In this tutorial we will create a GCC cross-compiler for your own operating system. This compiler is specially made to target exactly your operating system and is what allows you to leave the current operating system behind. You need a cross-compiler for operating systems development, unless you are developing on your own operating system.
Generally speaking, a cross-compiler is a compiler that runs on platform A (the host), but generates executables for platform B (the target). These two platforms may (but do not need to) differ in CPU, operating system, and/or executable format. In our case, the host platform is your current operating system, and the target platform is the operating system you are about to make. It is important to realize that these two platforms are not the same; your operating system is always going to be different from your current operating system. This is why we need to build a cross-compiler first, you will most certainly run into trouble otherwise.
Why do I need a Cross Compiler?
- Main article: Why do I need a Cross Compiler?
You need to use a cross-compiler unless you are developing on your own operating system. The compiler must know the correct target platform (CPU, operating system), otherwise you will run into trouble. If you use the compiler that comes with your system, then the compiler won't know it is compiling something else entirely. Some tutorials suggest using your system compiler and passing a lot of problematic options to the compiler. This will certainly give you a lot of problems in the future and the solution is build a cross-compiler. If you have already attempted to make an operating system without using a cross-compiler, please read the article Why do I need a Cross Compiler?.
Which compiler version do I want?
- Main article: Building GCC
The newest GCC is recommended as it is the latest and greatest release. For instance, you may run into trouble if you use GCC 4.6.3 to build a GCC 4.8.0 cross-compiler. If you are not using the latest major GCC release for your system compiler, we recommend that you build the newest GCC as your system compiler.
You can also use older releases as they are usually reasonably good. If your local system compiler isn't too terribly old (at least GCC 4.6.0), you may wish to save yourself the trouble and just pick the latest minor release (such as 4.6.3 if your system compiler is 4.6.1) for your cross-compiler.
You can view your current compiler version by invoking:
You may be able to use an older major GCC release to build a cross-compiler of a newer major GCC release. For instance, GCC 4.7.3 may be able to build a GCC 4.8.0 cross-compiler. However, if you want to use the latest and greatest GCC version for your cross-compiler, we recommend that you bootstrap the newest GCC as your system compiler first. Individuals using OS X 10.7 or earlier might want to invest in either building a system GCC (that outputs native Mach-O), or upgrading the local LLVM/Clang installation. Users with 10.8 and above should install the Command Line Tools from Apple's developer website and use Clang to cross-compile GCC.
Which Binutils version do I want?
- Main article: Cross-Compiler Successful Builds
We recommend that you use the latest and greatest Binutils release. Note, however, that not all combinations of GCC and Binutils work. If you run into trouble, use a Binutils that was released at roughly the same time as your desired compiler version. You probably need at least Binutils 2.22, or preferably the latest 2.23.2 release. It doesn't matter what Binutils version you have installed on your current operating system.
Deciding on the target platform
- Main article: Target Triplet
You should already know this. If you are following the Bare Bones tutorial, you wish to build a cross-compiler for i686-elf.
Preparing for the build
The GNU Compiler Collection is an advanced piece of software with dependencies. You need to install certain dependencies in order to build gcc. You need to install GNU make, GNU bison, flex, and of course an existing system compiler you wish to replace. In addition, you also need the packages GNU GMP, GNU MPFR, and MPC that are used by GCC for floating point support.
You need a host system with a working GCC installation, and enough memory as well as hard drive space. How much qualifies as "enough" is depending on the versions of the software involved, but GCC is a big piece of software, so don't be surprised when 128 or 256 MByte are not sufficient.
In short you need the following that you can install manually or through package management:
- A Unix-like environment (Windows users)
- GCC (existing release you wish to replace)
- G++ (if building a version of GCC >= 4.8.0)
- GNU Make
- GNU Bison
- GNU GMP
- GNU MPFR
- GNU MPC
- ISL (optional)
- CLooG (optional)
Downloading the Source Code
In addition, to build GCC you need to have installed GNU GMP, GNU MPFR, GNU Mpc and the ISL library. You may already have these libraries and the development files installed, but this tutorial builds them as part of GCC. If you don't need this, simply don't build them as part of GCC. Note that not all GMP, MPFR and Mpc combinations are compatible with a given GCC release. You also need Texinfo to build Binutils.
You can download GNU GMP from its website. (libgmp3-dev on apt-based systems, dev-libs/gmp on Gentoo, gmp-devel on Fedora, libgmp-devel on Cygwin)
You can download GNU MPFR from its website. (libmpfr-dev on apt-based systems, dev-libs/mpfr on Gentoo, mpfr-devel on Fedora, libmpfr-devel on Cygwin)
You can download ISL from its website (optional). (libisl-dev on apt-based systems, libisl-devel on Cygwin)
You can download ClooG from its website (optional). (libcloog-isl-dev on apt-based systems, libcloog-isl-devel on Cygwin)
You can download GNU Mpc from its website. (libmpc-dev on apt-based systems, dev-libs/mpc on Gentoo, libmpc-devel on Fedora, libmpc-devel on Cygwin)
You can download GNU Texinfo from its website. (texinfo on apt-based systems, texinfo on Arch Linux, sys-apps/texinfo on Gentoo, texinfo on Cygwin)
Download the needed source code into a suitable directory such as $HOME/src.An alternative way to download GMP, MPFR and Mpc is to go to the top level directory of the GCC source code (i.e. gcc-x.y.z/) and type the following:
Note: The versioning scheme used is that each fullstop separates a full number, i.e. Binutils 2.20.0 is newer than 2.9.0. This may be confusing, if you have not encountered this (quite common) versioning scheme yet, when looking at an alphanumerically sorted list of tarballs: The file at the bottom of the list is not the latest version! An easy way of getting the latest version is to sort by the last modified date and scrolling to the bottom.
Note: Version 5.x (or later) of Texinfo is known to be incompatible with the current Binutils 2.23.2 release (and older). You can check your current version using makeinfo --version. If your version is too new and you encounter problems during the build, you will need to either use Binutils 2.24 release (or newer) or install an older version of Texinfo - perhaps through building from source - and add it to your PATH prior and during the Binutils build.
Note: Version 0.13 (or later) of ISL is incompatible with the current CLooG 0.18.1 release (and older). Use version 0.12.2 of ISL or the build will fail.
Your distribution may ship its own patched GCC and Binutils that is customized to work on your particular Linux distribution. You may not be able to build a functional system compiler using the upstream sources you downloaded above. In that case, try a newer GCC release or get the patched source code. For instance, some GCC releases are known to not understand the new Debian multiarch directory structure. However, if the compiler we are about to build is a cross-compiler targetting another operating system (such as your new one), then this is much less a worry.
Note for all Gentoo users: Gentoo, being a source-based distribution, makes it almost ridiculously easy to set up a cross-development toolchain:
emerge -av crossdev crossdev --help PORTDIR_OVERLAY="/usr/local/crossdev" crossdev --stage1 --binutils <binutils-version> --gcc <gcc-version> --target <target>
This will install a GCC cross-compiler into a "slot", i.e. alongside already-existing compiler versions. You can install several cross-compilers that way, simply by changing target designations. An unfortunate downside is that it will also pull in gentoo patches and pass additional configure options that differ from the official GCC Cross-Compiler setup, and they might behave differently.
After the compilation ran, you can now use your cross-compiler by calling <target>-gcc. You can also use gcc-config to toggle between compiler versions should you need to do so. Don't replace your system compiler with a cross-compiler however. The package manager will also suggest updates as soon as they become available.
You can uninstall the cross-compiler by calling crossdev --clean <target>. Read the cross-development document for additional information.
Note that the version numbers to binutils and gcc are Gentoo package versions, i.e. there might be a suffix to the "official" (GNU) version that addresses additional patchsets supplied by the Gentoo maintainers. (For example, --binutils 2.24-r3 --gcc 4.8.3 is the latest stable package pair at the time of this writing.) You can omit the version numbers to use the latest package available.
Portage uses overlays to store packages that are not part of the original package management. Crossdev needs one overlay where it can store its binutils and gcc packages before building them. You can configure one properly, or you can use PORTDIR_OVERLAY to point at where it should keep its package manager files. Using PORTDIR_OVERLAY is not a good idea with existing overlays, but by then you should know how you have personally set them up earlier anyway and how to do it properly.
OS X Users
Additionally, OS X users need a replacement libiconv because the system libiconv is seriously out of date. OS X users can download the latest libiconv release by visiting the libiconv website or directly accessing the GNU main FTP mirror.
When compiling GCC 4.3 or higher on OS X 10.4 and 10.5, you may get unresolved symbol errors related to libiconv. This is because the version shipped with OS X is seriously out of date. Install a new version (compile it yourself or use MacPorts) and add --with-libiconv-prefix=/opt/local (or /usr/local if you compiled it yourself) to GCC's ./configure line. Alternatively you may place the libiconv source as gcc-x.y.z/libiconv and it will be compiled as part of the GCC compilation process. (This trick also works for MPFR, GMP, and Mpc).
The makefiles of Binutils and GCC use the $(CC) variable to invoke the compiler. On OS X, this resolves to GCC by default, which is actually not the "real thing", but Clang. Note that since at least OS X 10.8, Xcode's Command Line Tools package comes with Clang, and this version of Clang does indeed work to compile a working version of GCC, unlike what these instructions previously reflected.
Note that users running OS X 10.7 may need to find and install GCC, either from Homebrew, or from somewhere on Apple's website. Thus, the instructions below are really only relevant for these users, but your mileage may vary.
# This is only necessary for OS X users running 10.7 or below. export CC=/usr/bin/gcc-4.2 export CXX=/usr/bin/g++-4.2 export CPP=/usr/bin/cpp-4.2 export LD=/usr/bin/gcc-4.2
You might want to unset these exports once you compiled and installed the cross compiler, as it might confuse other builds. Do not make these permanent!
Note for Lion users: If you're on Lion (or above) chances are that you don't have the "real" GCC since Apple removed it from the Xcode package, but you can still install it. You can do it via Homebrew or by compiling from source, both are perfectly described on a StackExchange answer.
Note for Maverick users: You can build binutils-2.24 and gcc-4.8.3 (possible other version) with Xcode 5.1.1. Note that building GCC with LLVM is not officially supported and may cause interesting bugs, if you are willing to take this risk and save time building host-gcc just to compile a cross-gcc, follow this. Install GMP, MPFR, Mpc with MacPorts.
sudo port install gmp mpfr libmpc
../binutils-2.24/configure --prefix=$PREFIX \ --target=$TARGET \ --enable-interwork --enable-multilib \ --disable-nls --disable-werror
../gcc-4.8.3/configure --prefix=$PREFIX \ --target=$TARGET \ --disable-nls \ --enable-languages=c,c++ --without-headers \ --enable-interwork --enable-multilib \ --with-gmp=/usr --with-mpc=/opt/local --with-mpfr=/opt/local
- Note that there is issue with port's GMP, we use the version from OS X from /usr instead.
Windows users need to set up a Unix-like enviroment such as MinGW or Cygwin. It may well be worth looking into systems such as Linux and see if they fit your needs, as you commonly use a lot of Unix-like tools in operating systems development and this is much easier from a Unix-like operating system. If you have just installed the basic Cygwin package, you have to run the setup.exe again and install the following packages: GCC, G++, Make, Flex, Bison, Diffutils, libintl-devel, libgmp-devel, libmpfr-devel, libmpc-devel, Texinfo
MinGW + MSYS is an option, and as it addresses the native Windows API instead of a POSIX emulation layer, results in a slightly faster toolchain. Some software packages will not build properly under MSYS as they were not designed for use with Windows. As far as this tutorial is concerned, everything that applies to Cygwin also applies to MSYS unless otherwise specified. Make sure you install the C and C++ compilers, and the MSYS Basic System.
The "Windows Subsystem for Linux (Beta)", released with the Windows 10 Anniversary update is also an option for using a cross compiler. (Tested 08/08/2016 with GCC 6.1.0 and Binutils 2.27) This cross-compiler works reasonably fast, although being in beta state, it may not be ideal permanent development platform.
Cygwin note: Cygwin includes your Windows %PATH% in its bash $PATH. If you were using DJGPP before, this could result in confusion as e.g. calling GCC on the Cygwin bash command line would still call the DJGPP compiler. After uninstalling DJGPP, you should delete the DJGPP environment variable and clear the C:\djgpp entry (or wherever you installed it) from your %PATH%. Likewise, it might be a bad idea to mix build environments in your system PATH variable.
MinGW note: Some MinGW-specific information on building a cross-toolchain can be found on the hosted cross-compiler how-to page on the MinGW homepage.
Windows Subsystem for Linux (Beta) Note: You cannot have your cross compiler in the /mnt/c/ (or /mnt/"x") areas, as trying to compile your cross-compiler there will generate errors, whereas building to $HOME/opt/cross works perfectly. This is fixed with Windows Update KB3176929
We build a toolset running on your host that can turn source code into object files for your target system.
You need to decide where to install your new compiler. It is dangerous and a very bad idea to install it into system directories. You also need to decide whether the new compiler should be installed globally or just for you. If you want to install it just for you (recommended), installing into $HOME/opt/cross is normally a good idea. If you want to install it globally, installing it into /usr/local/cross is normally a good idea.
Please note that we build everything out of the source directory tree, as is considered good practice. Some packages only support building outside, some only inside and some both (but may not offer extensive checking with make). Building GCC inside the source directory tree fails miserably, at least for older versions.
export PREFIX="$HOME/opt/cross" export TARGET=i686-elf export PATH="$PREFIX/bin:$PATH"
We add the installation prefix to the PATH of the current shell session. This ensures that the compiler build is able to detect our new binutils once we have built them.
The prefix will configure the build process so that all the files of your cross-compiler environment end up in $HOME/opt/cross. You can change that prefix to whatever you like (e.g., /opt/cross or $HOME/cross would be options). If you have administrator access and wish to make the cross-compiler toolchain available to all users, you can install it into the /usr/local prefix - or perhaps a /usr/local/cross prefix if you are willing to change the system configuration such that this directory is in the search paths for all users. Technically, you could even install directly to /usr, so that your cross-compiler would reside alongside your system compiler, but that is not recommended for several reasons (like risking to overwrite your system compiler if you get TARGET wrong, or getting into conflict with your system's package management).
cd $HOME/src mkdir build-binutils cd build-binutils ../binutils-x.y.z/configure --target=$TARGET --prefix="$PREFIX" --with-sysroot --disable-nls --disable-werror make make install
This compiles the binutils (assembler, disassembler, and various other useful stuff), runnable on your system but handling code in the format specified by $TARGET.
--disable-nls tells binutils not to include native language support. This is basically optional, but reduces dependencies and compile time. It will also result in English-language diagnostics, which the people on the Forum understand when you ask your questions. ;-)
--with-sysroot tells binutils to enable sysroot support in the cross-compiler by pointing it to a default empty directory. By default the linker refuses to use sysroots for no good technical reason, while gcc is able to handle both cases at runtime. This will be useful later on.
- See also the offical instructions for configuring gcc.
Now, you can build GCC.
cd $HOME/src # The $PREFIX/bin dir _must_ be in the PATH. We did that above. which -- $TARGET-as || echo $TARGET-as is not in the PATH mkdir build-gcc cd build-gcc ../gcc-x.y.z/configure --target=$TARGET --prefix="$PREFIX" --disable-nls --enable-languages=c,c++ --without-headers make all-gcc make all-target-libgcc make install-gcc make install-target-libgcc
We build libgcc, a low-level support library that the compiler expects available at compile time. Linking against libgcc provides integer, floating point, decimal, stack unwinding (useful for exception handling) and other support functions. Note how we are not simply running make && make install as that would build way too much, not all components of gcc are ready to target your unfinished operating system.
--disable-nls is the same as for binutils above.
--without-headers tells GCC not to rely on any C library (standard or runtime) being present for the target.
--enable-languages tells GCC not to compile all the other language frontends it supports, but only C (and optionally C++).
It will take a while to build your cross-compiler.
If you are building a cross compiler for x86-64, you may want to consider building Libgcc without the "red zone": Libgcc_without_red_zone
Using the new Compiler
Now you have a "naked" cross-compiler. It does not have access to a C library or C runtime yet, so you cannot use any of the standard includes or create runnable binaries. But it is quite sufficient to compile the kernel you will be making shortly. Your toolset resides in $HOME/opt/cross (or what you set $PREFIX to). For example, you have a GCC executable installed as $HOME/opt/cross/bin/$TARGET-gcc, which creates programs for your TARGET.
You can now run your new compiler by invoking something like:
Note how this compiler is not able to compile normal C programs. The cross-compiler will spit errors whenever you want to #include any of the standard headers (except for a select few that actually are platform-independent, and generated by the compiler itself). This is quite correct - you don't have a standard library for the target system yet!
The C standard defines two different kinds of executing environments - "freestanding" and "hosted". While the definition might be rather fuzzy for the average application programmer, it is pretty clear-cut when you're doing OS development: A kernel is "freestanding", everything you do in user space is "hosted". A "freestanding" environment needs to provide only a subset of the C library: float.h, iso646.h, limits.h, stdalign.h, stdarg.h, stdbool.h, stddef.h, stdint.h and stdnoreturn.h (as of C11). All of these consist of typedef s and #define s "only", so you can implement them without a single .c file in sight.
To use your new compiler simply by invoking $TARGET-gcc, add $HOME/opt/cross/bin to your $PATH by typing:
This command will add your new compiler to your PATH for this shell session. If you wish to use it permanently, add the PATH command to your ~/.profile configuration shell script or similar. Consult your shell documentation for more information.
You can now move on to complete the Bare Bones tutorial variant that lead you here and complete it using your new cross-compiler. If you built a new GCC version as your system compiler and used it to build the cross-compiler, you can now safely uninstall it unless you wish to continue using it.
In general, verify that you read the instructions carefully and typed the commands precisely. Don't skip instructions. You will have to set your PATH variable again if you use a new shell instance, if you don't make it permanent by adding it to your shell profile. If a compilation seems to have gotten really messed up, type make distclean, and then start the make process over again. Ensure your un-archiever doesn't change newline characters.
ld: cannot find -lgcc
You specified that you want to link the GCC low-level runtime library into your executable through the -lgcc' switch, but forgot to build and properly install the library.
What's alphabetically on the top or bottom is not necessarily the latest version. After 2.9 comes 2.10, 2.11, 2.12 and then there are fifteen more releases that are all newer and progressively more likely to build or support your choice of GCC version.
- Cross-Compiler Successful Builds - combinations of GCC and Binutils which have been shown to work with this tutorial by OSDev.org members.
- Target Triplet - on target triplets and their use
- OS Specific Toolchain - going a step further and adding your own target.
- LLVM Cross-Compiler - some compilers make things much easier.
- Canadian Cross - making things yet more complicated.
- http://kegel.com/crosstool has a popular example of a script that automatically downloads, patches, and builds binutils, gcc, and glibc for known platforms.
- http://gcc.gnu.org/onlinedocs/gccint/Libgcc.html - Summary of the support functions you get when you link with libgcc.
- http://forums.gentoo.org/viewtopic.php?t=66125 - Compiling Windows applications under Linux
- http://www.libsdl.org/extras/win32/cross/README.txt - dito
- https://github.com/travisg/toolchains - Another script for building simple cross compilers
- https://www.youtube.com/watch?v=aESwsmnA7Ec - A walkthrough of how to build a cross-compiler using Cygwin on Windows.
- https://github.com/Cheapskate01/Cross-Compiler-Build-Script - A dead-simple script that Binutils and Gcc for you.
These were built by people in the OSdev community for their own building needs and shared at will, without guaranteeing any support or that it will even work on your setup. YMMV.
Latests versions for Linux (many arch)
For Linux x86_64 host
- i386-elf 4.9.1 target
- i686-elf 4.9.1 target
- x86_64-elf 4.9.1 target
- aarch64-elf 4.9.1 target
- arm-eabi 4.9.1 target
- m68k-elf 4.9.1 target
- powerpc-elf 4.9.1 target
- sparc-elf 4.9.1 target
- sh-elf 4.9.1 target
The packages from phillid.tk below have been shrunk to about 10 MiB for each pair of packages (GCC & Binutils). Please note that this has been achieved by enabling only the C front-end for GCC. If you're going to write your OS in any language but C or Assembly, these packages aren't for you. These are actually Pacman packages, but untarring them to / and rm-ing /.MTREE and other clutter dotfiles contained in the package will work the same.
For Windows host
For Windows Subsystem for Linux (Beta) host
- i686-elf 6.1.0 target (extracts to a directory called "cross", don't forget to install 'make' - I would recommend "apt-get install build-essential" to also add additional useful tools)
For OSX host
ARM prebuilt toolchains for multiple host platforms
ARM provides it's own prebuilt toolchain based upon GNU utilities for development targeting ARM systems.