The next section of this guide is meant to get
you up and running with LLVM, and to give you some basic information about
the LLVM environment. The first subsection gives
a short summary for those who are already familiar with the system and
want to get started as quickly as possible.
The later sections of this guide describe the general layout of the the LLVM source-tree, a simple example using the LLVM tool chain, and links to find more information about LLVM or to get
help via e-mail.
Here's the short story for getting up and running quickly with LLVM:
- Find the path to the CVS repository containing LLVM (we'll call this CVSROOTDIR).
- cd where-you-want-llvm-to-live
- cvs -d CVSROOTDIR checkout llvm
- cd llvm
- Edit Makefile.config to set local paths. This includes
setting the install location of the C frontend, and the various paths
to the C and C++ compilers used to build LLVM itself.
- Set your LLVM_LIB_SEARCH_PATH environment variable.
- gmake -k |& tee gnumake.out
# this is csh or tcsh syntax
See Setting up your environment on tips to
simplify working with the LLVM front-end and compiled tools. See the
other sub-sections below for other useful details in working with LLVM,
or go straight to Program Layout to learn about the
layout of the source code tree.
Through this manual, the following names are used to denote paths
specific to the local system and working environment. These are not
environment variables you need to set, but just strings used in the rest
of this document below.. In any of the examples below, simply replace
each of these names with the appropriate pathname on your local system.
All these paths are absolute:
Before checking out the source code, you will need to know the path to
CVS repository containing LLVM source code (we'll call this
CVSROOTDIR below). Ask the person responsible for your local LLVM
installation to give you this path.
To get a fresh copy of the entire source code, all you
need to do is check it out from CVS as follows:
- cd where-you-want-llvm-to-live
- cvs -d CVSROOTDIR checkout llvm
This will create an 'llvm' directory in the current
directory and fully populate it with the LLVM source code, Makefiles,
test directories, and local copies of documentation files.
The file llvm/Makefile.config
defines the following path variables,
which are specific to a particular installation of LLVM.
These should need to be modified only once after checking out a copy
of LLVM (if the default values do not already match your system):
- CXX = Path to C++ compiler to use.
- LLVM_OBJ_DIR = Path to the llvm directory where
object files should be placed.
(See the Section on
The location for LLVM object files
for more information.)
- LLVMGCCDIR = Path to the location of the LLVM front-end
binaries and associated libraries.
- PURIFY = Path to the purify program.
In addition to settings in this file, you must set a
LLVM_LIB_SEARCH_PATH environment variable in your startup scripts.
This environment variable is used to locate "system" libraries like
"-lc" and "-lm" when linking. This variable should be set
to the absolute path for the bytecode-libs subdirectory of the C front-end
install. For example,
/home/vadve/lattner/local/x86/llvm-gcc/bytecode-libs for the X86
version of the C front-end, on our research machines.
The LLVM make system sends most output files generated during the build
into the directory defined by the variable LLVM_OBJ_DIR in
llvm/Makefile.config. This can be either just your normal LLVM
source tree or some other directory writable by you. You may wish to put
object files on a different filesystem either to keep them from being backed
up or to speed up local builds.
If you do not wish to use a different location for object files (building
into the source tree directly), just set this variable to ".".
NOTE: This step is optional but will set up your environment so you
can use the compiled LLVM tools with as little hassle as
possible.)
Add the following lines to your .cshrc (or the corresponding
lines to your .profile if you use a bourne shell derivative).
# Make the C front end easy to use...
alias llvmgcc LLVMGCCDIR/bin/llvm-gcc
# Make the LLVM tools easy to use...
setenv PATH LLVM_OBJ_DIR/tools/Debug:${PATH}
The llvmgcc alias is useful because the C compiler is not
included in the CVS tree you just checked out.
The other LLVM tools are part of the LLVM
source base, and built when compiling LLVM. They will be built into the
LLVM_OBJ_DIR/tools/Debug directory.
Every directory in the LLVM source tree includes a Makefile to
build it, and any subdirectories that it contains. These makefiles require
that you use gmake, instead of make to build them, but can
otherwise be used freely. To build the entire LLVM system, just enter the
top level llvm directory and type gmake. A few minutes
later you will hopefully have a freshly compiled toolchain waiting for you
in llvm/tools/Debug. If you want to look at the libraries that
were compiled, look in llvm/lib/Debug.
If you get an error talking about a /localhome directory, follow
the instructions in the section about Setting Up Your
Environment.
One useful source of infomation about the LLVM sourcebase is the LLVM doxygen documentation, available at http://llvm.cs.uiuc.edu/doxygen/. The
following is a brief introduction to code layout:
Every directory checked out of CVS will contain a CVS directory,
for the most part these can just be ignored.
If you are building with the "BUILD_ROOT=." option enabled in the
Makefile.common file, most source directories will contain two
directories, Depend and Debug. The Depend
directory contains automatically generated dependance files which are used
during compilation to make sure that source files get rebuilt if a header
file they use is modified. The Debug directory holds the object
files, library files and executables that are used for building a debug
enabled build. The Release directory is created to hold the same
files when the ENABLE_OPTIMIZED=1 flag is passed to gmake,
causing an optimized built to be performed.
This directory contains public header files exported from the LLVM
library. The two main subdirectories of this directory are:
- llvm/include/llvm - This directory contains all of the LLVM
specific header files. This directory also has subdirectories for
different portions of LLVM: Analysis, CodeGen,
Reoptimizer, Target, Transforms, etc...
- llvm/include/Support - This directory contains generic
support libraries that are independant of LLVM, but are used by LLVM.
For example, some C++ STL utilities and a Command Line option processing
library.
This directory contains most source files of LLVM system. In LLVM almost all
code exists in libraries, making it very easy to share code among the
different tools.
- llvm/lib/VMCore/
- This directory holds the core LLVM
source files that implement core classes like Instruction and BasicBlock.
- llvm/lib/AsmParser/
- This directory holds the source code
for the LLVM assembly language parser library.
- llvm/lib/ByteCode/
- This directory holds code for reading
and write LLVM bytecode.
- llvm/lib/CWriter/
- This directory implements the LLVM to C
converter.
- llvm/lib/Analysis/
- This directory contains a variety of
different program analyses, such as Dominator Information, Call Graphs,
Induction Variables, Interval Identification, Natural Loop Identification,
etc...
- llvm/lib/Transforms/
- This directory contains the source
code for the LLVM to LLVM program transformations, such as Aggressive Dead
Code Elimination, Sparse Conditional Constant Propogation, Inlining, Loop
Invarient Code Motion, Dead Global Elimination, Pool Allocation, and many
others...
- llvm/lib/Target/
- This directory contains files that
describe various target architectures for code generation. For example,
the llvm/lib/Target/Sparc directory holds the Sparc machine
description.
- llvm/lib/CodeGen/
- This directory contains the major parts
of the code generator: Instruction Selector, Instruction Scheduling, and
Register Allocation.
- llvm/lib/Reoptimizer/
- This directory holds code related
to the runtime reoptimizer framework that is currently under development.
- llvm/lib/Support/
- This directory contains the source code
that corresponds to the header files located in
llvm/include/Support/.
This directory contains regression tests and source code that is used to
test the LLVM infrastructure...
The tools directory contains the executables built out of the
libraries above, which form the main part of the user interface. You can
always get help for a tool by typing tool_name --help. The
following is a brief introduction to the most important tools.
- as
- The assembler transforms the human readable
LLVM assembly to LLVM bytecode.
- dis
- The disassembler transforms the LLVM bytecode
to human readable LLVM assembly. Additionally it can convert LLVM
bytecode to C, which is enabled with the -c option.
- lli
- lli is the LLVM interpreter, which
can directly execute LLVM bytecode (although very slowly...). In addition
to a simple intepreter, lli is also has debugger and tracing
modes (entered by specifying -debug or -trace on the
command line, respectively).
- llc
- llc is the LLVM backend compiler,
which translates LLVM bytecode to a SPARC assembly file.
- llvmgcc
- llvmgcc is a GCC based C frontend
that has been retargeted to emit LLVM code as the machine code output. It
works just like any other GCC compiler, taking the typical -c, -S, -E,
-o options that are typically used. The source code for the
llvmgcc tool is currently not included in the LLVM cvs tree
because it is quite large and not very interesting.
- gccas
- This tool is invoked by the
llvmgcc frontend as the "assembler" part of the compiler. This
tool actually assembles LLVM assembly to LLVM bytecode,
performs a variety of optimizations,
and outputs LLVM bytecode. Thus when you invoke llvmgcc -c x.c -o
x.o, you are causing gccas to be run, which writes the
x.o file (which is an LLVM bytecode file that can be
disassembled or manipulated just like any other bytecode file). The
command line interface to gccas is designed to be as close as
possible to the system 'as' utility so that the gcc
frontend itself did not have to be modified to interface to a "wierd"
assembler.
- gccld
- gccld links together several LLVM
bytecode files into one bytecode file and does some optimization. It is
the linker invoked by the gcc frontend when multiple .o files need to be
linked together. Like gccas the command line interface of
gccld is designed to match the system linker, to aid
interfacing with the GCC frontend.
- opt
- opt reads LLVM bytecode, applies a
series of LLVM to LLVM transformations (which are specified on the command
line), and then outputs the resultant bytecode. The 'opt --help'
command is a good way to get a list of the program transformations
available in LLVM.
- analyze
- analyze is used to run a specific
analysis on an input LLVM bytecode file and print out the results. It is
primarily useful for debugging analyses, or familiarizing yourself with
what an analysis does.
- First, create a simple C file, name it 'hello.c':
#include <stdio.h>
int main() {
printf("hello world\n");
return 0;
}
- Next, compile the C file into a LLVM bytecode file:
% llvmgcc hello.c -o hello
This will create two result files: hello and
hello.bc. The hello.bc is the LLVM bytecode that
corresponds the the compiled program and the library facilities that it
required. hello is a simple shell script that runs the bytecode
file with lli, making the result directly executable.
- Run the program. To make sure the program ran, execute one of the
following commands:
% ./hello
or
% lli hello.bc
- Use the dis utility to take a look at the LLVM assembly
code:
% dis < hello.bc | less
- Compile the program to native Sparc assembly using the code
generator:
% llc hello.bc -o hello.s
- Assemble the native sparc assemble file into a program:
% /opt/SUNWspro/bin/cc -xarch=v9 hello.s -o hello.sparc
- Execute the native sparc program:
% ./hello.sparc
This document is just an introduction to how to use LLVM to do
some simple things... there are many more interesting and complicated things
that you can do that aren't documented here (but we'll gladly accept a patch
if you want to write something up!). For more information about LLVM, check
out:
If you have any questions or run into any snags (or you have any
additions...), please send an email to
Nicholas Hildenbrandt or
Chris Lattner.
Last modified: Sun May 11 16:49:46 CDT 2003