1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
4 <title>Getting Started with LLVM System</title>
8 <center><h1>Getting Started with the LLVM System<br><font size=3>By: <a
9 href="mailto:gshi1@uiuc.edu">Guochun Shi</a>,
10 <a href="mailto:sabre@nondot.org">Chris Lattner</a> and
11 <a href="http://www.cs.uiuc.edu/~vadve">Vikram Adve</a>
14 <!--=====================================================================-->
15 <h2><a name="Contents">Contents</a></h2>
16 <!--=====================================================================-->
19 <li><a href="#overview">Overview</a>
20 <li><a href="#starting">Getting started with LLVM</a>
22 <li><a href="#quickstart">Getting started quickly (a summary)</a>
23 <li><a href="#checkout">Checkout LLVM from CVS</a>
24 <li><a href="#terminology">Terminology and Notation</tt></a>
25 <li><a href="#objfiles">The location for object files</tt></a>
26 <li><a href="#config">Local Configuration Options</tt></a>
27 <li><a href="#environment">Setting up your environment</a>
28 <li><a href="#compile">Compiling the source code</a>
30 <li><a href="#layout">Program layout</a>
32 <li><a href="#cvsdir">CVS directories</a>
33 <li><a href="#dd"><tt>Depend</tt>, <tt>Debug</tt>, &
34 <tt>Release</tt> directories</a></li>
35 <li><a href="#include"><tt>llvm/include</tt></a>
36 <li><a href="#lib"><tt>llvm/lib</tt></a>
37 <li><a href="#test"><tt>llvm/test</tt></a>
38 <li><a href="#tools"><tt>llvm/tools</tt></a>
40 <li><a href="#tutorial">An example using the LLVM tool chain</a>
41 <li><a href="#links">Links</a>
45 <!--=====================================================================-->
47 <h2><a name="overview"><b>Overview</b></a></h2>
49 <!--=====================================================================-->
51 <p>The <a href"starting">next section</a> of this guide is meant to get
52 you up and running with LLVM, and to give you some basic information about
53 the LLVM environment. The <a href"#quickstart">first subsection</a> gives
54 a short summary for those who are already familiar with the system and
55 want to get started as quickly as possible.
57 <p>The later sections of this guide describe the <a
58 href"#layout">general layout</a> of the the LLVM source-tree, a <a
59 href="#tutorial">simple example</a> using the LLVM tool chain, and <a
60 href="#links">links</a> to find more information about LLVM or to get
63 <!--=====================================================================-->
65 <h2><a name="starting"><b>Getting Started</b></a></h2>
67 <!--=====================================================================-->
70 <!--=====================================================================-->
71 <h3><a name="quickstart"><b>Getting Started Quickly (A Summary)</b></a></h3>
72 <!--=====================================================================-->
74 Here's the short story for getting up and running quickly with LLVM:
76 <li>Find the path to the CVS repository containing LLVM (we'll call this <i>CVSROOTDIR</i>).
77 <li><tt>cd <i>where-you-want-llvm-to-live</i></tt>
78 <li><tt>cvs -d <i>CVSROOTDIR</i> checkout llvm</tt>
80 <li>Edit <tt>Makefile.config</tt> to set local paths. This includes
81 setting the install location of the C frontend, and the various paths
82 to the C and C++ compilers used to build LLVM itself.
83 <li>Set your LLVM_LIB_SEARCH_PATH environment variable.
84 <li><tt>gmake -k |& tee gnumake.out
85 # this is csh or tcsh syntax</tt>
88 <p>See <a href="#environment">Setting up your environment</a> on tips to
89 simplify working with the LLVM front-end and compiled tools. See the
90 other sub-sections below for other useful details in working with LLVM,
91 or go straight to <a href="#layout">Program Layout</a> to learn about the
92 layout of the source code tree.
94 <!------------------------------------------------------------------------->
95 <h3><a name="terminology">Terminology and Notation</a></h3>
96 <!------------------------------------------------------------------------->
98 <p>Through this manual, the following names are used to denote paths
99 specific to the local system and working environment. <i>These are not
100 environment variables you need to set, but just strings used in the rest
101 of this document below.</i>. In any of the examples below, simply replace
102 each of these names with the appropriate pathname on your local system.
103 All these paths are absolute:</p>
107 <!------------------------------------------------------------------------->
108 <h3><a name="checkout">Checkout LLVM from CVS</a></h3>
109 <!------------------------------------------------------------------------->
111 <p>Before checking out the source code, you will need to know the path to
112 CVS repository containing LLVM source code (we'll call this
113 <i>CVSROOTDIR</i> below). Ask the person responsible for your local LLVM
114 installation to give you this path.
116 <p>To get a fresh copy of the entire source code, all you
117 need to do is check it out from CVS as follows:
119 <li><tt>cd <i>where-you-want-llvm-to-live</i></tt>
120 <li><tt>cvs -d <i>CVSROOTDIR</i> checkout llvm</tt></p>
123 <p>This will create an '<tt>llvm</tt>' directory in the current
124 directory and fully populate it with the LLVM source code, Makefiles,
125 test directories, and local copies of documentation files.</p>
127 <!------------------------------------------------------------------------->
128 <h3><a name="config">Local Configuration Options</a></h3>
129 <!------------------------------------------------------------------------->
131 <p>The file <tt>llvm/Makefile.config</tt>
132 defines the following path variables,
133 which are specific to a particular installation of LLVM.
134 These should need to be modified only once after checking out a copy
135 of LLVM (if the default values do not already match your system):
138 <p><li><i>CXX</i> = Path to C++ compiler to use.
139 <p><li><i>LLVM_OBJ_DIR</i> = Path to the llvm directory where
140 object files should be placed.
141 (See the Section on <a href=#objfiles>
142 The location for LLVM object files</a>
143 for more information.)
144 <p><li><i>LLVMGCCDIR</i> = Path to the location of the LLVM front-end
145 binaries and associated libraries.
146 <p><li><i>PURIFY</i> = Path to the purify program.
149 In addition to settings in this file, you must set a
150 <tt>LLVM_LIB_SEARCH_PATH</tt> environment variable in your startup scripts.
151 This environment variable is used to locate "system" libraries like
152 "<tt>-lc</tt>" and "<tt>-lm</tt>" when linking. This variable should be set
153 to the absolute path for the bytecode-libs subdirectory of the C front-end
154 install. For example,
155 <tt>/home/vadve/lattner/local/x86/llvm-gcc/bytecode-libs</tt> for the X86
156 version of the C front-end, on our research machines.<p>
158 <!------------------------------------------------------------------------->
159 <h3><a name="objfiles">The location for LLVM object files</a></h3>
160 <!------------------------------------------------------------------------->
162 <p>The LLVM make system sends most output files generated during the build
163 into the directory defined by the variable LLVM_OBJ_DIR in
164 <tt>llvm/Makefile.config</tt>. This can be either just your normal LLVM
165 source tree or some other directory writable by you. You may wish to put
166 object files on a different filesystem either to keep them from being backed
167 up or to speed up local builds.
169 <p>If you do not wish to use a different location for object files (building
170 into the source tree directly), just set this variable to ".".<p>
172 <!------------------------------------------------------------------------->
173 <h3><a name="environment">Setting up your environment</a></h3>
174 <!------------------------------------------------------------------------->
176 <i>NOTE: This step is optional but will set up your environment so you
177 can use the compiled LLVM tools with as little hassle as
180 <p>Add the following lines to your <tt>.cshrc</tt> (or the corresponding
181 lines to your <tt>.profile</tt> if you use a bourne shell derivative).
184 # Make the C front end easy to use...
185 alias llvmgcc <i>LLVMGCCDIR</i><tt>/bin/llvm-gcc</tt>
187 # Make the LLVM tools easy to use...
188 setenv PATH <i>LLVM_OBJ_DIR</i>/llvm/tools/Debug:${PATH}
190 The <tt>llvmgcc</tt> alias is useful because the C compiler is not
191 included in the CVS tree you just checked out.
193 <p>The other <a href="#tools">LLVM tools</a> are part of the LLVM
194 source base, and built when compiling LLVM. They will be built into the
195 <tt><i>LLVM_OBJ_DIR</i>/tools/Debug</tt> directory.</p>
197 <!------------------------------------------------------------------------->
198 <h3><a name="compile">Compiling the source code</a></h3>
199 <!------------------------------------------------------------------------->
201 <p>Every directory in the LLVM source tree includes a <tt>Makefile</tt> to
202 build it, and any subdirectories that it contains. These makefiles require
203 that you use <tt>gmake</tt>, instead of <tt>make</tt> to build them, but can
204 otherwise be used freely. To build the entire LLVM system, just enter the
205 top level <tt>llvm</tt> directory and type <tt>gmake</tt>. A few minutes
206 later you will hopefully have a freshly compiled toolchain waiting for you
207 in <tt>llvm/tools/Debug</tt>. If you want to look at the libraries that
208 were compiled, look in <tt>llvm/lib/Debug</tt>.</p>
210 If you get an error talking about a <tt>/localhome</tt> directory, follow
211 the instructions in the section about <a href="#environment">Setting Up Your
216 <!--=====================================================================-->
218 <h2><a name="layout"><b>Program Layout</b></a></h2>
220 <!--=====================================================================-->
222 <p>One useful source of infomation about the LLVM sourcebase is the LLVM <a
223 href="http://www.doxygen.org">doxygen</a> documentation, available at <tt><a
224 href="http://llvm.cs.uiuc.edu/doxygen/">http://llvm.cs.uiuc.edu/doxygen/</a></tt>. The
225 following is a brief introduction to code layout:</p>
228 <!------------------------------------------------------------------------->
229 <h3><a name="cvsdir"><tt>CVS</tt> directories</a></h3>
230 <!------------------------------------------------------------------------->
232 Every directory checked out of CVS will contain a <tt>CVS</tt> directory,
233 for the most part these can just be ignored.
236 <!------------------------------------------------------------------------->
237 <h3><a name="ddr"><tt>Depend</tt>, <tt>Debug</tt>, & <tt>Release</tt>
239 <!------------------------------------------------------------------------->
241 If you are building with the "<tt>BUILD_ROOT=.</tt>" option enabled in the
242 <tt>Makefile.common</tt> file, most source directories will contain two
243 directories, <tt>Depend</tt> and <tt>Debug</tt>. The <tt>Depend</tt>
244 directory contains automatically generated dependance files which are used
245 during compilation to make sure that source files get rebuilt if a header
246 file they use is modified. The <tt>Debug</tt> directory holds the object
247 files, library files and executables that are used for building a debug
248 enabled build. The <tt>Release</tt> directory is created to hold the same
249 files when the <tt>ENABLE_OPTIMIZED=1</tt> flag is passed to <tt>gmake</tt>,
250 causing an optimized built to be performed.<p>
253 <!------------------------------------------------------------------------->
254 <h3><a name="include"><tt>llvm/include</tt></a></h3>
255 <!------------------------------------------------------------------------->
257 This directory contains public header files exported from the LLVM
258 library. The two main subdirectories of this directory are:<p>
261 <li><tt>llvm/include/llvm</tt> - This directory contains all of the LLVM
262 specific header files. This directory also has subdirectories for
263 different portions of LLVM: <tt>Analysis</tt>, <tt>CodeGen</tt>,
264 <tt>Reoptimizer</tt>, <tt>Target</tt>, <tt>Transforms</tt>, etc...
266 <li><tt>llvm/include/Support</tt> - This directory contains generic
267 support libraries that are independant of LLVM, but are used by LLVM.
268 For example, some C++ STL utilities and a Command Line option processing
272 <!------------------------------------------------------------------------->
273 <h3><a name="lib"><tt>llvm/lib</tt></a></h3>
274 <!------------------------------------------------------------------------->
276 This directory contains most source files of LLVM system. In LLVM almost all
277 code exists in libraries, making it very easy to share code among the
278 different <a href="#tools">tools</a>.<p>
281 <dt><tt>llvm/lib/VMCore/</tt><dd> This directory holds the core LLVM
282 source files that implement core classes like Instruction and BasicBlock.
284 <dt><tt>llvm/lib/AsmParser/</tt><dd> This directory holds the source code
285 for the LLVM assembly language parser library.
287 <dt><tt>llvm/lib/ByteCode/</tt><dd> This directory holds code for reading
288 and write LLVM bytecode.
290 <dt><tt>llvm/lib/CWriter/</tt><dd> This directory implements the LLVM to C
293 <dt><tt>llvm/lib/Analysis/</tt><dd> This directory contains a variety of
294 different program analyses, such as Dominator Information, Call Graphs,
295 Induction Variables, Interval Identification, Natural Loop Identification,
298 <dt><tt>llvm/lib/Transforms/</tt><dd> This directory contains the source
299 code for the LLVM to LLVM program transformations, such as Aggressive Dead
300 Code Elimination, Sparse Conditional Constant Propagation, Inlining, Loop
301 Invarient Code Motion, Dead Global Elimination, Pool Allocation, and many
304 <dt><tt>llvm/lib/Target/</tt><dd> This directory contains files that
305 describe various target architectures for code generation. For example,
306 the llvm/lib/Target/Sparc directory holds the Sparc machine
309 <dt><tt>llvm/lib/CodeGen/</tt><dd> This directory contains the major parts
310 of the code generator: Instruction Selector, Instruction Scheduling, and
313 <dt><tt>llvm/lib/Reoptimizer/</tt><dd> This directory holds code related
314 to the runtime reoptimizer framework that is currently under development.
316 <dt><tt>llvm/lib/Support/</tt><dd> This directory contains the source code
317 that corresponds to the header files located in
318 <tt>llvm/include/Support/</tt>.
321 <!------------------------------------------------------------------------->
322 <h3><a name="test"><tt>llvm/test</tt></a></h3>
323 <!------------------------------------------------------------------------->
325 <p>This directory contains regression tests and source code that is used to
326 test the LLVM infrastructure...</p>
328 <!------------------------------------------------------------------------->
329 <h3><a name="tools"><tt>llvm/tools</tt></a></h3>
330 <!------------------------------------------------------------------------->
332 <p>The <b>tools</b> directory contains the executables built out of the
333 libraries above, which form the main part of the user interface. You can
334 always get help for a tool by typing <tt>tool_name --help</tt>. The
335 following is a brief introduction to the most important tools.</p>
338 <dt><tt><b>as</b></tt><dd>The assembler transforms the human readable
339 LLVM assembly to LLVM bytecode.<p>
341 <dt><tt><b>dis</b></tt><dd>The disassembler transforms the LLVM bytecode
342 to human readable LLVM assembly. Additionally it can convert LLVM
343 bytecode to C, which is enabled with the <tt>-c</tt> option.<p>
345 <dt><tt><b>lli</b></tt><dd> <tt>lli</tt> is the LLVM interpreter, which
346 can directly execute LLVM bytecode (although very slowly...). In addition
347 to a simple intepreter, <tt>lli</tt> is also has debugger and tracing
348 modes (entered by specifying <tt>-debug</tt> or <tt>-trace</tt> on the
349 command line, respectively).<p>
351 <dt><tt><b>llc</b></tt><dd> <tt>llc</tt> is the LLVM backend compiler,
352 which translates LLVM bytecode to a SPARC assembly file.<p>
354 <dt><tt><b>llvmgcc</b></tt><dd> <tt>llvmgcc</tt> is a GCC based C frontend
355 that has been retargeted to emit LLVM code as the machine code output. It
356 works just like any other GCC compiler, taking the typical <tt>-c, -S, -E,
357 -o</tt> options that are typically used. The source code for the
358 <tt>llvmgcc</tt> tool is currently not included in the LLVM cvs tree
359 because it is quite large and not very interesting.<p>
362 <dt><tt><b>gccas</b></tt><dd> This tool is invoked by the
363 <tt>llvmgcc</tt> frontend as the "assembler" part of the compiler. This
364 tool actually assembles LLVM assembly to LLVM bytecode,
365 performs a variety of optimizations,
366 and outputs LLVM bytecode. Thus when you invoke <tt>llvmgcc -c x.c -o
367 x.o</tt>, you are causing <tt>gccas</tt> to be run, which writes the
368 <tt>x.o</tt> file (which is an LLVM bytecode file that can be
369 disassembled or manipulated just like any other bytecode file). The
370 command line interface to <tt>gccas</tt> is designed to be as close as
371 possible to the <b>system</b> '<tt>as</tt>' utility so that the gcc
372 frontend itself did not have to be modified to interface to a "wierd"
375 <dt><tt><b>gccld</b></tt><dd> <tt>gccld</tt> links together several LLVM
376 bytecode files into one bytecode file and does some optimization. It is
377 the linker invoked by the gcc frontend when multiple .o files need to be
378 linked together. Like <tt>gccas</tt> the command line interface of
379 <tt>gccld</tt> is designed to match the system linker, to aid
380 interfacing with the GCC frontend.<p>
383 <dt><tt><b>opt</b></tt><dd> <tt>opt</tt> reads LLVM bytecode, applies a
384 series of LLVM to LLVM transformations (which are specified on the command
385 line), and then outputs the resultant bytecode. The '<tt>opt --help</tt>'
386 command is a good way to get a list of the program transformations
387 available in LLVM.<p>
390 <dt><tt><b>analyze</b></tt><dd> <tt>analyze</tt> is used to run a specific
391 analysis on an input LLVM bytecode file and print out the results. It is
392 primarily useful for debugging analyses, or familiarizing yourself with
393 what an analysis does.<p>
397 <!--=====================================================================-->
398 <h2><a name="tutorial">An example using the LLVM tool chain</h2>
399 <!--=====================================================================-->
402 <li>First, create a simple C file, name it 'hello.c':
404 #include <stdio.h>
406 printf("hello world\n");
411 <li>Next, compile the C file into a LLVM bytecode file:<p>
413 <tt>% llvmgcc hello.c -o hello</tt><p>
415 This will create two result files: <tt>hello</tt> and
416 <tt>hello.bc</tt>. The <tt>hello.bc</tt> is the LLVM bytecode that
417 corresponds the the compiled program and the library facilities that it
418 required. <tt>hello</tt> is a simple shell script that runs the bytecode
419 file with <tt>lli</tt>, making the result directly executable.<p>
421 <li>Run the program. To make sure the program ran, execute one of the
422 following commands:<p>
424 <tt>% ./hello</tt><p>
428 <tt>% lli hello.bc</tt><p>
430 <li>Use the <tt>dis</tt> utility to take a look at the LLVM assembly
433 <tt>% dis < hello.bc | less</tt><p>
435 <li>Compile the program to native Sparc assembly using the code
438 <tt>% llc hello.bc -o hello.s</tt><p>
440 <li>Assemble the native sparc assemble file into a program:<p>
442 <tt>% /opt/SUNWspro/bin/cc -xarch=v9 hello.s -o hello.sparc</tt><p>
444 <li>Execute the native sparc program:<p>
446 <tt>% ./hello.sparc</tt><p>
451 <!--=====================================================================-->
452 <h2><a name="links">Links</a></h2>
453 <!--=====================================================================-->
455 <p>This document is just an <b>introduction</b> to how to use LLVM to do
456 some simple things... there are many more interesting and complicated things
457 that you can do that aren't documented here (but we'll gladly accept a patch
458 if you want to write something up!). For more information about LLVM, check
462 <li><a href="http://llvm.cs.uiuc.edu/">LLVM homepage</a></li>
463 <li><a href="http://llvm.cs.uiuc.edu/doxygen/">LLVM doxygen tree</a></li>
468 If you have any questions or run into any snags (or you have any
469 additions...), please send an email to
470 <a href="mailto:sabre@nondot.org">Chris Lattner</a>.</p>
472 <!-- Created: Mon Jul 1 02:29:02 CDT 2002 -->
474 Last modified: Tue Jun 3 22:06:43 CDT 2003