llvm/docs/CMake.html

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
                      "http://www.w3.org/TR/html4/strict.dtd">
<html>
<head>
  <title>Building LLVM with CMake</title>
  <link rel="stylesheet" href="llvm.css" type="text/css">
</head>

<h1>
  Building LLVM with CMake
</h1>

<ul>
  <li><a href="#intro">Introduction</a></li>
  <li><a href="#quickstart">Quick start</a></li>
  <li><a href="#usage">Basic CMake usage</a>
  <li><a href="#options">Options and variables</a>
    <ul>
    <li><a href="#freccmake">Frequently-used CMake variables</a></li>
    <li><a href="#llvmvars">LLVM-specific variables</a></li>
  </ul></li>
  <li><a href="#testing">Executing the test suite</a>
  <li><a href="#cross">Cross compiling</a>
  <li><a href="#embedding">Embedding LLVM in your project</a>
    <ul>
    <li><a href="#passdev">Developing LLVM pass out of source</a></li>
  </ul></li>
  <li><a href="#specifics">Compiler/Platform specific topics</a>
    <ul>
    <li><a href="#msvc">Microsoft Visual C++</a></li>
  </ul></li>
</ul>

<div class="doc_author">
<p>Written by <a href="mailto:ofv (a] wanadoo.es">Oscar Fuentes</a></p>
</div>

<!-- *********************************************************************** -->
<h2>
<a name="intro">Introduction</a>
</h2>
<!-- *********************************************************************** -->

<div>

  <p><a href="http://www.cmake.org/">CMake</a> is a cross-platform
    build-generator tool. CMake does not build the project, it generates
    the files needed by your build tool (GNU make, Visual Studio, etc) for
    building LLVM.</p>

  <p>If you are really anxious about getting a functional LLVM build,
    go to the <a href="#quickstart">Quick start</a> section. If you
    are a CMake novice, start on <a href="#usage">Basic CMake
      usage</a> and then go back to the <a href="#quickstart">Quick
      start</a> once you know what you are
    doing. The <a href="#options">Options and variables</a> section
    is a reference for customizing your build. If you already have
    experience with CMake, this is the recommended starting point.
</div>

<!-- *********************************************************************** -->
<h2>
<a name="quickstart">Quick start</a>
</h2>
<!-- *********************************************************************** -->

<div>

<p> We use here the command-line, non-interactive CMake interface </p>

<ol>

  <li><p><a href="http://www.cmake.org/cmake/resources/software.html">Download</a>
      and install CMake. Version 2.8 is the minimum required.</p>

  <li><p>Open a shell. Your development tools must be reachable from this
      shell through the PATH environment variable.</p>

  <li><p>Create a directory for containing the build. It is not
      supported to build LLVM on the source directory. cd to this
      directory:</p>
    <div class="doc_code">
      <p><tt>mkdir mybuilddir</tt></p>
      <p><tt>cd mybuilddir</tt></p>
    </div>

  <li><p>Execute this command on the shell
      replacing <i>path/to/llvm/source/root</i> with the path to the
      root of your LLVM source tree:</p>
    <div class="doc_code">
      <p><tt>cmake path/to/llvm/source/root</tt></p>
    </div>

    <p>CMake will detect your development environment, perform a
      series of test and generate the files required for building
      LLVM. CMake will use default values for all build
      parameters. See the <a href="#options">Options and variables</a>
      section for fine-tuning your build</p>

    <p>This can fail if CMake can't detect your toolset, or if it
      thinks that the environment is not sane enough. On this case
      make sure that the toolset that you intend to use is the only
      one reachable from the shell and that the shell itself is the
      correct one for you development environment. CMake will refuse
      to build MinGW makefiles if you have a POSIX shell reachable
      through the PATH environment variable, for instance. You can
      force CMake to use a given build tool, see
      the <a href="#usage">Usage</a> section.</p>

</ol>

</div>

<!-- *********************************************************************** -->
<h2>
  <a name="usage">Basic CMake usage</a>
</h2>
<!-- *********************************************************************** -->

<div>

  <p>This section explains basic aspects of CMake, mostly for
    explaining those options which you may need on your day-to-day
    usage.</p>

  <p>CMake comes with extensive documentation in the form of html
    files and on the cmake executable itself. Execute <i>cmake
    --help</i> for further help options.</p>

  <p>CMake requires to know for which build tool it shall generate
    files (GNU make, Visual Studio, Xcode, etc). If not specified on
    the command line, it tries to guess it based on you
    environment. Once identified the build tool, CMake uses the
    corresponding <i>Generator</i> for creating files for your build
    tool. You can explicitly specify the generator with the command
    line option <i>-G "Name of the generator"</i>. For knowing the
    available generators on your platform, execute</p>

    <div class="doc_code">
      <p><tt>cmake --help</tt></p>
    </div>

    <p>This will list the generator's names at the end of the help
      text. Generator's names are case-sensitive. Example:</p>

    <div class="doc_code">
      <p><tt>cmake -G "Visual Studio 8 2005" path/to/llvm/source/root</tt></p>
    </div>

    <p>For a given development platform there can be more than one
      adequate generator. If you use Visual Studio "NMake Makefiles"
      is a generator you can use for building with NMake. By default,
      CMake chooses the more specific generator supported by your
      development environment. If you want an alternative generator,
      you must tell this to CMake with the <i>-G</i> option.</p>

    <p>TODO: explain variables and cache. Move explanation here from
      #options section.</p>

</div>

<!-- *********************************************************************** -->
<h2>
  <a name="options">Options and variables</a>
</h2>
<!-- *********************************************************************** -->

<div>

  <p>Variables customize how the build will be generated. Options are
    boolean variables, with possible values ON/OFF. Options and
    variables are defined on the CMake command line like this:</p>

  <div class="doc_code">
    <p><tt>cmake -DVARIABLE=value path/to/llvm/source</tt></p>
  </div>

  <p>You can set a variable after the initial CMake invocation for
    changing its value. You can also undefine a variable:</p>

  <div class="doc_code">
    <p><tt>cmake -UVARIABLE path/to/llvm/source</tt></p>
  </div>

  <p>Variables are stored on the CMake cache. This is a file
    named <tt>CMakeCache.txt</tt> on the root of the build
    directory. Do not hand-edit it.</p>

  <p>Variables are listed here appending its type after a colon. It is
    correct to write the variable and the type on the CMake command
    line:</p>

  <div class="doc_code">
    <p><tt>cmake -DVARIABLE:TYPE=value path/to/llvm/source</tt></p>
  </div>

<!-- ======================================================================= -->
<h3>
  <a name="freccmake">Frequently-used CMake variables</a>
</h3>

<div>

<p>Here are listed some of the CMake variables that are used often,
  along with a brief explanation and LLVM-specific notes. For full
  documentation, check the CMake docs or execute <i>cmake
  --help-variable VARIABLE_NAME</i>.</p>

<dl>
  <dt><b>CMAKE_BUILD_TYPE</b>:STRING</dt>

  <dd>Sets the build type for <i>make</i> based generators. Possible
    values are Release, Debug, RelWithDebInfo and MinSizeRel. On
    systems like Visual Studio the user sets the build type with the IDE
    settings.</dd>

  <dt><b>CMAKE_INSTALL_PREFIX</b>:PATH</dt>
  <dd>Path where LLVM will be installed if "make install" is invoked
    or the "INSTALL" target is built.</dd>

  <dt><b>LLVM_LIBDIR_SUFFIX</b>:STRING</dt>
  <dd>Extra suffix to append to the directory where libraries are to
    be installed. On a 64-bit architecture, one could use
    -DLLVM_LIBDIR_SUFFIX=64 to install libraries to /usr/lib64.</dd>

  <dt><b>CMAKE_C_FLAGS</b>:STRING</dt>
  <dd>Extra flags to use when compiling C source files.</dd>

  <dt><b>CMAKE_CXX_FLAGS</b>:STRING</dt>
  <dd>Extra flags to use when compiling C++ source files.</dd>

  <dt><b>BUILD_SHARED_LIBS</b>:BOOL</dt>
  <dd>Flag indicating is shared libraries will be built. Its default
    value is OFF. Shared libraries are not supported on Windows and
    not recommended in the other OSes.</dd>
</dl>

</div>

<!-- ======================================================================= -->
<h3>
  <a name="llvmvars">LLVM-specific variables</a>
</h3>

<div>

<dl>
  <dt><b>LLVM_TARGETS_TO_BUILD</b>:STRING</dt>
  <dd>Semicolon-separated list of targets to build, or <i>all</i> for
    building all targets. Case-sensitive. For Visual C++ defaults
    to <i>X86</i>. On the other cases defaults to <i>all</i>. Example:
    <i>-DLLVM_TARGETS_TO_BUILD="X86;PowerPC;Alpha"</i>.</dd>

  <dt><b>LLVM_BUILD_TOOLS</b>:BOOL</dt>
  <dd>Build LLVM tools. Defaults to ON. Targets for building each tool
    are generated in any case. You can build an tool separately by
    invoking its target. For example, you can build <i>llvm-as</i>
    with a makefile-based system executing <i>make llvm-as</i> on the
    root of your build directory.</dd>

  <dt><b>LLVM_INCLUDE_TOOLS</b>:BOOL</dt>
  <dd>Generate build targets for the LLVM tools. Defaults to
    ON. You can use that option for disabling the generation of build
    targets for the LLVM tools.</dd>

  <dt><b>LLVM_BUILD_EXAMPLES</b>:BOOL</dt>
  <dd>Build LLVM examples. Defaults to OFF. Targets for building each
    example are generated in any case. See documentation
    for <i>LLVM_BUILD_TOOLS</i> above for more details.</dd>

  <dt><b>LLVM_INCLUDE_EXAMPLES</b>:BOOL</dt>
  <dd>Generate build targets for the LLVM examples. Defaults to
    ON. You can use that option for disabling the generation of build
    targets for the LLVM examples.</dd>

  <dt><b>LLVM_BUILD_TESTS</b>:BOOL</dt>
  <dd>Build LLVM unit tests. Defaults to OFF. Targets for building
    each unit test are generated in any case. You can build a specific
    unit test with the target <i>UnitTestNameTests</i> (where at this
    time <i>UnitTestName</i> can be ADT, Analysis, ExecutionEngine,
    JIT, Support, Transform, VMCore; see the subdirectories
    of <i>unittests</i> for an updated list.) It is possible to build
    all unit tests with the target <i>UnitTests</i>.</dd>

  <dt><b>LLVM_INCLUDE_TESTS</b>:BOOL</dt>
  <dd>Generate build targets for the LLVM unit tests. Defaults to
    ON. You can use that option for disabling the generation of build
    targets for the LLVM unit tests.</dd>

  <dt><b>LLVM_APPEND_VC_REV</b>:BOOL</dt>
  <dd>Append version control revision info (svn revision number or git
    revision id) to LLVM version string (stored in the PACKAGE_VERSION
    macro). For this to work cmake must be invoked before the
    build. Defaults to OFF.</dd>

  <dt><b>LLVM_ENABLE_THREADS</b>:BOOL</dt>
  <dd>Build with threads support, if available. Defaults to ON.</dd>

  <dt><b>LLVM_ENABLE_ASSERTIONS</b>:BOOL</dt>
  <dd>Enables code assertions. Defaults to OFF if and only if
    CMAKE_BUILD_TYPE is <i>Release</i>.</dd>

  <dt><b>LLVM_ENABLE_PIC</b>:BOOL</dt>
  <dd>Add the <i>-fPIC</i> flag for the compiler command-line, if the
    compiler supports this flag. Some systems, like Windows, do not
    need this flag. Defaults to ON.</dd>

  <dt><b>LLVM_ENABLE_WARNINGS</b>:BOOL</dt>
  <dd>Enable all compiler warnings. Defaults to ON.</dd>

  <dt><b>LLVM_ENABLE_PEDANTIC</b>:BOOL</dt>
  <dd>Enable pedantic mode. This disable compiler specific extensions, is
    possible. Defaults to ON.</dd>

  <dt><b>LLVM_ENABLE_WERROR</b>:BOOL</dt>
  <dd>Stop and fail build, if a compiler warning is
    triggered. Defaults to OFF.</dd>

  <dt><b>LLVM_BUILD_32_BITS</b>:BOOL</dt>
  <dd>Build 32-bits executables and libraries on 64-bits systems. This
    option is available only on some 64-bits unix systems. Defaults to
    OFF.</dd>

  <dt><b>LLVM_TARGET_ARCH</b>:STRING</dt>
  <dd>LLVM target to use for native code generation. This is required
    for JIT generation. It defaults to "host", meaning that it shall
    pick the architecture of the machine where LLVM is being built. If
    you are cross-compiling, set it to the target architecture
    name.</dd>

  <dt><b>LLVM_TABLEGEN</b>:STRING</dt>
  <dd>Full path to a native TableGen executable (usually
    named <i>tblgen</i>). This is intented for cross-compiling: if the
    user sets this variable, no native TableGen will be created.</dd>

  <dt><b>LLVM_LIT_ARGS</b>:STRING</dt>
  <dd>Arguments given to lit.
    <tt>make check</tt> and <tt>make clang-test</tt> are affected.
    By default, <tt>&quot;-sv --no-progress-bar&quot;</tt>
    on Visual C++ and Xcode,
    <tt>&quot;-sv&quot;</tt> on others.</dd>

  <dt><b>LLVM_LIT_TOOLS_DIR</b>:PATH</dt>
  <dd>The path to GnuWin32 tools for tests. Valid on Windows host.
    Defaults to "", then Lit seeks tools according to %PATH%.
    Lit can find tools(eg. grep, sort, &amp;c) on LLVM_LIT_TOOLS_DIR at first,
    without specifying GnuWin32 to %PATH%.</dd>

  <dt><b>LLVM_ENABLE_FFI</b>:BOOL</dt>
  <dd>Indicates whether LLVM Interpreter will be linked with Foreign
    Function Interface library. If the library or its headers are
    installed on a custom location, you can set the variables
    FFI_INCLUDE_DIR and FFI_LIBRARY_DIR. Defaults to OFF.</dd>

  <dt><b>LLVM_CLANG_SOURCE_DIR</b>:PATH</dt>
  <dd>Path to Clang's source directory. Defaults to tools/clang.
    Clang will not be built when it is empty or it does not point valid
    path.</dd>
</dl>

</div>

</div>

<!-- *********************************************************************** -->
<h2>
  <a name="testing">Executing the test suite</a>
</h2>
<!-- *********************************************************************** -->

<div>

<p>Testing is performed when the <i>check</i> target is built. For
  instance, if you are using makefiles, execute this command while on
  the top level of your build directory:</p>

<div class="doc_code">
  <p><tt>make check</tt></p>
</div>

<p>On Visual Studio, you may run tests to build the project "check".</p>

</div>

<!-- *********************************************************************** -->
<h2>
  <a name="cross">Cross compiling</a>
</h2>
<!-- *********************************************************************** -->

<div>

<p>See <a href="http://www.vtk.org/Wiki/CMake_Cross_Compiling">this
    wiki page</a> for generic instructions on how to cross-compile
    with CMake. It goes into detailed explanations and may seem
    daunting, but it is not. On the wiki page there are several
    examples including toolchain files. Go directly to
    <a href="http://www.vtk.org/Wiki/CMake_Cross_Compiling#Information_how_to_set_up_various_cross_compiling_toolchains">this
    section</a> for a quick solution.</p>

<p>Also see the <a href="#llvmvars">LLVM-specific variables</a>
  section for variables used when cross-compiling.</p>

</div>

<!-- *********************************************************************** -->
<h2>
  <a name="embedding">Embedding LLVM in your project</a>
</h2>
<!-- *********************************************************************** -->

<div>

  <p>The most difficult part of adding LLVM to the build of a project
    is to determine the set of LLVM libraries corresponding to the set
    of required LLVM features. What follows is an example of how to
    obtain this information:</p>

  <div class="doc_code">
    <pre>
    <b># A convenience variable:</b>
    set(LLVM_ROOT "" CACHE PATH "Root of LLVM install.")
    <b># A bit of a sanity check:</b>
    if( NOT EXISTS ${LLVM_ROOT}/include/llvm )
    message(FATAL_ERROR "LLVM_ROOT (${LLVM_ROOT}) is not a valid LLVM install")
    endif()
    <b># We incorporate the CMake features provided by LLVM:</b>
    set(CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH} "${LLVM_ROOT}/share/llvm/cmake")
    include(LLVMConfig)
    <b># Now set the header and library paths:</b>
    include_directories( ${LLVM_INCLUDE_DIRS} )
    link_directories( ${LLVM_LIBRARY_DIRS} )
    add_definitions( ${LLVM_DEFINITIONS} )
    <b># Let's suppose we want to build a JIT compiler with support for
    # binary code (no interpreter):</b>
    llvm_map_components_to_libraries(REQ_LLVM_LIBRARIES jit native)
    <b># Finally, we link the LLVM libraries to our executable:</b>
    target_link_libraries(mycompiler ${REQ_LLVM_LIBRARIES})
    </pre>
  </div>

  <p>This assumes that LLVM_ROOT points to an install of LLVM. The
    procedure works too for uninstalled builds although we need to take
    care to add an <i>include_directories</i> for the location of the
    headers on the LLVM source directory (if we are building
    out-of-source.)</p>

  <p>Alternativaly, you can utilize CMake's <i>find_package</i>
    functionality. Here is an equivalent variant of snippet shown above:</p>

  <div class="doc_code">
    <pre>
    find_package(LLVM)

    if( NOT LLVM_FOUND )
      message(FATAL_ERROR "LLVM package can't be found. Set CMAKE_PREFIX_PATH variable to LLVM's installation prefix.")
    endif()

    include_directories( ${LLVM_INCLUDE_DIRS} )
    link_directories( ${LLVM_LIBRARY_DIRS} )

    llvm_map_components_to_libraries(REQ_LLVM_LIBRARIES jit native)

    target_link_libraries(mycompiler ${REQ_LLVM_LIBRARIES})
    </pre>
  </div>

<!-- ======================================================================= -->
<h3>
  <a name="passdev">Developing LLVM pass out of source</a>
</h3>

<div>

  <p>It is possible to develop LLVM passes against installed LLVM.
     An example of project layout provided below:</p>

  <div class="doc_code">
    <pre>
      &lt;project dir&gt;/
          |
          CMakeLists.txt
          &lt;pass name&gt;/
              |
              CMakeLists.txt
              Pass.cpp
              ...
    </pre>
  </div>

  <p>Contents of &lt;project dir&gt;/CMakeLists.txt:</p>

  <div class="doc_code">
    <pre>
    find_package(LLVM)

    <b># Define add_llvm_* macro's.</b>
    include(AddLLVM)

    add_definitions(${LLVM_DEFINITIONS})
    include_directories(${LLVM_INCLUDE_DIRS})
    link_directories(${LLVM_LIBRARY_DIRS})

    add_subdirectory(&lt;pass name&gt;)
    </pre>
  </div>

  <p>Contents of &lt;project dir&gt;/&lt;pass name&gt;/CMakeLists.txt:</p>

  <div class="doc_code">
    <pre>
    add_llvm_loadable_module(LLVMPassname
      Pass.cpp
      )
    </pre>
  </div>

  <p>When you are done developing your pass, you may wish to integrate it
     into LLVM source tree. You can achieve it in two easy steps:<br>
     1. Copying &lt;pass name&gt; folder into &lt;LLVM root&gt;/lib/Transform directory.<br>
     2. Adding "add_subdirectory(&lt;pass name&gt;)" line into &lt;LLVM root&gt;/lib/Transform/CMakeLists.txt</p>
</div>
<!-- *********************************************************************** -->

</div>

<!-- *********************************************************************** -->
<h2>
  <a name="specifics">Compiler/Platform specific topics</a>
</h2>
<!-- *********************************************************************** -->

<div>

<p>Notes for specific compilers and/or platforms.</p>

<h3>
  <a name="msvc">Microsoft Visual C++</a>
</h3>

<div>

<dl>
  <dt><b>LLVM_COMPILER_JOBS</b>:STRING</dt>
  <dd>Specifies the maximum number of parallell compiler jobs to use
    per project when building with msbuild or Visual Studio. Only supported for
    Visual Studio 2008 and Visual Studio 2010 CMake generators. 0 means use all
    processors. Default is 0.</dd>
</dl>

</div>

</div>

<!-- *********************************************************************** -->

<hr>
<address>
  <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
  src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
  <a href="http://validator.w3.org/check/referer"><img
  src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>

  <a href="mailto:ofv (a] wanadoo.es">Oscar Fuentes</a><br>
  <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
  Last modified: $Date: 2010-08-09 03:59:36 +0100 (Mon, 9 Aug 2010) $
</address>

</body>
</html>