Contents|Index|Previous|Next

SPARC Options

The following -m switches are supported on the SPARC.

-mno-app-regs
-mapp-regs   
Specify -mapp-regs to generate output using the global registers 2 through 4, which the SPARC SVR4 Application Binary Interface (ABI) reserves for applications. This is the default.
To be fully SVR4 ABI compliant at the cost of some performance loss, specify -mno-app-regs. You should compile libraries and system software with this option.
-mfpu
-mhard-float    
Generate output containing floating point instructions. This is the default.
-mno-fpu
-msoft-float    
Generate output containing library calls for floating point.
-msoft-float changes the calling convention in the output file; therefore, it is only useful if you compile all of a program with this option. In particular, you need to compile libgcc.a, the library that comes with GNU CC, with -msoft-float in order for this to work.
-mhard-quad-float   
Generate output containing quad-word (long double) floating point instructions.
-msoft-quad-float   
Generate output containing library calls for quad-word (long double) floating point instructions. The functions called are those specified in the SPARC ABI. This is the default.
As of this writing, there are no SPARC implementations that have hardware support for the quad-word floating point instructions. They all invoke a trap handler for one of these instructions, and then the trap handler emulates the effect of the instruction. Because of the trap handler overhead, this is much slower than calling the ABI library routines. Thus the -msoft-quad-float option is the default.
-mno-epilogue
-mepilogue   
With -mepilogue (the default), the compiler always emits code for function exit at the end of each function. Any function exit in the middle of the function (such as a return statement in C) will generate a jump to the exit code at the end of the function.
With -mno-epilogue, the compiler tries to emit exit code inline at every function exit.
-mno-flat
-mflat   
With -mflat, the compiler does not generate save/restore instructions and will use a flat or single register window calling convention. This model uses %i7 as the frame pointer and is compatible with the normal register window model. Code from either may be intermixed. The local registers and the input registers (0-5) are still treated as “call saved” registers and will be saved on the stack as necessary.
With -mno-flat (the default), the compiler emits save/restore instructions (except for leaf functions) and is the normal mode of operation.
-mno-unaligned-doubles
-munaligned-doubles    
Assume that doubles have 8 byte alignment. This is the default.
With -munaligned-doubles, GNU CC assumes that doubles have 8 byte alignment only if they are contained in another type, or if they have an absolute address.
Otherwise, it assumes they have 4 byte alignment. Specifying this option avoids some rare compatibility problems with code generated by other compilers. It is not the default because it results in a performance loss, especially for floating point code.
-mv8
-msparclite
These two options select variations on the SPARC architecture.
By default (unless specifically configured for the Fujitsu SPARClite), GCC generates code for the v7 variant of the SPARC architecture.
-mv8 will give you SPARC v8 code. The only difference from v7 code is that the compiler emits the integer multiply and integer divide instructions which exist in SPARC v8 but not in SPARC v7.
-msparclite will give you SPARClite code. This adds the integer multiply, integer divide step and scan (ffs) instructions which exist in SPARClite but not in SPARC v7.
These options are deprecated and will be deleted in GNU CC 2.9. They have been replaced with -mcpu=xxx.
-mcypress
-msupersparc    
These two options select the processor for which the code is optimized.
With -mcypress (the default), the compiler optimizes code for the Cypress CY7C602 chip, as used in the SparcStation/ SparcServer 3xx series. This is also appropriate for the older SparcStation 1, 2, IPX, etc.
With -msupersparc the compiler optimizes code for the SuperSparc CPU, as used in the SparcStation 10, 1000 and 2000 series. This flag also enables use of the full SPARC v8 instruction set. These options are deprecated and will be deleted in GNU CC 2.9. They have been replaced with -mcpu=xxx.
-mcpu=cpu_type
Set architecture type and instruction scheduling parameters for machine type cpu_type. Supported values for cpu_type are common, cypress, v8, supersparc, sparclite, f930, f934, sparclet, 90c701, v8plus, v9, and ultrasparc. Specifying ‘v9’ is only supported on true 64 bit targets.
-mtune=cpu_type
Set the instruction scheduling parameters for machine type cpu_type, but do not set the architecture type like -mcpu=cpu_type would. The same values for -mcpu=cpu_type are used for -tune=cpu_type.

The following -m switches are supported in addition to the previous switches on SPARCLET processor.

-mlittle-endian
Generate code for a processor running in little-endian mode.

-mlive-g0
Treat register
%g0 as a normal register. GCC will continue to clobber it as necessary but will not assume it always reads as 0.

-mbroken-saverestore
Generate code that does not use non-trivial forms of the
save and restore instructions. Early versions of the SPARCLET processor do not correctly handle save and restore instructions used with arguments. They correctly handle them used without arguments. A save instruction used without arguments increments the current window pointer but does not allocate a new stack frame. It is assumed that the window overflow trap handler will properly handle this case as will interrupt handlers.

The following -m switches are supported in addition to the previous switches on SPARC V9 processors in 64 bit environments.
-mlittle-endian
Generate code for a processor running in little-endian mode.
-mmedlow   
Generate code for the Medium/Low code model: assume a 32 bit address space. Programs are statically linked, PIC is not supported. Pointers are still 64 bits. It is very likely that a future version of GCC will rename this option.
-mmedany   
Generate code for the Medium/Anywhere code model: assume a 32 bit text and a 32 bit data segment, both starting anywhere (determined at link time). Programs are statically linked, PIC is not supported. Pointers are still 64 bits.
It is very likely that a future version of GCC will rename this option.
-mfullany   
Generate code for the Full/Anywhere code model: assume a full 64 bit address space. PIC is not supported.
It is very likely that a future version of GCC will rename this option.
-mint64   
Types long and int are 64 bits.
-mlong32   
Types long and int are 32 bits.
-mlong64
-mint32    
Type
long is 64 bits, and type int is 32 bits.
-mstack-bias
-mno-stack-bias   
With -mstack-bias, assume that the stack pointer, and frame pointer if present, are offset by -2047 which must be added back when making stack frame references. Otherwise, assume no such offset is present.