Main Content

Generic target options

Specify size of data types and endianness by creating your own target processor

Description

If a target processor is not directly supported by Polyspace®, you can create your own target. You specify the target mcpu representing a generic "Micro Controller/Processor Unit" and then explicitly specify sizes of fundamental data types, endianness and other characteristics.

Settings

In the user interface of the Polyspace desktop products, the Generic target options dialog box opens when you set the Target processor type to mcpu. The Target processor type option is available on the Target & Compiler node in the Configuration pane.

Example Generic target options dialog box.

Use the dialog box to specify the name of a new mcpu target, for example My_target. That new target is added to the Target processor type option list.

Default characteristics of a new target: listed as type [size]

  • char [8]

  • short [16]

  • int [16]

  • long [32]

  • long long [32]

  • float [32]

  • double [32]

  • long double [32]

  • pointer [16]

  • alignment [32]

  • char is signed

  • endianness is little-endian

Command-Line Options

When using the command line, use -target mcpu along with these target specification options.

OptionDescriptionAvailable WithExample
-little-endian

Little-endian architectures are Less Significant byte First (LSF). For example: i386.

Specifies that the less significant byte of a short integer (e.g. 0x00FF) is stored at the first byte (0xFF) and the most significant byte (0x00) at the second byte.

All targetspolyspace-bug-finder -target mcpu -little-endian
-big-endian

Big-endian architectures are Most Significant byte First (MSF). For example: SPARC, m68k.

Specifies that the most significant byte of a short integer (e.g. 0x00FF) is stored at the first byte (0x00) and the less significant byte (0xFF) at the second byte.

All targetspolyspace-bug-finder -target mcpu -big-endian
-default-sign-of-char [signed | unsigned]

Specify default sign of char.

signed: Specifies that char is signed, overriding target’s default.

unsigned: Specifies that char is unsigned, overriding target’s default.

All targetspolyspace-bug-finder -default-sign-of-char unsigned -target mcpu
-char-is-16bits

char defined as 16 bits and all objects have a minimum alignment of 16 bits

Incompatible with -short-is-8bits and -align 8

mcpupolyspace-bug-finder -target mcpu -char-is-16bits
-short-is-8bitsDefine short as 8 bits, regardless of signmcpupolyspace-bug-finder -target mcpu -short-is-8bits
-int-is-32bitsDefine int as 32 bits, regardless of sign. Alignment is also set to 32 bits.mcpu, hc08, hc12, mpc5xxpolyspace-bug-finder -target mcpu -int-is-32bits
-long-is-32bits

Define long as 32 bits, regardless of sign. Alignment is also set to 32 bits.

If your project sets int to 64 bits, you cannot use this option.

All targetspolyspace-bug-finder -target mcpu -long-is-32bits
-long-long-is-64bitsDefine long long as 64 bits, regardless of sign. Alignment is also set to 64 bits.mcpupolyspace-bug-finder -target mcpu -long-long-is-64bits
-double-is-64bitsDefine double and long double as 64 bits, regardless of sign.mcpu, sharc21x61, hc08, hc12, mpc5xxpolyspace-bug-finder -target mcpu -double-is-64bits
-pointer-is-24bitsDefine pointer as 24 bits, regardless of sign.c18polyspace-bug-finder -target c18 -pointer-is-24bits
-pointer-is-32bitsDefine pointer as 32 bits, regardless of sign.mcpupolyspace-bug-finder -target mcpu -pointer-is-32bits
-align [128|64|32|16|8]

Specifies the largest alignment of scalar objects, structures, classes, unions, or arrays to the 128, 64, 32, 16, or 8 bit boundaries.

For instance, if the alignment of basic types in an array or struct is always 8, the array or struct storage is strictly determined by the size of the individual data objects without member and end padding.

All targetspolyspace-bug-finder -target mcpu -align 16

See also:

You can also use the option -custom-target to specify sizes in bytes of fundamental data types, signedness of plain char, alignment of structures and underlying types of standard typedef-s such as size_t, wchar_t and ptrdiff_t.

Examples

GCC Toolchains

If you use any of these GCC toolchains for your software development, you can setup your Polyspace analysis so that your code compiles with Polyspace:

  • ARM® Ltd. GNU® Arm Embedded Toolchain

  • HighTec EDV-Systeme

  • Linaro® GNU cross-toolchain

  • Melexis®

  • MENTOR® Embedded Sourcery™ CodeBench

  • QNX® Software Development Platform

  • Rowley Associates' CrossWorks

  • STMicroelectronics® TrueSTUDIO® for STM32

  • Texas Instruments® Code Composer Studio™

  • Wind River® GNU Compiler

Use polyspace-configure on a build command that uses one of these toolchains and extract information about your compiler configuration. The command creates a Polyspace project by default. To generate an options file that you then pass to Polyspace at the command line, run polyspace-configure with the option -output-options-file.

Alternatively, to manually set the details of your compiler configuration at the command line:

Targets for GCC and Clang Based Compilers

If you select one of the gnu#.x or clang#.x compilers with the option Compiler (-compiler), you can specify one of the supported target processor types. See Target processor type (-target). If a target processor type is not directly listed as supported, you can create the target by using the generic target options.

These tables show examples of targets that you can create:

GCC Based Compiler Targets

TargetOptions
ARM 
-target mcpu
-int-is-32bits
-long-long-is-64bits
-double-is-64bits
-pointer-is-32bits
-enum-type-definition auto-signed-first
-wchar-t-type-is unsigned-int
ARM64 
-custom-target false,8,2,4,-1,8,8,4,8,16,8,16,1,little,unsigned_long,long,unsigned_int
-D__aarch64__
MSP430 
-target mcpu
-long-long-is-64bits
-double-is-64bits
-wchar-t-type-is signed-long
-align 16
RISC-V
-custom-target false,8,2,4,-1,4,8,4,8,16,4,16,1,big,unsigned_int,int,int
-D__riscv
-D__riscv_xlen=xlen_size
-Dfloat_abi_macro

Here:

  • xlen_size is 32 or 64, depending on the target.

  • Depending on the hardware architecture of the target platform, float_abi_macro is one of these values:

    • __riscv_float_abi_soft

    • __riscv_float_abi_single

    • __riscv_float_abi_double

    • __riscv_float_abi_quad

PowerPC 
-target mcpu 
-int-is-32bits 
-long-long-is-64bits 
-double-is-64bits 
-pointer-is-32bits 
-wchar-t-type-is signed-int
Tricore 
-target mcpu
-int-is-32bits 
-long-long-is-64bits 
-double-is-64bits 
-pointer-is-32bits 
-enum-type-definition auto-signed-first 
-wchar-t-type-is signed-int

Clang Based Compiler Targets

TargetOptions
Cadence Tensilica Xtensa 
-compiler clang10.x
  -custom-target false,8,2,4,-1,4,8,4,8,8,4,8,1,little,unsigned_int,int,unsigned_short
  -D__XTENSA__
  -D__XCC__

Emulate Microchip MPLAB XC16 and XC32 Compilers

If you build your source code using Microchip MPLAB XC16 or XC32 compilers, you can set up your Polyspace analysis so that your code compiles with Polyspace. Enter these options at the command line or specify them in the Configuration pane of the Polyspace desktop user interface.

CompilerTarget Processor FamiliesOptions
MPLAB XC16

PIC24

dsPIC

 
-compiler gnu4.6 
-D__XC__ 
-D__XC16__ 
-target=mcpu 
-wchar-t-type-is unsigned-int 
-align 16 
-long-long-is-64bits
MPLAB XC32PIC32 
-compiler gnu4.8 
-custom-target true,8,2,4,-1,4,8,4,4,8,4,8,1,big,unsigned_long,long,int 
-D__PIC32M 
-D__PIC32MX 
-D__PIC32MX__ 
-D__XC32 
-D__XC32__ 
-D__XC 
-D__XC__ 
-D__mips=32 
-D__mips__ 
-D_mips

The set of macros specified with the option Preprocessor definitions (-D) is a minimal set. Specify additional macros as needed to ensure your code compiles with Polyspace.

Tips

If you use Polyspace as You Code extensions in IDEs, enter this option in an analysis options file. See options file.

See Also

Topics