Customizing LLVMC: Reference Manual

Note: This document is a work-in-progress. Additions and clarifications are welcome.

LLVMC is a generic compiler driver, designed to be customizable and extensible. It plays the same role for LLVM as the gcc program does for GCC - LLVMC's job is essentially to transform a set of input files into a set of targets depending on configuration rules and user options. What makes LLVMC different is that these transformation rules are completely customizable - in fact, LLVMC knows nothing about the specifics of transformation (even the command-line options are mostly not hard-coded) and regards the transformation structure as an abstract graph. This makes it possible to adapt LLVMC for other purposes - for example, as a build tool for game resources.

Because LLVMC employs TableGen [1] as its configuration language, you need to be familiar with it to customize LLVMC.

Written by Mikhail Glushenkov

LLVMC tries hard to be as compatible with gcc as possible, although there are some small differences. Most of the time, however, you shouldn't be able to notice them:

$ # This works as expected:
$ llvmc2 -O3 -Wall hello.cpp
$ ./a.out
hello

One nice feature of LLVMC is that one doesn't have to distinguish between different compilers for different languages (think g++ and gcc) - the right toolchain is chosen automatically based on input language names (which are, in turn, determined from file extensions). If you want to force files ending with ".c" to compile as C++, use the -x option, just like you would do it with gcc:

$ llvmc2 -x c hello.cpp
$ # hello.cpp is really a C file
$ ./a.out
hello

On the other hand, when using LLVMC as a linker to combine several C++ object files you should provide the --linker option since it's impossible for LLVMC to choose the right linker in that case:

$ llvmc2 -c hello.cpp
$ llvmc2 hello.o
[A lot of link-time errors skipped]
$ llvmc2 --linker=c++ hello.o
$ ./a.out
hello

LLVMC has some built-in options that can't be overridden in the configuration files:

At the time of writing LLVMC does not support on-the-fly reloading of configuration, so to customize LLVMC you'll have to recompile the source code (which lives under $LLVM_DIR/tools/llvmc2). The default configuration files are Common.td (contains common definitions, don't forget to include it in your configuration files), Tools.td (tool descriptions) and Graph.td (compilation graph definition).

To compile LLVMC with your own configuration file (say,``MyGraph.td``), run make like this:

$ cd $LLVM_DIR/tools/llvmc2
$ make GRAPH=MyGraph.td TOOLNAME=my_llvmc

This will build an executable named my_llvmc. There are also several sample configuration files in the llvmc2/examples subdirectory that should help to get you started.

Internally, LLVMC stores information about possible source transformations in form of a graph. Nodes in this graph represent tools, and edges between two nodes represent a transformation path. A special "root" node is used to mark entry points for the transformations. LLVMC also assigns a weight to each edge (more on this later) to choose between several alternative edges.

The definition of the compilation graph (see file Graph.td) is just a list of edges:

def CompilationGraph : CompilationGraph<[
    Edge<root, llvm_gcc_c>,
    Edge<root, llvm_gcc_assembler>,
    ...

    Edge<llvm_gcc_c, llc>,
    Edge<llvm_gcc_cpp, llc>,
    ...

    OptionalEdge<llvm_gcc_c, opt, [(switch_on "opt")]>,
    OptionalEdge<llvm_gcc_cpp, opt, [(switch_on "opt")]>,
    ...

    OptionalEdge<llvm_gcc_assembler, llvm_gcc_cpp_linker,
        (case (input_languages_contain "c++"), (inc_weight),
              (or (parameter_equals "linker", "g++"),
                  (parameter_equals "linker", "c++")), (inc_weight))>,
    ...

    ]>;

As you can see, the edges can be either default or optional, where optional edges are differentiated by sporting a case expression used to calculate the edge's weight.

The default edges are assigned a weight of 1, and optional edges get a weight of 0 + 2*N where N is the number of tests that evaluated to true in the case expression. It is also possible to provide an integer parameter to inc_weight and dec_weight - in this case, the weight is increased (or decreased) by the provided value instead of the default 2.

When passing an input file through the graph, LLVMC picks the edge with the maximum weight. To avoid ambiguity, there should be only one default edge between two nodes (with the exception of the root node, which gets a special treatment - there you are allowed to specify one default edge per language).

To get a visual representation of the compilation graph (useful for debugging), run llvmc2 --view-graph. You will need dot and gsview installed for this to work properly.

As was said earlier, nodes in the compilation graph represent tools, which are described separately. A tool definition looks like this (taken from the Tools.td file):

def llvm_gcc_cpp : Tool<[
    (in_language "c++"),
    (out_language "llvm-assembler"),
    (output_suffix "bc"),
    (cmd_line "llvm-g++ -c $INFILE -o $OUTFILE -emit-llvm"),
    (sink)
    ]>;

This defines a new tool called llvm_gcc_cpp, which is an alias for llvm-g++. As you can see, a tool definition is just a list of properties; most of them should be self-explanatory. The sink property means that this tool should be passed all command-line options that lack explicit descriptions.

The complete list of the currently implemented tool properties follows:

The next tool definition is slightly more complex:

def llvm_gcc_linker : Tool<[
    (in_language "object-code"),
    (out_language "executable"),
    (output_suffix "out"),
    (cmd_line "llvm-gcc $INFILE -o $OUTFILE"),
    (join),
    (prefix_list_option "L", (forward),
                        (help "add a directory to link path")),
    (prefix_list_option "l", (forward),
                        (help "search a library when linking")),
    (prefix_list_option "Wl", (unpack_values),
                        (help "pass options to linker"))
    ]>;

This tool has a "join" property, which means that it behaves like a linker. This tool also defines several command-line options: -l, -L and -Wl which have their usual meaning. An option has two attributes: a name and a (possibly empty) list of properties. All currently implemented option types and properties are described below:

It can be handy to have all information about options gathered in a single place to provide an overview. This can be achieved by using a so-called OptionList:

def Options : OptionList<[
(switch_option "E", (help "Help string")),
(alias_option "quiet", "q")
...
]>;

OptionList is also a good place to specify option aliases.

Tool-specific option properties like append_cmd have (obviously) no meaning in the context of OptionList, so the only properties allowed there are help and required.

Option lists are used at the file scope. See file examples/Clang.td for an example of OptionList usage.

Normally, LLVMC executes programs from the system PATH. Sometimes, this is not sufficient: for example, we may want to specify tool names in the configuration file. This can be achieved via the mechanism of hooks - to compile LLVMC with your hooks, just drop a .cpp file into tools/llvmc2 directory. Hooks should live in the hooks namespace and have the signature std::string hooks::MyHookName (void). They can be used from the cmd_line tool property:

(cmd_line "$CALL(MyHook)/path/to/file -o $CALL(AnotherHook)")

It is also possible to use environment variables in the same manner:

(cmd_line "$ENV(VAR1)/path/to/file -o $ENV(VAR2)")

To change the command line string based on user-provided options use the case expression (documented below):

(cmd_line
  (case
    (switch_on "E"),
       "llvm-g++ -E -x c $INFILE -o $OUTFILE",
    (default),
       "llvm-g++ -c -x c $INFILE -o $OUTFILE -emit-llvm"))

The 'case' construct can be used to calculate weights of the optional edges and to choose between several alternative command line strings in the cmd_line tool property. It is designed after the similarly-named construct in functional languages and takes the form (case (test_1), statement_1, (test_2), statement_2, ... (test_N), statement_N). The statements are evaluated only if the corresponding tests evaluate to true.

Examples:

// Increases edge weight by 5 if "-A" is provided on the
// command-line, and by 5 more if "-B" is also provided.
(case
    (switch_on "A"), (inc_weight 5),
    (switch_on "B"), (inc_weight 5))

// Evaluates to "cmdline1" if option "-A" is provided on the
// command line, otherwise to "cmdline2"
(case
    (switch_on "A"), "cmdline1",
    (switch_on "B"), "cmdline2",
    (default), "cmdline3")

Note the slight difference in 'case' expression handling in contexts of edge weights and command line specification - in the second example the value of the "B" switch is never checked when switch "A" is enabled, and the whole expression always evaluates to "cmdline1" in that case.

Case expressions can also be nested, i.e. the following is legal:

(case (switch_on "E"), (case (switch_on "o"), ..., (default), ...)
      (default), ...)

You should, however, try to avoid doing that because it hurts readability. It is usually better to split tool descriptions and/or use TableGen inheritance instead.

One last thing that you will need to modify when adding support for a new language to LLVMC is the language map, which defines mappings from file extensions to language names. It is used to choose the proper toolchain(s) for a given input file set. Language map definition is located in the file Tools.td and looks like this:

def LanguageMap : LanguageMap<
    [LangToSuffixes<"c++", ["cc", "cp", "cxx", "cpp", "CPP", "c++", "C"]>,
     LangToSuffixes<"c", ["c"]>,
     ...
    ]>;
[1]TableGen Fundamentals http://llvm.cs.uiuc.edu/docs/TableGenFundamentals.html

Valid CSS! Valid XHTML 1.0! The LLVM Compiler Infrastructure
Last modified: $Date$