1 ===================================================================
2 How To Cross-Compile Clang/LLVM using Clang/LLVM
3 ===================================================================
8 This document contains information about building LLVM and
9 Clang on host machine, targeting another platform.
11 For more information on how to use Clang as a cross-compiler,
12 please check http://clang.llvm.org/docs/CrossCompilation.html.
14 TODO: Add MIPS and other platforms to this document.
16 Cross-Compiling from x86_64 to ARM
17 ==================================
19 In this use case, we'll be using CMake and Ninja, on a Debian-based Linux
20 system, cross-compiling from an x86_64 host (most Intel and AMD chips
21 nowadays) to a hard-float ARM target (most ARM targets nowadays).
23 The packages you'll need are:
26 * ninja-build (from backports in Ubuntu)
27 * gcc-4.7-arm-linux-gnueabihf
28 * gcc-4.7-multilib-arm-linux-gnueabihf
29 * binutils-arm-linux-gnueabihf
31 * libsfgcc1-armhf-cross
32 * libstdc++6-armhf-cross
33 * libstdc++6-4.7-dev-armhf-cross
38 For more information on how to configure CMake for LLVM/Clang,
41 The CMake options you need to add are:
42 * -DCMAKE_CROSSCOMPILING=True
43 * -DCMAKE_INSTALL_PREFIX=<install-dir>
44 * -DLLVM_TABLEGEN=<path-to-host-bin>/llvm-tblgen
45 * -DCLANG_TABLEGEN=<path-to-host-bin>/clang-tblgen
46 * -DLLVM_DEFAULT_TARGET_TRIPLE=arm-linux-gnueabihf
47 * -DLLVM_TARGET_ARCH=ARM
48 * -DLLVM_TARGETS_TO_BUILD=ARM
49 * -DCMAKE_CXX_FLAGS='-target armv7a-linux-gnueabihf -mcpu=cortex-a9
50 -I/usr/arm-linux-gnueabihf/include/c++/4.7.2/arm-linux-gnueabihf/
51 -I/usr/arm-linux-gnueabihf/include/ -mfloat-abi=hard
52 -ccc-gcc-name arm-linux-gnueabihf-gcc'
54 The TableGen options are required to compile it with the host compiler,
55 so you'll need to compile LLVM (or at least `llvm-tblgen`) to your host
56 platform before you start. The CXX flags define the target, cpu (which
57 defaults to fpu=VFP3 with NEON), and forcing the hard-float ABI. If you're
58 using Clang as a cross-compiler, you will *also* have to set ``-ccc-gcc-name``,
59 to make sure it picks the correct linker.
61 Most of the time, what you want is to have a native compiler to the
62 platform itself, but not others. It might not even be feasible to
63 produce x86 binaries from ARM targets, so there's no point in compiling
64 all back-ends. For that reason, you should also set the "TARGETS_TO_BUILD"
65 to only build the ARM back-end.
67 You must set the CMAKE_INSTALL_PREFIX, otherwise a ``ninja install``
68 will copy ARM binaries to your root filesystem, which is not what you
74 There are some bugs in current LLVM, which require some fiddling before
77 #. If you're using Clang as the cross-compiler, there is a problem in
78 the LLVM ARM back-end that is producing absolute relocations on
79 position-independent code (R_ARM_THM_MOVW_ABS_NC), so for now, you
84 -DLLVM_ENABLE_PIC=False
86 This is not a problem, since Clang/LLVM libraries are statically
87 linked anyway, it shouldn't affect much.
89 #. The ARM libraries won't be installed in your system, and possibly
90 not easily installable anyway, so you'll have to build/download
91 them separately. But the CMake prepare step, which check for
92 dependencies, will check the `host` libraries, not the `target`
95 A quick way of getting the libraries is to download them from
96 a distribution repository, like Debian (http://packages.debian.org/wheezy/),
97 and download the missing libraries. Note that the `libXXX`
98 will have the shared objects (.so) and the `libXXX-dev` will
99 give you the headers and the static (.a) library. Just in
102 The ones you need for ARM are: ``libtinfo``, ``zlib1g``,
103 ``libxml2`` and ``liblzma``. In the Debian repository you'll
104 find downloads for all architectures.
106 After you download and unpack all `.deb` packages, copy all
107 ``.so`` and ``.a`` to a directory, make the appropriate
108 symbolic links (if necessary), and add the relevant ``-L``
109 and ``-I`` paths to -DCMAKE_CXX_FLAGS above.
112 Running CMake and Building
113 --------------------------
115 Finally, if you're using your platform compiler, run:
119 $ cmake -G Ninja <source-dir> <options above>
121 If you're using Clang as the cross-compiler, run:
125 $ CC='clang' CXX='clang++' cmake -G Ninja <source-dir> <options above>
127 If you have clang/clang++ on the path, it should just work, and special
128 Ninja files will be created in the build directory. I strongly suggest
129 you to run cmake on a separate build directory, *not* inside the
132 To build, simply type:
138 It should automatically find out how many cores you have, what are
139 the rules that needs building and will build the whole thing.
141 You can't run ``ninja check-all`` on this tree because the created
142 binaries are targeted to ARM, not x86_64.
147 After the LLVM/Clang has built successfully, you should install it
154 which will create a sysroot on the install-dir. You can then TarGz
155 that directory into a binary with the full triple name (for easy
156 identification), like:
160 $ ln -sf <install-dir> arm-linux-gnueabihf-clang
161 $ tar zchf arm-linux-gnueabihf-clang.tar.gz arm-linux-gnueabihf-clang
163 If you copy that TarBall to your target board, you'll be able to use
164 it for running the test-suite, for example. Follow the guidelines at
165 http://llvm.org/docs/lnt/quickstart.html, unpack the TarBall in the
166 test directory, and use options:
170 $ ./sandbox/bin/python sandbox/bin/lnt runtest nt \
172 --test-suite `pwd`/test-suite \
173 --cc `pwd`/arm-linux-gnueabihf-clang/bin/clang \
174 --cxx `pwd`/arm-linux-gnueabihf-clang/bin/clang++
176 Remember to add the ``-jN`` options to ``lnt`` to the number of CPUs
177 on your board. Also, the path to your clang has to be absolute, so
178 you'll need the `pwd` trick above.