Do not pack VLIW instructions.
1
1 '-mno-eflags'
1
1 Do not mark ABI switches in e_flags.
1
1 '-mcond-move'
1
1 Enable the use of conditional-move instructions (default).
1
1 This switch is mainly for debugging the compiler and will likely be
1 removed in a future version.
1
1 '-mno-cond-move'
1
1 Disable the use of conditional-move instructions.
1
1 This switch is mainly for debugging the compiler and will likely be
1 removed in a future version.
1
1 '-mscc'
1
1 Enable the use of conditional set instructions (default).
1
1 This switch is mainly for debugging the compiler and will likely be
1 removed in a future version.
1
1 '-mno-scc'
1
1 Disable the use of conditional set instructions.
1
1 This switch is mainly for debugging the compiler and will likely be
1 removed in a future version.
1
1 '-mcond-exec'
1
1 Enable the use of conditional execution (default).
1
1 This switch is mainly for debugging the compiler and will likely be
1 removed in a future version.
1
1 '-mno-cond-exec'
1
1 Disable the use of conditional execution.
1
1 This switch is mainly for debugging the compiler and will likely be
1 removed in a future version.
1
1 '-mvliw-branch'
1
1 Run a pass to pack branches into VLIW instructions (default).
1
1 This switch is mainly for debugging the compiler and will likely be
1 removed in a future version.
1
1 '-mno-vliw-branch'
1
1 Do not run a pass to pack branches into VLIW instructions.
1
1 This switch is mainly for debugging the compiler and will likely be
1 removed in a future version.
1
1 '-mmulti-cond-exec'
1
1 Enable optimization of '&&' and '||' in conditional execution
1 (default).
1
1 This switch is mainly for debugging the compiler and will likely be
1 removed in a future version.
1
1 '-mno-multi-cond-exec'
1
1 Disable optimization of '&&' and '||' in conditional execution.
1
1 This switch is mainly for debugging the compiler and will likely be
1 removed in a future version.
1
1 '-mnested-cond-exec'
1
1 Enable nested conditional execution optimizations (default).
1
1 This switch is mainly for debugging the compiler and will likely be
1 removed in a future version.
1
1 '-mno-nested-cond-exec'
1
1 Disable nested conditional execution optimizations.
1
1 This switch is mainly for debugging the compiler and will likely be
1 removed in a future version.
1
1 '-moptimize-membar'
1
1 This switch removes redundant 'membar' instructions from the
1 compiler-generated code. It is enabled by default.
1
1 '-mno-optimize-membar'
1
1 This switch disables the automatic removal of redundant 'membar'
1 instructions from the generated code.
1
1 '-mtomcat-stats'
1
1 Cause gas to print out tomcat statistics.
1
1 '-mcpu=CPU'
1
1 Select the processor type for which to generate code. Possible
1 values are 'frv', 'fr550', 'tomcat', 'fr500', 'fr450', 'fr405',
1 'fr400', 'fr300' and 'simple'.
1