Supported architectures
During the analysis, TrustInSoft CI faithfully emulates a target hardware architecture, respecting properties such as endianness, integer types and alignment constraints.
For each analysis, you can specify the target architecture of your choice by setting the
machdep field in the Analyses configuration file to the corresponding architecture name. If machdep is not provided, TrustInSoft CI emulates, by default, a 32-bit x86 processor.Supported architectures
The table below provides a high-level description and the endianness of each supported architecture. To get further information on the data types, visit the Fundamental data types section.
With the exception of
apple_ppc_32, x86_win32 andx86_win64, all target architectures may be specified with the gcc_ prefix, in which case gcc language extensions are allowed, and the__int128integer type is available on a 64-bitmachdep.Architecture name in machdep
Endianness
Description
x86_16Little
An x86 processor in 16-bit mode, in “tiny” or “small” memory model (using 16-bit “near” pointers). The long double type is the 80387 80-bit native format.
x86_16_hugeLittle
An x86 processor in 16-bit mode, in “large” or “huge” memory model (using 32-bit “far” pointers). The long double type is the 80387 80-bit native format.
x86_32Little
An x86 processor in 32-bit mode. The long double type is the 80387 80-bit native format.
x86_win32Little
An x86 processor in 32-bit mode, using the Win32 ABI.
x86_64Little
An x86-64 processor in 64-bit mode. The long double type is the IEEE754 quad-precision type.
x86_win64Little
An x86 processor in 64-bit mode, using the Win64 ABI. The long type is 32 bits.
ppc_32Big
A 32-bit PowerPC processor. The char type is unsigned; the long double type is the IEEE754 quad-precision type.
ppc_64Big
A 64-bit PowerPC processor. The char type is unsigned; the long double type is the IEEE754 quad-precision type.
apple_ppc_32Big
Same as
ppc_32, but forcing the “char” type to be signed, as done by the Apple toolchain (as opposed to the default PowerPC use of unsigned char type), and allowing gcc language extensions.arm_eabiLittle
A 32-bit ARM processor, using the extended ABI (EABI). The char type is unsigned.
armeb_eabiBig
A 32-bit ARM processor, using the extended ABI (EABI). The char type is unsigned.
aarch64Little
A 64-bit ARMv8 processor (AArch64). The char type is unsigned; the long double type is the IEEE754 quad-precision type.
aarch64ebBig
A 64-bit ARMv8 processor (AArch64). The char type is unsigned; the long double type is the IEEE754 quad-precision type.
sparc_32Big
A 32-bit SparcV8 processor.
sparc_64Big
AA 64-bit SparcV9 processor. The long double type is the IEEE754 quad-precision type.
mips_o32Big
A 32-bit MIPS processor, using the old ABI (o32).
mipsel_o32Little
A 32-bit MIPS processor, using the old ABI (o32).
mips_n32Big
A 64-bit MIPS processor, using the new 32-bit ABI (n32). The long double type is the IEEE754 quad-precision type.
mipsel_n32Little
A 64-bit MIPS processor, using the new 32-bit ABI (n32). The long double type is the IEEE754 quad-precision type.
Fundamental data types
The table below provides the default characteristics of the fundamental data types.
Please refer to both the above table and the below CSV file to get their exact characteristics for each architecture.
TrustInSoft-CI-Supported-Architectures-2020-11-25.csv
5KB
Text
Data types per architecture
Data type
Description
char
is 8 bits, and defaults to signed
short
is 16 bits
int
is 16 bits on 16-bit
machdeps (x86_16, x86_16_huge), 32 bits otherwiselong
is 64 bits on 64-bit
machdeps, 32 bits otherwiselong long
is 64 bits
float
is 32 bits (IEEE754 single-precision type)
double
is 64 bits (IEEE754 double-precision type)
long double
is identical to double
Last modified 10mo ago
Copy link