Naming conventions for cross-compilation toolchains

Naming convention

The naming convention for the cross-compilation toolchain is: arch [-vendor] [-os] [-(gnu)eabi]

Depending on whether the operating system is supported or not, ARM GCC can be divided into those that support the operating system and those that do not.

arm-none-eabi: This does not have an operating system, so it cannot support functions that are closely related to the operating system, such as fork(2). It uses newlib, a C library specifically for embedded systems. arm-none-linux-eabi: For Linux, it uses Glibc

Examples 

1. arm-none-eabi-gcc

(ARM architecture, no vendor, not target an operating system, complies with the ARM EABI) is used to compile bare-metal systems for the ARM architecture (including the ARM Linux boot and kernel, but not for compiling Linux applications). It is generally suitable for chips with ARM7, Cortex-M and Cortex-R cores, so it does not support functions closely related to the operating system, such as fork(2). It uses newlib, a C library specifically for embedded systems.

2. arm-none-linux-gnueabi-gcc

(ARM architecture, no vendor, creates binaries that run on the Linux operating system, and uses the GNU EABI)

Mainly used for Linux systems based on ARM architecture, it can be used to compile ARM architecture u-boot, Linux kernel, Linux applications, etc. arm-none-linux-gnueabi is based on GCC, uses the Glibc library, and is optimized by Codesourcery. The floating-point operation of arm-none-linux-gnueabi-xxx cross-compilation tool is excellent. Generally, ARM9, ARM11, Cortex-A cores with Linux operating systems will use it.

3. arm-eabi-gcc

Android ARM compiler.

4. armcc

The compiler tool launched by ARM has similar functions to arm-none-eabi. It can compile bare metal programs (u-boot, kernel), but cannot compile Linux applications. armcc is usually used together with ARM development tools. The compilers in KeilMDK, ADS, RVDS and DS-5 are all armcc, so armcc compilers are all charged (except the patriotic version, hehe~~).

5. arm-none-uclinuxeabi-gcc and arm-none-symbianelf-gcc

arm-none-uclinuxeabi is for uCLinux, using Glibc.

arm-none-symbianelf is used for Symbian, I have never used it and don’t know what the C library is.

ABI and EABI

ABI: Application Binary Interface (ABI) for the ARM Architecture. In computers, the application binary interface describes the low-level interface between an application (or other type) and the operating system or other applications.

EABI: Embedded ABI. The Embedded Application Binary Interface specifies standard conventions for file formats, data types, register usage, stack organization optimization, and parameters in an embedded software. Developers using their own assembly language can also use EABI as an interface to the assembly language generated by a compatible compiler.

The main difference between the two is that ABI is on computers and EABI is on embedded platforms (such as ARM, MIPS, etc.).

arm-linux-gnueabi-gcc and arm-linux-gnueabihf-gcc

The two cross compilers are suitable for two different architectures, armel and armhf. Armel and armhf adopt different strategies for floating-point operations (only arm with fpu can support these two floating-point operation strategies).

In fact, the two cross compilers are just different in the default value of gcc option -mfloat-abi. gcc option -mfloat-abi has three values: soft, softfp, and hard (the latter two require the arm to have an fpu floating point unit, and soft is compatible with the latter two, but softfp and hard are incompatible with each other): soft: do not use the fpu for floating point calculations, even if there is an fpu floating point unit, but use software mode. 

softfp: The default value used by the armel architecture (the corresponding compiler is arm-linux-gnueabi-gcc). It uses fpu for calculations, but passes parameters using ordinary registers. In this way, when an interrupt occurs, only ordinary registers need to be saved, and the interrupt load is small, but the parameters need to be converted into floating point before calculation. 

hard: The default value used by the armhf architecture (the corresponding compiler arm-linux-gnueabihf-gcc). It uses the fpu for calculations and the floating-point registers in the fpu for parameter transfer, eliminating the need for conversion. It has the best performance but a high interrupt load.

Save the C file contents used in the following test as mfloat.c:

1. Compile with arm-linux-gnueabihf-gcc and use the "-v" option to get more detailed information:

# arm-linux-gnueabihf-gcc -v mfloat.c COLLECT_GCC_OPTIONS='-v' '-march=armv7-a' '-mfloat-abi=hard' '-mfpu=vfpv3-d16′ '-mthumb' -mfloat-abi=hard

It can be seen that hard hardware floating-point mode is used.

2. Compile using arm-linux-gnueabi-gcc:

# arm-linux-gnueabi-gcc -v mfloat.c COLLECT_GCC_OPTIONS='-v' '-march=armv7-a' '-mfloat-abi=softfp' '-mfpu=vfpv3-d16′ '-mthumb' -mfloat-abi=softfp

It can be seen that softfp mode is used.