Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / llvm / 0e2d3ac0f8794af9f4502df131f78cea7b8c2518 / . / docs / ProgrammersManual.html
blob: 9661b7693574c076f0d8b87f311b5f642b75fa2e [file] [log] [blame]
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3<html>
4<head>
5 <title>LLVM Programmer's Manual</title>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]6 <link rel="stylesheet" href="llvm.css" type="text/css">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]7</head>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]8<body>
9
10<div class="doc_title">
11 LLVM Programmer's Manual
12</div>
13
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]14<ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]15 <li><a href="#introduction">Introduction</a></li>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]16 <li><a href="#general">General Information</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]17 <ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]18 <li><a href="#stl">The C++ Standard Template Library</a></li>
19<!--
20 <li>The <tt>-time-passes</tt> option</li>
21 <li>How to use the LLVM Makefile system</li>
22 <li>How to write a regression test</li>
Chris Lattner61db4652004-12-08 19:05:44 +0000[diff] [blame]23
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]24-->
Chris Lattner84b7f8d2003-08-01 22:20:59 +0000[diff] [blame]25 </ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]26 </li>
27 <li><a href="#apis">Important and useful LLVM APIs</a>
28 <ul>
29 <li><a href="#isa">The <tt>isa&lt;&gt;</tt>, <tt>cast&lt;&gt;</tt>
30and <tt>dyn_cast&lt;&gt;</tt> templates</a> </li>
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]31 <li><a href="#DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]32option</a>
33 <ul>
34 <li><a href="#DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt>
35and the <tt>-debug-only</tt> option</a> </li>
36 </ul>
37 </li>
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]38 <li><a href="#Statistic">The <tt>Statistic</tt> class &amp; <tt>-stats</tt>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]39option</a></li>
40<!--
41 <li>The <tt>InstVisitor</tt> template
42 <li>The general graph API
43-->
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]44 <li><a href="#ViewGraph">Viewing graphs while debugging code</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]45 </ul>
46 </li>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]47 <li><a href="#datastructure">Picking the Right Data Structure for a Task</a>
48 <ul>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]49 <li><a href="#ds_sequential">Sequential Containers (std::vector, std::list, etc)</a>
50 <ul>
51 <li><a href="#dss_fixedarrays">Fixed Size Arrays</a></li>
52 <li><a href="#dss_heaparrays">Heap Allocated Arrays</a></li>
53 <li><a href="#dss_smallvector">"llvm/ADT/SmallVector.h"</a></li>
54 <li><a href="#dss_vector">&lt;vector&gt;</a></li>
55 <li><a href="#dss_deque">&lt;deque&gt;</a></li>
56 <li><a href="#dss_list">&lt;list&gt;</a></li>
57 <li><a href="#dss_ilist">llvm/ADT/ilist</a></li>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]58 <li><a href="#dss_other">Other Sequential Container Options</a></li>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]59 </ul></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]60 <li><a href="#ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
61 <ul>
62 <li><a href="#dss_sortedvectorset">A sorted 'vector'</a></li>
63 <li><a href="#dss_smallset">"llvm/ADT/SmallSet.h"</a></li>
64 <li><a href="#dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a></li>
65 <li><a href="#dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a></li>
66 <li><a href="#dss_set">&lt;set&gt;</a></li>
67 <li><a href="#dss_setvector">"llvm/ADT/SetVector.h"</a></li>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]68 <li><a href="#dss_uniquevector">"llvm/ADT/UniqueVector.h"</a></li>
69 <li><a href="#dss_otherset">Other Set-Like ContainerOptions</a></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]70 </ul></li>
Chris Lattnerf3692522007-02-03 19:51:56 +0000[diff] [blame]71 <li><a href="#ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
72 <ul>
73 <li><a href="#dss_sortedvectormap">A sorted 'vector'</a></li>
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]74 <li><a href="#dss_stringmap">"llvm/ADT/StringMap.h"</a></li>
Chris Lattnerf3692522007-02-03 19:51:56 +0000[diff] [blame]75 <li><a href="#dss_indexedmap">"llvm/ADT/IndexedMap.h"</a></li>
76 <li><a href="#dss_densemap">"llvm/ADT/DenseMap.h"</a></li>
77 <li><a href="#dss_map">&lt;map&gt;</a></li>
78 <li><a href="#dss_othermap">Other Map-Like Container Options</a></li>
79 </ul></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]80 </ul>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]81 </li>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]82 <li><a href="#common">Helpful Hints for Common Operations</a>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]83 <ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]84 <li><a href="#inspection">Basic Inspection and Traversal Routines</a>
85 <ul>
86 <li><a href="#iterate_function">Iterating over the <tt>BasicBlock</tt>s
87in a <tt>Function</tt></a> </li>
88 <li><a href="#iterate_basicblock">Iterating over the <tt>Instruction</tt>s
89in a <tt>BasicBlock</tt></a> </li>
90 <li><a href="#iterate_institer">Iterating over the <tt>Instruction</tt>s
91in a <tt>Function</tt></a> </li>
92 <li><a href="#iterate_convert">Turning an iterator into a
93class pointer</a> </li>
94 <li><a href="#iterate_complex">Finding call sites: a more
95complex example</a> </li>
96 <li><a href="#calls_and_invokes">Treating calls and invokes
97the same way</a> </li>
98 <li><a href="#iterate_chains">Iterating over def-use &amp;
99use-def chains</a> </li>
100 </ul>
101 </li>
102 <li><a href="#simplechanges">Making simple changes</a>
103 <ul>
104 <li><a href="#schanges_creating">Creating and inserting new
105 <tt>Instruction</tt>s</a> </li>
106 <li><a href="#schanges_deleting">Deleting <tt>Instruction</tt>s</a> </li>
107 <li><a href="#schanges_replacing">Replacing an <tt>Instruction</tt>
108with another <tt>Value</tt></a> </li>
109 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]110 </li>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]111<!--
112 <li>Working with the Control Flow Graph
113 <ul>
114 <li>Accessing predecessors and successors of a <tt>BasicBlock</tt>
115 <li>
116 <li>
117 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]118-->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]119 </ul>
120 </li>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]121
122 <li><a href="#advanced">Advanced Topics</a>
123 <ul>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]124 <li><a href="#TypeResolve">LLVM Type Resolution</a>
125 <ul>
126 <li><a href="#BuildRecType">Basic Recursive Type Construction</a></li>
127 <li><a href="#refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a></li>
128 <li><a href="#PATypeHolder">The PATypeHolder Class</a></li>
129 <li><a href="#AbstractTypeUser">The AbstractTypeUser Class</a></li>
130 </ul></li>
131
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]132 <li><a href="#SymbolTable">The <tt>ValueSymbolTable</tt> and <tt>TypeSymbolTable</tt> classes </a></li>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]133 </ul></li>
134
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]135 <li><a href="#coreclasses">The Core LLVM Class Hierarchy Reference</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]136 <ul>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]137 <li><a href="#Type">The <tt>Type</tt> class</a> </li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]138 <li><a href="#Module">The <tt>Module</tt> class</a></li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]139 <li><a href="#Value">The <tt>Value</tt> class</a>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]140 <ul>
141 <li><a href="#User">The <tt>User</tt> class</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]142 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]143 <li><a href="#Instruction">The <tt>Instruction</tt> class</a></li>
144 <li><a href="#Constant">The <tt>Constant</tt> class</a>
145 <ul>
146 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]147 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]148 <li><a href="#Function">The <tt>Function</tt> class</a></li>
149 <li><a href="#GlobalVariable">The <tt>GlobalVariable</tt> class</a></li>
150 </ul>
151 </li>
152 </ul>
153 </li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]154 </ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]155 </li>
156 <li><a href="#BasicBlock">The <tt>BasicBlock</tt> class</a></li>
157 <li><a href="#Argument">The <tt>Argument</tt> class</a></li>
158 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]159 </li>
160 </ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]161 </li>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]162</ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]163
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]164<div class="doc_author">
165 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
Chris Lattner94c43592004-05-26 16:52:55 +0000[diff] [blame]166 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>,
167 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a>, and
168 <a href="mailto:rspencer@x10sys.com">Reid Spencer</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]169</div>
170
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]171<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]172<div class="doc_section">
173 <a name="introduction">Introduction </a>
174</div>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]175<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]176
177<div class="doc_text">
178
179<p>This document is meant to highlight some of the important classes and
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]180interfaces available in the LLVM source-base. This manual is not
181intended to explain what LLVM is, how it works, and what LLVM code looks
182like. It assumes that you know the basics of LLVM and are interested
183in writing transformations or otherwise analyzing or manipulating the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]184code.</p>
185
186<p>This document should get you oriented so that you can find your
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]187way in the continuously growing source code that makes up the LLVM
188infrastructure. Note that this manual is not intended to serve as a
189replacement for reading the source code, so if you think there should be
190a method in one of these classes to do something, but it's not listed,
191check the source. Links to the <a href="/doxygen/">doxygen</a> sources
192are provided to make this as easy as possible.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]193
194<p>The first section of this document describes general information that is
195useful to know when working in the LLVM infrastructure, and the second describes
196the Core LLVM classes. In the future this manual will be extended with
197information describing how to use extension libraries, such as dominator
198information, CFG traversal routines, and useful utilities like the <tt><a
199href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.</p>
200
201</div>
202
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]203<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]204<div class="doc_section">
205 <a name="general">General Information</a>
206</div>
207<!-- *********************************************************************** -->
208
209<div class="doc_text">
210
211<p>This section contains general information that is useful if you are working
212in the LLVM source-base, but that isn't specific to any particular API.</p>
213
214</div>
215
216<!-- ======================================================================= -->
217<div class="doc_subsection">
218 <a name="stl">The C++ Standard Template Library</a>
219</div>
220
221<div class="doc_text">
222
223<p>LLVM makes heavy use of the C++ Standard Template Library (STL),
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]224perhaps much more than you are used to, or have seen before. Because of
225this, you might want to do a little background reading in the
226techniques used and capabilities of the library. There are many good
227pages that discuss the STL, and several books on the subject that you
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]228can get, so it will not be discussed in this document.</p>
229
230<p>Here are some useful links:</p>
231
232<ol>
233
234<li><a href="http://www.dinkumware.com/refxcpp.html">Dinkumware C++ Library
235reference</a> - an excellent reference for the STL and other parts of the
236standard C++ library.</li>
237
238<li><a href="http://www.tempest-sw.com/cpp/">C++ In a Nutshell</a> - This is an
Tanya Lattner09cf73c2004-06-22 04:24:55 +0000[diff] [blame]239O'Reilly book in the making. It has a decent
240Standard Library
241Reference that rivals Dinkumware's, and is unfortunately no longer free since the book has been
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]242published.</li>
243
244<li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
245Questions</a></li>
246
247<li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
248Contains a useful <a
249href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
250STL</a>.</li>
251
252<li><a href="http://www.research.att.com/%7Ebs/C++.html">Bjarne Stroustrup's C++
253Page</a></li>
254
Tanya Lattner79445ba2004-12-08 18:34:56 +0000[diff] [blame]255<li><a href="http://64.78.49.204/">
Reid Spencer096603a2004-05-26 08:41:35 +0000[diff] [blame]256Bruce Eckel's Thinking in C++, 2nd ed. Volume 2 Revision 4.0 (even better, get
257the book).</a></li>
258
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]259</ol>
260
261<p>You are also encouraged to take a look at the <a
262href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
263to write maintainable code more than where to put your curly braces.</p>
264
265</div>
266
267<!-- ======================================================================= -->
268<div class="doc_subsection">
269 <a name="stl">Other useful references</a>
270</div>
271
272<div class="doc_text">
273
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]274<ol>
275<li><a href="http://www.psc.edu/%7Esemke/cvs_branches.html">CVS
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]276Branch and Tag Primer</a></li>
Misha Brukmana0f71e42004-06-18 18:39:00 +0000[diff] [blame]277<li><a href="http://www.fortran-2000.com/ArnaudRecipes/sharedlib.html">Using
278static and shared libraries across platforms</a></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]279</ol>
280
281</div>
282
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]283<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]284<div class="doc_section">
285 <a name="apis">Important and useful LLVM APIs</a>
286</div>
287<!-- *********************************************************************** -->
288
289<div class="doc_text">
290
291<p>Here we highlight some LLVM APIs that are generally useful and good to
292know about when writing transformations.</p>
293
294</div>
295
296<!-- ======================================================================= -->
297<div class="doc_subsection">
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]298 <a name="isa">The <tt>isa&lt;&gt;</tt>, <tt>cast&lt;&gt;</tt> and
299 <tt>dyn_cast&lt;&gt;</tt> templates</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]300</div>
301
302<div class="doc_text">
303
304<p>The LLVM source-base makes extensive use of a custom form of RTTI.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]305These templates have many similarities to the C++ <tt>dynamic_cast&lt;&gt;</tt>
306operator, but they don't have some drawbacks (primarily stemming from
307the fact that <tt>dynamic_cast&lt;&gt;</tt> only works on classes that
308have a v-table). Because they are used so often, you must know what they
309do and how they work. All of these templates are defined in the <a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]310 href="/doxygen/Casting_8h-source.html"><tt>llvm/Support/Casting.h</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]311file (note that you very rarely have to include this file directly).</p>
312
313<dl>
314 <dt><tt>isa&lt;&gt;</tt>: </dt>
315
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]316 <dd><p>The <tt>isa&lt;&gt;</tt> operator works exactly like the Java
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]317 "<tt>instanceof</tt>" operator. It returns true or false depending on whether
318 a reference or pointer points to an instance of the specified class. This can
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]319 be very useful for constraint checking of various sorts (example below).</p>
320 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]321
322 <dt><tt>cast&lt;&gt;</tt>: </dt>
323
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]324 <dd><p>The <tt>cast&lt;&gt;</tt> operator is a "checked cast" operation. It
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]325 converts a pointer or reference from a base class to a derived cast, causing
326 an assertion failure if it is not really an instance of the right type. This
327 should be used in cases where you have some information that makes you believe
328 that something is of the right type. An example of the <tt>isa&lt;&gt;</tt>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]329 and <tt>cast&lt;&gt;</tt> template is:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]330
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]331<div class="doc_code">
332<pre>
333static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
334 if (isa&lt;<a href="#Constant">Constant</a>&gt;(V) || isa&lt;<a href="#Argument">Argument</a>&gt;(V) || isa&lt;<a href="#GlobalValue">GlobalValue</a>&gt;(V))
335 return true;
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]336
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]337 // <i>Otherwise, it must be an instruction...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]338 return !L-&gt;contains(cast&lt;<a href="#Instruction">Instruction</a>&gt;(V)-&gt;getParent());
339}
340</pre>
341</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]342
343 <p>Note that you should <b>not</b> use an <tt>isa&lt;&gt;</tt> test followed
344 by a <tt>cast&lt;&gt;</tt>, for that use the <tt>dyn_cast&lt;&gt;</tt>
345 operator.</p>
346
347 </dd>
348
349 <dt><tt>dyn_cast&lt;&gt;</tt>:</dt>
350
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]351 <dd><p>The <tt>dyn_cast&lt;&gt;</tt> operator is a "checking cast" operation.
352 It checks to see if the operand is of the specified type, and if so, returns a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]353 pointer to it (this operator does not work with references). If the operand is
354 not of the correct type, a null pointer is returned. Thus, this works very
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]355 much like the <tt>dynamic_cast&lt;&gt;</tt> operator in C++, and should be
356 used in the same circumstances. Typically, the <tt>dyn_cast&lt;&gt;</tt>
357 operator is used in an <tt>if</tt> statement or some other flow control
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]358 statement like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]359
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]360<div class="doc_code">
361<pre>
362if (<a href="#AllocationInst">AllocationInst</a> *AI = dyn_cast&lt;<a href="#AllocationInst">AllocationInst</a>&gt;(Val)) {
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]363 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]364}
365</pre>
366</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]367
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]368 <p>This form of the <tt>if</tt> statement effectively combines together a call
369 to <tt>isa&lt;&gt;</tt> and a call to <tt>cast&lt;&gt;</tt> into one
370 statement, which is very convenient.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]371
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]372 <p>Note that the <tt>dyn_cast&lt;&gt;</tt> operator, like C++'s
373 <tt>dynamic_cast&lt;&gt;</tt> or Java's <tt>instanceof</tt> operator, can be
374 abused. In particular, you should not use big chained <tt>if/then/else</tt>
375 blocks to check for lots of different variants of classes. If you find
376 yourself wanting to do this, it is much cleaner and more efficient to use the
377 <tt>InstVisitor</tt> class to dispatch over the instruction type directly.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]378
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]379 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]380
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]381 <dt><tt>cast_or_null&lt;&gt;</tt>: </dt>
382
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]383 <dd><p>The <tt>cast_or_null&lt;&gt;</tt> operator works just like the
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]384 <tt>cast&lt;&gt;</tt> operator, except that it allows for a null pointer as an
385 argument (which it then propagates). This can sometimes be useful, allowing
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]386 you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]387
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]388 <dt><tt>dyn_cast_or_null&lt;&gt;</tt>: </dt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]389
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]390 <dd><p>The <tt>dyn_cast_or_null&lt;&gt;</tt> operator works just like the
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]391 <tt>dyn_cast&lt;&gt;</tt> operator, except that it allows for a null pointer
392 as an argument (which it then propagates). This can sometimes be useful,
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]393 allowing you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]394
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]395</dl>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]396
397<p>These five templates can be used with any classes, whether they have a
398v-table or not. To add support for these templates, you simply need to add
399<tt>classof</tt> static methods to the class you are interested casting
400to. Describing this is currently outside the scope of this document, but there
401are lots of examples in the LLVM source base.</p>
402
403</div>
404
405<!-- ======================================================================= -->
406<div class="doc_subsection">
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]407 <a name="DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt> option</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]408</div>
409
410<div class="doc_text">
411
412<p>Often when working on your pass you will put a bunch of debugging printouts
413and other code into your pass. After you get it working, you want to remove
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]414it, but you may need it again in the future (to work out new bugs that you run
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]415across).</p>
416
417<p> Naturally, because of this, you don't want to delete the debug printouts,
418but you don't want them to always be noisy. A standard compromise is to comment
419them out, allowing you to enable them if you need them in the future.</p>
420
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]421<p>The "<tt><a href="/doxygen/Debug_8h-source.html">llvm/Support/Debug.h</a></tt>"
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]422file provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to
423this problem. Basically, you can put arbitrary code into the argument of the
424<tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
425tool) is run with the '<tt>-debug</tt>' command line argument:</p>
426
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]427<div class="doc_code">
428<pre>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]429DOUT &lt;&lt; "I am here!\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]430</pre>
431</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]432
433<p>Then you can run your pass like this:</p>
434
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]435<div class="doc_code">
436<pre>
437$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]438<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]439$ opt &lt; a.bc &gt; /dev/null -mypass -debug
440I am here!
441</pre>
442</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]443
444<p>Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you
445to not have to create "yet another" command line option for the debug output for
446your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
447so they do not cause a performance impact at all (for the same reason, they
448should also not contain side-effects!).</p>
449
450<p>One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
451enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
452"<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
453program hasn't been started yet, you can always just run it with
454<tt>-debug</tt>.</p>
455
456</div>
457
458<!-- _______________________________________________________________________ -->
459<div class="doc_subsubsection">
Chris Lattnerc9151082005-04-26 22:57:07 +0000[diff] [blame]460 <a name="DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt> and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]461 the <tt>-debug-only</tt> option</a>
462</div>
463
464<div class="doc_text">
465
466<p>Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
467just turns on <b>too much</b> information (such as when working on the code
468generator). If you want to enable debug information with more fine-grained
469control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
470option as follows:</p>
471
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]472<div class="doc_code">
473<pre>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]474DOUT &lt;&lt; "No debug type\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]475#undef DEBUG_TYPE
476#define DEBUG_TYPE "foo"
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]477DOUT &lt;&lt; "'foo' debug type\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]478#undef DEBUG_TYPE
479#define DEBUG_TYPE "bar"
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]480DOUT &lt;&lt; "'bar' debug type\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]481#undef DEBUG_TYPE
482#define DEBUG_TYPE ""
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]483DOUT &lt;&lt; "No debug type (2)\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]484</pre>
485</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]486
487<p>Then you can run your pass like this:</p>
488
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]489<div class="doc_code">
490<pre>
491$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]492<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]493$ opt &lt; a.bc &gt; /dev/null -mypass -debug
494No debug type
495'foo' debug type
496'bar' debug type
497No debug type (2)
498$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=foo
499'foo' debug type
500$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=bar
501'bar' debug type
502</pre>
503</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]504
505<p>Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of
506a file, to specify the debug type for the entire module (if you do this before
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]507you <tt>#include "llvm/Support/Debug.h"</tt>, you don't have to insert the ugly
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]508<tt>#undef</tt>'s). Also, you should use names more meaningful than "foo" and
509"bar", because there is no system in place to ensure that names do not
510conflict. If two different modules use the same string, they will all be turned
511on when the name is specified. This allows, for example, all debug information
512for instruction scheduling to be enabled with <tt>-debug-type=InstrSched</tt>,
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]513even if the source lives in multiple files.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]514
515</div>
516
517<!-- ======================================================================= -->
518<div class="doc_subsection">
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]519 <a name="Statistic">The <tt>Statistic</tt> class &amp; <tt>-stats</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]520 option</a>
521</div>
522
523<div class="doc_text">
524
525<p>The "<tt><a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]526href="/doxygen/Statistic_8h-source.html">llvm/ADT/Statistic.h</a></tt>" file
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]527provides a class named <tt>Statistic</tt> that is used as a unified way to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]528keep track of what the LLVM compiler is doing and how effective various
529optimizations are. It is useful to see what optimizations are contributing to
530making a particular program run faster.</p>
531
532<p>Often you may run your pass on some big program, and you're interested to see
533how many times it makes a certain transformation. Although you can do this with
534hand inspection, or some ad-hoc method, this is a real pain and not very useful
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]535for big programs. Using the <tt>Statistic</tt> class makes it very easy to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]536keep track of this information, and the calculated information is presented in a
537uniform manner with the rest of the passes being executed.</p>
538
539<p>There are many examples of <tt>Statistic</tt> uses, but the basics of using
540it are as follows:</p>
541
542<ol>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]543 <li><p>Define your statistic like this:</p>
544
545<div class="doc_code">
546<pre>
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]547#define <a href="#DEBUG_TYPE">DEBUG_TYPE</a> "mypassname" <i>// This goes before any #includes.</i>
548STATISTIC(NumXForms, "The # of times I did stuff");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]549</pre>
550</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]551
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]552 <p>The <tt>STATISTIC</tt> macro defines a static variable, whose name is
553 specified by the first argument. The pass name is taken from the DEBUG_TYPE
554 macro, and the description is taken from the second argument. The variable
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]555 defined ("NumXForms" in this case) acts like an unsigned integer.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]556
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]557 <li><p>Whenever you make a transformation, bump the counter:</p>
558
559<div class="doc_code">
560<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]561++NumXForms; // <i>I did stuff!</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]562</pre>
563</div>
564
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]565 </li>
566 </ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]567
568 <p>That's all you have to do. To get '<tt>opt</tt>' to print out the
569 statistics gathered, use the '<tt>-stats</tt>' option:</p>
570
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]571<div class="doc_code">
572<pre>
573$ opt -stats -mypassname &lt; program.bc &gt; /dev/null
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]574<i>... statistics output ...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]575</pre>
576</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]577
Reid Spencer6b6c73e2007-02-09 16:00:28 +0000[diff] [blame]578 <p> When running <tt>opt</tt> on a C file from the SPEC benchmark
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]579suite, it gives a report that looks like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]580
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]581<div class="doc_code">
582<pre>
583 7646 bytecodewriter - Number of normal instructions
584 725 bytecodewriter - Number of oversized instructions
585 129996 bytecodewriter - Number of bytecode bytes written
586 2817 raise - Number of insts DCEd or constprop'd
587 3213 raise - Number of cast-of-self removed
588 5046 raise - Number of expression trees converted
589 75 raise - Number of other getelementptr's formed
590 138 raise - Number of load/store peepholes
591 42 deadtypeelim - Number of unused typenames removed from symtab
592 392 funcresolve - Number of varargs functions resolved
593 27 globaldce - Number of global variables removed
594 2 adce - Number of basic blocks removed
595 134 cee - Number of branches revectored
596 49 cee - Number of setcc instruction eliminated
597 532 gcse - Number of loads removed
598 2919 gcse - Number of instructions removed
599 86 indvars - Number of canonical indvars added
600 87 indvars - Number of aux indvars removed
601 25 instcombine - Number of dead inst eliminate
602 434 instcombine - Number of insts combined
603 248 licm - Number of load insts hoisted
604 1298 licm - Number of insts hoisted to a loop pre-header
605 3 licm - Number of insts hoisted to multiple loop preds (bad, no loop pre-header)
606 75 mem2reg - Number of alloca's promoted
607 1444 cfgsimplify - Number of blocks simplified
608</pre>
609</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]610
611<p>Obviously, with so many optimizations, having a unified framework for this
612stuff is very nice. Making your pass fit well into the framework makes it more
613maintainable and useful.</p>
614
615</div>
616
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]617<!-- ======================================================================= -->
618<div class="doc_subsection">
619 <a name="ViewGraph">Viewing graphs while debugging code</a>
620</div>
621
622<div class="doc_text">
623
624<p>Several of the important data structures in LLVM are graphs: for example
625CFGs made out of LLVM <a href="#BasicBlock">BasicBlock</a>s, CFGs made out of
626LLVM <a href="CodeGenerator.html#machinebasicblock">MachineBasicBlock</a>s, and
627<a href="CodeGenerator.html#selectiondag_intro">Instruction Selection
628DAGs</a>. In many cases, while debugging various parts of the compiler, it is
629nice to instantly visualize these graphs.</p>
630
631<p>LLVM provides several callbacks that are available in a debug build to do
632exactly that. If you call the <tt>Function::viewCFG()</tt> method, for example,
633the current LLVM tool will pop up a window containing the CFG for the function
634where each basic block is a node in the graph, and each node contains the
635instructions in the block. Similarly, there also exists
636<tt>Function::viewCFGOnly()</tt> (does not include the instructions), the
637<tt>MachineFunction::viewCFG()</tt> and <tt>MachineFunction::viewCFGOnly()</tt>,
638and the <tt>SelectionDAG::viewGraph()</tt> methods. Within GDB, for example,
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]639you can usually use something like <tt>call DAG.viewGraph()</tt> to pop
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]640up a window. Alternatively, you can sprinkle calls to these functions in your
641code in places you want to debug.</p>
642
643<p>Getting this to work requires a small amount of configuration. On Unix
644systems with X11, install the <a href="http://www.graphviz.org">graphviz</a>
645toolkit, and make sure 'dot' and 'gv' are in your path. If you are running on
646Mac OS/X, download and install the Mac OS/X <a
647href="http://www.pixelglow.com/graphviz/">Graphviz program</a>, and add
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]648<tt>/Applications/Graphviz.app/Contents/MacOS/</tt> (or wherever you install
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]649it) to your path. Once in your system and path are set up, rerun the LLVM
650configure script and rebuild LLVM to enable this functionality.</p>
651
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]652<p><tt>SelectionDAG</tt> has been extended to make it easier to locate
653<i>interesting</i> nodes in large complex graphs. From gdb, if you
654<tt>call DAG.setGraphColor(<i>node</i>, "<i>color</i>")</tt>, then the
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]655next <tt>call DAG.viewGraph()</tt> would highlight the node in the
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]656specified color (choices of colors can be found at <a
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]657href="http://www.graphviz.org/doc/info/colors.html">colors</a>.) More
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]658complex node attributes can be provided with <tt>call
659DAG.setGraphAttrs(<i>node</i>, "<i>attributes</i>")</tt> (choices can be
660found at <a href="http://www.graphviz.org/doc/info/attrs.html">Graph
661Attributes</a>.) If you want to restart and clear all the current graph
662attributes, then you can <tt>call DAG.clearGraphAttrs()</tt>. </p>
663
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]664</div>
665
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]666<!-- *********************************************************************** -->
667<div class="doc_section">
668 <a name="datastructure">Picking the Right Data Structure for a Task</a>
669</div>
670<!-- *********************************************************************** -->
671
672<div class="doc_text">
673
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]674<p>LLVM has a plethora of data structures in the <tt>llvm/ADT/</tt> directory,
675 and we commonly use STL data structures. This section describes the trade-offs
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]676 you should consider when you pick one.</p>
677
678<p>
679The first step is a choose your own adventure: do you want a sequential
680container, a set-like container, or a map-like container? The most important
681thing when choosing a container is the algorithmic properties of how you plan to
682access the container. Based on that, you should use:</p>
683
684<ul>
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]685<li>a <a href="#ds_map">map-like</a> container if you need efficient look-up
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]686 of an value based on another value. Map-like containers also support
687 efficient queries for containment (whether a key is in the map). Map-like
688 containers generally do not support efficient reverse mapping (values to
689 keys). If you need that, use two maps. Some map-like containers also
690 support efficient iteration through the keys in sorted order. Map-like
691 containers are the most expensive sort, only use them if you need one of
692 these capabilities.</li>
693
694<li>a <a href="#ds_set">set-like</a> container if you need to put a bunch of
695 stuff into a container that automatically eliminates duplicates. Some
696 set-like containers support efficient iteration through the elements in
697 sorted order. Set-like containers are more expensive than sequential
698 containers.
699</li>
700
701<li>a <a href="#ds_sequential">sequential</a> container provides
702 the most efficient way to add elements and keeps track of the order they are
703 added to the collection. They permit duplicates and support efficient
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]704 iteration, but do not support efficient look-up based on a key.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]705</li>
706
707</ul>
708
709<p>
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]710Once the proper category of container is determined, you can fine tune the
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]711memory use, constant factors, and cache behaviors of access by intelligently
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]712picking a member of the category. Note that constant factors and cache behavior
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]713can be a big deal. If you have a vector that usually only contains a few
714elements (but could contain many), for example, it's much better to use
715<a href="#dss_smallvector">SmallVector</a> than <a href="#dss_vector">vector</a>
716. Doing so avoids (relatively) expensive malloc/free calls, which dwarf the
717cost of adding the elements to the container. </p>
718
719</div>
720
721<!-- ======================================================================= -->
722<div class="doc_subsection">
723 <a name="ds_sequential">Sequential Containers (std::vector, std::list, etc)</a>
724</div>
725
726<div class="doc_text">
727There are a variety of sequential containers available for you, based on your
728needs. Pick the first in this section that will do what you want.
729</div>
730
731<!-- _______________________________________________________________________ -->
732<div class="doc_subsubsection">
733 <a name="dss_fixedarrays">Fixed Size Arrays</a>
734</div>
735
736<div class="doc_text">
737<p>Fixed size arrays are very simple and very fast. They are good if you know
738exactly how many elements you have, or you have a (low) upper bound on how many
739you have.</p>
740</div>
741
742<!-- _______________________________________________________________________ -->
743<div class="doc_subsubsection">
744 <a name="dss_heaparrays">Heap Allocated Arrays</a>
745</div>
746
747<div class="doc_text">
748<p>Heap allocated arrays (new[] + delete[]) are also simple. They are good if
749the number of elements is variable, if you know how many elements you will need
750before the array is allocated, and if the array is usually large (if not,
751consider a <a href="#dss_smallvector">SmallVector</a>). The cost of a heap
752allocated array is the cost of the new/delete (aka malloc/free). Also note that
753if you are allocating an array of a type with a constructor, the constructor and
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]754destructors will be run for every element in the array (re-sizable vectors only
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]755construct those elements actually used).</p>
756</div>
757
758<!-- _______________________________________________________________________ -->
759<div class="doc_subsubsection">
760 <a name="dss_smallvector">"llvm/ADT/SmallVector.h"</a>
761</div>
762
763<div class="doc_text">
764<p><tt>SmallVector&lt;Type, N&gt;</tt> is a simple class that looks and smells
765just like <tt>vector&lt;Type&gt;</tt>:
766it supports efficient iteration, lays out elements in memory order (so you can
767do pointer arithmetic between elements), supports efficient push_back/pop_back
768operations, supports efficient random access to its elements, etc.</p>
769
770<p>The advantage of SmallVector is that it allocates space for
771some number of elements (N) <b>in the object itself</b>. Because of this, if
772the SmallVector is dynamically smaller than N, no malloc is performed. This can
773be a big win in cases where the malloc/free call is far more expensive than the
774code that fiddles around with the elements.</p>
775
776<p>This is good for vectors that are "usually small" (e.g. the number of
777predecessors/successors of a block is usually less than 8). On the other hand,
778this makes the size of the SmallVector itself large, so you don't want to
779allocate lots of them (doing so will waste a lot of space). As such,
780SmallVectors are most useful when on the stack.</p>
781
782<p>SmallVector also provides a nice portable and efficient replacement for
783<tt>alloca</tt>.</p>
784
785</div>
786
787<!-- _______________________________________________________________________ -->
788<div class="doc_subsubsection">
789 <a name="dss_vector">&lt;vector&gt;</a>
790</div>
791
792<div class="doc_text">
793<p>
794std::vector is well loved and respected. It is useful when SmallVector isn't:
795when the size of the vector is often large (thus the small optimization will
796rarely be a benefit) or if you will be allocating many instances of the vector
797itself (which would waste space for elements that aren't in the container).
798vector is also useful when interfacing with code that expects vectors :).
799</p>
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]800
801<p>One worthwhile note about std::vector: avoid code like this:</p>
802
803<div class="doc_code">
804<pre>
805for ( ... ) {
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]806 std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]807 use V;
808}
809</pre>
810</div>
811
812<p>Instead, write this as:</p>
813
814<div class="doc_code">
815<pre>
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]816std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]817for ( ... ) {
818 use V;
819 V.clear();
820}
821</pre>
822</div>
823
824<p>Doing so will save (at least) one heap allocation and free per iteration of
825the loop.</p>
826
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]827</div>
828
829<!-- _______________________________________________________________________ -->
830<div class="doc_subsubsection">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]831 <a name="dss_deque">&lt;deque&gt;</a>
832</div>
833
834<div class="doc_text">
835<p>std::deque is, in some senses, a generalized version of std::vector. Like
836std::vector, it provides constant time random access and other similar
837properties, but it also provides efficient access to the front of the list. It
838does not guarantee continuity of elements within memory.</p>
839
840<p>In exchange for this extra flexibility, std::deque has significantly higher
841constant factor costs than std::vector. If possible, use std::vector or
842something cheaper.</p>
843</div>
844
845<!-- _______________________________________________________________________ -->
846<div class="doc_subsubsection">
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]847 <a name="dss_list">&lt;list&gt;</a>
848</div>
849
850<div class="doc_text">
851<p>std::list is an extremely inefficient class that is rarely useful.
852It performs a heap allocation for every element inserted into it, thus having an
853extremely high constant factor, particularly for small data types. std::list
854also only supports bidirectional iteration, not random access iteration.</p>
855
856<p>In exchange for this high cost, std::list supports efficient access to both
857ends of the list (like std::deque, but unlike std::vector or SmallVector). In
858addition, the iterator invalidation characteristics of std::list are stronger
859than that of a vector class: inserting or removing an element into the list does
860not invalidate iterator or pointers to other elements in the list.</p>
861</div>
862
863<!-- _______________________________________________________________________ -->
864<div class="doc_subsubsection">
865 <a name="dss_ilist">llvm/ADT/ilist</a>
866</div>
867
868<div class="doc_text">
869<p><tt>ilist&lt;T&gt;</tt> implements an 'intrusive' doubly-linked list. It is
870intrusive, because it requires the element to store and provide access to the
871prev/next pointers for the list.</p>
872
873<p>ilist has the same drawbacks as std::list, and additionally requires an
874ilist_traits implementation for the element type, but it provides some novel
875characteristics. In particular, it can efficiently store polymorphic objects,
876the traits class is informed when an element is inserted or removed from the
877list, and ilists are guaranteed to support a constant-time splice operation.
878</p>
879
880<p>These properties are exactly what we want for things like Instructions and
881basic blocks, which is why these are implemented with ilists.</p>
882</div>
883
884<!-- _______________________________________________________________________ -->
885<div class="doc_subsubsection">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]886 <a name="dss_other">Other Sequential Container options</a>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]887</div>
888
889<div class="doc_text">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]890<p>Other STL containers are available, such as std::string.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]891
892<p>There are also various STL adapter classes such as std::queue,
893std::priority_queue, std::stack, etc. These provide simplified access to an
894underlying container but don't affect the cost of the container itself.</p>
895
896</div>
897
898
899<!-- ======================================================================= -->
900<div class="doc_subsection">
901 <a name="ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
902</div>
903
904<div class="doc_text">
905
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]906<p>Set-like containers are useful when you need to canonicalize multiple values
907into a single representation. There are several different choices for how to do
908this, providing various trade-offs.</p>
909
910</div>
911
912
913<!-- _______________________________________________________________________ -->
914<div class="doc_subsubsection">
915 <a name="dss_sortedvectorset">A sorted 'vector'</a>
916</div>
917
918<div class="doc_text">
919
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]920<p>If you intend to insert a lot of elements, then do a lot of queries, a
921great approach is to use a vector (or other sequential container) with
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]922std::sort+std::unique to remove duplicates. This approach works really well if
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]923your usage pattern has these two distinct phases (insert then query), and can be
924coupled with a good choice of <a href="#ds_sequential">sequential container</a>.
925</p>
926
927<p>
928This combination provides the several nice properties: the result data is
929contiguous in memory (good for cache locality), has few allocations, is easy to
930address (iterators in the final vector are just indices or pointers), and can be
931efficiently queried with a standard binary or radix search.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]932
933</div>
934
935<!-- _______________________________________________________________________ -->
936<div class="doc_subsubsection">
937 <a name="dss_smallset">"llvm/ADT/SmallSet.h"</a>
938</div>
939
940<div class="doc_text">
941
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]942<p>If you have a set-like data structure that is usually small and whose elements
Chris Lattner4ddfac12007-02-03 07:59:51 +0000[diff] [blame]943are reasonably small, a <tt>SmallSet&lt;Type, N&gt;</tt> is a good choice. This set
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]944has space for N elements in place (thus, if the set is dynamically smaller than
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]945N, no malloc traffic is required) and accesses them with a simple linear search.
946When the set grows beyond 'N' elements, it allocates a more expensive representation that
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]947guarantees efficient access (for most types, it falls back to std::set, but for
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]948pointers it uses something far better, <a
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]949href="#dss_smallptrset">SmallPtrSet</a>).</p>
950
951<p>The magic of this class is that it handles small sets extremely efficiently,
952but gracefully handles extremely large sets without loss of efficiency. The
953drawback is that the interface is quite small: it supports insertion, queries
954and erasing, but does not support iteration.</p>
955
956</div>
957
958<!-- _______________________________________________________________________ -->
959<div class="doc_subsubsection">
960 <a name="dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a>
961</div>
962
963<div class="doc_text">
964
965<p>SmallPtrSet has all the advantages of SmallSet (and a SmallSet of pointers is
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]966transparently implemented with a SmallPtrSet), but also supports iterators. If
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]967more than 'N' insertions are performed, a single quadratically
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]968probed hash table is allocated and grows as needed, providing extremely
969efficient access (constant time insertion/deleting/queries with low constant
970factors) and is very stingy with malloc traffic.</p>
971
972<p>Note that, unlike std::set, the iterators of SmallPtrSet are invalidated
973whenever an insertion occurs. Also, the values visited by the iterators are not
974visited in sorted order.</p>
975
976</div>
977
978<!-- _______________________________________________________________________ -->
979<div class="doc_subsubsection">
980 <a name="dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a>
981</div>
982
983<div class="doc_text">
984
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]985<p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]986FoldingSet is an aggregate class that is really good at uniquing
987expensive-to-create or polymorphic objects. It is a combination of a chained
988hash table with intrusive links (uniqued objects are required to inherit from
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]989FoldingSetNode) that uses <a href="#dss_smallvector">SmallVector</a> as part of
990its ID process.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]991
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]992<p>Consider a case where you want to implement a "getOrCreateFoo" method for
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]993a complex object (for example, a node in the code generator). The client has a
994description of *what* it wants to generate (it knows the opcode and all the
995operands), but we don't want to 'new' a node, then try inserting it into a set
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]996only to find out it already exists, at which point we would have to delete it
997and return the node that already exists.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]998</p>
999
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1000<p>To support this style of client, FoldingSet perform a query with a
1001FoldingSetNodeID (which wraps SmallVector) that can be used to describe the
1002element that we want to query for. The query either returns the element
1003matching the ID or it returns an opaque ID that indicates where insertion should
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1004take place. Construction of the ID usually does not require heap traffic.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1005
1006<p>Because FoldingSet uses intrusive links, it can support polymorphic objects
1007in the set (for example, you can have SDNode instances mixed with LoadSDNodes).
1008Because the elements are individually allocated, pointers to the elements are
1009stable: inserting or removing elements does not invalidate any pointers to other
1010elements.
1011</p>
1012
1013</div>
1014
1015<!-- _______________________________________________________________________ -->
1016<div class="doc_subsubsection">
1017 <a name="dss_set">&lt;set&gt;</a>
1018</div>
1019
1020<div class="doc_text">
1021
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1022<p><tt>std::set</tt> is a reasonable all-around set class, which is decent at
1023many things but great at nothing. std::set allocates memory for each element
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1024inserted (thus it is very malloc intensive) and typically stores three pointers
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1025per element in the set (thus adding a large amount of per-element space
1026overhead). It offers guaranteed log(n) performance, which is not particularly
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1027fast from a complexity standpoint (particularly if the elements of the set are
1028expensive to compare, like strings), and has extremely high constant factors for
1029lookup, insertion and removal.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1030
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1031<p>The advantages of std::set are that its iterators are stable (deleting or
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1032inserting an element from the set does not affect iterators or pointers to other
1033elements) and that iteration over the set is guaranteed to be in sorted order.
1034If the elements in the set are large, then the relative overhead of the pointers
1035and malloc traffic is not a big deal, but if the elements of the set are small,
1036std::set is almost never a good choice.</p>
1037
1038</div>
1039
1040<!-- _______________________________________________________________________ -->
1041<div class="doc_subsubsection">
1042 <a name="dss_setvector">"llvm/ADT/SetVector.h"</a>
1043</div>
1044
1045<div class="doc_text">
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1046<p>LLVM's SetVector&lt;Type&gt; is an adapter class that combines your choice of
1047a set-like container along with a <a href="#ds_sequential">Sequential
1048Container</a>. The important property
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1049that this provides is efficient insertion with uniquing (duplicate elements are
1050ignored) with iteration support. It implements this by inserting elements into
1051both a set-like container and the sequential container, using the set-like
1052container for uniquing and the sequential container for iteration.
1053</p>
1054
1055<p>The difference between SetVector and other sets is that the order of
1056iteration is guaranteed to match the order of insertion into the SetVector.
1057This property is really important for things like sets of pointers. Because
1058pointer values are non-deterministic (e.g. vary across runs of the program on
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1059different machines), iterating over the pointers in the set will
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1060not be in a well-defined order.</p>
1061
1062<p>
1063The drawback of SetVector is that it requires twice as much space as a normal
1064set and has the sum of constant factors from the set-like container and the
1065sequential container that it uses. Use it *only* if you need to iterate over
1066the elements in a deterministic order. SetVector is also expensive to delete
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1067elements out of (linear time), unless you use it's "pop_back" method, which is
1068faster.
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1069</p>
1070
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1071<p>SetVector is an adapter class that defaults to using std::vector and std::set
1072for the underlying containers, so it is quite expensive. However,
1073<tt>"llvm/ADT/SetVector.h"</tt> also provides a SmallSetVector class, which
1074defaults to using a SmallVector and SmallSet of a specified size. If you use
1075this, and if your sets are dynamically smaller than N, you will save a lot of
1076heap traffic.</p>
1077
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1078</div>
1079
1080<!-- _______________________________________________________________________ -->
1081<div class="doc_subsubsection">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1082 <a name="dss_uniquevector">"llvm/ADT/UniqueVector.h"</a>
1083</div>
1084
1085<div class="doc_text">
1086
1087<p>
1088UniqueVector is similar to <a href="#dss_setvector">SetVector</a>, but it
1089retains a unique ID for each element inserted into the set. It internally
1090contains a map and a vector, and it assigns a unique ID for each value inserted
1091into the set.</p>
1092
1093<p>UniqueVector is very expensive: its cost is the sum of the cost of
1094maintaining both the map and vector, it has high complexity, high constant
1095factors, and produces a lot of malloc traffic. It should be avoided.</p>
1096
1097</div>
1098
1099
1100<!-- _______________________________________________________________________ -->
1101<div class="doc_subsubsection">
1102 <a name="dss_otherset">Other Set-Like Container Options</a>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1103</div>
1104
1105<div class="doc_text">
1106
1107<p>
1108The STL provides several other options, such as std::multiset and the various
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1109"hash_set" like containers (whether from C++ TR1 or from the SGI library).</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1110
1111<p>std::multiset is useful if you're not interested in elimination of
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1112duplicates, but has all the drawbacks of std::set. A sorted vector (where you
1113don't delete duplicate entries) or some other approach is almost always
1114better.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1115
1116<p>The various hash_set implementations (exposed portably by
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1117"llvm/ADT/hash_set") is a simple chained hashtable. This algorithm is as malloc
1118intensive as std::set (performing an allocation for each element inserted,
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1119thus having really high constant factors) but (usually) provides O(1)
1120insertion/deletion of elements. This can be useful if your elements are large
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1121(thus making the constant-factor cost relatively low) or if comparisons are
1122expensive. Element iteration does not visit elements in a useful order.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1123
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1124</div>
1125
1126<!-- ======================================================================= -->
1127<div class="doc_subsection">
1128 <a name="ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
1129</div>
1130
1131<div class="doc_text">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1132Map-like containers are useful when you want to associate data to a key. As
1133usual, there are a lot of different ways to do this. :)
1134</div>
1135
1136<!-- _______________________________________________________________________ -->
1137<div class="doc_subsubsection">
1138 <a name="dss_sortedvectormap">A sorted 'vector'</a>
1139</div>
1140
1141<div class="doc_text">
1142
1143<p>
1144If your usage pattern follows a strict insert-then-query approach, you can
1145trivially use the same approach as <a href="#dss_sortedvectorset">sorted vectors
1146for set-like containers</a>. The only difference is that your query function
1147(which uses std::lower_bound to get efficient log(n) lookup) should only compare
1148the key, not both the key and value. This yields the same advantages as sorted
1149vectors for sets.
1150</p>
1151</div>
1152
1153<!-- _______________________________________________________________________ -->
1154<div class="doc_subsubsection">
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1155 <a name="dss_stringmap">"llvm/ADT/StringMap.h"</a>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1156</div>
1157
1158<div class="doc_text">
1159
1160<p>
1161Strings are commonly used as keys in maps, and they are difficult to support
1162efficiently: they are variable length, inefficient to hash and compare when
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1163long, expensive to copy, etc. StringMap is a specialized container designed to
1164cope with these issues. It supports mapping an arbitrary range of bytes to an
1165arbitrary other object.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1166
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1167<p>The StringMap implementation uses a quadratically-probed hash table, where
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1168the buckets store a pointer to the heap allocated entries (and some other
1169stuff). The entries in the map must be heap allocated because the strings are
1170variable length. The string data (key) and the element object (value) are
1171stored in the same allocation with the string data immediately after the element
1172object. This container guarantees the "<tt>(char*)(&amp;Value+1)</tt>" points
1173to the key string for a value.</p>
1174
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1175<p>The StringMap is very fast for several reasons: quadratic probing is very
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1176cache efficient for lookups, the hash value of strings in buckets is not
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1177recomputed when lookup up an element, StringMap rarely has to touch the
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1178memory for unrelated objects when looking up a value (even when hash collisions
1179happen), hash table growth does not recompute the hash values for strings
1180already in the table, and each pair in the map is store in a single allocation
1181(the string data is stored in the same allocation as the Value of a pair).</p>
1182
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1183<p>StringMap also provides query methods that take byte ranges, so it only ever
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1184copies a string if a value is inserted into the table.</p>
1185</div>
1186
1187<!-- _______________________________________________________________________ -->
1188<div class="doc_subsubsection">
1189 <a name="dss_indexedmap">"llvm/ADT/IndexedMap.h"</a>
1190</div>
1191
1192<div class="doc_text">
1193<p>
1194IndexedMap is a specialized container for mapping small dense integers (or
1195values that can be mapped to small dense integers) to some other type. It is
1196internally implemented as a vector with a mapping function that maps the keys to
1197the dense integer range.
1198</p>
1199
1200<p>
1201This is useful for cases like virtual registers in the LLVM code generator: they
1202have a dense mapping that is offset by a compile-time constant (the first
1203virtual register ID).</p>
1204
1205</div>
1206
1207<!-- _______________________________________________________________________ -->
1208<div class="doc_subsubsection">
1209 <a name="dss_densemap">"llvm/ADT/DenseMap.h"</a>
1210</div>
1211
1212<div class="doc_text">
1213
1214<p>
1215DenseMap is a simple quadratically probed hash table. It excels at supporting
1216small keys and values: it uses a single allocation to hold all of the pairs that
1217are currently inserted in the map. DenseMap is a great way to map pointers to
1218pointers, or map other small types to each other.
1219</p>
1220
1221<p>
1222There are several aspects of DenseMap that you should be aware of, however. The
1223iterators in a densemap are invalidated whenever an insertion occurs, unlike
1224map. Also, because DenseMap allocates space for a large number of key/value
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1225pairs (it starts with 64 by default), it will waste a lot of space if your keys
1226or values are large. Finally, you must implement a partial specialization of
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1227DenseMapKeyInfo for the key that you want, if it isn't already supported. This
1228is required to tell DenseMap about two special marker values (which can never be
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1229inserted into the map) that it needs internally.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1230
1231</div>
1232
1233<!-- _______________________________________________________________________ -->
1234<div class="doc_subsubsection">
1235 <a name="dss_map">&lt;map&gt;</a>
1236</div>
1237
1238<div class="doc_text">
1239
1240<p>
1241std::map has similar characteristics to <a href="#dss_set">std::set</a>: it uses
1242a single allocation per pair inserted into the map, it offers log(n) lookup with
1243an extremely large constant factor, imposes a space penalty of 3 pointers per
1244pair in the map, etc.</p>
1245
1246<p>std::map is most useful when your keys or values are very large, if you need
1247to iterate over the collection in sorted order, or if you need stable iterators
1248into the map (i.e. they don't get invalidated if an insertion or deletion of
1249another element takes place).</p>
1250
1251</div>
1252
1253<!-- _______________________________________________________________________ -->
1254<div class="doc_subsubsection">
1255 <a name="dss_othermap">Other Map-Like Container Options</a>
1256</div>
1257
1258<div class="doc_text">
1259
1260<p>
1261The STL provides several other options, such as std::multimap and the various
1262"hash_map" like containers (whether from C++ TR1 or from the SGI library).</p>
1263
1264<p>std::multimap is useful if you want to map a key to multiple values, but has
1265all the drawbacks of std::map. A sorted vector or some other approach is almost
1266always better.</p>
1267
1268<p>The various hash_map implementations (exposed portably by
1269"llvm/ADT/hash_map") are simple chained hash tables. This algorithm is as
1270malloc intensive as std::map (performing an allocation for each element
1271inserted, thus having really high constant factors) but (usually) provides O(1)
1272insertion/deletion of elements. This can be useful if your elements are large
1273(thus making the constant-factor cost relatively low) or if comparisons are
1274expensive. Element iteration does not visit elements in a useful order.</p>
1275
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1276</div>
1277
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]1278
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1279<!-- *********************************************************************** -->
1280<div class="doc_section">
1281 <a name="common">Helpful Hints for Common Operations</a>
1282</div>
1283<!-- *********************************************************************** -->
1284
1285<div class="doc_text">
1286
1287<p>This section describes how to perform some very simple transformations of
1288LLVM code. This is meant to give examples of common idioms used, showing the
1289practical side of LLVM transformations. <p> Because this is a "how-to" section,
1290you should also read about the main classes that you will be working with. The
1291<a href="#coreclasses">Core LLVM Class Hierarchy Reference</a> contains details
1292and descriptions of the main classes that you should know about.</p>
1293
1294</div>
1295
1296<!-- NOTE: this section should be heavy on example code -->
1297<!-- ======================================================================= -->
1298<div class="doc_subsection">
1299 <a name="inspection">Basic Inspection and Traversal Routines</a>
1300</div>
1301
1302<div class="doc_text">
1303
1304<p>The LLVM compiler infrastructure have many different data structures that may
1305be traversed. Following the example of the C++ standard template library, the
1306techniques used to traverse these various data structures are all basically the
1307same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
1308method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
1309function returns an iterator pointing to one past the last valid element of the
1310sequence, and there is some <tt>XXXiterator</tt> data type that is common
1311between the two operations.</p>
1312
1313<p>Because the pattern for iteration is common across many different aspects of
1314the program representation, the standard template library algorithms may be used
1315on them, and it is easier to remember how to iterate. First we show a few common
1316examples of the data structures that need to be traversed. Other data
1317structures are traversed in very similar ways.</p>
1318
1319</div>
1320
1321<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1322<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1323 <a name="iterate_function">Iterating over the </a><a
1324 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
1325 href="#Function"><tt>Function</tt></a>
1326</div>
1327
1328<div class="doc_text">
1329
1330<p>It's quite common to have a <tt>Function</tt> instance that you'd like to
1331transform in some way; in particular, you'd like to manipulate its
1332<tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over all of
1333the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>. The following is
1334an example that prints the name of a <tt>BasicBlock</tt> and the number of
1335<tt>Instruction</tt>s it contains:</p>
1336
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1337<div class="doc_code">
1338<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1339// <i>func is a pointer to a Function instance</i>
1340for (Function::iterator i = func-&gt;begin(), e = func-&gt;end(); i != e; ++i)
1341 // <i>Print out the name of the basic block if it has one, and then the</i>
1342 // <i>number of instructions that it contains</i>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1343 llvm::cerr &lt;&lt; "Basic block (name=" &lt;&lt; i-&gt;getName() &lt;&lt; ") has "
1344 &lt;&lt; i-&gt;size() &lt;&lt; " instructions.\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1345</pre>
1346</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1347
1348<p>Note that i can be used as if it were a pointer for the purposes of
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]1349invoking member functions of the <tt>Instruction</tt> class. This is
1350because the indirection operator is overloaded for the iterator
Chris Lattner7496ec52003-08-05 22:54:23 +0000[diff] [blame]1351classes. In the above code, the expression <tt>i-&gt;size()</tt> is
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1352exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.</p>
1353
1354</div>
1355
1356<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1357<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1358 <a name="iterate_basicblock">Iterating over the </a><a
1359 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1360 href="#BasicBlock"><tt>BasicBlock</tt></a>
1361</div>
1362
1363<div class="doc_text">
1364
1365<p>Just like when dealing with <tt>BasicBlock</tt>s in <tt>Function</tt>s, it's
1366easy to iterate over the individual instructions that make up
1367<tt>BasicBlock</tt>s. Here's a code snippet that prints out each instruction in
1368a <tt>BasicBlock</tt>:</p>
1369
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1370<div class="doc_code">
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1371<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1372// <i>blk is a pointer to a BasicBlock instance</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1373for (BasicBlock::iterator i = blk-&gt;begin(), e = blk-&gt;end(); i != e; ++i)
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1374 // <i>The next statement works since operator&lt;&lt;(ostream&amp;,...)</i>
1375 // <i>is overloaded for Instruction&amp;</i>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1376 llvm::cerr &lt;&lt; *i &lt;&lt; "\n";
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1377</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1378</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1379
1380<p>However, this isn't really the best way to print out the contents of a
1381<tt>BasicBlock</tt>! Since the ostream operators are overloaded for virtually
1382anything you'll care about, you could have just invoked the print routine on the
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1383basic block itself: <tt>llvm::cerr &lt;&lt; *blk &lt;&lt; "\n";</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1384
1385</div>
1386
1387<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1388<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1389 <a name="iterate_institer">Iterating over the </a><a
1390 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1391 href="#Function"><tt>Function</tt></a>
1392</div>
1393
1394<div class="doc_text">
1395
1396<p>If you're finding that you commonly iterate over a <tt>Function</tt>'s
1397<tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s <tt>Instruction</tt>s,
1398<tt>InstIterator</tt> should be used instead. You'll need to include <a
1399href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>,
1400and then instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1401small example that shows how to dump all instructions in a function to the standard error stream:<p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1402
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1403<div class="doc_code">
1404<pre>
1405#include "<a href="/doxygen/InstIterator_8h-source.html">llvm/Support/InstIterator.h</a>"
1406
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1407// <i>F is a pointer to a Function instance</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1408for (inst_iterator i = inst_begin(F), e = inst_end(F); i != e; ++i)
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1409 llvm::cerr &lt;&lt; *i &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1410</pre>
1411</div>
1412
1413<p>Easy, isn't it? You can also use <tt>InstIterator</tt>s to fill a
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1414work list with its initial contents. For example, if you wanted to
1415initialize a work list to contain all instructions in a <tt>Function</tt>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1416F, all you would need to do is something like:</p>
1417
1418<div class="doc_code">
1419<pre>
1420std::set&lt;Instruction*&gt; worklist;
1421worklist.insert(inst_begin(F), inst_end(F));
1422</pre>
1423</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1424
1425<p>The STL set <tt>worklist</tt> would now contain all instructions in the
1426<tt>Function</tt> pointed to by F.</p>
1427
1428</div>
1429
1430<!-- _______________________________________________________________________ -->
1431<div class="doc_subsubsection">
1432 <a name="iterate_convert">Turning an iterator into a class pointer (and
1433 vice-versa)</a>
1434</div>
1435
1436<div class="doc_text">
1437
1438<p>Sometimes, it'll be useful to grab a reference (or pointer) to a class
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]1439instance when all you've got at hand is an iterator. Well, extracting
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1440a reference or a pointer from an iterator is very straight-forward.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1441Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and <tt>j</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1442is a <tt>BasicBlock::const_iterator</tt>:</p>
1443
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1444<div class="doc_code">
1445<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1446Instruction&amp; inst = *i; // <i>Grab reference to instruction reference</i>
1447Instruction* pinst = &amp;*i; // <i>Grab pointer to instruction reference</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1448const Instruction&amp; inst = *j;
1449</pre>
1450</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1451
1452<p>However, the iterators you'll be working with in the LLVM framework are
1453special: they will automatically convert to a ptr-to-instance type whenever they
1454need to. Instead of dereferencing the iterator and then taking the address of
1455the result, you can simply assign the iterator to the proper pointer type and
1456you get the dereference and address-of operation as a result of the assignment
1457(behind the scenes, this is a result of overloading casting mechanisms). Thus
1458the last line of the last example,</p>
1459
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1460<div class="doc_code">
1461<pre>
1462Instruction* pinst = &amp;*i;
1463</pre>
1464</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1465
1466<p>is semantically equivalent to</p>
1467
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1468<div class="doc_code">
1469<pre>
1470Instruction* pinst = i;
1471</pre>
1472</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1473
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1474<p>It's also possible to turn a class pointer into the corresponding iterator,
1475and this is a constant time operation (very efficient). The following code
1476snippet illustrates use of the conversion constructors provided by LLVM
1477iterators. By using these, you can explicitly grab the iterator of something
1478without actually obtaining it via iteration over some structure:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1479
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1480<div class="doc_code">
1481<pre>
1482void printNextInstruction(Instruction* inst) {
1483 BasicBlock::iterator it(inst);
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1484 ++it; // <i>After this line, it refers to the instruction after *inst</i>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1485 if (it != inst-&gt;getParent()-&gt;end()) llvm::cerr &lt;&lt; *it &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1486}
1487</pre>
1488</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1489
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1490</div>
1491
1492<!--_______________________________________________________________________-->
1493<div class="doc_subsubsection">
1494 <a name="iterate_complex">Finding call sites: a slightly more complex
1495 example</a>
1496</div>
1497
1498<div class="doc_text">
1499
1500<p>Say that you're writing a FunctionPass and would like to count all the
1501locations in the entire module (that is, across every <tt>Function</tt>) where a
1502certain function (i.e., some <tt>Function</tt>*) is already in scope. As you'll
1503learn later, you may want to use an <tt>InstVisitor</tt> to accomplish this in a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1504much more straight-forward manner, but this example will allow us to explore how
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1505you'd do it if you didn't have <tt>InstVisitor</tt> around. In pseudo-code, this
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1506is what we want to do:</p>
1507
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1508<div class="doc_code">
1509<pre>
1510initialize callCounter to zero
1511for each Function f in the Module
1512 for each BasicBlock b in f
1513 for each Instruction i in b
1514 if (i is a CallInst and calls the given function)
1515 increment callCounter
1516</pre>
1517</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1518
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1519<p>And the actual code is (remember, because we're writing a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1520<tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply has to
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1521override the <tt>runOnFunction</tt> method):</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1522
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1523<div class="doc_code">
1524<pre>
1525Function* targetFunc = ...;
1526
1527class OurFunctionPass : public FunctionPass {
1528 public:
1529 OurFunctionPass(): callCounter(0) { }
1530
1531 virtual runOnFunction(Function&amp; F) {
1532 for (Function::iterator b = F.begin(), be = F.end(); b != be; ++b) {
1533 for (BasicBlock::iterator i = b-&gt;begin(); ie = b-&gt;end(); i != ie; ++i) {
1534 if (<a href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a>&lt;<a
1535 href="#CallInst">CallInst</a>&gt;(&amp;*i)) {
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1536 // <i>We know we've encountered a call instruction, so we</i>
1537 // <i>need to determine if it's a call to the</i>
1538 // <i>function pointed to by m_func or not</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1539
1540 if (callInst-&gt;getCalledFunction() == targetFunc)
1541 ++callCounter;
1542 }
1543 }
1544 }
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1545 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1546
1547 private:
1548 unsigned callCounter;
1549};
1550</pre>
1551</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1552
1553</div>
1554
Brian Gaekef1972c62003-11-07 19:25:45 +0000[diff] [blame]1555<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1556<div class="doc_subsubsection">
1557 <a name="calls_and_invokes">Treating calls and invokes the same way</a>
1558</div>
1559
1560<div class="doc_text">
1561
1562<p>You may have noticed that the previous example was a bit oversimplified in
1563that it did not deal with call sites generated by 'invoke' instructions. In
1564this, and in other situations, you may find that you want to treat
1565<tt>CallInst</tt>s and <tt>InvokeInst</tt>s the same way, even though their
1566most-specific common base class is <tt>Instruction</tt>, which includes lots of
1567less closely-related things. For these cases, LLVM provides a handy wrapper
1568class called <a
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]1569href="http://llvm.org/doxygen/classllvm_1_1CallSite.html"><tt>CallSite</tt></a>.
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1570It is essentially a wrapper around an <tt>Instruction</tt> pointer, with some
1571methods that provide functionality common to <tt>CallInst</tt>s and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1572<tt>InvokeInst</tt>s.</p>
1573
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1574<p>This class has "value semantics": it should be passed by value, not by
1575reference and it should not be dynamically allocated or deallocated using
1576<tt>operator new</tt> or <tt>operator delete</tt>. It is efficiently copyable,
1577assignable and constructable, with costs equivalents to that of a bare pointer.
1578If you look at its definition, it has only a single pointer member.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1579
1580</div>
1581
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]1582<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1583<div class="doc_subsubsection">
1584 <a name="iterate_chains">Iterating over def-use &amp; use-def chains</a>
1585</div>
1586
1587<div class="doc_text">
1588
1589<p>Frequently, we might have an instance of the <a
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]1590href="/doxygen/classllvm_1_1Value.html">Value Class</a> and we want to
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]1591determine which <tt>User</tt>s use the <tt>Value</tt>. The list of all
1592<tt>User</tt>s of a particular <tt>Value</tt> is called a <i>def-use</i> chain.
1593For example, let's say we have a <tt>Function*</tt> named <tt>F</tt> to a
1594particular function <tt>foo</tt>. Finding all of the instructions that
1595<i>use</i> <tt>foo</tt> is as simple as iterating over the <i>def-use</i> chain
1596of <tt>F</tt>:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1597
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1598<div class="doc_code">
1599<pre>
1600Function* F = ...;
1601
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1602for (Value::use_iterator i = F-&gt;use_begin(), e = F-&gt;use_end(); i != e; ++i)
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1603 if (Instruction *Inst = dyn_cast&lt;Instruction&gt;(*i)) {
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1604 llvm::cerr &lt;&lt; "F is used in instruction:\n";
1605 llvm::cerr &lt;&lt; *Inst &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1606 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1607</pre>
1608</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1609
1610<p>Alternately, it's common to have an instance of the <a
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]1611href="/doxygen/classllvm_1_1User.html">User Class</a> and need to know what
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1612<tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used by a
1613<tt>User</tt> is known as a <i>use-def</i> chain. Instances of class
1614<tt>Instruction</tt> are common <tt>User</tt>s, so we might want to iterate over
1615all of the values that a particular instruction uses (that is, the operands of
1616the particular <tt>Instruction</tt>):</p>
1617
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1618<div class="doc_code">
1619<pre>
1620Instruction* pi = ...;
1621
1622for (User::op_iterator i = pi-&gt;op_begin(), e = pi-&gt;op_end(); i != e; ++i) {
1623 Value* v = *i;
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1624 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1625}
1626</pre>
1627</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1628
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]1629<!--
1630 def-use chains ("finding all users of"): Value::use_begin/use_end
1631 use-def chains ("finding all values used"): User::op_begin/op_end [op=operand]
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1632-->
1633
1634</div>
1635
1636<!-- ======================================================================= -->
1637<div class="doc_subsection">
1638 <a name="simplechanges">Making simple changes</a>
1639</div>
1640
1641<div class="doc_text">
1642
1643<p>There are some primitive transformation operations present in the LLVM
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]1644infrastructure that are worth knowing about. When performing
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1645transformations, it's fairly common to manipulate the contents of basic
1646blocks. This section describes some of the common methods for doing so
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1647and gives example code.</p>
1648
1649</div>
1650
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1651<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1652<div class="doc_subsubsection">
1653 <a name="schanges_creating">Creating and inserting new
1654 <tt>Instruction</tt>s</a>
1655</div>
1656
1657<div class="doc_text">
1658
1659<p><i>Instantiating Instructions</i></p>
1660
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1661<p>Creation of <tt>Instruction</tt>s is straight-forward: simply call the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1662constructor for the kind of instruction to instantiate and provide the necessary
1663parameters. For example, an <tt>AllocaInst</tt> only <i>requires</i> a
1664(const-ptr-to) <tt>Type</tt>. Thus:</p>
1665
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1666<div class="doc_code">
1667<pre>
1668AllocaInst* ai = new AllocaInst(Type::IntTy);
1669</pre>
1670</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1671
1672<p>will create an <tt>AllocaInst</tt> instance that represents the allocation of
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1673one integer in the current stack frame, at run time. Each <tt>Instruction</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1674subclass is likely to have varying default parameters which change the semantics
1675of the instruction, so refer to the <a
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]1676href="/doxygen/classllvm_1_1Instruction.html">doxygen documentation for the subclass of
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1677Instruction</a> that you're interested in instantiating.</p>
1678
1679<p><i>Naming values</i></p>
1680
1681<p>It is very useful to name the values of instructions when you're able to, as
1682this facilitates the debugging of your transformations. If you end up looking
1683at generated LLVM machine code, you definitely want to have logical names
1684associated with the results of instructions! By supplying a value for the
1685<tt>Name</tt> (default) parameter of the <tt>Instruction</tt> constructor, you
1686associate a logical name with the result of the instruction's execution at
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1687run time. For example, say that I'm writing a transformation that dynamically
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1688allocates space for an integer on the stack, and that integer is going to be
1689used as some kind of index by some other code. To accomplish this, I place an
1690<tt>AllocaInst</tt> at the first point in the first <tt>BasicBlock</tt> of some
1691<tt>Function</tt>, and I'm intending to use it within the same
1692<tt>Function</tt>. I might do:</p>
1693
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1694<div class="doc_code">
1695<pre>
1696AllocaInst* pa = new AllocaInst(Type::IntTy, 0, "indexLoc");
1697</pre>
1698</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1699
1700<p>where <tt>indexLoc</tt> is now the logical name of the instruction's
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1701execution value, which is a pointer to an integer on the run time stack.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1702
1703<p><i>Inserting instructions</i></p>
1704
1705<p>There are essentially two ways to insert an <tt>Instruction</tt>
1706into an existing sequence of instructions that form a <tt>BasicBlock</tt>:</p>
1707
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]1708<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1709 <li>Insertion into an explicit instruction list
1710
1711 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within that
1712 <tt>BasicBlock</tt>, and a newly-created instruction we wish to insert
1713 before <tt>*pi</tt>, we do the following: </p>
1714
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1715<div class="doc_code">
1716<pre>
1717BasicBlock *pb = ...;
1718Instruction *pi = ...;
1719Instruction *newInst = new Instruction(...);
1720
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1721pb-&gt;getInstList().insert(pi, newInst); // <i>Inserts newInst before pi in pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1722</pre>
1723</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]1724
1725 <p>Appending to the end of a <tt>BasicBlock</tt> is so common that
1726 the <tt>Instruction</tt> class and <tt>Instruction</tt>-derived
1727 classes provide constructors which take a pointer to a
1728 <tt>BasicBlock</tt> to be appended to. For example code that
1729 looked like: </p>
1730
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1731<div class="doc_code">
1732<pre>
1733BasicBlock *pb = ...;
1734Instruction *newInst = new Instruction(...);
1735
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1736pb-&gt;getInstList().push_back(newInst); // <i>Appends newInst to pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1737</pre>
1738</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]1739
1740 <p>becomes: </p>
1741
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1742<div class="doc_code">
1743<pre>
1744BasicBlock *pb = ...;
1745Instruction *newInst = new Instruction(..., pb);
1746</pre>
1747</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]1748
1749 <p>which is much cleaner, especially if you are creating
1750 long instruction streams.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1751
1752 <li>Insertion into an implicit instruction list
1753
1754 <p><tt>Instruction</tt> instances that are already in <tt>BasicBlock</tt>s
1755 are implicitly associated with an existing instruction list: the instruction
1756 list of the enclosing basic block. Thus, we could have accomplished the same
1757 thing as the above code without being given a <tt>BasicBlock</tt> by doing:
1758 </p>
1759
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1760<div class="doc_code">
1761<pre>
1762Instruction *pi = ...;
1763Instruction *newInst = new Instruction(...);
1764
1765pi-&gt;getParent()-&gt;getInstList().insert(pi, newInst);
1766</pre>
1767</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1768
1769 <p>In fact, this sequence of steps occurs so frequently that the
1770 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes provide
1771 constructors which take (as a default parameter) a pointer to an
1772 <tt>Instruction</tt> which the newly-created <tt>Instruction</tt> should
1773 precede. That is, <tt>Instruction</tt> constructors are capable of
1774 inserting the newly-created instance into the <tt>BasicBlock</tt> of a
1775 provided instruction, immediately before that instruction. Using an
1776 <tt>Instruction</tt> constructor with a <tt>insertBefore</tt> (default)
1777 parameter, the above code becomes:</p>
1778
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1779<div class="doc_code">
1780<pre>
1781Instruction* pi = ...;
1782Instruction* newInst = new Instruction(..., pi);
1783</pre>
1784</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1785
1786 <p>which is much cleaner, especially if you're creating a lot of
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1787 instructions and adding them to <tt>BasicBlock</tt>s.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1788</ul>
1789
1790</div>
1791
1792<!--_______________________________________________________________________-->
1793<div class="doc_subsubsection">
1794 <a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a>
1795</div>
1796
1797<div class="doc_text">
1798
1799<p>Deleting an instruction from an existing sequence of instructions that form a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1800<a href="#BasicBlock"><tt>BasicBlock</tt></a> is very straight-forward. First,
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1801you must have a pointer to the instruction that you wish to delete. Second, you
1802need to obtain the pointer to that instruction's basic block. You use the
1803pointer to the basic block to get its list of instructions and then use the
1804erase function to remove your instruction. For example:</p>
1805
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1806<div class="doc_code">
1807<pre>
1808<a href="#Instruction">Instruction</a> *I = .. ;
1809<a href="#BasicBlock">BasicBlock</a> *BB = I-&gt;getParent();
1810
1811BB-&gt;getInstList().erase(I);
1812</pre>
1813</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1814
1815</div>
1816
1817<!--_______________________________________________________________________-->
1818<div class="doc_subsubsection">
1819 <a name="schanges_replacing">Replacing an <tt>Instruction</tt> with another
1820 <tt>Value</tt></a>
1821</div>
1822
1823<div class="doc_text">
1824
1825<p><i>Replacing individual instructions</i></p>
1826
1827<p>Including "<a href="/doxygen/BasicBlockUtils_8h-source.html">llvm/Transforms/Utils/BasicBlockUtils.h</a>"
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1828permits use of two very useful replace functions: <tt>ReplaceInstWithValue</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1829and <tt>ReplaceInstWithInst</tt>.</p>
1830
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1831<h4><a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a></h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1832
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1833<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1834 <li><tt>ReplaceInstWithValue</tt>
1835
1836 <p>This function replaces all uses (within a basic block) of a given
1837 instruction with a value, and then removes the original instruction. The
1838 following example illustrates the replacement of the result of a particular
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]1839 <tt>AllocaInst</tt> that allocates memory for a single integer with a null
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1840 pointer to an integer.</p>
1841
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1842<div class="doc_code">
1843<pre>
1844AllocaInst* instToReplace = ...;
1845BasicBlock::iterator ii(instToReplace);
1846
1847ReplaceInstWithValue(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
1848 Constant::getNullValue(PointerType::get(Type::IntTy)));
1849</pre></div></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1850
1851 <li><tt>ReplaceInstWithInst</tt>
1852
1853 <p>This function replaces a particular instruction with another
1854 instruction. The following example illustrates the replacement of one
1855 <tt>AllocaInst</tt> with another.</p>
1856
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1857<div class="doc_code">
1858<pre>
1859AllocaInst* instToReplace = ...;
1860BasicBlock::iterator ii(instToReplace);
1861
1862ReplaceInstWithInst(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
1863 new AllocaInst(Type::IntTy, 0, "ptrToReplacedInt"));
1864</pre></div></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1865</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1866
1867<p><i>Replacing multiple uses of <tt>User</tt>s and <tt>Value</tt>s</i></p>
1868
1869<p>You can use <tt>Value::replaceAllUsesWith</tt> and
1870<tt>User::replaceUsesOfWith</tt> to change more than one use at a time. See the
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]1871doxygen documentation for the <a href="/doxygen/classllvm_1_1Value.html">Value Class</a>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]1872and <a href="/doxygen/classllvm_1_1User.html">User Class</a>, respectively, for more
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1873information.</p>
1874
1875<!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
1876include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
1877ReplaceInstWithValue, ReplaceInstWithInst -->
1878
1879</div>
1880
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]1881<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1882<div class="doc_section">
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]1883 <a name="advanced">Advanced Topics</a>
1884</div>
1885<!-- *********************************************************************** -->
1886
1887<div class="doc_text">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]1888<p>
1889This section describes some of the advanced or obscure API's that most clients
1890do not need to be aware of. These API's tend manage the inner workings of the
1891LLVM system, and only need to be accessed in unusual circumstances.
1892</p>
1893</div>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]1894
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]1895<!-- ======================================================================= -->
1896<div class="doc_subsection">
1897 <a name="TypeResolve">LLVM Type Resolution</a>
1898</div>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]1899
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]1900<div class="doc_text">
1901
1902<p>
1903The LLVM type system has a very simple goal: allow clients to compare types for
1904structural equality with a simple pointer comparison (aka a shallow compare).
1905This goal makes clients much simpler and faster, and is used throughout the LLVM
1906system.
1907</p>
1908
1909<p>
1910Unfortunately achieving this goal is not a simple matter. In particular,
1911recursive types and late resolution of opaque types makes the situation very
1912difficult to handle. Fortunately, for the most part, our implementation makes
1913most clients able to be completely unaware of the nasty internal details. The
1914primary case where clients are exposed to the inner workings of it are when
1915building a recursive type. In addition to this case, the LLVM bytecode reader,
1916assembly parser, and linker also have to be aware of the inner workings of this
1917system.
1918</p>
1919
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]1920<p>
1921For our purposes below, we need three concepts. First, an "Opaque Type" is
1922exactly as defined in the <a href="LangRef.html#t_opaque">language
1923reference</a>. Second an "Abstract Type" is any type which includes an
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]1924opaque type as part of its type graph (for example "<tt>{ opaque, i32 }</tt>").
1925Third, a concrete type is a type that is not an abstract type (e.g. "<tt>{ i32,
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]1926float }</tt>").
1927</p>
1928
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]1929</div>
1930
1931<!-- ______________________________________________________________________ -->
1932<div class="doc_subsubsection">
1933 <a name="BuildRecType">Basic Recursive Type Construction</a>
1934</div>
1935
1936<div class="doc_text">
1937
1938<p>
1939Because the most common question is "how do I build a recursive type with LLVM",
1940we answer it now and explain it as we go. Here we include enough to cause this
1941to be emitted to an output .ll file:
1942</p>
1943
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1944<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]1945<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]1946%mylist = type { %mylist*, i32 }
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]1947</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1948</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]1949
1950<p>
1951To build this, use the following LLVM APIs:
1952</p>
1953
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1954<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]1955<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1956// <i>Create the initial outer struct</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1957<a href="#PATypeHolder">PATypeHolder</a> StructTy = OpaqueType::get();
1958std::vector&lt;const Type*&gt; Elts;
1959Elts.push_back(PointerType::get(StructTy));
1960Elts.push_back(Type::IntTy);
1961StructType *NewSTy = StructType::get(Elts);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]1962
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]1963// <i>At this point, NewSTy = "{ opaque*, i32 }". Tell VMCore that</i>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1964// <i>the struct and the opaque type are actually the same.</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1965cast&lt;OpaqueType&gt;(StructTy.get())-&gt;<a href="#refineAbstractTypeTo">refineAbstractTypeTo</a>(NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]1966
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1967// <i>NewSTy is potentially invalidated, but StructTy (a <a href="#PATypeHolder">PATypeHolder</a>) is</i>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1968// <i>kept up-to-date</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1969NewSTy = cast&lt;StructType&gt;(StructTy.get());
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]1970
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1971// <i>Add a name for the type to the module symbol table (optional)</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1972MyModule-&gt;addTypeName("mylist", NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]1973</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1974</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]1975
1976<p>
1977This code shows the basic approach used to build recursive types: build a
1978non-recursive type using 'opaque', then use type unification to close the cycle.
1979The type unification step is performed by the <tt><a
Chris Lattneraff26d12007-02-03 03:06:52 +0000[diff] [blame]1980href="#refineAbstractTypeTo">refineAbstractTypeTo</a></tt> method, which is
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]1981described next. After that, we describe the <a
1982href="#PATypeHolder">PATypeHolder class</a>.
1983</p>
1984
1985</div>
1986
1987<!-- ______________________________________________________________________ -->
1988<div class="doc_subsubsection">
1989 <a name="refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a>
1990</div>
1991
1992<div class="doc_text">
1993<p>
1994The <tt>refineAbstractTypeTo</tt> method starts the type unification process.
1995While this method is actually a member of the DerivedType class, it is most
1996often used on OpaqueType instances. Type unification is actually a recursive
1997process. After unification, types can become structurally isomorphic to
1998existing types, and all duplicates are deleted (to preserve pointer equality).
1999</p>
2000
2001<p>
2002In the example above, the OpaqueType object is definitely deleted.
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2003Additionally, if there is an "{ \2*, i32}" type already created in the system,
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2004the pointer and struct type created are <b>also</b> deleted. Obviously whenever
2005a type is deleted, any "Type*" pointers in the program are invalidated. As
2006such, it is safest to avoid having <i>any</i> "Type*" pointers to abstract types
2007live across a call to <tt>refineAbstractTypeTo</tt> (note that non-abstract
2008types can never move or be deleted). To deal with this, the <a
2009href="#PATypeHolder">PATypeHolder</a> class is used to maintain a stable
2010reference to a possibly refined type, and the <a
2011href="#AbstractTypeUser">AbstractTypeUser</a> class is used to update more
2012complex datastructures.
2013</p>
2014
2015</div>
2016
2017<!-- ______________________________________________________________________ -->
2018<div class="doc_subsubsection">
2019 <a name="PATypeHolder">The PATypeHolder Class</a>
2020</div>
2021
2022<div class="doc_text">
2023<p>
2024PATypeHolder is a form of a "smart pointer" for Type objects. When VMCore
2025happily goes about nuking types that become isomorphic to existing types, it
2026automatically updates all PATypeHolder objects to point to the new type. In the
2027example above, this allows the code to maintain a pointer to the resultant
2028resolved recursive type, even though the Type*'s are potentially invalidated.
2029</p>
2030
2031<p>
2032PATypeHolder is an extremely light-weight object that uses a lazy union-find
2033implementation to update pointers. For example the pointer from a Value to its
2034Type is maintained by PATypeHolder objects.
2035</p>
2036
2037</div>
2038
2039<!-- ______________________________________________________________________ -->
2040<div class="doc_subsubsection">
2041 <a name="AbstractTypeUser">The AbstractTypeUser Class</a>
2042</div>
2043
2044<div class="doc_text">
2045
2046<p>
2047Some data structures need more to perform more complex updates when types get
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2048resolved. To support this, a class can derive from the AbstractTypeUser class.
2049This class
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2050allows it to get callbacks when certain types are resolved. To register to get
2051callbacks for a particular type, the DerivedType::{add/remove}AbstractTypeUser
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]2052methods can be called on a type. Note that these methods only work for <i>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2053 abstract</i> types. Concrete types (those that do not include any opaque
2054objects) can never be refined.
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2055</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2056</div>
2057
2058
2059<!-- ======================================================================= -->
2060<div class="doc_subsection">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2061 <a name="SymbolTable">The <tt>ValueSymbolTable</tt> and
2062 <tt>TypeSymbolTable</tt> classes</a>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2063</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2064
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2065<div class="doc_text">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2066<p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1ValueSymbolTable.html">
2067ValueSymbolTable</a></tt> class provides a symbol table that the <a
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2068href="#Function"><tt>Function</tt></a> and <a href="#Module">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2069<tt>Module</tt></a> classes use for naming value definitions. The symbol table
2070can provide a name for any <a href="#Value"><tt>Value</tt></a>.
2071The <tt><a href="http://llvm.org/doxygen/classllvm_1_1TypeSymbolTable.html">
2072TypeSymbolTable</a></tt> class is used by the <tt>Module</tt> class to store
2073names for types.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2074
Reid Spencera6362242007-01-07 00:41:39 +0000[diff] [blame]2075<p>Note that the <tt>SymbolTable</tt> class should not be directly accessed
2076by most clients. It should only be used when iteration over the symbol table
2077names themselves are required, which is very special purpose. Note that not
2078all LLVM
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2079<a href="#Value">Value</a>s have names, and those without names (i.e. they have
2080an empty name) do not exist in the symbol table.
2081</p>
2082
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2083<p>These symbol tables support iteration over the values/types in the symbol
2084table with <tt>begin/end/iterator</tt> and supports querying to see if a
2085specific name is in the symbol table (with <tt>lookup</tt>). The
2086<tt>ValueSymbolTable</tt> class exposes no public mutator methods, instead,
2087simply call <tt>setName</tt> on a value, which will autoinsert it into the
2088appropriate symbol table. For types, use the Module::addTypeName method to
2089insert entries into the symbol table.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2090
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2091</div>
2092
2093
2094
2095<!-- *********************************************************************** -->
2096<div class="doc_section">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2097 <a name="coreclasses">The Core LLVM Class Hierarchy Reference </a>
2098</div>
2099<!-- *********************************************************************** -->
2100
2101<div class="doc_text">
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2102<p><tt>#include "<a href="/doxygen/Type_8h-source.html">llvm/Type.h</a>"</tt>
2103<br>doxygen info: <a href="/doxygen/classllvm_1_1Type.html">Type Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2104
2105<p>The Core LLVM classes are the primary means of representing the program
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2106being inspected or transformed. The core LLVM classes are defined in
2107header files in the <tt>include/llvm/</tt> directory, and implemented in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2108the <tt>lib/VMCore</tt> directory.</p>
2109
2110</div>
2111
2112<!-- ======================================================================= -->
2113<div class="doc_subsection">
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2114 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
2115</div>
2116
2117<div class="doc_text">
2118
2119 <p><tt>Type</tt> is a superclass of all type classes. Every <tt>Value</tt> has
2120 a <tt>Type</tt>. <tt>Type</tt> cannot be instantiated directly but only
2121 through its subclasses. Certain primitive types (<tt>VoidType</tt>,
2122 <tt>LabelType</tt>, <tt>FloatType</tt> and <tt>DoubleType</tt>) have hidden
2123 subclasses. They are hidden because they offer no useful functionality beyond
2124 what the <tt>Type</tt> class offers except to distinguish themselves from
2125 other subclasses of <tt>Type</tt>.</p>
2126 <p>All other types are subclasses of <tt>DerivedType</tt>. Types can be
2127 named, but this is not a requirement. There exists exactly
2128 one instance of a given shape at any one time. This allows type equality to
2129 be performed with address equality of the Type Instance. That is, given two
2130 <tt>Type*</tt> values, the types are identical if the pointers are identical.
2131 </p>
2132</div>
2133
2134<!-- _______________________________________________________________________ -->
2135<div class="doc_subsubsection">
2136 <a name="m_Value">Important Public Methods</a>
2137</div>
2138
2139<div class="doc_text">
2140
2141<ul>
Chris Lattner8f79df32007-01-15 01:55:32 +0000[diff] [blame]2142 <li><tt>bool isInteger() const</tt>: Returns true for any integer type.</li>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2143
2144 <li><tt>bool isFloatingPoint()</tt>: Return true if this is one of the two
2145 floating point types.</li>
2146
2147 <li><tt>bool isAbstract()</tt>: Return true if the type is abstract (contains
2148 an OpaqueType anywhere in its definition).</li>
2149
2150 <li><tt>bool isSized()</tt>: Return true if the type has known size. Things
2151 that don't have a size are abstract types, labels and void.</li>
2152
2153</ul>
2154</div>
2155
2156<!-- _______________________________________________________________________ -->
2157<div class="doc_subsubsection">
2158 <a name="m_Value">Important Derived Types</a>
2159</div>
2160<div class="doc_text">
2161<dl>
2162 <dt><tt>IntegerType</tt></dt>
2163 <dd>Subclass of DerivedType that represents integer types of any bit width.
2164 Any bit width between <tt>IntegerType::MIN_INT_BITS</tt> (1) and
2165 <tt>IntegerType::MAX_INT_BITS</tt> (~8 million) can be represented.
2166 <ul>
2167 <li><tt>static const IntegerType* get(unsigned NumBits)</tt>: get an integer
2168 type of a specific bit width.</li>
2169 <li><tt>unsigned getBitWidth() const</tt>: Get the bit width of an integer
2170 type.</li>
2171 </ul>
2172 </dd>
2173 <dt><tt>SequentialType</tt></dt>
2174 <dd>This is subclassed by ArrayType and PointerType
2175 <ul>
2176 <li><tt>const Type * getElementType() const</tt>: Returns the type of each
2177 of the elements in the sequential type. </li>
2178 </ul>
2179 </dd>
2180 <dt><tt>ArrayType</tt></dt>
2181 <dd>This is a subclass of SequentialType and defines the interface for array
2182 types.
2183 <ul>
2184 <li><tt>unsigned getNumElements() const</tt>: Returns the number of
2185 elements in the array. </li>
2186 </ul>
2187 </dd>
2188 <dt><tt>PointerType</tt></dt>
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]2189 <dd>Subclass of SequentialType for pointer types.</dd>
Reid Spencer9d6565a2007-02-15 02:26:10 +0000[diff] [blame]2190 <dt><tt>VectorType</tt></dt>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2191 <dd>Subclass of SequentialType for vector types. A
2192 vector type is similar to an ArrayType but is distinguished because it is
2193 a first class type wherease ArrayType is not. Vector types are used for
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2194 vector operations and are usually small vectors of of an integer or floating
2195 point type.</dd>
2196 <dt><tt>StructType</tt></dt>
2197 <dd>Subclass of DerivedTypes for struct types.</dd>
2198 <dt><tt>FunctionType</tt></dt>
2199 <dd>Subclass of DerivedTypes for function types.
2200 <ul>
2201 <li><tt>bool isVarArg() const</tt>: Returns true if its a vararg
2202 function</li>
2203 <li><tt> const Type * getReturnType() const</tt>: Returns the
2204 return type of the function.</li>
2205 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns
2206 the type of the ith parameter.</li>
2207 <li><tt> const unsigned getNumParams() const</tt>: Returns the
2208 number of formal parameters.</li>
2209 </ul>
2210 </dd>
2211 <dt><tt>OpaqueType</tt></dt>
2212 <dd>Sublcass of DerivedType for abstract types. This class
2213 defines no content and is used as a placeholder for some other type. Note
2214 that OpaqueType is used (temporarily) during type resolution for forward
2215 references of types. Once the referenced type is resolved, the OpaqueType
2216 is replaced with the actual type. OpaqueType can also be used for data
2217 abstraction. At link time opaque types can be resolved to actual types
2218 of the same name.</dd>
2219</dl>
2220</div>
2221
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2222
2223
2224<!-- ======================================================================= -->
2225<div class="doc_subsection">
2226 <a name="Module">The <tt>Module</tt> class</a>
2227</div>
2228
2229<div class="doc_text">
2230
2231<p><tt>#include "<a
2232href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt><br> doxygen info:
2233<a href="/doxygen/classllvm_1_1Module.html">Module Class</a></p>
2234
2235<p>The <tt>Module</tt> class represents the top level structure present in LLVM
2236programs. An LLVM module is effectively either a translation unit of the
2237original program or a combination of several translation units merged by the
2238linker. The <tt>Module</tt> class keeps track of a list of <a
2239href="#Function"><tt>Function</tt></a>s, a list of <a
2240href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
2241href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
2242helpful member functions that try to make common operations easy.</p>
2243
2244</div>
2245
2246<!-- _______________________________________________________________________ -->
2247<div class="doc_subsubsection">
2248 <a name="m_Module">Important Public Members of the <tt>Module</tt> class</a>
2249</div>
2250
2251<div class="doc_text">
2252
2253<ul>
2254 <li><tt>Module::Module(std::string name = "")</tt></li>
2255</ul>
2256
2257<p>Constructing a <a href="#Module">Module</a> is easy. You can optionally
2258provide a name for it (probably based on the name of the translation unit).</p>
2259
2260<ul>
2261 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
2262 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
2263
2264 <tt>begin()</tt>, <tt>end()</tt>
2265 <tt>size()</tt>, <tt>empty()</tt>
2266
2267 <p>These are forwarding methods that make it easy to access the contents of
2268 a <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
2269 list.</p></li>
2270
2271 <li><tt>Module::FunctionListType &amp;getFunctionList()</tt>
2272
2273 <p> Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
2274 necessary to use when you need to update the list or perform a complex
2275 action that doesn't have a forwarding method.</p>
2276
2277 <p><!-- Global Variable --></p></li>
2278</ul>
2279
2280<hr>
2281
2282<ul>
2283 <li><tt>Module::global_iterator</tt> - Typedef for global variable list iterator<br>
2284
2285 <tt>Module::const_global_iterator</tt> - Typedef for const_iterator.<br>
2286
2287 <tt>global_begin()</tt>, <tt>global_end()</tt>
2288 <tt>global_size()</tt>, <tt>global_empty()</tt>
2289
2290 <p> These are forwarding methods that make it easy to access the contents of
2291 a <tt>Module</tt> object's <a
2292 href="#GlobalVariable"><tt>GlobalVariable</tt></a> list.</p></li>
2293
2294 <li><tt>Module::GlobalListType &amp;getGlobalList()</tt>
2295
2296 <p>Returns the list of <a
2297 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s. This is necessary to
2298 use when you need to update the list or perform a complex action that
2299 doesn't have a forwarding method.</p>
2300
2301 <p><!-- Symbol table stuff --> </p></li>
2302</ul>
2303
2304<hr>
2305
2306<ul>
2307 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
2308
2309 <p>Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
2310 for this <tt>Module</tt>.</p>
2311
2312 <p><!-- Convenience methods --></p></li>
2313</ul>
2314
2315<hr>
2316
2317<ul>
2318 <li><tt><a href="#Function">Function</a> *getFunction(const std::string
2319 &amp;Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt>
2320
2321 <p>Look up the specified function in the <tt>Module</tt> <a
2322 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
2323 <tt>null</tt>.</p></li>
2324
2325 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const
2326 std::string &amp;Name, const <a href="#FunctionType">FunctionType</a> *T)</tt>
2327
2328 <p>Look up the specified function in the <tt>Module</tt> <a
2329 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
2330 external declaration for the function and return it.</p></li>
2331
2332 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt>
2333
2334 <p>If there is at least one entry in the <a
2335 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
2336 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
2337 string.</p></li>
2338
2339 <li><tt>bool addTypeName(const std::string &amp;Name, const <a
2340 href="#Type">Type</a> *Ty)</tt>
2341
2342 <p>Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
2343 mapping <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this
2344 name, true is returned and the <a
2345 href="#SymbolTable"><tt>SymbolTable</tt></a> is not modified.</p></li>
2346</ul>
2347
2348</div>
2349
2350
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2351<!-- ======================================================================= -->
2352<div class="doc_subsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2353 <a name="Value">The <tt>Value</tt> class</a>
2354</div>
2355
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2356<div class="doc_text">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2357
2358<p><tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt>
2359<br>
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]2360doxygen info: <a href="/doxygen/classllvm_1_1Value.html">Value Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2361
2362<p>The <tt>Value</tt> class is the most important class in the LLVM Source
2363base. It represents a typed value that may be used (among other things) as an
2364operand to an instruction. There are many different types of <tt>Value</tt>s,
2365such as <a href="#Constant"><tt>Constant</tt></a>s,<a
2366href="#Argument"><tt>Argument</tt></a>s. Even <a
2367href="#Instruction"><tt>Instruction</tt></a>s and <a
2368href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.</p>
2369
2370<p>A particular <tt>Value</tt> may be used many times in the LLVM representation
2371for a program. For example, an incoming argument to a function (represented
2372with an instance of the <a href="#Argument">Argument</a> class) is "used" by
2373every instruction in the function that references the argument. To keep track
2374of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
2375href="#User"><tt>User</tt></a>s that is using it (the <a
2376href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
2377graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
2378def-use information in the program, and is accessible through the <tt>use_</tt>*
2379methods, shown below.</p>
2380
2381<p>Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed,
2382and this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
2383method. In addition, all LLVM values can be named. The "name" of the
2384<tt>Value</tt> is a symbolic string printed in the LLVM code:</p>
2385
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2386<div class="doc_code">
2387<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2388%<b>foo</b> = add i32 1, 2
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2389</pre>
2390</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2391
2392<p><a name="#nameWarning">The name of this instruction is "foo".</a> <b>NOTE</b>
2393that the name of any value may be missing (an empty string), so names should
2394<b>ONLY</b> be used for debugging (making the source code easier to read,
2395debugging printouts), they should not be used to keep track of values or map
2396between them. For this purpose, use a <tt>std::map</tt> of pointers to the
2397<tt>Value</tt> itself instead.</p>
2398
2399<p>One important aspect of LLVM is that there is no distinction between an SSA
2400variable and the operation that produces it. Because of this, any reference to
2401the value produced by an instruction (or the value available as an incoming
Chris Lattnerd5fc4fc2004-03-18 14:58:55 +0000[diff] [blame]2402argument, for example) is represented as a direct pointer to the instance of
2403the class that
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2404represents this value. Although this may take some getting used to, it
2405simplifies the representation and makes it easier to manipulate.</p>
2406
2407</div>
2408
2409<!-- _______________________________________________________________________ -->
2410<div class="doc_subsubsection">
2411 <a name="m_Value">Important Public Members of the <tt>Value</tt> class</a>
2412</div>
2413
2414<div class="doc_text">
2415
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2416<ul>
2417 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the
2418use-list<br>
2419 <tt>Value::use_const_iterator</tt> - Typedef for const_iterator over
2420the use-list<br>
2421 <tt>unsigned use_size()</tt> - Returns the number of users of the
2422value.<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]2423 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2424 <tt>use_iterator use_begin()</tt> - Get an iterator to the start of
2425the use-list.<br>
2426 <tt>use_iterator use_end()</tt> - Get an iterator to the end of the
2427use-list.<br>
2428 <tt><a href="#User">User</a> *use_back()</tt> - Returns the last
2429element in the list.
2430 <p> These methods are the interface to access the def-use
2431information in LLVM. As with all other iterators in LLVM, the naming
2432conventions follow the conventions defined by the <a href="#stl">STL</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2433 </li>
2434 <li><tt><a href="#Type">Type</a> *getType() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2435 <p>This method returns the Type of the Value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2436 </li>
2437 <li><tt>bool hasName() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]2438 <tt>std::string getName() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2439 <tt>void setName(const std::string &amp;Name)</tt>
2440 <p> This family of methods is used to access and assign a name to a <tt>Value</tt>,
2441be aware of the <a href="#nameWarning">precaution above</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2442 </li>
2443 <li><tt>void replaceAllUsesWith(Value *V)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2444
2445 <p>This method traverses the use list of a <tt>Value</tt> changing all <a
2446 href="#User"><tt>User</tt>s</a> of the current value to refer to
2447 "<tt>V</tt>" instead. For example, if you detect that an instruction always
2448 produces a constant value (for example through constant folding), you can
2449 replace all uses of the instruction with the constant like this:</p>
2450
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2451<div class="doc_code">
2452<pre>
2453Inst-&gt;replaceAllUsesWith(ConstVal);
2454</pre>
2455</div>
2456
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2457</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2458
2459</div>
2460
2461<!-- ======================================================================= -->
2462<div class="doc_subsection">
2463 <a name="User">The <tt>User</tt> class</a>
2464</div>
2465
2466<div class="doc_text">
2467
2468<p>
2469<tt>#include "<a href="/doxygen/User_8h-source.html">llvm/User.h</a>"</tt><br>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]2470doxygen info: <a href="/doxygen/classllvm_1_1User.html">User Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2471Superclass: <a href="#Value"><tt>Value</tt></a></p>
2472
2473<p>The <tt>User</tt> class is the common base class of all LLVM nodes that may
2474refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
2475that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
2476referring to. The <tt>User</tt> class itself is a subclass of
2477<tt>Value</tt>.</p>
2478
2479<p>The operands of a <tt>User</tt> point directly to the LLVM <a
2480href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
2481Single Assignment (SSA) form, there can only be one definition referred to,
2482allowing this direct connection. This connection provides the use-def
2483information in LLVM.</p>
2484
2485</div>
2486
2487<!-- _______________________________________________________________________ -->
2488<div class="doc_subsubsection">
2489 <a name="m_User">Important Public Members of the <tt>User</tt> class</a>
2490</div>
2491
2492<div class="doc_text">
2493
2494<p>The <tt>User</tt> class exposes the operand list in two ways: through
2495an index access interface and through an iterator based interface.</p>
2496
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2497<ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2498 <li><tt>Value *getOperand(unsigned i)</tt><br>
2499 <tt>unsigned getNumOperands()</tt>
2500 <p> These two methods expose the operands of the <tt>User</tt> in a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2501convenient form for direct access.</p></li>
2502
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2503 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand
2504list<br>
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]2505 <tt>op_iterator op_begin()</tt> - Get an iterator to the start of
2506the operand list.<br>
2507 <tt>op_iterator op_end()</tt> - Get an iterator to the end of the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2508operand list.
2509 <p> Together, these methods make up the iterator based interface to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2510the operands of a <tt>User</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2511</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2512
2513</div>
2514
2515<!-- ======================================================================= -->
2516<div class="doc_subsection">
2517 <a name="Instruction">The <tt>Instruction</tt> class</a>
2518</div>
2519
2520<div class="doc_text">
2521
2522<p><tt>#include "</tt><tt><a
2523href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt><br>
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]2524doxygen info: <a href="/doxygen/classllvm_1_1Instruction.html">Instruction Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2525Superclasses: <a href="#User"><tt>User</tt></a>, <a
2526href="#Value"><tt>Value</tt></a></p>
2527
2528<p>The <tt>Instruction</tt> class is the common base class for all LLVM
2529instructions. It provides only a few methods, but is a very commonly used
2530class. The primary data tracked by the <tt>Instruction</tt> class itself is the
2531opcode (instruction type) and the parent <a
2532href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
2533into. To represent a specific type of instruction, one of many subclasses of
2534<tt>Instruction</tt> are used.</p>
2535
2536<p> Because the <tt>Instruction</tt> class subclasses the <a
2537href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
2538way as for other <a href="#User"><tt>User</tt></a>s (with the
2539<tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
2540<tt>op_begin()</tt>/<tt>op_end()</tt> methods).</p> <p> An important file for
2541the <tt>Instruction</tt> class is the <tt>llvm/Instruction.def</tt> file. This
2542file contains some meta-data about the various different types of instructions
2543in LLVM. It describes the enum values that are used as opcodes (for example
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]2544<tt>Instruction::Add</tt> and <tt>Instruction::ICmp</tt>), as well as the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2545concrete sub-classes of <tt>Instruction</tt> that implement the instruction (for
2546example <tt><a href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]2547href="#CmpInst">CmpInst</a></tt>). Unfortunately, the use of macros in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2548this file confuses doxygen, so these enum values don't show up correctly in the
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]2549<a href="/doxygen/classllvm_1_1Instruction.html">doxygen output</a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2550
2551</div>
2552
2553<!-- _______________________________________________________________________ -->
2554<div class="doc_subsubsection">
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]2555 <a name="s_Instruction">Important Subclasses of the <tt>Instruction</tt>
2556 class</a>
2557</div>
2558<div class="doc_text">
2559 <ul>
2560 <li><tt><a name="BinaryOperator">BinaryOperator</a></tt>
2561 <p>This subclasses represents all two operand instructions whose operands
2562 must be the same type, except for the comparison instructions.</p></li>
2563 <li><tt><a name="CastInst">CastInst</a></tt>
2564 <p>This subclass is the parent of the 12 casting instructions. It provides
2565 common operations on cast instructions.</p>
2566 <li><tt><a name="CmpInst">CmpInst</a></tt>
2567 <p>This subclass respresents the two comparison instructions,
2568 <a href="LangRef.html#i_icmp">ICmpInst</a> (integer opreands), and
2569 <a href="LangRef.html#i_fcmp">FCmpInst</a> (floating point operands).</p>
2570 <li><tt><a name="TerminatorInst">TerminatorInst</a></tt>
2571 <p>This subclass is the parent of all terminator instructions (those which
2572 can terminate a block).</p>
2573 </ul>
2574 </div>
2575
2576<!-- _______________________________________________________________________ -->
2577<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2578 <a name="m_Instruction">Important Public Members of the <tt>Instruction</tt>
2579 class</a>
2580</div>
2581
2582<div class="doc_text">
2583
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2584<ul>
2585 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2586 <p>Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that
2587this <tt>Instruction</tt> is embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2588 <li><tt>bool mayWriteToMemory()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2589 <p>Returns true if the instruction writes to memory, i.e. it is a
2590 <tt>call</tt>,<tt>free</tt>,<tt>invoke</tt>, or <tt>store</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2591 <li><tt>unsigned getOpcode()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2592 <p>Returns the opcode for the <tt>Instruction</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2593 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2594 <p>Returns another instance of the specified instruction, identical
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2595in all ways to the original except that the instruction has no parent
2596(ie it's not embedded into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>),
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2597and it has no name</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2598</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2599
2600</div>
2601
2602<!-- ======================================================================= -->
2603<div class="doc_subsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2604 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2605</div>
2606
2607<div class="doc_text">
2608
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2609<p>Constant represents a base class for different types of constants. It
2610is subclassed by ConstantInt, ConstantArray, etc. for representing
2611the various types of Constants. <a href="#GlobalValue">GlobalValue</a> is also
2612a subclass, which represents the address of a global variable or function.
2613</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2614
2615</div>
2616
2617<!-- _______________________________________________________________________ -->
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2618<div class="doc_subsubsection">Important Subclasses of Constant </div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2619<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2620<ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2621 <li>ConstantInt : This subclass of Constant represents an integer constant of
2622 any width.
2623 <ul>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]2624 <li><tt>const APInt&amp; getValue() const</tt>: Returns the underlying
2625 value of this constant, an APInt value.</li>
2626 <li><tt>int64_t getSExtValue() const</tt>: Converts the underlying APInt
2627 value to an int64_t via sign extension. If the value (not the bit width)
2628 of the APInt is too large to fit in an int64_t, an assertion will result.
2629 For this reason, use of this method is discouraged.</li>
2630 <li><tt>uint64_t getZExtValue() const</tt>: Converts the underlying APInt
2631 value to a uint64_t via zero extension. IF the value (not the bit width)
2632 of the APInt is too large to fit in a uint64_t, an assertion will result.
Reid Spencer4474d872007-03-02 01:31:31 +0000[diff] [blame]2633 For this reason, use of this method is discouraged.</li>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]2634 <li><tt>static ConstantInt* get(const APInt&amp; Val)</tt>: Returns the
2635 ConstantInt object that represents the value provided by <tt>Val</tt>.
2636 The type is implied as the IntegerType that corresponds to the bit width
2637 of <tt>Val</tt>.</li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2638 <li><tt>static ConstantInt* get(const Type *Ty, uint64_t Val)</tt>:
2639 Returns the ConstantInt object that represents the value provided by
2640 <tt>Val</tt> for integer type <tt>Ty</tt>.</li>
2641 </ul>
2642 </li>
2643 <li>ConstantFP : This class represents a floating point constant.
2644 <ul>
2645 <li><tt>double getValue() const</tt>: Returns the underlying value of
2646 this constant. </li>
2647 </ul>
2648 </li>
2649 <li>ConstantArray : This represents a constant array.
2650 <ul>
2651 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
2652 a vector of component constants that makeup this array. </li>
2653 </ul>
2654 </li>
2655 <li>ConstantStruct : This represents a constant struct.
2656 <ul>
2657 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
2658 a vector of component constants that makeup this array. </li>
2659 </ul>
2660 </li>
2661 <li>GlobalValue : This represents either a global variable or a function. In
2662 either case, the value is a constant fixed address (after linking).
2663 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2664</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2665</div>
2666
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2667
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2668<!-- ======================================================================= -->
2669<div class="doc_subsection">
2670 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
2671</div>
2672
2673<div class="doc_text">
2674
2675<p><tt>#include "<a
2676href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt><br>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]2677doxygen info: <a href="/doxygen/classllvm_1_1GlobalValue.html">GlobalValue
2678Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]2679Superclasses: <a href="#Constant"><tt>Constant</tt></a>,
2680<a href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2681
2682<p>Global values (<a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
2683href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
2684visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
2685Because they are visible at global scope, they are also subject to linking with
2686other globals defined in different translation units. To control the linking
2687process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
2688<tt>GlobalValue</tt>s know whether they have internal or external linkage, as
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]2689defined by the <tt>LinkageTypes</tt> enumeration.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2690
2691<p>If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
2692<tt>static</tt> in C), it is not visible to code outside the current translation
2693unit, and does not participate in linking. If it has external linkage, it is
2694visible to external code, and does participate in linking. In addition to
2695linkage information, <tt>GlobalValue</tt>s keep track of which <a
2696href="#Module"><tt>Module</tt></a> they are currently part of.</p>
2697
2698<p>Because <tt>GlobalValue</tt>s are memory objects, they are always referred to
2699by their <b>address</b>. As such, the <a href="#Type"><tt>Type</tt></a> of a
2700global is always a pointer to its contents. It is important to remember this
2701when using the <tt>GetElementPtrInst</tt> instruction because this pointer must
2702be dereferenced first. For example, if you have a <tt>GlobalVariable</tt> (a
2703subclass of <tt>GlobalValue)</tt> that is an array of 24 ints, type <tt>[24 x
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2704i32]</tt>, then the <tt>GlobalVariable</tt> is a pointer to that array. Although
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2705the address of the first element of this array and the value of the
2706<tt>GlobalVariable</tt> are the same, they have different types. The
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2707<tt>GlobalVariable</tt>'s type is <tt>[24 x i32]</tt>. The first element's type
2708is <tt>i32.</tt> Because of this, accessing a global value requires you to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2709dereference the pointer with <tt>GetElementPtrInst</tt> first, then its elements
2710can be accessed. This is explained in the <a href="LangRef.html#globalvars">LLVM
2711Language Reference Manual</a>.</p>
2712
2713</div>
2714
2715<!-- _______________________________________________________________________ -->
2716<div class="doc_subsubsection">
2717 <a name="m_GlobalValue">Important Public Members of the <tt>GlobalValue</tt>
2718 class</a>
2719</div>
2720
2721<div class="doc_text">
2722
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2723<ul>
2724 <li><tt>bool hasInternalLinkage() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]2725 <tt>bool hasExternalLinkage() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2726 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt>
2727 <p> These methods manipulate the linkage characteristics of the <tt>GlobalValue</tt>.</p>
2728 <p> </p>
2729 </li>
2730 <li><tt><a href="#Module">Module</a> *getParent()</tt>
2731 <p> This returns the <a href="#Module"><tt>Module</tt></a> that the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2732GlobalValue is currently embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2733</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2734
2735</div>
2736
2737<!-- ======================================================================= -->
2738<div class="doc_subsection">
2739 <a name="Function">The <tt>Function</tt> class</a>
2740</div>
2741
2742<div class="doc_text">
2743
2744<p><tt>#include "<a
2745href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt><br> doxygen
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]2746info: <a href="/doxygen/classllvm_1_1Function.html">Function Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]2747Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
2748<a href="#Constant"><tt>Constant</tt></a>,
2749<a href="#User"><tt>User</tt></a>,
2750<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2751
2752<p>The <tt>Function</tt> class represents a single procedure in LLVM. It is
2753actually one of the more complex classes in the LLVM heirarchy because it must
2754keep track of a large amount of data. The <tt>Function</tt> class keeps track
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]2755of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal
2756<a href="#Argument"><tt>Argument</tt></a>s, and a
2757<a href="#SymbolTable"><tt>SymbolTable</tt></a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2758
2759<p>The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most
2760commonly used part of <tt>Function</tt> objects. The list imposes an implicit
2761ordering of the blocks in the function, which indicate how the code will be
2762layed out by the backend. Additionally, the first <a
2763href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
2764<tt>Function</tt>. It is not legal in LLVM to explicitly branch to this initial
2765block. There are no implicit exit nodes, and in fact there may be multiple exit
2766nodes from a single <tt>Function</tt>. If the <a
2767href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
2768the <tt>Function</tt> is actually a function declaration: the actual body of the
2769function hasn't been linked in yet.</p>
2770
2771<p>In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
2772<tt>Function</tt> class also keeps track of the list of formal <a
2773href="#Argument"><tt>Argument</tt></a>s that the function receives. This
2774container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
2775nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
2776the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.</p>
2777
2778<p>The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used
2779LLVM feature that is only used when you have to look up a value by name. Aside
2780from that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used
2781internally to make sure that there are not conflicts between the names of <a
2782href="#Instruction"><tt>Instruction</tt></a>s, <a
2783href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
2784href="#Argument"><tt>Argument</tt></a>s in the function body.</p>
2785
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]2786<p>Note that <tt>Function</tt> is a <a href="#GlobalValue">GlobalValue</a>
2787and therefore also a <a href="#Constant">Constant</a>. The value of the function
2788is its address (after linking) which is guaranteed to be constant.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2789</div>
2790
2791<!-- _______________________________________________________________________ -->
2792<div class="doc_subsubsection">
2793 <a name="m_Function">Important Public Members of the <tt>Function</tt>
2794 class</a>
2795</div>
2796
2797<div class="doc_text">
2798
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2799<ul>
2800 <li><tt>Function(const </tt><tt><a href="#FunctionType">FunctionType</a>
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]2801 *Ty, LinkageTypes Linkage, const std::string &amp;N = "", Module* Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2802
2803 <p>Constructor used when you need to create new <tt>Function</tt>s to add
2804 the the program. The constructor must specify the type of the function to
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]2805 create and what type of linkage the function should have. The <a
2806 href="#FunctionType"><tt>FunctionType</tt></a> argument
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2807 specifies the formal arguments and return value for the function. The same
2808 <a href="#FunctionTypel"><tt>FunctionType</tt></a> value can be used to
2809 create multiple functions. The <tt>Parent</tt> argument specifies the Module
2810 in which the function is defined. If this argument is provided, the function
2811 will automatically be inserted into that module's list of
2812 functions.</p></li>
2813
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2814 <li><tt>bool isExternal()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2815
2816 <p>Return whether or not the <tt>Function</tt> has a body defined. If the
2817 function is "external", it does not have a body, and thus must be resolved
2818 by linking with a function defined in a different translation unit.</p></li>
2819
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2820 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]2821 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2822
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]2823 <tt>begin()</tt>, <tt>end()</tt>
2824 <tt>size()</tt>, <tt>empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2825
2826 <p>These are forwarding methods that make it easy to access the contents of
2827 a <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
2828 list.</p></li>
2829
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2830 <li><tt>Function::BasicBlockListType &amp;getBasicBlockList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2831
2832 <p>Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This
2833 is necessary to use when you need to update the list or perform a complex
2834 action that doesn't have a forwarding method.</p></li>
2835
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]2836 <li><tt>Function::arg_iterator</tt> - Typedef for the argument list
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2837iterator<br>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]2838 <tt>Function::const_arg_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2839
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]2840 <tt>arg_begin()</tt>, <tt>arg_end()</tt>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]2841 <tt>arg_size()</tt>, <tt>arg_empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2842
2843 <p>These are forwarding methods that make it easy to access the contents of
2844 a <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a>
2845 list.</p></li>
2846
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2847 <li><tt>Function::ArgumentListType &amp;getArgumentList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2848
2849 <p>Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
2850 necessary to use when you need to update the list or perform a complex
2851 action that doesn't have a forwarding method.</p></li>
2852
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2853 <li><tt><a href="#BasicBlock">BasicBlock</a> &amp;getEntryBlock()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2854
2855 <p>Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
2856 function. Because the entry block for the function is always the first
2857 block, this returns the first block of the <tt>Function</tt>.</p></li>
2858
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2859 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
2860 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2861
2862 <p>This traverses the <a href="#Type"><tt>Type</tt></a> of the
2863 <tt>Function</tt> and returns the return type of the function, or the <a
2864 href="#FunctionType"><tt>FunctionType</tt></a> of the actual
2865 function.</p></li>
2866
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2867 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2868
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2869 <p> Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2870 for this <tt>Function</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2871</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2872
2873</div>
2874
2875<!-- ======================================================================= -->
2876<div class="doc_subsection">
2877 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
2878</div>
2879
2880<div class="doc_text">
2881
2882<p><tt>#include "<a
2883href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt>
2884<br>
Tanya Lattnera3da7772004-06-22 08:02:25 +0000[diff] [blame]2885doxygen info: <a href="/doxygen/classllvm_1_1GlobalVariable.html">GlobalVariable
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]2886 Class</a><br>
2887Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
2888<a href="#Constant"><tt>Constant</tt></a>,
2889<a href="#User"><tt>User</tt></a>,
2890<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2891
2892<p>Global variables are represented with the (suprise suprise)
2893<tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are also
2894subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such are
2895always referenced by their address (global values must live in memory, so their
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]2896"name" refers to their constant address). See
2897<a href="#GlobalValue"><tt>GlobalValue</tt></a> for more on this. Global
2898variables may have an initial value (which must be a
2899<a href="#Constant"><tt>Constant</tt></a>), and if they have an initializer,
2900they may be marked as "constant" themselves (indicating that their contents
2901never change at runtime).</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2902</div>
2903
2904<!-- _______________________________________________________________________ -->
2905<div class="doc_subsubsection">
2906 <a name="m_GlobalVariable">Important Public Members of the
2907 <tt>GlobalVariable</tt> class</a>
2908</div>
2909
2910<div class="doc_text">
2911
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2912<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2913 <li><tt>GlobalVariable(const </tt><tt><a href="#Type">Type</a> *Ty, bool
2914 isConstant, LinkageTypes&amp; Linkage, <a href="#Constant">Constant</a>
2915 *Initializer = 0, const std::string &amp;Name = "", Module* Parent = 0)</tt>
2916
2917 <p>Create a new global variable of the specified type. If
2918 <tt>isConstant</tt> is true then the global variable will be marked as
2919 unchanging for the program. The Linkage parameter specifies the type of
2920 linkage (internal, external, weak, linkonce, appending) for the variable. If
2921 the linkage is InternalLinkage, WeakLinkage, or LinkOnceLinkage,&nbsp; then
2922 the resultant global variable will have internal linkage. AppendingLinkage
2923 concatenates together all instances (in different translation units) of the
2924 variable into a single variable but is only applicable to arrays. &nbsp;See
2925 the <a href="LangRef.html#modulestructure">LLVM Language Reference</a> for
2926 further details on linkage types. Optionally an initializer, a name, and the
2927 module to put the variable into may be specified for the global variable as
2928 well.</p></li>
2929
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2930 <li><tt>bool isConstant() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2931
2932 <p>Returns true if this is a global variable that is known not to
2933 be modified at runtime.</p></li>
2934
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2935 <li><tt>bool hasInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2936
2937 <p>Returns true if this <tt>GlobalVariable</tt> has an intializer.</p></li>
2938
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2939 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2940
2941 <p>Returns the intial value for a <tt>GlobalVariable</tt>. It is not legal
2942 to call this method if there is no initializer.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2943</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2944
2945</div>
2946
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2947
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2948<!-- ======================================================================= -->
2949<div class="doc_subsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2950 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2951</div>
2952
2953<div class="doc_text">
2954
2955<p><tt>#include "<a
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2956href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt><br>
2957doxygen info: <a href="/doxygen/structllvm_1_1BasicBlock.html">BasicBlock
2958Class</a><br>
2959Superclass: <a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2960
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2961<p>This class represents a single entry multiple exit section of the code,
2962commonly known as a basic block by the compiler community. The
2963<tt>BasicBlock</tt> class maintains a list of <a
2964href="#Instruction"><tt>Instruction</tt></a>s, which form the body of the block.
2965Matching the language definition, the last element of this list of instructions
2966is always a terminator instruction (a subclass of the <a
2967href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).</p>
2968
2969<p>In addition to tracking the list of instructions that make up the block, the
2970<tt>BasicBlock</tt> class also keeps track of the <a
2971href="#Function"><tt>Function</tt></a> that it is embedded into.</p>
2972
2973<p>Note that <tt>BasicBlock</tt>s themselves are <a
2974href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
2975like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
2976<tt>label</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2977
2978</div>
2979
2980<!-- _______________________________________________________________________ -->
2981<div class="doc_subsubsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2982 <a name="m_BasicBlock">Important Public Members of the <tt>BasicBlock</tt>
2983 class</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2984</div>
2985
2986<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2987<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2988
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2989<li><tt>BasicBlock(const std::string &amp;Name = "", </tt><tt><a
2990 href="#Function">Function</a> *Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2991
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2992<p>The <tt>BasicBlock</tt> constructor is used to create new basic blocks for
2993insertion into a function. The constructor optionally takes a name for the new
2994block, and a <a href="#Function"><tt>Function</tt></a> to insert it into. If
2995the <tt>Parent</tt> parameter is specified, the new <tt>BasicBlock</tt> is
2996automatically inserted at the end of the specified <a
2997href="#Function"><tt>Function</tt></a>, if not specified, the BasicBlock must be
2998manually inserted into the <a href="#Function"><tt>Function</tt></a>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2999
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3000<li><tt>BasicBlock::iterator</tt> - Typedef for instruction list iterator<br>
3001<tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
3002<tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
3003<tt>size()</tt>, <tt>empty()</tt>
3004STL-style functions for accessing the instruction list.
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3005
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3006<p>These methods and typedefs are forwarding functions that have the same
3007semantics as the standard library methods of the same names. These methods
3008expose the underlying instruction list of a basic block in a way that is easy to
3009manipulate. To get the full complement of container operations (including
3010operations to update the list), you must use the <tt>getInstList()</tt>
3011method.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3012
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3013<li><tt>BasicBlock::InstListType &amp;getInstList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3014
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3015<p>This method is used to get access to the underlying container that actually
3016holds the Instructions. This method must be used when there isn't a forwarding
3017function in the <tt>BasicBlock</tt> class for the operation that you would like
3018to perform. Because there are no forwarding functions for "updating"
3019operations, you need to use this if you want to update the contents of a
3020<tt>BasicBlock</tt>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3021
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3022<li><tt><a href="#Function">Function</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3023
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3024<p> Returns a pointer to <a href="#Function"><tt>Function</tt></a> the block is
3025embedded into, or a null pointer if it is homeless.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3026
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3027<li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3028
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3029<p> Returns a pointer to the terminator instruction that appears at the end of
3030the <tt>BasicBlock</tt>. If there is no terminator instruction, or if the last
3031instruction in the block is not a terminator, then a null pointer is
3032returned.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3033
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3034</ul>
3035
3036</div>
3037
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3038
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3039<!-- ======================================================================= -->
3040<div class="doc_subsection">
3041 <a name="Argument">The <tt>Argument</tt> class</a>
3042</div>
3043
3044<div class="doc_text">
3045
3046<p>This subclass of Value defines the interface for incoming formal
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]3047arguments to a function. A Function maintains a list of its formal
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3048arguments. An argument has a pointer to the parent Function.</p>
3049
3050</div>
3051
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3052<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3053<hr>
3054<address>
3055 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
3056 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
3057 <a href="http://validator.w3.org/check/referer"><img
3058 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
3059
3060 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
3061 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]3062 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3063 Last modified: $Date$
3064</address>
3065
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3066</body>
3067</html>