<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
"http://www.w3.org/TR/html4/strict.dtd">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<title>LLVM Link Time Optimization: Design and Implementation</title>
<link rel="stylesheet" href="llvm.css" type="text/css">
</head>
<h1>
LLVM Link Time Optimization: Design and Implementation
</h1>
<ul>
<li><a href="#desc">Description</a></li>
<li><a href="#design">Design Philosophy</a>
<ul>
<li><a href="#example1">Example of link time optimization</a></li>
<li><a href="#alternative_approaches">Alternative Approaches</a></li>
</ul></li>
<li><a href="#multiphase">Multi-phase communication between LLVM and linker</a>
<ul>
<li><a href="#phase1">Phase 1 : Read LLVM Bitcode Files</a></li>
<li><a href="#phase2">Phase 2 : Symbol Resolution</a></li>
<li><a href="#phase3">Phase 3 : Optimize Bitcode Files</a></li>
<li><a href="#phase4">Phase 4 : Symbol Resolution after optimization</a></li>
</ul></li>
<li><a href="#lto">libLTO</a>
<ul>
<li><a href="#lto_module_t">lto_module_t</a></li>
<li><a href="#lto_code_gen_t">lto_code_gen_t</a></li>
</ul>
</ul>
<div class="doc_author">
<p>Written by Devang Patel and Nick Kledzik</p>
</div>
<!-- *********************************************************************** -->
<h2>
<a name="desc">Description</a>
</h2>
<!-- *********************************************************************** -->
<div>
<p>
LLVM features powerful intermodular optimizations which can be used at link
time. Link Time Optimization (LTO) is another name for intermodular optimization
when performed during the link stage. This document describes the interface
and design between the LTO optimizer and the linker.</p>
</div>
<!-- *********************************************************************** -->
<h2>
<a name="design">Design Philosophy</a>
</h2>
<!-- *********************************************************************** -->
<div>
<p>
The LLVM Link Time Optimizer provides complete transparency, while doing
intermodular optimization, in the compiler tool chain. Its main goal is to let
the developer take advantage of intermodular optimizations without making any
significant changes to the developer's makefiles or build system. This is
achieved through tight integration with the linker. In this model, the linker
treates LLVM bitcode files like native object files and allows mixing and
matching among them. The linker uses <a href="#lto">libLTO</a>, a shared
object, to handle LLVM bitcode files. This tight integration between
the linker and LLVM optimizer helps to do optimizations that are not possible
in other models. The linker input allows the optimizer to avoid relying on
conservative escape analysis.
</p>
<!-- ======================================================================= -->
<h3>
<a name="example1">Example of link time optimization</a>
</h3>
<div>
<p>The following example illustrates the advantages of LTO's integrated
approach and clean interface. This example requires a system linker which
supports LTO through the interface described in this document. Here,
clang transparently invokes system linker. </p>
<ul>
<li> Input source file <tt>a.c</tt> is compiled into LLVM bitcode form.
<li> Input source file <tt>main.c</tt> is compiled into native object code.
</ul>
<pre class="doc_code">
--- a.h ---
extern int foo1(void);
extern void foo2(void);
extern void foo4(void);
--- a.c ---
#include "a.h"
static signed int i = 0;
void foo2(void) {
i = -1;
}
static int foo3() {
foo4();
return 10;
}
int foo1(void) {
int data = 0;
if (i < 0)
data = foo3();
data = data + 42;
return data;
}
--- main.c ---
#include <stdio.h>
#include "a.h"
void foo4(void) {
printf("Hi\n");
}
int main() {
return foo1();
}
--- command lines ---
$ clang -emit-llvm -c a.c -o a.o # <-- a.o is LLVM bitcode file
$ clang -c main.c -o main.o # <-- main.o is native object file
$ clang a.o main.o -o main # <-- standard link command without any modifications
</pre>
<ul>
<li>In this example, the linker recognizes that <tt>foo2()</tt> is an
externally visible symbol defined in LLVM bitcode file. The linker
completes its usual symbol resolution pass and finds that <tt>foo2()</tt>
is not used anywhere. This information is used by the LLVM optimizer and
it removes <tt>foo2()</tt>.</li>
<li>As soon as <tt>foo2()</tt> is removed, the optimizer recognizes that condition
<tt>i < 0</tt> is always false, which means <tt>foo3()</tt> is never
used. Hence, the optimizer also removes <tt>foo3()</tt>.</li>
<li>And this in turn, enables linker to remove <tt>foo4()</tt>.</li>
</ul>
<p>This example illustrates the advantage of tight integration with the
linker. Here, the optimizer can not remove <tt>foo3()</tt> without the
linker's input.</p>
</div>
<!-- ======================================================================= -->
<h3>
<a name="alternative_approaches">Alternative Approaches</a>
</h3>
<div>
<dl>
<dt><b>Compiler driver invokes link time optimizer separately.</b></dt>
<dd>In this model the link time optimizer is not able to take advantage of
information collected during the linker's normal symbol resolution phase.
In the above example, the optimizer can not remove <tt>foo2()</tt> without
the linker's input because it is externally visible. This in turn prohibits
the optimizer from removing <tt>foo3()</tt>.</dd>
<dt><b>Use separate tool to collect symbol information from all object
files.</b></dt>
<dd>In this model, a new, separate, tool or library replicates the linker's
capability to collect information for link time optimization. Not only is
this code duplication difficult to justify, but it also has several other
disadvantages. For example, the linking semantics and the features
provided by the linker on various platform are not unique. This means,
this new tool needs to support all such features and platforms in one
super tool or a separate tool per platform is required. This increases
maintenance cost for link time optimizer significantly, which is not
necessary. This approach also requires staying synchronized with linker
developements on various platforms, which is not the main focus of the link
time optimizer. Finally, this approach increases end user's build time due
to the duplication of work done by this separate tool and the linker itself.
</dd>
</dl>
</div>
</div>
<!-- *********************************************************************** -->
<h2>
<a name="multiphase">Multi-phase communication between libLTO and linker</a>
</h2>
<div>
<p>The linker collects information about symbol defininitions and uses in
various link objects which is more accurate than any information collected
by other tools during typical build cycles. The linker collects this
information by looking at the definitions and uses of symbols in native .o
files and using symbol visibility information. The linker also uses
user-supplied information, such as a list of exported symbols. LLVM
optimizer collects control flow information, data flow information and knows
much more about program structure from the optimizer's point of view.
Our goal is to take advantage of tight integration between the linker and
the optimizer by sharing this information during various linking phases.
</p>
<!-- ======================================================================= -->
<h3>
<a name="phase1">Phase 1 : Read LLVM Bitcode Files</a>
</h3>
<div>
<p>The linker first reads all object files in natural order and collects
symbol information. This includes native object files as well as LLVM bitcode
files. To minimize the cost to the linker in the case that all .o files
are native object files, the linker only calls <tt>lto_module_create()</tt>
when a supplied object file is found to not be a native object file. If
<tt>lto_module_create()</tt> returns that the file is an LLVM bitcode file,
the linker
then iterates over the module using <tt>lto_module_get_symbol_name()</tt> and
<tt>lto_module_get_symbol_attribute()</tt> to get all symbols defined and
referenced.
This information is added to the linker's global symbol table.
</p>
<p>The lto* functions are all implemented in a shared object libLTO. This
allows the LLVM LTO code to be updated independently of the linker tool.
On platforms that support it, the shared object is lazily loaded.
</p>
</div>
<!-- ======================================================================= -->
<h3>
<a name="phase2">Phase 2 : Symbol Resolution</a>
</h3>
<div>
<p>In this stage, the linker resolves symbols using global symbol table.
It may report undefined symbol errors, read archive members, replace
weak symbols, etc. The linker is able to do this seamlessly even though it
does not know the exact content of input LLVM bitcode files. If dead code
stripping is enabled then the linker collects the list of live symbols.
</p>
</div>
<!-- ======================================================================= -->
<h3>
<a name="phase3">Phase 3 : Optimize Bitcode Files</a>
</h3>
<div>
<p>After symbol resolution, the linker tells the LTO shared object which
symbols are needed by native object files. In the example above, the linker
reports that only <tt>foo1()</tt> is used by native object files using
<tt>lto_codegen_add_must_preserve_symbol()</tt>. Next the linker invokes
the LLVM optimizer and code generators using <tt>lto_codegen_compile()</tt>
which returns a native object file creating by merging the LLVM bitcode files
and applying various optimization passes.
</p>
</div>
<!-- ======================================================================= -->
<h3>
<a name="phase4">Phase 4 : Symbol Resolution after optimization</a>
</h3>
<div>
<p>In this phase, the linker reads optimized a native object file and
updates the internal global symbol table to reflect any changes. The linker
also collects information about any changes in use of external symbols by
LLVM bitcode files. In the example above, the linker notes that
<tt>foo4()</tt> is not used any more. If dead code stripping is enabled then
the linker refreshes the live symbol information appropriately and performs
dead code stripping.</p>
<p>After this phase, the linker continues linking as if it never saw LLVM
bitcode files.</p>
</div>
</div>
<!-- *********************************************************************** -->
<h2>
<a name="lto">libLTO</a>
</h2>
<div>
<p><tt>libLTO</tt> is a shared object that is part of the LLVM tools, and
is intended for use by a linker. <tt>libLTO</tt> provides an abstract C
interface to use the LLVM interprocedural optimizer without exposing details
of LLVM's internals. The intention is to keep the interface as stable as
possible even when the LLVM optimizer continues to evolve. It should even
be possible for a completely different compilation technology to provide
a different libLTO that works with their object files and the standard
linker tool.</p>
<!-- ======================================================================= -->
<h3>
<a name="lto_module_t">lto_module_t</a>
</h3>
<div>
<p>A non-native object file is handled via an <tt>lto_module_t</tt>.
The following functions allow the linker to check if a file (on disk
or in a memory buffer) is a file which libLTO can process:</p>
<pre class="doc_code">
lto_module_is_object_file(const char*)
lto_module_is_object_file_for_target(const char*, const char*)
lto_module_is_object_file_in_memory(const void*, size_t)
lto_module_is_object_file_in_memory_for_target(const void*, size_t, const char*)
</pre>
<p>If the object file can be processed by libLTO, the linker creates a
<tt>lto_module_t</tt> by using one of</p>
<pre class="doc_code">
lto_module_create(const char*)
lto_module_create_from_memory(const void*, size_t)
</pre>
<p>and when done, the handle is released via</p>
<pre class="doc_code">
lto_module_dispose(lto_module_t)
</pre>
<p>The linker can introspect the non-native object file by getting the number of
symbols and getting the name and attributes of each symbol via:</p>
<pre class="doc_code">
lto_module_get_num_symbols(lto_module_t)
lto_module_get_symbol_name(lto_module_t, unsigned int)
lto_module_get_symbol_attribute(lto_module_t, unsigned int)
</pre>
<p>The attributes of a symbol include the alignment, visibility, and kind.</p>
</div>
<!-- ======================================================================= -->
<h3>
<a name="lto_code_gen_t">lto_code_gen_t</a>
</h3>
<div>
<p>Once the linker has loaded each non-native object files into an
<tt>lto_module_t</tt>, it can request libLTO to process them all and
generate a native object file. This is done in a couple of steps.
First, a code generator is created with:</p>
<pre class="doc_code">lto_codegen_create()</pre>
<p>Then, each non-native object file is added to the code generator with:</p>
<pre class="doc_code">
lto_codegen_add_module(lto_code_gen_t, lto_module_t)
</pre>
<p>The linker then has the option of setting some codegen options. Whether or
not to generate DWARF debug info is set with:</p>
<pre class="doc_code">lto_codegen_set_debug_model(lto_code_gen_t)</pre>
<p>Which kind of position independence is set with:</p>
<pre class="doc_code">lto_codegen_set_pic_model(lto_code_gen_t) </pre>
<p>And each symbol that is referenced by a native object file or otherwise must
not be optimized away is set with:</p>
<pre class="doc_code">
lto_codegen_add_must_preserve_symbol(lto_code_gen_t, const char*)
</pre>
<p>After all these settings are done, the linker requests that a native object
file be created from the modules with the settings using:</p>
<pre class="doc_code">lto_codegen_compile(lto_code_gen_t, size*)</pre>
<p>which returns a pointer to a buffer containing the generated native
object file. The linker then parses that and links it with the rest
of the native object files.</p>
</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>
Devang Patel and Nick Kledzik<br>
<a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
Last modified: $Date$
</address>
</body>
</html>