Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / da021aa4104db5eef83e08350eb065af259cd8e9 / . / docs / ProgrammersManual.html
blob: 716d364ed5699a2c73ca2c84ca93dbf3aee094be [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>
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]65 <li><a href="#dss_denseset">"llvm/ADT/DenseSet.h"</a></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]66 <li><a href="#dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a></li>
67 <li><a href="#dss_set">&lt;set&gt;</a></li>
68 <li><a href="#dss_setvector">"llvm/ADT/SetVector.h"</a></li>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]69 <li><a href="#dss_uniquevector">"llvm/ADT/UniqueVector.h"</a></li>
70 <li><a href="#dss_otherset">Other Set-Like ContainerOptions</a></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]71 </ul></li>
Chris Lattnerf3692522007-02-03 19:51:56 +0000[diff] [blame]72 <li><a href="#ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
73 <ul>
74 <li><a href="#dss_sortedvectormap">A sorted 'vector'</a></li>
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]75 <li><a href="#dss_stringmap">"llvm/ADT/StringMap.h"</a></li>
Chris Lattnerf3692522007-02-03 19:51:56 +0000[diff] [blame]76 <li><a href="#dss_indexedmap">"llvm/ADT/IndexedMap.h"</a></li>
77 <li><a href="#dss_densemap">"llvm/ADT/DenseMap.h"</a></li>
78 <li><a href="#dss_map">&lt;map&gt;</a></li>
79 <li><a href="#dss_othermap">Other Map-Like Container Options</a></li>
80 </ul></li>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]81 <li><a href="#ds_bit">BitVector-like containers</a>
82 <ul>
83 <li><a href="#dss_bitvector">A dense bitvector</a></li>
84 <li><a href="#dss_sparsebitvector">A sparse bitvector</a></li>
85 </ul></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]86 </ul>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]87 </li>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]88 <li><a href="#common">Helpful Hints for Common Operations</a>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]89 <ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]90 <li><a href="#inspection">Basic Inspection and Traversal Routines</a>
91 <ul>
92 <li><a href="#iterate_function">Iterating over the <tt>BasicBlock</tt>s
93in a <tt>Function</tt></a> </li>
94 <li><a href="#iterate_basicblock">Iterating over the <tt>Instruction</tt>s
95in a <tt>BasicBlock</tt></a> </li>
96 <li><a href="#iterate_institer">Iterating over the <tt>Instruction</tt>s
97in a <tt>Function</tt></a> </li>
98 <li><a href="#iterate_convert">Turning an iterator into a
99class pointer</a> </li>
100 <li><a href="#iterate_complex">Finding call sites: a more
101complex example</a> </li>
102 <li><a href="#calls_and_invokes">Treating calls and invokes
103the same way</a> </li>
104 <li><a href="#iterate_chains">Iterating over def-use &amp;
105use-def chains</a> </li>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]106 <li><a href="#iterate_preds">Iterating over predecessors &amp;
107successors of blocks</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]108 </ul>
109 </li>
110 <li><a href="#simplechanges">Making simple changes</a>
111 <ul>
112 <li><a href="#schanges_creating">Creating and inserting new
113 <tt>Instruction</tt>s</a> </li>
114 <li><a href="#schanges_deleting">Deleting <tt>Instruction</tt>s</a> </li>
115 <li><a href="#schanges_replacing">Replacing an <tt>Instruction</tt>
116with another <tt>Value</tt></a> </li>
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]117 <li><a href="#schanges_deletingGV">Deleting <tt>GlobalVariable</tt>s</a> </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]118 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]119 </li>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]120<!--
121 <li>Working with the Control Flow Graph
122 <ul>
123 <li>Accessing predecessors and successors of a <tt>BasicBlock</tt>
124 <li>
125 <li>
126 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]127-->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]128 </ul>
129 </li>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]130
131 <li><a href="#advanced">Advanced Topics</a>
132 <ul>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]133 <li><a href="#TypeResolve">LLVM Type Resolution</a>
134 <ul>
135 <li><a href="#BuildRecType">Basic Recursive Type Construction</a></li>
136 <li><a href="#refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a></li>
137 <li><a href="#PATypeHolder">The PATypeHolder Class</a></li>
138 <li><a href="#AbstractTypeUser">The AbstractTypeUser Class</a></li>
139 </ul></li>
140
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]141 <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]142 </ul></li>
143
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]144 <li><a href="#coreclasses">The Core LLVM Class Hierarchy Reference</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]145 <ul>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]146 <li><a href="#Type">The <tt>Type</tt> class</a> </li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]147 <li><a href="#Module">The <tt>Module</tt> class</a></li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]148 <li><a href="#Value">The <tt>Value</tt> class</a>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]149 <ul>
150 <li><a href="#User">The <tt>User</tt> class</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]151 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]152 <li><a href="#Instruction">The <tt>Instruction</tt> class</a></li>
153 <li><a href="#Constant">The <tt>Constant</tt> class</a>
154 <ul>
155 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]156 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]157 <li><a href="#Function">The <tt>Function</tt> class</a></li>
158 <li><a href="#GlobalVariable">The <tt>GlobalVariable</tt> class</a></li>
159 </ul>
160 </li>
161 </ul>
162 </li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]163 </ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]164 </li>
165 <li><a href="#BasicBlock">The <tt>BasicBlock</tt> class</a></li>
166 <li><a href="#Argument">The <tt>Argument</tt> class</a></li>
167 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]168 </li>
169 </ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]170 </li>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]171</ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]172
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]173<div class="doc_author">
174 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
Chris Lattner94c43592004-05-26 16:52:55 +0000[diff] [blame]175 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>,
176 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a>, and
177 <a href="mailto:rspencer@x10sys.com">Reid Spencer</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]178</div>
179
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]180<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]181<div class="doc_section">
182 <a name="introduction">Introduction </a>
183</div>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]184<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]185
186<div class="doc_text">
187
188<p>This document is meant to highlight some of the important classes and
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]189interfaces available in the LLVM source-base. This manual is not
190intended to explain what LLVM is, how it works, and what LLVM code looks
191like. It assumes that you know the basics of LLVM and are interested
192in writing transformations or otherwise analyzing or manipulating the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]193code.</p>
194
195<p>This document should get you oriented so that you can find your
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]196way in the continuously growing source code that makes up the LLVM
197infrastructure. Note that this manual is not intended to serve as a
198replacement for reading the source code, so if you think there should be
199a method in one of these classes to do something, but it's not listed,
200check the source. Links to the <a href="/doxygen/">doxygen</a> sources
201are provided to make this as easy as possible.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]202
203<p>The first section of this document describes general information that is
204useful to know when working in the LLVM infrastructure, and the second describes
205the Core LLVM classes. In the future this manual will be extended with
206information describing how to use extension libraries, such as dominator
207information, CFG traversal routines, and useful utilities like the <tt><a
208href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.</p>
209
210</div>
211
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]212<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]213<div class="doc_section">
214 <a name="general">General Information</a>
215</div>
216<!-- *********************************************************************** -->
217
218<div class="doc_text">
219
220<p>This section contains general information that is useful if you are working
221in the LLVM source-base, but that isn't specific to any particular API.</p>
222
223</div>
224
225<!-- ======================================================================= -->
226<div class="doc_subsection">
227 <a name="stl">The C++ Standard Template Library</a>
228</div>
229
230<div class="doc_text">
231
232<p>LLVM makes heavy use of the C++ Standard Template Library (STL),
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]233perhaps much more than you are used to, or have seen before. Because of
234this, you might want to do a little background reading in the
235techniques used and capabilities of the library. There are many good
236pages that discuss the STL, and several books on the subject that you
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]237can get, so it will not be discussed in this document.</p>
238
239<p>Here are some useful links:</p>
240
241<ol>
242
243<li><a href="http://www.dinkumware.com/refxcpp.html">Dinkumware C++ Library
244reference</a> - an excellent reference for the STL and other parts of the
245standard C++ library.</li>
246
247<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]248O'Reilly book in the making. It has a decent
249Standard Library
250Reference 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]251published.</li>
252
253<li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
254Questions</a></li>
255
256<li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
257Contains a useful <a
258href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
259STL</a>.</li>
260
261<li><a href="http://www.research.att.com/%7Ebs/C++.html">Bjarne Stroustrup's C++
262Page</a></li>
263
Tanya Lattner79445ba2004-12-08 18:34:56 +0000[diff] [blame]264<li><a href="http://64.78.49.204/">
Reid Spencer096603a2004-05-26 08:41:35 +0000[diff] [blame]265Bruce Eckel's Thinking in C++, 2nd ed. Volume 2 Revision 4.0 (even better, get
266the book).</a></li>
267
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]268</ol>
269
270<p>You are also encouraged to take a look at the <a
271href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
272to write maintainable code more than where to put your curly braces.</p>
273
274</div>
275
276<!-- ======================================================================= -->
277<div class="doc_subsection">
278 <a name="stl">Other useful references</a>
279</div>
280
281<div class="doc_text">
282
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]283<ol>
284<li><a href="http://www.psc.edu/%7Esemke/cvs_branches.html">CVS
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]285Branch and Tag Primer</a></li>
Misha Brukmana0f71e42004-06-18 18:39:00 +0000[diff] [blame]286<li><a href="http://www.fortran-2000.com/ArnaudRecipes/sharedlib.html">Using
287static and shared libraries across platforms</a></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]288</ol>
289
290</div>
291
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]292<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]293<div class="doc_section">
294 <a name="apis">Important and useful LLVM APIs</a>
295</div>
296<!-- *********************************************************************** -->
297
298<div class="doc_text">
299
300<p>Here we highlight some LLVM APIs that are generally useful and good to
301know about when writing transformations.</p>
302
303</div>
304
305<!-- ======================================================================= -->
306<div class="doc_subsection">
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]307 <a name="isa">The <tt>isa&lt;&gt;</tt>, <tt>cast&lt;&gt;</tt> and
308 <tt>dyn_cast&lt;&gt;</tt> templates</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]309</div>
310
311<div class="doc_text">
312
313<p>The LLVM source-base makes extensive use of a custom form of RTTI.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]314These templates have many similarities to the C++ <tt>dynamic_cast&lt;&gt;</tt>
315operator, but they don't have some drawbacks (primarily stemming from
316the fact that <tt>dynamic_cast&lt;&gt;</tt> only works on classes that
317have a v-table). Because they are used so often, you must know what they
318do and how they work. All of these templates are defined in the <a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]319 href="/doxygen/Casting_8h-source.html"><tt>llvm/Support/Casting.h</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]320file (note that you very rarely have to include this file directly).</p>
321
322<dl>
323 <dt><tt>isa&lt;&gt;</tt>: </dt>
324
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]325 <dd><p>The <tt>isa&lt;&gt;</tt> operator works exactly like the Java
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]326 "<tt>instanceof</tt>" operator. It returns true or false depending on whether
327 a reference or pointer points to an instance of the specified class. This can
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]328 be very useful for constraint checking of various sorts (example below).</p>
329 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]330
331 <dt><tt>cast&lt;&gt;</tt>: </dt>
332
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]333 <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]334 converts a pointer or reference from a base class to a derived cast, causing
335 an assertion failure if it is not really an instance of the right type. This
336 should be used in cases where you have some information that makes you believe
337 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]338 and <tt>cast&lt;&gt;</tt> template is:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]339
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]340<div class="doc_code">
341<pre>
342static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
343 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))
344 return true;
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]345
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]346 // <i>Otherwise, it must be an instruction...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]347 return !L-&gt;contains(cast&lt;<a href="#Instruction">Instruction</a>&gt;(V)-&gt;getParent());
348}
349</pre>
350</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]351
352 <p>Note that you should <b>not</b> use an <tt>isa&lt;&gt;</tt> test followed
353 by a <tt>cast&lt;&gt;</tt>, for that use the <tt>dyn_cast&lt;&gt;</tt>
354 operator.</p>
355
356 </dd>
357
358 <dt><tt>dyn_cast&lt;&gt;</tt>:</dt>
359
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]360 <dd><p>The <tt>dyn_cast&lt;&gt;</tt> operator is a "checking cast" operation.
361 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]362 pointer to it (this operator does not work with references). If the operand is
363 not of the correct type, a null pointer is returned. Thus, this works very
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]364 much like the <tt>dynamic_cast&lt;&gt;</tt> operator in C++, and should be
365 used in the same circumstances. Typically, the <tt>dyn_cast&lt;&gt;</tt>
366 operator is used in an <tt>if</tt> statement or some other flow control
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]367 statement like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]368
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]369<div class="doc_code">
370<pre>
371if (<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]372 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]373}
374</pre>
375</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]376
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]377 <p>This form of the <tt>if</tt> statement effectively combines together a call
378 to <tt>isa&lt;&gt;</tt> and a call to <tt>cast&lt;&gt;</tt> into one
379 statement, which is very convenient.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]380
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]381 <p>Note that the <tt>dyn_cast&lt;&gt;</tt> operator, like C++'s
382 <tt>dynamic_cast&lt;&gt;</tt> or Java's <tt>instanceof</tt> operator, can be
383 abused. In particular, you should not use big chained <tt>if/then/else</tt>
384 blocks to check for lots of different variants of classes. If you find
385 yourself wanting to do this, it is much cleaner and more efficient to use the
386 <tt>InstVisitor</tt> class to dispatch over the instruction type directly.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]387
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]388 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]389
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]390 <dt><tt>cast_or_null&lt;&gt;</tt>: </dt>
391
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]392 <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]393 <tt>cast&lt;&gt;</tt> operator, except that it allows for a null pointer as an
394 argument (which it then propagates). This can sometimes be useful, allowing
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]395 you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]396
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]397 <dt><tt>dyn_cast_or_null&lt;&gt;</tt>: </dt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]398
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]399 <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]400 <tt>dyn_cast&lt;&gt;</tt> operator, except that it allows for a null pointer
401 as an argument (which it then propagates). This can sometimes be useful,
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]402 allowing you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]403
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]404</dl>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]405
406<p>These five templates can be used with any classes, whether they have a
407v-table or not. To add support for these templates, you simply need to add
408<tt>classof</tt> static methods to the class you are interested casting
409to. Describing this is currently outside the scope of this document, but there
410are lots of examples in the LLVM source base.</p>
411
412</div>
413
414<!-- ======================================================================= -->
415<div class="doc_subsection">
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]416 <a name="DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt> option</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]417</div>
418
419<div class="doc_text">
420
421<p>Often when working on your pass you will put a bunch of debugging printouts
422and other code into your pass. After you get it working, you want to remove
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]423it, 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]424across).</p>
425
426<p> Naturally, because of this, you don't want to delete the debug printouts,
427but you don't want them to always be noisy. A standard compromise is to comment
428them out, allowing you to enable them if you need them in the future.</p>
429
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]430<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]431file provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to
432this problem. Basically, you can put arbitrary code into the argument of the
433<tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
434tool) is run with the '<tt>-debug</tt>' command line argument:</p>
435
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]436<div class="doc_code">
437<pre>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]438DOUT &lt;&lt; "I am here!\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]439</pre>
440</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]441
442<p>Then you can run your pass like this:</p>
443
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]444<div class="doc_code">
445<pre>
446$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]447<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]448$ opt &lt; a.bc &gt; /dev/null -mypass -debug
449I am here!
450</pre>
451</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]452
453<p>Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you
454to not have to create "yet another" command line option for the debug output for
455your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
456so they do not cause a performance impact at all (for the same reason, they
457should also not contain side-effects!).</p>
458
459<p>One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
460enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
461"<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
462program hasn't been started yet, you can always just run it with
463<tt>-debug</tt>.</p>
464
465</div>
466
467<!-- _______________________________________________________________________ -->
468<div class="doc_subsubsection">
Chris Lattnerc9151082005-04-26 22:57:07 +0000[diff] [blame]469 <a name="DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt> and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]470 the <tt>-debug-only</tt> option</a>
471</div>
472
473<div class="doc_text">
474
475<p>Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
476just turns on <b>too much</b> information (such as when working on the code
477generator). If you want to enable debug information with more fine-grained
478control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
479option as follows:</p>
480
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]481<div class="doc_code">
482<pre>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]483DOUT &lt;&lt; "No debug type\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]484#undef DEBUG_TYPE
485#define DEBUG_TYPE "foo"
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]486DOUT &lt;&lt; "'foo' debug type\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]487#undef DEBUG_TYPE
488#define DEBUG_TYPE "bar"
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]489DOUT &lt;&lt; "'bar' debug type\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]490#undef DEBUG_TYPE
491#define DEBUG_TYPE ""
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]492DOUT &lt;&lt; "No debug type (2)\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]493</pre>
494</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]495
496<p>Then you can run your pass like this:</p>
497
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]498<div class="doc_code">
499<pre>
500$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]501<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]502$ opt &lt; a.bc &gt; /dev/null -mypass -debug
503No debug type
504'foo' debug type
505'bar' debug type
506No debug type (2)
507$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=foo
508'foo' debug type
509$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=bar
510'bar' debug type
511</pre>
512</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]513
514<p>Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of
515a 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]516you <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]517<tt>#undef</tt>'s). Also, you should use names more meaningful than "foo" and
518"bar", because there is no system in place to ensure that names do not
519conflict. If two different modules use the same string, they will all be turned
520on when the name is specified. This allows, for example, all debug information
521for instruction scheduling to be enabled with <tt>-debug-type=InstrSched</tt>,
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]522even if the source lives in multiple files.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]523
524</div>
525
526<!-- ======================================================================= -->
527<div class="doc_subsection">
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]528 <a name="Statistic">The <tt>Statistic</tt> class &amp; <tt>-stats</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]529 option</a>
530</div>
531
532<div class="doc_text">
533
534<p>The "<tt><a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]535href="/doxygen/Statistic_8h-source.html">llvm/ADT/Statistic.h</a></tt>" file
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]536provides a class named <tt>Statistic</tt> that is used as a unified way to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]537keep track of what the LLVM compiler is doing and how effective various
538optimizations are. It is useful to see what optimizations are contributing to
539making a particular program run faster.</p>
540
541<p>Often you may run your pass on some big program, and you're interested to see
542how many times it makes a certain transformation. Although you can do this with
543hand 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]544for big programs. Using the <tt>Statistic</tt> class makes it very easy to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]545keep track of this information, and the calculated information is presented in a
546uniform manner with the rest of the passes being executed.</p>
547
548<p>There are many examples of <tt>Statistic</tt> uses, but the basics of using
549it are as follows:</p>
550
551<ol>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]552 <li><p>Define your statistic like this:</p>
553
554<div class="doc_code">
555<pre>
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]556#define <a href="#DEBUG_TYPE">DEBUG_TYPE</a> "mypassname" <i>// This goes before any #includes.</i>
557STATISTIC(NumXForms, "The # of times I did stuff");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]558</pre>
559</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]560
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]561 <p>The <tt>STATISTIC</tt> macro defines a static variable, whose name is
562 specified by the first argument. The pass name is taken from the DEBUG_TYPE
563 macro, and the description is taken from the second argument. The variable
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]564 defined ("NumXForms" in this case) acts like an unsigned integer.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]565
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]566 <li><p>Whenever you make a transformation, bump the counter:</p>
567
568<div class="doc_code">
569<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]570++NumXForms; // <i>I did stuff!</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]571</pre>
572</div>
573
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]574 </li>
575 </ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]576
577 <p>That's all you have to do. To get '<tt>opt</tt>' to print out the
578 statistics gathered, use the '<tt>-stats</tt>' option:</p>
579
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]580<div class="doc_code">
581<pre>
582$ opt -stats -mypassname &lt; program.bc &gt; /dev/null
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]583<i>... statistics output ...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]584</pre>
585</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]586
Reid Spencer6b6c73e2007-02-09 16:00:28 +0000[diff] [blame]587 <p> When running <tt>opt</tt> on a C file from the SPEC benchmark
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]588suite, it gives a report that looks like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]589
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]590<div class="doc_code">
591<pre>
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]592 7646 bitcodewriter - Number of normal instructions
593 725 bitcodewriter - Number of oversized instructions
594 129996 bitcodewriter - Number of bitcode bytes written
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]595 2817 raise - Number of insts DCEd or constprop'd
596 3213 raise - Number of cast-of-self removed
597 5046 raise - Number of expression trees converted
598 75 raise - Number of other getelementptr's formed
599 138 raise - Number of load/store peepholes
600 42 deadtypeelim - Number of unused typenames removed from symtab
601 392 funcresolve - Number of varargs functions resolved
602 27 globaldce - Number of global variables removed
603 2 adce - Number of basic blocks removed
604 134 cee - Number of branches revectored
605 49 cee - Number of setcc instruction eliminated
606 532 gcse - Number of loads removed
607 2919 gcse - Number of instructions removed
608 86 indvars - Number of canonical indvars added
609 87 indvars - Number of aux indvars removed
610 25 instcombine - Number of dead inst eliminate
611 434 instcombine - Number of insts combined
612 248 licm - Number of load insts hoisted
613 1298 licm - Number of insts hoisted to a loop pre-header
614 3 licm - Number of insts hoisted to multiple loop preds (bad, no loop pre-header)
615 75 mem2reg - Number of alloca's promoted
616 1444 cfgsimplify - Number of blocks simplified
617</pre>
618</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]619
620<p>Obviously, with so many optimizations, having a unified framework for this
621stuff is very nice. Making your pass fit well into the framework makes it more
622maintainable and useful.</p>
623
624</div>
625
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]626<!-- ======================================================================= -->
627<div class="doc_subsection">
628 <a name="ViewGraph">Viewing graphs while debugging code</a>
629</div>
630
631<div class="doc_text">
632
633<p>Several of the important data structures in LLVM are graphs: for example
634CFGs made out of LLVM <a href="#BasicBlock">BasicBlock</a>s, CFGs made out of
635LLVM <a href="CodeGenerator.html#machinebasicblock">MachineBasicBlock</a>s, and
636<a href="CodeGenerator.html#selectiondag_intro">Instruction Selection
637DAGs</a>. In many cases, while debugging various parts of the compiler, it is
638nice to instantly visualize these graphs.</p>
639
640<p>LLVM provides several callbacks that are available in a debug build to do
641exactly that. If you call the <tt>Function::viewCFG()</tt> method, for example,
642the current LLVM tool will pop up a window containing the CFG for the function
643where each basic block is a node in the graph, and each node contains the
644instructions in the block. Similarly, there also exists
645<tt>Function::viewCFGOnly()</tt> (does not include the instructions), the
646<tt>MachineFunction::viewCFG()</tt> and <tt>MachineFunction::viewCFGOnly()</tt>,
647and the <tt>SelectionDAG::viewGraph()</tt> methods. Within GDB, for example,
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]648you can usually use something like <tt>call DAG.viewGraph()</tt> to pop
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]649up a window. Alternatively, you can sprinkle calls to these functions in your
650code in places you want to debug.</p>
651
652<p>Getting this to work requires a small amount of configuration. On Unix
653systems with X11, install the <a href="http://www.graphviz.org">graphviz</a>
654toolkit, and make sure 'dot' and 'gv' are in your path. If you are running on
655Mac OS/X, download and install the Mac OS/X <a
656href="http://www.pixelglow.com/graphviz/">Graphviz program</a>, and add
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]657<tt>/Applications/Graphviz.app/Contents/MacOS/</tt> (or wherever you install
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]658it) to your path. Once in your system and path are set up, rerun the LLVM
659configure script and rebuild LLVM to enable this functionality.</p>
660
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]661<p><tt>SelectionDAG</tt> has been extended to make it easier to locate
662<i>interesting</i> nodes in large complex graphs. From gdb, if you
663<tt>call DAG.setGraphColor(<i>node</i>, "<i>color</i>")</tt>, then the
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]664next <tt>call DAG.viewGraph()</tt> would highlight the node in the
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]665specified color (choices of colors can be found at <a
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]666href="http://www.graphviz.org/doc/info/colors.html">colors</a>.) More
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]667complex node attributes can be provided with <tt>call
668DAG.setGraphAttrs(<i>node</i>, "<i>attributes</i>")</tt> (choices can be
669found at <a href="http://www.graphviz.org/doc/info/attrs.html">Graph
670Attributes</a>.) If you want to restart and clear all the current graph
671attributes, then you can <tt>call DAG.clearGraphAttrs()</tt>. </p>
672
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]673</div>
674
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]675<!-- *********************************************************************** -->
676<div class="doc_section">
677 <a name="datastructure">Picking the Right Data Structure for a Task</a>
678</div>
679<!-- *********************************************************************** -->
680
681<div class="doc_text">
682
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]683<p>LLVM has a plethora of data structures in the <tt>llvm/ADT/</tt> directory,
684 and we commonly use STL data structures. This section describes the trade-offs
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]685 you should consider when you pick one.</p>
686
687<p>
688The first step is a choose your own adventure: do you want a sequential
689container, a set-like container, or a map-like container? The most important
690thing when choosing a container is the algorithmic properties of how you plan to
691access the container. Based on that, you should use:</p>
692
693<ul>
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]694<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]695 of an value based on another value. Map-like containers also support
696 efficient queries for containment (whether a key is in the map). Map-like
697 containers generally do not support efficient reverse mapping (values to
698 keys). If you need that, use two maps. Some map-like containers also
699 support efficient iteration through the keys in sorted order. Map-like
700 containers are the most expensive sort, only use them if you need one of
701 these capabilities.</li>
702
703<li>a <a href="#ds_set">set-like</a> container if you need to put a bunch of
704 stuff into a container that automatically eliminates duplicates. Some
705 set-like containers support efficient iteration through the elements in
706 sorted order. Set-like containers are more expensive than sequential
707 containers.
708</li>
709
710<li>a <a href="#ds_sequential">sequential</a> container provides
711 the most efficient way to add elements and keeps track of the order they are
712 added to the collection. They permit duplicates and support efficient
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]713 iteration, but do not support efficient look-up based on a key.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]714</li>
715
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]716<li>a <a href="#ds_bit">bit</a> container provides an efficient way to store and
717 perform set operations on sets of numeric id's, while automatically
718 eliminating duplicates. Bit containers require a maximum of 1 bit for each
719 identifier you want to store.
720</li>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]721</ul>
722
723<p>
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]724Once the proper category of container is determined, you can fine tune the
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]725memory use, constant factors, and cache behaviors of access by intelligently
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]726picking a member of the category. Note that constant factors and cache behavior
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]727can be a big deal. If you have a vector that usually only contains a few
728elements (but could contain many), for example, it's much better to use
729<a href="#dss_smallvector">SmallVector</a> than <a href="#dss_vector">vector</a>
730. Doing so avoids (relatively) expensive malloc/free calls, which dwarf the
731cost of adding the elements to the container. </p>
732
733</div>
734
735<!-- ======================================================================= -->
736<div class="doc_subsection">
737 <a name="ds_sequential">Sequential Containers (std::vector, std::list, etc)</a>
738</div>
739
740<div class="doc_text">
741There are a variety of sequential containers available for you, based on your
742needs. Pick the first in this section that will do what you want.
743</div>
744
745<!-- _______________________________________________________________________ -->
746<div class="doc_subsubsection">
747 <a name="dss_fixedarrays">Fixed Size Arrays</a>
748</div>
749
750<div class="doc_text">
751<p>Fixed size arrays are very simple and very fast. They are good if you know
752exactly how many elements you have, or you have a (low) upper bound on how many
753you have.</p>
754</div>
755
756<!-- _______________________________________________________________________ -->
757<div class="doc_subsubsection">
758 <a name="dss_heaparrays">Heap Allocated Arrays</a>
759</div>
760
761<div class="doc_text">
762<p>Heap allocated arrays (new[] + delete[]) are also simple. They are good if
763the number of elements is variable, if you know how many elements you will need
764before the array is allocated, and if the array is usually large (if not,
765consider a <a href="#dss_smallvector">SmallVector</a>). The cost of a heap
766allocated array is the cost of the new/delete (aka malloc/free). Also note that
767if you are allocating an array of a type with a constructor, the constructor and
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]768destructors will be run for every element in the array (re-sizable vectors only
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]769construct those elements actually used).</p>
770</div>
771
772<!-- _______________________________________________________________________ -->
773<div class="doc_subsubsection">
774 <a name="dss_smallvector">"llvm/ADT/SmallVector.h"</a>
775</div>
776
777<div class="doc_text">
778<p><tt>SmallVector&lt;Type, N&gt;</tt> is a simple class that looks and smells
779just like <tt>vector&lt;Type&gt;</tt>:
780it supports efficient iteration, lays out elements in memory order (so you can
781do pointer arithmetic between elements), supports efficient push_back/pop_back
782operations, supports efficient random access to its elements, etc.</p>
783
784<p>The advantage of SmallVector is that it allocates space for
785some number of elements (N) <b>in the object itself</b>. Because of this, if
786the SmallVector is dynamically smaller than N, no malloc is performed. This can
787be a big win in cases where the malloc/free call is far more expensive than the
788code that fiddles around with the elements.</p>
789
790<p>This is good for vectors that are "usually small" (e.g. the number of
791predecessors/successors of a block is usually less than 8). On the other hand,
792this makes the size of the SmallVector itself large, so you don't want to
793allocate lots of them (doing so will waste a lot of space). As such,
794SmallVectors are most useful when on the stack.</p>
795
796<p>SmallVector also provides a nice portable and efficient replacement for
797<tt>alloca</tt>.</p>
798
799</div>
800
801<!-- _______________________________________________________________________ -->
802<div class="doc_subsubsection">
803 <a name="dss_vector">&lt;vector&gt;</a>
804</div>
805
806<div class="doc_text">
807<p>
808std::vector is well loved and respected. It is useful when SmallVector isn't:
809when the size of the vector is often large (thus the small optimization will
810rarely be a benefit) or if you will be allocating many instances of the vector
811itself (which would waste space for elements that aren't in the container).
812vector is also useful when interfacing with code that expects vectors :).
813</p>
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]814
815<p>One worthwhile note about std::vector: avoid code like this:</p>
816
817<div class="doc_code">
818<pre>
819for ( ... ) {
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]820 std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]821 use V;
822}
823</pre>
824</div>
825
826<p>Instead, write this as:</p>
827
828<div class="doc_code">
829<pre>
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]830std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]831for ( ... ) {
832 use V;
833 V.clear();
834}
835</pre>
836</div>
837
838<p>Doing so will save (at least) one heap allocation and free per iteration of
839the loop.</p>
840
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]841</div>
842
843<!-- _______________________________________________________________________ -->
844<div class="doc_subsubsection">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]845 <a name="dss_deque">&lt;deque&gt;</a>
846</div>
847
848<div class="doc_text">
849<p>std::deque is, in some senses, a generalized version of std::vector. Like
850std::vector, it provides constant time random access and other similar
851properties, but it also provides efficient access to the front of the list. It
852does not guarantee continuity of elements within memory.</p>
853
854<p>In exchange for this extra flexibility, std::deque has significantly higher
855constant factor costs than std::vector. If possible, use std::vector or
856something cheaper.</p>
857</div>
858
859<!-- _______________________________________________________________________ -->
860<div class="doc_subsubsection">
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]861 <a name="dss_list">&lt;list&gt;</a>
862</div>
863
864<div class="doc_text">
865<p>std::list is an extremely inefficient class that is rarely useful.
866It performs a heap allocation for every element inserted into it, thus having an
867extremely high constant factor, particularly for small data types. std::list
868also only supports bidirectional iteration, not random access iteration.</p>
869
870<p>In exchange for this high cost, std::list supports efficient access to both
871ends of the list (like std::deque, but unlike std::vector or SmallVector). In
872addition, the iterator invalidation characteristics of std::list are stronger
873than that of a vector class: inserting or removing an element into the list does
874not invalidate iterator or pointers to other elements in the list.</p>
875</div>
876
877<!-- _______________________________________________________________________ -->
878<div class="doc_subsubsection">
879 <a name="dss_ilist">llvm/ADT/ilist</a>
880</div>
881
882<div class="doc_text">
883<p><tt>ilist&lt;T&gt;</tt> implements an 'intrusive' doubly-linked list. It is
884intrusive, because it requires the element to store and provide access to the
885prev/next pointers for the list.</p>
886
887<p>ilist has the same drawbacks as std::list, and additionally requires an
888ilist_traits implementation for the element type, but it provides some novel
889characteristics. In particular, it can efficiently store polymorphic objects,
890the traits class is informed when an element is inserted or removed from the
891list, and ilists are guaranteed to support a constant-time splice operation.
892</p>
893
894<p>These properties are exactly what we want for things like Instructions and
895basic blocks, which is why these are implemented with ilists.</p>
896</div>
897
898<!-- _______________________________________________________________________ -->
899<div class="doc_subsubsection">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]900 <a name="dss_other">Other Sequential Container options</a>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]901</div>
902
903<div class="doc_text">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]904<p>Other STL containers are available, such as std::string.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]905
906<p>There are also various STL adapter classes such as std::queue,
907std::priority_queue, std::stack, etc. These provide simplified access to an
908underlying container but don't affect the cost of the container itself.</p>
909
910</div>
911
912
913<!-- ======================================================================= -->
914<div class="doc_subsection">
915 <a name="ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
916</div>
917
918<div class="doc_text">
919
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]920<p>Set-like containers are useful when you need to canonicalize multiple values
921into a single representation. There are several different choices for how to do
922this, providing various trade-offs.</p>
923
924</div>
925
926
927<!-- _______________________________________________________________________ -->
928<div class="doc_subsubsection">
929 <a name="dss_sortedvectorset">A sorted 'vector'</a>
930</div>
931
932<div class="doc_text">
933
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]934<p>If you intend to insert a lot of elements, then do a lot of queries, a
935great approach is to use a vector (or other sequential container) with
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]936std::sort+std::unique to remove duplicates. This approach works really well if
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]937your usage pattern has these two distinct phases (insert then query), and can be
938coupled with a good choice of <a href="#ds_sequential">sequential container</a>.
939</p>
940
941<p>
942This combination provides the several nice properties: the result data is
943contiguous in memory (good for cache locality), has few allocations, is easy to
944address (iterators in the final vector are just indices or pointers), and can be
945efficiently queried with a standard binary or radix search.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]946
947</div>
948
949<!-- _______________________________________________________________________ -->
950<div class="doc_subsubsection">
951 <a name="dss_smallset">"llvm/ADT/SmallSet.h"</a>
952</div>
953
954<div class="doc_text">
955
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]956<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]957are 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]958has space for N elements in place (thus, if the set is dynamically smaller than
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]959N, no malloc traffic is required) and accesses them with a simple linear search.
960When the set grows beyond 'N' elements, it allocates a more expensive representation that
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]961guarantees efficient access (for most types, it falls back to std::set, but for
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]962pointers it uses something far better, <a
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]963href="#dss_smallptrset">SmallPtrSet</a>).</p>
964
965<p>The magic of this class is that it handles small sets extremely efficiently,
966but gracefully handles extremely large sets without loss of efficiency. The
967drawback is that the interface is quite small: it supports insertion, queries
968and erasing, but does not support iteration.</p>
969
970</div>
971
972<!-- _______________________________________________________________________ -->
973<div class="doc_subsubsection">
974 <a name="dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a>
975</div>
976
977<div class="doc_text">
978
979<p>SmallPtrSet has all the advantages of SmallSet (and a SmallSet of pointers is
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]980transparently implemented with a SmallPtrSet), but also supports iterators. If
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]981more than 'N' insertions are performed, a single quadratically
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]982probed hash table is allocated and grows as needed, providing extremely
983efficient access (constant time insertion/deleting/queries with low constant
984factors) and is very stingy with malloc traffic.</p>
985
986<p>Note that, unlike std::set, the iterators of SmallPtrSet are invalidated
987whenever an insertion occurs. Also, the values visited by the iterators are not
988visited in sorted order.</p>
989
990</div>
991
992<!-- _______________________________________________________________________ -->
993<div class="doc_subsubsection">
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]994 <a name="dss_denseset">"llvm/ADT/DenseSet.h"</a>
995</div>
996
997<div class="doc_text">
998
999<p>
1000DenseSet is a simple quadratically probed hash table. It excels at supporting
1001small values: it uses a single allocation to hold all of the pairs that
1002are currently inserted in the set. DenseSet is a great way to unique small
1003values that are not simple pointers (use <a
1004href="#dss_smallptrset">SmallPtrSet</a> for pointers). Note that DenseSet has
1005the same requirements for the value type that <a
1006href="#dss_densemap">DenseMap</a> has.
1007</p>
1008
1009</div>
1010
1011<!-- _______________________________________________________________________ -->
1012<div class="doc_subsubsection">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1013 <a name="dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a>
1014</div>
1015
1016<div class="doc_text">
1017
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1018<p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1019FoldingSet is an aggregate class that is really good at uniquing
1020expensive-to-create or polymorphic objects. It is a combination of a chained
1021hash table with intrusive links (uniqued objects are required to inherit from
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1022FoldingSetNode) that uses <a href="#dss_smallvector">SmallVector</a> as part of
1023its ID process.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1024
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1025<p>Consider a case where you want to implement a "getOrCreateFoo" method for
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1026a complex object (for example, a node in the code generator). The client has a
1027description of *what* it wants to generate (it knows the opcode and all the
1028operands), 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]1029only to find out it already exists, at which point we would have to delete it
1030and return the node that already exists.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1031</p>
1032
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1033<p>To support this style of client, FoldingSet perform a query with a
1034FoldingSetNodeID (which wraps SmallVector) that can be used to describe the
1035element that we want to query for. The query either returns the element
1036matching the ID or it returns an opaque ID that indicates where insertion should
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1037take place. Construction of the ID usually does not require heap traffic.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1038
1039<p>Because FoldingSet uses intrusive links, it can support polymorphic objects
1040in the set (for example, you can have SDNode instances mixed with LoadSDNodes).
1041Because the elements are individually allocated, pointers to the elements are
1042stable: inserting or removing elements does not invalidate any pointers to other
1043elements.
1044</p>
1045
1046</div>
1047
1048<!-- _______________________________________________________________________ -->
1049<div class="doc_subsubsection">
1050 <a name="dss_set">&lt;set&gt;</a>
1051</div>
1052
1053<div class="doc_text">
1054
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1055<p><tt>std::set</tt> is a reasonable all-around set class, which is decent at
1056many things but great at nothing. std::set allocates memory for each element
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1057inserted (thus it is very malloc intensive) and typically stores three pointers
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1058per element in the set (thus adding a large amount of per-element space
1059overhead). It offers guaranteed log(n) performance, which is not particularly
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1060fast from a complexity standpoint (particularly if the elements of the set are
1061expensive to compare, like strings), and has extremely high constant factors for
1062lookup, insertion and removal.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1063
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1064<p>The advantages of std::set are that its iterators are stable (deleting or
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1065inserting an element from the set does not affect iterators or pointers to other
1066elements) and that iteration over the set is guaranteed to be in sorted order.
1067If the elements in the set are large, then the relative overhead of the pointers
1068and malloc traffic is not a big deal, but if the elements of the set are small,
1069std::set is almost never a good choice.</p>
1070
1071</div>
1072
1073<!-- _______________________________________________________________________ -->
1074<div class="doc_subsubsection">
1075 <a name="dss_setvector">"llvm/ADT/SetVector.h"</a>
1076</div>
1077
1078<div class="doc_text">
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1079<p>LLVM's SetVector&lt;Type&gt; is an adapter class that combines your choice of
1080a set-like container along with a <a href="#ds_sequential">Sequential
1081Container</a>. The important property
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1082that this provides is efficient insertion with uniquing (duplicate elements are
1083ignored) with iteration support. It implements this by inserting elements into
1084both a set-like container and the sequential container, using the set-like
1085container for uniquing and the sequential container for iteration.
1086</p>
1087
1088<p>The difference between SetVector and other sets is that the order of
1089iteration is guaranteed to match the order of insertion into the SetVector.
1090This property is really important for things like sets of pointers. Because
1091pointer values are non-deterministic (e.g. vary across runs of the program on
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1092different machines), iterating over the pointers in the set will
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1093not be in a well-defined order.</p>
1094
1095<p>
1096The drawback of SetVector is that it requires twice as much space as a normal
1097set and has the sum of constant factors from the set-like container and the
1098sequential container that it uses. Use it *only* if you need to iterate over
1099the elements in a deterministic order. SetVector is also expensive to delete
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1100elements out of (linear time), unless you use it's "pop_back" method, which is
1101faster.
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1102</p>
1103
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1104<p>SetVector is an adapter class that defaults to using std::vector and std::set
1105for the underlying containers, so it is quite expensive. However,
1106<tt>"llvm/ADT/SetVector.h"</tt> also provides a SmallSetVector class, which
1107defaults to using a SmallVector and SmallSet of a specified size. If you use
1108this, and if your sets are dynamically smaller than N, you will save a lot of
1109heap traffic.</p>
1110
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1111</div>
1112
1113<!-- _______________________________________________________________________ -->
1114<div class="doc_subsubsection">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1115 <a name="dss_uniquevector">"llvm/ADT/UniqueVector.h"</a>
1116</div>
1117
1118<div class="doc_text">
1119
1120<p>
1121UniqueVector is similar to <a href="#dss_setvector">SetVector</a>, but it
1122retains a unique ID for each element inserted into the set. It internally
1123contains a map and a vector, and it assigns a unique ID for each value inserted
1124into the set.</p>
1125
1126<p>UniqueVector is very expensive: its cost is the sum of the cost of
1127maintaining both the map and vector, it has high complexity, high constant
1128factors, and produces a lot of malloc traffic. It should be avoided.</p>
1129
1130</div>
1131
1132
1133<!-- _______________________________________________________________________ -->
1134<div class="doc_subsubsection">
1135 <a name="dss_otherset">Other Set-Like Container Options</a>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1136</div>
1137
1138<div class="doc_text">
1139
1140<p>
1141The STL provides several other options, such as std::multiset and the various
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1142"hash_set" like containers (whether from C++ TR1 or from the SGI library).</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1143
1144<p>std::multiset is useful if you're not interested in elimination of
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1145duplicates, but has all the drawbacks of std::set. A sorted vector (where you
1146don't delete duplicate entries) or some other approach is almost always
1147better.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1148
1149<p>The various hash_set implementations (exposed portably by
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1150"llvm/ADT/hash_set") is a simple chained hashtable. This algorithm is as malloc
1151intensive as std::set (performing an allocation for each element inserted,
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1152thus having really high constant factors) but (usually) provides O(1)
1153insertion/deletion of elements. This can be useful if your elements are large
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1154(thus making the constant-factor cost relatively low) or if comparisons are
1155expensive. Element iteration does not visit elements in a useful order.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1156
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1157</div>
1158
1159<!-- ======================================================================= -->
1160<div class="doc_subsection">
1161 <a name="ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
1162</div>
1163
1164<div class="doc_text">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1165Map-like containers are useful when you want to associate data to a key. As
1166usual, there are a lot of different ways to do this. :)
1167</div>
1168
1169<!-- _______________________________________________________________________ -->
1170<div class="doc_subsubsection">
1171 <a name="dss_sortedvectormap">A sorted 'vector'</a>
1172</div>
1173
1174<div class="doc_text">
1175
1176<p>
1177If your usage pattern follows a strict insert-then-query approach, you can
1178trivially use the same approach as <a href="#dss_sortedvectorset">sorted vectors
1179for set-like containers</a>. The only difference is that your query function
1180(which uses std::lower_bound to get efficient log(n) lookup) should only compare
1181the key, not both the key and value. This yields the same advantages as sorted
1182vectors for sets.
1183</p>
1184</div>
1185
1186<!-- _______________________________________________________________________ -->
1187<div class="doc_subsubsection">
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1188 <a name="dss_stringmap">"llvm/ADT/StringMap.h"</a>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1189</div>
1190
1191<div class="doc_text">
1192
1193<p>
1194Strings are commonly used as keys in maps, and they are difficult to support
1195efficiently: they are variable length, inefficient to hash and compare when
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1196long, expensive to copy, etc. StringMap is a specialized container designed to
1197cope with these issues. It supports mapping an arbitrary range of bytes to an
1198arbitrary other object.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1199
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1200<p>The StringMap implementation uses a quadratically-probed hash table, where
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1201the buckets store a pointer to the heap allocated entries (and some other
1202stuff). The entries in the map must be heap allocated because the strings are
1203variable length. The string data (key) and the element object (value) are
1204stored in the same allocation with the string data immediately after the element
1205object. This container guarantees the "<tt>(char*)(&amp;Value+1)</tt>" points
1206to the key string for a value.</p>
1207
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1208<p>The StringMap is very fast for several reasons: quadratic probing is very
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1209cache efficient for lookups, the hash value of strings in buckets is not
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1210recomputed when lookup up an element, StringMap rarely has to touch the
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1211memory for unrelated objects when looking up a value (even when hash collisions
1212happen), hash table growth does not recompute the hash values for strings
1213already in the table, and each pair in the map is store in a single allocation
1214(the string data is stored in the same allocation as the Value of a pair).</p>
1215
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1216<p>StringMap also provides query methods that take byte ranges, so it only ever
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1217copies a string if a value is inserted into the table.</p>
1218</div>
1219
1220<!-- _______________________________________________________________________ -->
1221<div class="doc_subsubsection">
1222 <a name="dss_indexedmap">"llvm/ADT/IndexedMap.h"</a>
1223</div>
1224
1225<div class="doc_text">
1226<p>
1227IndexedMap is a specialized container for mapping small dense integers (or
1228values that can be mapped to small dense integers) to some other type. It is
1229internally implemented as a vector with a mapping function that maps the keys to
1230the dense integer range.
1231</p>
1232
1233<p>
1234This is useful for cases like virtual registers in the LLVM code generator: they
1235have a dense mapping that is offset by a compile-time constant (the first
1236virtual register ID).</p>
1237
1238</div>
1239
1240<!-- _______________________________________________________________________ -->
1241<div class="doc_subsubsection">
1242 <a name="dss_densemap">"llvm/ADT/DenseMap.h"</a>
1243</div>
1244
1245<div class="doc_text">
1246
1247<p>
1248DenseMap is a simple quadratically probed hash table. It excels at supporting
1249small keys and values: it uses a single allocation to hold all of the pairs that
1250are currently inserted in the map. DenseMap is a great way to map pointers to
1251pointers, or map other small types to each other.
1252</p>
1253
1254<p>
1255There are several aspects of DenseMap that you should be aware of, however. The
1256iterators in a densemap are invalidated whenever an insertion occurs, unlike
1257map. Also, because DenseMap allocates space for a large number of key/value
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1258pairs (it starts with 64 by default), it will waste a lot of space if your keys
1259or values are large. Finally, you must implement a partial specialization of
Chris Lattner76c1b972007-09-17 18:34:04 +0000[diff] [blame]1260DenseMapInfo for the key that you want, if it isn't already supported. This
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1261is required to tell DenseMap about two special marker values (which can never be
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1262inserted into the map) that it needs internally.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1263
1264</div>
1265
1266<!-- _______________________________________________________________________ -->
1267<div class="doc_subsubsection">
1268 <a name="dss_map">&lt;map&gt;</a>
1269</div>
1270
1271<div class="doc_text">
1272
1273<p>
1274std::map has similar characteristics to <a href="#dss_set">std::set</a>: it uses
1275a single allocation per pair inserted into the map, it offers log(n) lookup with
1276an extremely large constant factor, imposes a space penalty of 3 pointers per
1277pair in the map, etc.</p>
1278
1279<p>std::map is most useful when your keys or values are very large, if you need
1280to iterate over the collection in sorted order, or if you need stable iterators
1281into the map (i.e. they don't get invalidated if an insertion or deletion of
1282another element takes place).</p>
1283
1284</div>
1285
1286<!-- _______________________________________________________________________ -->
1287<div class="doc_subsubsection">
1288 <a name="dss_othermap">Other Map-Like Container Options</a>
1289</div>
1290
1291<div class="doc_text">
1292
1293<p>
1294The STL provides several other options, such as std::multimap and the various
1295"hash_map" like containers (whether from C++ TR1 or from the SGI library).</p>
1296
1297<p>std::multimap is useful if you want to map a key to multiple values, but has
1298all the drawbacks of std::map. A sorted vector or some other approach is almost
1299always better.</p>
1300
1301<p>The various hash_map implementations (exposed portably by
1302"llvm/ADT/hash_map") are simple chained hash tables. This algorithm is as
1303malloc intensive as std::map (performing an allocation for each element
1304inserted, thus having really high constant factors) but (usually) provides O(1)
1305insertion/deletion of elements. This can be useful if your elements are large
1306(thus making the constant-factor cost relatively low) or if comparisons are
1307expensive. Element iteration does not visit elements in a useful order.</p>
1308
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1309</div>
1310
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1311<!-- ======================================================================= -->
1312<div class="doc_subsection">
1313 <a name="ds_bit">Bit storage containers (BitVector, SparseBitVector)</a>
1314</div>
1315
1316<div class="doc_text">
Chris Lattner7086ce72007-09-25 22:37:50 +0000[diff] [blame]1317<p>Unlike the other containers, there are only two bit storage containers, and
1318choosing when to use each is relatively straightforward.</p>
1319
1320<p>One additional option is
1321<tt>std::vector&lt;bool&gt;</tt>: we discourage its use for two reasons 1) the
1322implementation in many common compilers (e.g. commonly available versions of
1323GCC) is extremely inefficient and 2) the C++ standards committee is likely to
1324deprecate this container and/or change it significantly somehow. In any case,
1325please don't use it.</p>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1326</div>
1327
1328<!-- _______________________________________________________________________ -->
1329<div class="doc_subsubsection">
1330 <a name="dss_bitvector">BitVector</a>
1331</div>
1332
1333<div class="doc_text">
1334<p> The BitVector container provides a fixed size set of bits for manipulation.
1335It supports individual bit setting/testing, as well as set operations. The set
1336operations take time O(size of bitvector), but operations are performed one word
1337at a time, instead of one bit at a time. This makes the BitVector very fast for
1338set operations compared to other containers. Use the BitVector when you expect
1339the number of set bits to be high (IE a dense set).
1340</p>
1341</div>
1342
1343<!-- _______________________________________________________________________ -->
1344<div class="doc_subsubsection">
1345 <a name="dss_sparsebitvector">SparseBitVector</a>
1346</div>
1347
1348<div class="doc_text">
1349<p> The SparseBitVector container is much like BitVector, with one major
1350difference: Only the bits that are set, are stored. This makes the
1351SparseBitVector much more space efficient than BitVector when the set is sparse,
1352as well as making set operations O(number of set bits) instead of O(size of
1353universe). The downside to the SparseBitVector is that setting and testing of random bits is O(N), and on large SparseBitVectors, this can be slower than BitVector. In our implementation, setting or testing bits in sorted order
1354(either forwards or reverse) is O(1) worst case. Testing and setting bits within 128 bits (depends on size) of the current bit is also O(1). As a general statement, testing/setting bits in a SparseBitVector is O(distance away from last set bit).
1355</p>
1356</div>
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]1357
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1358<!-- *********************************************************************** -->
1359<div class="doc_section">
1360 <a name="common">Helpful Hints for Common Operations</a>
1361</div>
1362<!-- *********************************************************************** -->
1363
1364<div class="doc_text">
1365
1366<p>This section describes how to perform some very simple transformations of
1367LLVM code. This is meant to give examples of common idioms used, showing the
1368practical side of LLVM transformations. <p> Because this is a "how-to" section,
1369you should also read about the main classes that you will be working with. The
1370<a href="#coreclasses">Core LLVM Class Hierarchy Reference</a> contains details
1371and descriptions of the main classes that you should know about.</p>
1372
1373</div>
1374
1375<!-- NOTE: this section should be heavy on example code -->
1376<!-- ======================================================================= -->
1377<div class="doc_subsection">
1378 <a name="inspection">Basic Inspection and Traversal Routines</a>
1379</div>
1380
1381<div class="doc_text">
1382
1383<p>The LLVM compiler infrastructure have many different data structures that may
1384be traversed. Following the example of the C++ standard template library, the
1385techniques used to traverse these various data structures are all basically the
1386same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
1387method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
1388function returns an iterator pointing to one past the last valid element of the
1389sequence, and there is some <tt>XXXiterator</tt> data type that is common
1390between the two operations.</p>
1391
1392<p>Because the pattern for iteration is common across many different aspects of
1393the program representation, the standard template library algorithms may be used
1394on them, and it is easier to remember how to iterate. First we show a few common
1395examples of the data structures that need to be traversed. Other data
1396structures are traversed in very similar ways.</p>
1397
1398</div>
1399
1400<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1401<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1402 <a name="iterate_function">Iterating over the </a><a
1403 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
1404 href="#Function"><tt>Function</tt></a>
1405</div>
1406
1407<div class="doc_text">
1408
1409<p>It's quite common to have a <tt>Function</tt> instance that you'd like to
1410transform in some way; in particular, you'd like to manipulate its
1411<tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over all of
1412the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>. The following is
1413an example that prints the name of a <tt>BasicBlock</tt> and the number of
1414<tt>Instruction</tt>s it contains:</p>
1415
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1416<div class="doc_code">
1417<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1418// <i>func is a pointer to a Function instance</i>
1419for (Function::iterator i = func-&gt;begin(), e = func-&gt;end(); i != e; ++i)
1420 // <i>Print out the name of the basic block if it has one, and then the</i>
1421 // <i>number of instructions that it contains</i>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1422 llvm::cerr &lt;&lt; "Basic block (name=" &lt;&lt; i-&gt;getName() &lt;&lt; ") has "
1423 &lt;&lt; i-&gt;size() &lt;&lt; " instructions.\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1424</pre>
1425</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1426
1427<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]1428invoking member functions of the <tt>Instruction</tt> class. This is
1429because the indirection operator is overloaded for the iterator
Chris Lattner7496ec52003-08-05 22:54:23 +0000[diff] [blame]1430classes. In the above code, the expression <tt>i-&gt;size()</tt> is
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1431exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.</p>
1432
1433</div>
1434
1435<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1436<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1437 <a name="iterate_basicblock">Iterating over the </a><a
1438 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1439 href="#BasicBlock"><tt>BasicBlock</tt></a>
1440</div>
1441
1442<div class="doc_text">
1443
1444<p>Just like when dealing with <tt>BasicBlock</tt>s in <tt>Function</tt>s, it's
1445easy to iterate over the individual instructions that make up
1446<tt>BasicBlock</tt>s. Here's a code snippet that prints out each instruction in
1447a <tt>BasicBlock</tt>:</p>
1448
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1449<div class="doc_code">
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1450<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1451// <i>blk is a pointer to a BasicBlock instance</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1452for (BasicBlock::iterator i = blk-&gt;begin(), e = blk-&gt;end(); i != e; ++i)
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1453 // <i>The next statement works since operator&lt;&lt;(ostream&amp;,...)</i>
1454 // <i>is overloaded for Instruction&amp;</i>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1455 llvm::cerr &lt;&lt; *i &lt;&lt; "\n";
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1456</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1457</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1458
1459<p>However, this isn't really the best way to print out the contents of a
1460<tt>BasicBlock</tt>! Since the ostream operators are overloaded for virtually
1461anything you'll care about, you could have just invoked the print routine on the
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1462basic block itself: <tt>llvm::cerr &lt;&lt; *blk &lt;&lt; "\n";</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1463
1464</div>
1465
1466<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1467<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1468 <a name="iterate_institer">Iterating over the </a><a
1469 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1470 href="#Function"><tt>Function</tt></a>
1471</div>
1472
1473<div class="doc_text">
1474
1475<p>If you're finding that you commonly iterate over a <tt>Function</tt>'s
1476<tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s <tt>Instruction</tt>s,
1477<tt>InstIterator</tt> should be used instead. You'll need to include <a
1478href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>,
1479and then instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1480small 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]1481
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1482<div class="doc_code">
1483<pre>
1484#include "<a href="/doxygen/InstIterator_8h-source.html">llvm/Support/InstIterator.h</a>"
1485
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1486// <i>F is a pointer to a Function instance</i>
Chris Lattnerda021aa2008-06-04 18:20:42 +0000[diff] [blame^]1487for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
1488 llvm::cerr &lt;&lt; *I &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1489</pre>
1490</div>
1491
1492<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]1493work list with its initial contents. For example, if you wanted to
1494initialize a work list to contain all instructions in a <tt>Function</tt>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1495F, all you would need to do is something like:</p>
1496
1497<div class="doc_code">
1498<pre>
1499std::set&lt;Instruction*&gt; worklist;
Chris Lattnerda021aa2008-06-04 18:20:42 +0000[diff] [blame^]1500// or better yet, SmallPtrSet&lt;Instruction*, 64&gt; worklist;
1501
1502for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
1503 worklist.insert(&amp;*I);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1504</pre>
1505</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1506
1507<p>The STL set <tt>worklist</tt> would now contain all instructions in the
1508<tt>Function</tt> pointed to by F.</p>
1509
1510</div>
1511
1512<!-- _______________________________________________________________________ -->
1513<div class="doc_subsubsection">
1514 <a name="iterate_convert">Turning an iterator into a class pointer (and
1515 vice-versa)</a>
1516</div>
1517
1518<div class="doc_text">
1519
1520<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]1521instance when all you've got at hand is an iterator. Well, extracting
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1522a reference or a pointer from an iterator is very straight-forward.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1523Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and <tt>j</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1524is a <tt>BasicBlock::const_iterator</tt>:</p>
1525
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1526<div class="doc_code">
1527<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1528Instruction&amp; inst = *i; // <i>Grab reference to instruction reference</i>
1529Instruction* pinst = &amp;*i; // <i>Grab pointer to instruction reference</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1530const Instruction&amp; inst = *j;
1531</pre>
1532</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1533
1534<p>However, the iterators you'll be working with in the LLVM framework are
1535special: they will automatically convert to a ptr-to-instance type whenever they
1536need to. Instead of dereferencing the iterator and then taking the address of
1537the result, you can simply assign the iterator to the proper pointer type and
1538you get the dereference and address-of operation as a result of the assignment
1539(behind the scenes, this is a result of overloading casting mechanisms). Thus
1540the last line of the last example,</p>
1541
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1542<div class="doc_code">
1543<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1544Instruction *pinst = &amp;*i;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1545</pre>
1546</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1547
1548<p>is semantically equivalent to</p>
1549
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1550<div class="doc_code">
1551<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1552Instruction *pinst = i;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1553</pre>
1554</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1555
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1556<p>It's also possible to turn a class pointer into the corresponding iterator,
1557and this is a constant time operation (very efficient). The following code
1558snippet illustrates use of the conversion constructors provided by LLVM
1559iterators. By using these, you can explicitly grab the iterator of something
1560without actually obtaining it via iteration over some structure:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1561
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1562<div class="doc_code">
1563<pre>
1564void printNextInstruction(Instruction* inst) {
1565 BasicBlock::iterator it(inst);
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1566 ++it; // <i>After this line, it refers to the instruction after *inst</i>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1567 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]1568}
1569</pre>
1570</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1571
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1572</div>
1573
1574<!--_______________________________________________________________________-->
1575<div class="doc_subsubsection">
1576 <a name="iterate_complex">Finding call sites: a slightly more complex
1577 example</a>
1578</div>
1579
1580<div class="doc_text">
1581
1582<p>Say that you're writing a FunctionPass and would like to count all the
1583locations in the entire module (that is, across every <tt>Function</tt>) where a
1584certain function (i.e., some <tt>Function</tt>*) is already in scope. As you'll
1585learn 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]1586much more straight-forward manner, but this example will allow us to explore how
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1587you'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]1588is what we want to do:</p>
1589
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1590<div class="doc_code">
1591<pre>
1592initialize callCounter to zero
1593for each Function f in the Module
1594 for each BasicBlock b in f
1595 for each Instruction i in b
1596 if (i is a CallInst and calls the given function)
1597 increment callCounter
1598</pre>
1599</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1600
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1601<p>And the actual code is (remember, because we're writing a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1602<tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply has to
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1603override the <tt>runOnFunction</tt> method):</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1604
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1605<div class="doc_code">
1606<pre>
1607Function* targetFunc = ...;
1608
1609class OurFunctionPass : public FunctionPass {
1610 public:
1611 OurFunctionPass(): callCounter(0) { }
1612
1613 virtual runOnFunction(Function&amp; F) {
1614 for (Function::iterator b = F.begin(), be = F.end(); b != be; ++b) {
1615 for (BasicBlock::iterator i = b-&gt;begin(); ie = b-&gt;end(); i != ie; ++i) {
1616 if (<a href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a>&lt;<a
1617 href="#CallInst">CallInst</a>&gt;(&amp;*i)) {
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1618 // <i>We know we've encountered a call instruction, so we</i>
1619 // <i>need to determine if it's a call to the</i>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1620 // <i>function pointed to by m_func or not.</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1621 if (callInst-&gt;getCalledFunction() == targetFunc)
1622 ++callCounter;
1623 }
1624 }
1625 }
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1626 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1627
1628 private:
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1629 unsigned callCounter;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1630};
1631</pre>
1632</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1633
1634</div>
1635
Brian Gaekef1972c62003-11-07 19:25:45 +0000[diff] [blame]1636<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1637<div class="doc_subsubsection">
1638 <a name="calls_and_invokes">Treating calls and invokes the same way</a>
1639</div>
1640
1641<div class="doc_text">
1642
1643<p>You may have noticed that the previous example was a bit oversimplified in
1644that it did not deal with call sites generated by 'invoke' instructions. In
1645this, and in other situations, you may find that you want to treat
1646<tt>CallInst</tt>s and <tt>InvokeInst</tt>s the same way, even though their
1647most-specific common base class is <tt>Instruction</tt>, which includes lots of
1648less closely-related things. For these cases, LLVM provides a handy wrapper
1649class called <a
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]1650href="http://llvm.org/doxygen/classllvm_1_1CallSite.html"><tt>CallSite</tt></a>.
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1651It is essentially a wrapper around an <tt>Instruction</tt> pointer, with some
1652methods that provide functionality common to <tt>CallInst</tt>s and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1653<tt>InvokeInst</tt>s.</p>
1654
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1655<p>This class has "value semantics": it should be passed by value, not by
1656reference and it should not be dynamically allocated or deallocated using
1657<tt>operator new</tt> or <tt>operator delete</tt>. It is efficiently copyable,
1658assignable and constructable, with costs equivalents to that of a bare pointer.
1659If you look at its definition, it has only a single pointer member.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1660
1661</div>
1662
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]1663<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1664<div class="doc_subsubsection">
1665 <a name="iterate_chains">Iterating over def-use &amp; use-def chains</a>
1666</div>
1667
1668<div class="doc_text">
1669
1670<p>Frequently, we might have an instance of the <a
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]1671href="/doxygen/classllvm_1_1Value.html">Value Class</a> and we want to
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]1672determine which <tt>User</tt>s use the <tt>Value</tt>. The list of all
1673<tt>User</tt>s of a particular <tt>Value</tt> is called a <i>def-use</i> chain.
1674For example, let's say we have a <tt>Function*</tt> named <tt>F</tt> to a
1675particular function <tt>foo</tt>. Finding all of the instructions that
1676<i>use</i> <tt>foo</tt> is as simple as iterating over the <i>def-use</i> chain
1677of <tt>F</tt>:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1678
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1679<div class="doc_code">
1680<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1681Function *F = ...;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1682
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1683for (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]1684 if (Instruction *Inst = dyn_cast&lt;Instruction&gt;(*i)) {
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1685 llvm::cerr &lt;&lt; "F is used in instruction:\n";
1686 llvm::cerr &lt;&lt; *Inst &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1687 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1688</pre>
1689</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1690
1691<p>Alternately, it's common to have an instance of the <a
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]1692href="/doxygen/classllvm_1_1User.html">User Class</a> and need to know what
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1693<tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used by a
1694<tt>User</tt> is known as a <i>use-def</i> chain. Instances of class
1695<tt>Instruction</tt> are common <tt>User</tt>s, so we might want to iterate over
1696all of the values that a particular instruction uses (that is, the operands of
1697the particular <tt>Instruction</tt>):</p>
1698
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1699<div class="doc_code">
1700<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1701Instruction *pi = ...;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1702
1703for (User::op_iterator i = pi-&gt;op_begin(), e = pi-&gt;op_end(); i != e; ++i) {
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1704 Value *v = *i;
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1705 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1706}
1707</pre>
1708</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1709
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]1710<!--
1711 def-use chains ("finding all users of"): Value::use_begin/use_end
1712 use-def chains ("finding all values used"): User::op_begin/op_end [op=operand]
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1713-->
1714
1715</div>
1716
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1717<!--_______________________________________________________________________-->
1718<div class="doc_subsubsection">
1719 <a name="iterate_preds">Iterating over predecessors &amp;
1720successors of blocks</a>
1721</div>
1722
1723<div class="doc_text">
1724
1725<p>Iterating over the predecessors and successors of a block is quite easy
1726with the routines defined in <tt>"llvm/Support/CFG.h"</tt>. Just use code like
1727this to iterate over all predecessors of BB:</p>
1728
1729<div class="doc_code">
1730<pre>
1731#include "llvm/Support/CFG.h"
1732BasicBlock *BB = ...;
1733
1734for (pred_iterator PI = pred_begin(BB), E = pred_end(BB); PI != E; ++PI) {
1735 BasicBlock *Pred = *PI;
1736 // <i>...</i>
1737}
1738</pre>
1739</div>
1740
1741<p>Similarly, to iterate over successors use
1742succ_iterator/succ_begin/succ_end.</p>
1743
1744</div>
1745
1746
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1747<!-- ======================================================================= -->
1748<div class="doc_subsection">
1749 <a name="simplechanges">Making simple changes</a>
1750</div>
1751
1752<div class="doc_text">
1753
1754<p>There are some primitive transformation operations present in the LLVM
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]1755infrastructure that are worth knowing about. When performing
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1756transformations, it's fairly common to manipulate the contents of basic
1757blocks. This section describes some of the common methods for doing so
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1758and gives example code.</p>
1759
1760</div>
1761
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1762<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1763<div class="doc_subsubsection">
1764 <a name="schanges_creating">Creating and inserting new
1765 <tt>Instruction</tt>s</a>
1766</div>
1767
1768<div class="doc_text">
1769
1770<p><i>Instantiating Instructions</i></p>
1771
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1772<p>Creation of <tt>Instruction</tt>s is straight-forward: simply call the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1773constructor for the kind of instruction to instantiate and provide the necessary
1774parameters. For example, an <tt>AllocaInst</tt> only <i>requires</i> a
1775(const-ptr-to) <tt>Type</tt>. Thus:</p>
1776
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1777<div class="doc_code">
1778<pre>
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]1779AllocaInst* ai = new AllocaInst(Type::Int32Ty);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1780</pre>
1781</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1782
1783<p>will create an <tt>AllocaInst</tt> instance that represents the allocation of
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1784one integer in the current stack frame, at run time. Each <tt>Instruction</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1785subclass is likely to have varying default parameters which change the semantics
1786of the instruction, so refer to the <a
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]1787href="/doxygen/classllvm_1_1Instruction.html">doxygen documentation for the subclass of
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1788Instruction</a> that you're interested in instantiating.</p>
1789
1790<p><i>Naming values</i></p>
1791
1792<p>It is very useful to name the values of instructions when you're able to, as
1793this facilitates the debugging of your transformations. If you end up looking
1794at generated LLVM machine code, you definitely want to have logical names
1795associated with the results of instructions! By supplying a value for the
1796<tt>Name</tt> (default) parameter of the <tt>Instruction</tt> constructor, you
1797associate a logical name with the result of the instruction's execution at
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1798run time. For example, say that I'm writing a transformation that dynamically
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1799allocates space for an integer on the stack, and that integer is going to be
1800used as some kind of index by some other code. To accomplish this, I place an
1801<tt>AllocaInst</tt> at the first point in the first <tt>BasicBlock</tt> of some
1802<tt>Function</tt>, and I'm intending to use it within the same
1803<tt>Function</tt>. I might do:</p>
1804
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1805<div class="doc_code">
1806<pre>
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]1807AllocaInst* pa = new AllocaInst(Type::Int32Ty, 0, "indexLoc");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1808</pre>
1809</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1810
1811<p>where <tt>indexLoc</tt> is now the logical name of the instruction's
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1812execution value, which is a pointer to an integer on the run time stack.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1813
1814<p><i>Inserting instructions</i></p>
1815
1816<p>There are essentially two ways to insert an <tt>Instruction</tt>
1817into an existing sequence of instructions that form a <tt>BasicBlock</tt>:</p>
1818
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]1819<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1820 <li>Insertion into an explicit instruction list
1821
1822 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within that
1823 <tt>BasicBlock</tt>, and a newly-created instruction we wish to insert
1824 before <tt>*pi</tt>, we do the following: </p>
1825
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1826<div class="doc_code">
1827<pre>
1828BasicBlock *pb = ...;
1829Instruction *pi = ...;
1830Instruction *newInst = new Instruction(...);
1831
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1832pb-&gt;getInstList().insert(pi, newInst); // <i>Inserts newInst before pi in pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1833</pre>
1834</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]1835
1836 <p>Appending to the end of a <tt>BasicBlock</tt> is so common that
1837 the <tt>Instruction</tt> class and <tt>Instruction</tt>-derived
1838 classes provide constructors which take a pointer to a
1839 <tt>BasicBlock</tt> to be appended to. For example code that
1840 looked like: </p>
1841
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1842<div class="doc_code">
1843<pre>
1844BasicBlock *pb = ...;
1845Instruction *newInst = new Instruction(...);
1846
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1847pb-&gt;getInstList().push_back(newInst); // <i>Appends newInst to pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1848</pre>
1849</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]1850
1851 <p>becomes: </p>
1852
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1853<div class="doc_code">
1854<pre>
1855BasicBlock *pb = ...;
1856Instruction *newInst = new Instruction(..., pb);
1857</pre>
1858</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]1859
1860 <p>which is much cleaner, especially if you are creating
1861 long instruction streams.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1862
1863 <li>Insertion into an implicit instruction list
1864
1865 <p><tt>Instruction</tt> instances that are already in <tt>BasicBlock</tt>s
1866 are implicitly associated with an existing instruction list: the instruction
1867 list of the enclosing basic block. Thus, we could have accomplished the same
1868 thing as the above code without being given a <tt>BasicBlock</tt> by doing:
1869 </p>
1870
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1871<div class="doc_code">
1872<pre>
1873Instruction *pi = ...;
1874Instruction *newInst = new Instruction(...);
1875
1876pi-&gt;getParent()-&gt;getInstList().insert(pi, newInst);
1877</pre>
1878</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1879
1880 <p>In fact, this sequence of steps occurs so frequently that the
1881 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes provide
1882 constructors which take (as a default parameter) a pointer to an
1883 <tt>Instruction</tt> which the newly-created <tt>Instruction</tt> should
1884 precede. That is, <tt>Instruction</tt> constructors are capable of
1885 inserting the newly-created instance into the <tt>BasicBlock</tt> of a
1886 provided instruction, immediately before that instruction. Using an
1887 <tt>Instruction</tt> constructor with a <tt>insertBefore</tt> (default)
1888 parameter, the above code becomes:</p>
1889
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1890<div class="doc_code">
1891<pre>
1892Instruction* pi = ...;
1893Instruction* newInst = new Instruction(..., pi);
1894</pre>
1895</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1896
1897 <p>which is much cleaner, especially if you're creating a lot of
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1898 instructions and adding them to <tt>BasicBlock</tt>s.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1899</ul>
1900
1901</div>
1902
1903<!--_______________________________________________________________________-->
1904<div class="doc_subsubsection">
1905 <a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a>
1906</div>
1907
1908<div class="doc_text">
1909
1910<p>Deleting an instruction from an existing sequence of instructions that form a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1911<a href="#BasicBlock"><tt>BasicBlock</tt></a> is very straight-forward. First,
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1912you must have a pointer to the instruction that you wish to delete. Second, you
1913need to obtain the pointer to that instruction's basic block. You use the
1914pointer to the basic block to get its list of instructions and then use the
1915erase function to remove your instruction. For example:</p>
1916
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1917<div class="doc_code">
1918<pre>
1919<a href="#Instruction">Instruction</a> *I = .. ;
Chris Lattner9f8ec252008-02-15 22:57:17 +0000[diff] [blame]1920I-&gt;eraseFromParent();
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1921</pre>
1922</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1923
1924</div>
1925
1926<!--_______________________________________________________________________-->
1927<div class="doc_subsubsection">
1928 <a name="schanges_replacing">Replacing an <tt>Instruction</tt> with another
1929 <tt>Value</tt></a>
1930</div>
1931
1932<div class="doc_text">
1933
1934<p><i>Replacing individual instructions</i></p>
1935
1936<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]1937permits use of two very useful replace functions: <tt>ReplaceInstWithValue</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1938and <tt>ReplaceInstWithInst</tt>.</p>
1939
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1940<h4><a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a></h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1941
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1942<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1943 <li><tt>ReplaceInstWithValue</tt>
1944
1945 <p>This function replaces all uses (within a basic block) of a given
1946 instruction with a value, and then removes the original instruction. The
1947 following example illustrates the replacement of the result of a particular
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]1948 <tt>AllocaInst</tt> that allocates memory for a single integer with a null
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1949 pointer to an integer.</p>
1950
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1951<div class="doc_code">
1952<pre>
1953AllocaInst* instToReplace = ...;
1954BasicBlock::iterator ii(instToReplace);
1955
1956ReplaceInstWithValue(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]1957 Constant::getNullValue(PointerType::get(Type::Int32Ty)));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1958</pre></div></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1959
1960 <li><tt>ReplaceInstWithInst</tt>
1961
1962 <p>This function replaces a particular instruction with another
1963 instruction. The following example illustrates the replacement of one
1964 <tt>AllocaInst</tt> with another.</p>
1965
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1966<div class="doc_code">
1967<pre>
1968AllocaInst* instToReplace = ...;
1969BasicBlock::iterator ii(instToReplace);
1970
1971ReplaceInstWithInst(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]1972 new AllocaInst(Type::Int32Ty, 0, "ptrToReplacedInt"));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1973</pre></div></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1974</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1975
1976<p><i>Replacing multiple uses of <tt>User</tt>s and <tt>Value</tt>s</i></p>
1977
1978<p>You can use <tt>Value::replaceAllUsesWith</tt> and
1979<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]1980doxygen documentation for the <a href="/doxygen/classllvm_1_1Value.html">Value Class</a>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]1981and <a href="/doxygen/classllvm_1_1User.html">User Class</a>, respectively, for more
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1982information.</p>
1983
1984<!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
1985include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
1986ReplaceInstWithValue, ReplaceInstWithInst -->
1987
1988</div>
1989
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]1990<!--_______________________________________________________________________-->
1991<div class="doc_subsubsection">
1992 <a name="schanges_deletingGV">Deleting <tt>GlobalVariable</tt>s</a>
1993</div>
1994
1995<div class="doc_text">
1996
Tanya Lattnerc5dfcdb2007-06-20 20:46:37 +0000[diff] [blame]1997<p>Deleting a global variable from a module is just as easy as deleting an
1998Instruction. First, you must have a pointer to the global variable that you wish
1999 to delete. You use this pointer to erase it from its parent, the module.
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2000 For example:</p>
2001
2002<div class="doc_code">
2003<pre>
2004<a href="#GlobalVariable">GlobalVariable</a> *GV = .. ;
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2005
Tanya Lattnerc5dfcdb2007-06-20 20:46:37 +0000[diff] [blame]2006GV-&gt;eraseFromParent();
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2007</pre>
2008</div>
2009
2010</div>
2011
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]2012<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2013<div class="doc_section">
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2014 <a name="advanced">Advanced Topics</a>
2015</div>
2016<!-- *********************************************************************** -->
2017
2018<div class="doc_text">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2019<p>
2020This section describes some of the advanced or obscure API's that most clients
2021do not need to be aware of. These API's tend manage the inner workings of the
2022LLVM system, and only need to be accessed in unusual circumstances.
2023</p>
2024</div>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2025
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2026<!-- ======================================================================= -->
2027<div class="doc_subsection">
2028 <a name="TypeResolve">LLVM Type Resolution</a>
2029</div>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2030
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2031<div class="doc_text">
2032
2033<p>
2034The LLVM type system has a very simple goal: allow clients to compare types for
2035structural equality with a simple pointer comparison (aka a shallow compare).
2036This goal makes clients much simpler and faster, and is used throughout the LLVM
2037system.
2038</p>
2039
2040<p>
2041Unfortunately achieving this goal is not a simple matter. In particular,
2042recursive types and late resolution of opaque types makes the situation very
2043difficult to handle. Fortunately, for the most part, our implementation makes
2044most clients able to be completely unaware of the nasty internal details. The
2045primary case where clients are exposed to the inner workings of it are when
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]2046building a recursive type. In addition to this case, the LLVM bitcode reader,
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2047assembly parser, and linker also have to be aware of the inner workings of this
2048system.
2049</p>
2050
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]2051<p>
2052For our purposes below, we need three concepts. First, an "Opaque Type" is
2053exactly as defined in the <a href="LangRef.html#t_opaque">language
2054reference</a>. Second an "Abstract Type" is any type which includes an
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2055opaque type as part of its type graph (for example "<tt>{ opaque, i32 }</tt>").
2056Third, 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]2057float }</tt>").
2058</p>
2059
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2060</div>
2061
2062<!-- ______________________________________________________________________ -->
2063<div class="doc_subsubsection">
2064 <a name="BuildRecType">Basic Recursive Type Construction</a>
2065</div>
2066
2067<div class="doc_text">
2068
2069<p>
2070Because the most common question is "how do I build a recursive type with LLVM",
2071we answer it now and explain it as we go. Here we include enough to cause this
2072to be emitted to an output .ll file:
2073</p>
2074
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2075<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2076<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2077%mylist = type { %mylist*, i32 }
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2078</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2079</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2080
2081<p>
2082To build this, use the following LLVM APIs:
2083</p>
2084
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2085<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2086<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2087// <i>Create the initial outer struct</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2088<a href="#PATypeHolder">PATypeHolder</a> StructTy = OpaqueType::get();
2089std::vector&lt;const Type*&gt; Elts;
2090Elts.push_back(PointerType::get(StructTy));
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2091Elts.push_back(Type::Int32Ty);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2092StructType *NewSTy = StructType::get(Elts);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2093
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2094// <i>At this point, NewSTy = "{ opaque*, i32 }". Tell VMCore that</i>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2095// <i>the struct and the opaque type are actually the same.</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2096cast&lt;OpaqueType&gt;(StructTy.get())-&gt;<a href="#refineAbstractTypeTo">refineAbstractTypeTo</a>(NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2097
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2098// <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]2099// <i>kept up-to-date</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2100NewSTy = cast&lt;StructType&gt;(StructTy.get());
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2101
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2102// <i>Add a name for the type to the module symbol table (optional)</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2103MyModule-&gt;addTypeName("mylist", NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2104</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2105</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2106
2107<p>
2108This code shows the basic approach used to build recursive types: build a
2109non-recursive type using 'opaque', then use type unification to close the cycle.
2110The type unification step is performed by the <tt><a
Chris Lattneraff26d12007-02-03 03:06:52 +0000[diff] [blame]2111href="#refineAbstractTypeTo">refineAbstractTypeTo</a></tt> method, which is
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2112described next. After that, we describe the <a
2113href="#PATypeHolder">PATypeHolder class</a>.
2114</p>
2115
2116</div>
2117
2118<!-- ______________________________________________________________________ -->
2119<div class="doc_subsubsection">
2120 <a name="refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a>
2121</div>
2122
2123<div class="doc_text">
2124<p>
2125The <tt>refineAbstractTypeTo</tt> method starts the type unification process.
2126While this method is actually a member of the DerivedType class, it is most
2127often used on OpaqueType instances. Type unification is actually a recursive
2128process. After unification, types can become structurally isomorphic to
2129existing types, and all duplicates are deleted (to preserve pointer equality).
2130</p>
2131
2132<p>
2133In the example above, the OpaqueType object is definitely deleted.
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2134Additionally, if there is an "{ \2*, i32}" type already created in the system,
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2135the pointer and struct type created are <b>also</b> deleted. Obviously whenever
2136a type is deleted, any "Type*" pointers in the program are invalidated. As
2137such, it is safest to avoid having <i>any</i> "Type*" pointers to abstract types
2138live across a call to <tt>refineAbstractTypeTo</tt> (note that non-abstract
2139types can never move or be deleted). To deal with this, the <a
2140href="#PATypeHolder">PATypeHolder</a> class is used to maintain a stable
2141reference to a possibly refined type, and the <a
2142href="#AbstractTypeUser">AbstractTypeUser</a> class is used to update more
2143complex datastructures.
2144</p>
2145
2146</div>
2147
2148<!-- ______________________________________________________________________ -->
2149<div class="doc_subsubsection">
2150 <a name="PATypeHolder">The PATypeHolder Class</a>
2151</div>
2152
2153<div class="doc_text">
2154<p>
2155PATypeHolder is a form of a "smart pointer" for Type objects. When VMCore
2156happily goes about nuking types that become isomorphic to existing types, it
2157automatically updates all PATypeHolder objects to point to the new type. In the
2158example above, this allows the code to maintain a pointer to the resultant
2159resolved recursive type, even though the Type*'s are potentially invalidated.
2160</p>
2161
2162<p>
2163PATypeHolder is an extremely light-weight object that uses a lazy union-find
2164implementation to update pointers. For example the pointer from a Value to its
2165Type is maintained by PATypeHolder objects.
2166</p>
2167
2168</div>
2169
2170<!-- ______________________________________________________________________ -->
2171<div class="doc_subsubsection">
2172 <a name="AbstractTypeUser">The AbstractTypeUser Class</a>
2173</div>
2174
2175<div class="doc_text">
2176
2177<p>
2178Some data structures need more to perform more complex updates when types get
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2179resolved. To support this, a class can derive from the AbstractTypeUser class.
2180This class
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2181allows it to get callbacks when certain types are resolved. To register to get
2182callbacks for a particular type, the DerivedType::{add/remove}AbstractTypeUser
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]2183methods can be called on a type. Note that these methods only work for <i>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2184 abstract</i> types. Concrete types (those that do not include any opaque
2185objects) can never be refined.
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2186</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2187</div>
2188
2189
2190<!-- ======================================================================= -->
2191<div class="doc_subsection">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2192 <a name="SymbolTable">The <tt>ValueSymbolTable</tt> and
2193 <tt>TypeSymbolTable</tt> classes</a>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2194</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2195
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2196<div class="doc_text">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2197<p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1ValueSymbolTable.html">
2198ValueSymbolTable</a></tt> class provides a symbol table that the <a
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2199href="#Function"><tt>Function</tt></a> and <a href="#Module">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2200<tt>Module</tt></a> classes use for naming value definitions. The symbol table
2201can provide a name for any <a href="#Value"><tt>Value</tt></a>.
2202The <tt><a href="http://llvm.org/doxygen/classllvm_1_1TypeSymbolTable.html">
2203TypeSymbolTable</a></tt> class is used by the <tt>Module</tt> class to store
2204names for types.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2205
Reid Spencera6362242007-01-07 00:41:39 +0000[diff] [blame]2206<p>Note that the <tt>SymbolTable</tt> class should not be directly accessed
2207by most clients. It should only be used when iteration over the symbol table
2208names themselves are required, which is very special purpose. Note that not
2209all LLVM
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2210<a href="#Value">Value</a>s have names, and those without names (i.e. they have
2211an empty name) do not exist in the symbol table.
2212</p>
2213
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2214<p>These symbol tables support iteration over the values/types in the symbol
2215table with <tt>begin/end/iterator</tt> and supports querying to see if a
2216specific name is in the symbol table (with <tt>lookup</tt>). The
2217<tt>ValueSymbolTable</tt> class exposes no public mutator methods, instead,
2218simply call <tt>setName</tt> on a value, which will autoinsert it into the
2219appropriate symbol table. For types, use the Module::addTypeName method to
2220insert entries into the symbol table.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2221
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2222</div>
2223
2224
2225
2226<!-- *********************************************************************** -->
2227<div class="doc_section">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2228 <a name="coreclasses">The Core LLVM Class Hierarchy Reference </a>
2229</div>
2230<!-- *********************************************************************** -->
2231
2232<div class="doc_text">
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2233<p><tt>#include "<a href="/doxygen/Type_8h-source.html">llvm/Type.h</a>"</tt>
2234<br>doxygen info: <a href="/doxygen/classllvm_1_1Type.html">Type Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2235
2236<p>The Core LLVM classes are the primary means of representing the program
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2237being inspected or transformed. The core LLVM classes are defined in
2238header files in the <tt>include/llvm/</tt> directory, and implemented in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2239the <tt>lib/VMCore</tt> directory.</p>
2240
2241</div>
2242
2243<!-- ======================================================================= -->
2244<div class="doc_subsection">
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2245 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
2246</div>
2247
2248<div class="doc_text">
2249
2250 <p><tt>Type</tt> is a superclass of all type classes. Every <tt>Value</tt> has
2251 a <tt>Type</tt>. <tt>Type</tt> cannot be instantiated directly but only
2252 through its subclasses. Certain primitive types (<tt>VoidType</tt>,
2253 <tt>LabelType</tt>, <tt>FloatType</tt> and <tt>DoubleType</tt>) have hidden
2254 subclasses. They are hidden because they offer no useful functionality beyond
2255 what the <tt>Type</tt> class offers except to distinguish themselves from
2256 other subclasses of <tt>Type</tt>.</p>
2257 <p>All other types are subclasses of <tt>DerivedType</tt>. Types can be
2258 named, but this is not a requirement. There exists exactly
2259 one instance of a given shape at any one time. This allows type equality to
2260 be performed with address equality of the Type Instance. That is, given two
2261 <tt>Type*</tt> values, the types are identical if the pointers are identical.
2262 </p>
2263</div>
2264
2265<!-- _______________________________________________________________________ -->
2266<div class="doc_subsubsection">
2267 <a name="m_Value">Important Public Methods</a>
2268</div>
2269
2270<div class="doc_text">
2271
2272<ul>
Chris Lattner8f79df32007-01-15 01:55:32 +0000[diff] [blame]2273 <li><tt>bool isInteger() const</tt>: Returns true for any integer type.</li>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2274
2275 <li><tt>bool isFloatingPoint()</tt>: Return true if this is one of the two
2276 floating point types.</li>
2277
2278 <li><tt>bool isAbstract()</tt>: Return true if the type is abstract (contains
2279 an OpaqueType anywhere in its definition).</li>
2280
2281 <li><tt>bool isSized()</tt>: Return true if the type has known size. Things
2282 that don't have a size are abstract types, labels and void.</li>
2283
2284</ul>
2285</div>
2286
2287<!-- _______________________________________________________________________ -->
2288<div class="doc_subsubsection">
2289 <a name="m_Value">Important Derived Types</a>
2290</div>
2291<div class="doc_text">
2292<dl>
2293 <dt><tt>IntegerType</tt></dt>
2294 <dd>Subclass of DerivedType that represents integer types of any bit width.
2295 Any bit width between <tt>IntegerType::MIN_INT_BITS</tt> (1) and
2296 <tt>IntegerType::MAX_INT_BITS</tt> (~8 million) can be represented.
2297 <ul>
2298 <li><tt>static const IntegerType* get(unsigned NumBits)</tt>: get an integer
2299 type of a specific bit width.</li>
2300 <li><tt>unsigned getBitWidth() const</tt>: Get the bit width of an integer
2301 type.</li>
2302 </ul>
2303 </dd>
2304 <dt><tt>SequentialType</tt></dt>
2305 <dd>This is subclassed by ArrayType and PointerType
2306 <ul>
2307 <li><tt>const Type * getElementType() const</tt>: Returns the type of each
2308 of the elements in the sequential type. </li>
2309 </ul>
2310 </dd>
2311 <dt><tt>ArrayType</tt></dt>
2312 <dd>This is a subclass of SequentialType and defines the interface for array
2313 types.
2314 <ul>
2315 <li><tt>unsigned getNumElements() const</tt>: Returns the number of
2316 elements in the array. </li>
2317 </ul>
2318 </dd>
2319 <dt><tt>PointerType</tt></dt>
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]2320 <dd>Subclass of SequentialType for pointer types.</dd>
Reid Spencer9d6565a2007-02-15 02:26:10 +0000[diff] [blame]2321 <dt><tt>VectorType</tt></dt>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2322 <dd>Subclass of SequentialType for vector types. A
2323 vector type is similar to an ArrayType but is distinguished because it is
2324 a first class type wherease ArrayType is not. Vector types are used for
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2325 vector operations and are usually small vectors of of an integer or floating
2326 point type.</dd>
2327 <dt><tt>StructType</tt></dt>
2328 <dd>Subclass of DerivedTypes for struct types.</dd>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]2329 <dt><tt><a name="FunctionType">FunctionType</a></tt></dt>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2330 <dd>Subclass of DerivedTypes for function types.
2331 <ul>
2332 <li><tt>bool isVarArg() const</tt>: Returns true if its a vararg
2333 function</li>
2334 <li><tt> const Type * getReturnType() const</tt>: Returns the
2335 return type of the function.</li>
2336 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns
2337 the type of the ith parameter.</li>
2338 <li><tt> const unsigned getNumParams() const</tt>: Returns the
2339 number of formal parameters.</li>
2340 </ul>
2341 </dd>
2342 <dt><tt>OpaqueType</tt></dt>
2343 <dd>Sublcass of DerivedType for abstract types. This class
2344 defines no content and is used as a placeholder for some other type. Note
2345 that OpaqueType is used (temporarily) during type resolution for forward
2346 references of types. Once the referenced type is resolved, the OpaqueType
2347 is replaced with the actual type. OpaqueType can also be used for data
2348 abstraction. At link time opaque types can be resolved to actual types
2349 of the same name.</dd>
2350</dl>
2351</div>
2352
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2353
2354
2355<!-- ======================================================================= -->
2356<div class="doc_subsection">
2357 <a name="Module">The <tt>Module</tt> class</a>
2358</div>
2359
2360<div class="doc_text">
2361
2362<p><tt>#include "<a
2363href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt><br> doxygen info:
2364<a href="/doxygen/classllvm_1_1Module.html">Module Class</a></p>
2365
2366<p>The <tt>Module</tt> class represents the top level structure present in LLVM
2367programs. An LLVM module is effectively either a translation unit of the
2368original program or a combination of several translation units merged by the
2369linker. The <tt>Module</tt> class keeps track of a list of <a
2370href="#Function"><tt>Function</tt></a>s, a list of <a
2371href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
2372href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
2373helpful member functions that try to make common operations easy.</p>
2374
2375</div>
2376
2377<!-- _______________________________________________________________________ -->
2378<div class="doc_subsubsection">
2379 <a name="m_Module">Important Public Members of the <tt>Module</tt> class</a>
2380</div>
2381
2382<div class="doc_text">
2383
2384<ul>
2385 <li><tt>Module::Module(std::string name = "")</tt></li>
2386</ul>
2387
2388<p>Constructing a <a href="#Module">Module</a> is easy. You can optionally
2389provide a name for it (probably based on the name of the translation unit).</p>
2390
2391<ul>
2392 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
2393 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
2394
2395 <tt>begin()</tt>, <tt>end()</tt>
2396 <tt>size()</tt>, <tt>empty()</tt>
2397
2398 <p>These are forwarding methods that make it easy to access the contents of
2399 a <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
2400 list.</p></li>
2401
2402 <li><tt>Module::FunctionListType &amp;getFunctionList()</tt>
2403
2404 <p> Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
2405 necessary to use when you need to update the list or perform a complex
2406 action that doesn't have a forwarding method.</p>
2407
2408 <p><!-- Global Variable --></p></li>
2409</ul>
2410
2411<hr>
2412
2413<ul>
2414 <li><tt>Module::global_iterator</tt> - Typedef for global variable list iterator<br>
2415
2416 <tt>Module::const_global_iterator</tt> - Typedef for const_iterator.<br>
2417
2418 <tt>global_begin()</tt>, <tt>global_end()</tt>
2419 <tt>global_size()</tt>, <tt>global_empty()</tt>
2420
2421 <p> These are forwarding methods that make it easy to access the contents of
2422 a <tt>Module</tt> object's <a
2423 href="#GlobalVariable"><tt>GlobalVariable</tt></a> list.</p></li>
2424
2425 <li><tt>Module::GlobalListType &amp;getGlobalList()</tt>
2426
2427 <p>Returns the list of <a
2428 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s. This is necessary to
2429 use when you need to update the list or perform a complex action that
2430 doesn't have a forwarding method.</p>
2431
2432 <p><!-- Symbol table stuff --> </p></li>
2433</ul>
2434
2435<hr>
2436
2437<ul>
2438 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
2439
2440 <p>Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
2441 for this <tt>Module</tt>.</p>
2442
2443 <p><!-- Convenience methods --></p></li>
2444</ul>
2445
2446<hr>
2447
2448<ul>
2449 <li><tt><a href="#Function">Function</a> *getFunction(const std::string
2450 &amp;Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt>
2451
2452 <p>Look up the specified function in the <tt>Module</tt> <a
2453 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
2454 <tt>null</tt>.</p></li>
2455
2456 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const
2457 std::string &amp;Name, const <a href="#FunctionType">FunctionType</a> *T)</tt>
2458
2459 <p>Look up the specified function in the <tt>Module</tt> <a
2460 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
2461 external declaration for the function and return it.</p></li>
2462
2463 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt>
2464
2465 <p>If there is at least one entry in the <a
2466 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
2467 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
2468 string.</p></li>
2469
2470 <li><tt>bool addTypeName(const std::string &amp;Name, const <a
2471 href="#Type">Type</a> *Ty)</tt>
2472
2473 <p>Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
2474 mapping <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this
2475 name, true is returned and the <a
2476 href="#SymbolTable"><tt>SymbolTable</tt></a> is not modified.</p></li>
2477</ul>
2478
2479</div>
2480
2481
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2482<!-- ======================================================================= -->
2483<div class="doc_subsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2484 <a name="Value">The <tt>Value</tt> class</a>
2485</div>
2486
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2487<div class="doc_text">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2488
2489<p><tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt>
2490<br>
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]2491doxygen info: <a href="/doxygen/classllvm_1_1Value.html">Value Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2492
2493<p>The <tt>Value</tt> class is the most important class in the LLVM Source
2494base. It represents a typed value that may be used (among other things) as an
2495operand to an instruction. There are many different types of <tt>Value</tt>s,
2496such as <a href="#Constant"><tt>Constant</tt></a>s,<a
2497href="#Argument"><tt>Argument</tt></a>s. Even <a
2498href="#Instruction"><tt>Instruction</tt></a>s and <a
2499href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.</p>
2500
2501<p>A particular <tt>Value</tt> may be used many times in the LLVM representation
2502for a program. For example, an incoming argument to a function (represented
2503with an instance of the <a href="#Argument">Argument</a> class) is "used" by
2504every instruction in the function that references the argument. To keep track
2505of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
2506href="#User"><tt>User</tt></a>s that is using it (the <a
2507href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
2508graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
2509def-use information in the program, and is accessible through the <tt>use_</tt>*
2510methods, shown below.</p>
2511
2512<p>Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed,
2513and this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
2514method. In addition, all LLVM values can be named. The "name" of the
2515<tt>Value</tt> is a symbolic string printed in the LLVM code:</p>
2516
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2517<div class="doc_code">
2518<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2519%<b>foo</b> = add i32 1, 2
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2520</pre>
2521</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2522
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]2523<p><a name="nameWarning">The name of this instruction is "foo".</a> <b>NOTE</b>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2524that the name of any value may be missing (an empty string), so names should
2525<b>ONLY</b> be used for debugging (making the source code easier to read,
2526debugging printouts), they should not be used to keep track of values or map
2527between them. For this purpose, use a <tt>std::map</tt> of pointers to the
2528<tt>Value</tt> itself instead.</p>
2529
2530<p>One important aspect of LLVM is that there is no distinction between an SSA
2531variable and the operation that produces it. Because of this, any reference to
2532the value produced by an instruction (or the value available as an incoming
Chris Lattnerd5fc4fc2004-03-18 14:58:55 +0000[diff] [blame]2533argument, for example) is represented as a direct pointer to the instance of
2534the class that
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2535represents this value. Although this may take some getting used to, it
2536simplifies the representation and makes it easier to manipulate.</p>
2537
2538</div>
2539
2540<!-- _______________________________________________________________________ -->
2541<div class="doc_subsubsection">
2542 <a name="m_Value">Important Public Members of the <tt>Value</tt> class</a>
2543</div>
2544
2545<div class="doc_text">
2546
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2547<ul>
2548 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the
2549use-list<br>
2550 <tt>Value::use_const_iterator</tt> - Typedef for const_iterator over
2551the use-list<br>
2552 <tt>unsigned use_size()</tt> - Returns the number of users of the
2553value.<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]2554 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2555 <tt>use_iterator use_begin()</tt> - Get an iterator to the start of
2556the use-list.<br>
2557 <tt>use_iterator use_end()</tt> - Get an iterator to the end of the
2558use-list.<br>
2559 <tt><a href="#User">User</a> *use_back()</tt> - Returns the last
2560element in the list.
2561 <p> These methods are the interface to access the def-use
2562information in LLVM. As with all other iterators in LLVM, the naming
2563conventions follow the conventions defined by the <a href="#stl">STL</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2564 </li>
2565 <li><tt><a href="#Type">Type</a> *getType() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2566 <p>This method returns the Type of the Value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2567 </li>
2568 <li><tt>bool hasName() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]2569 <tt>std::string getName() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2570 <tt>void setName(const std::string &amp;Name)</tt>
2571 <p> This family of methods is used to access and assign a name to a <tt>Value</tt>,
2572be aware of the <a href="#nameWarning">precaution above</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2573 </li>
2574 <li><tt>void replaceAllUsesWith(Value *V)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2575
2576 <p>This method traverses the use list of a <tt>Value</tt> changing all <a
2577 href="#User"><tt>User</tt>s</a> of the current value to refer to
2578 "<tt>V</tt>" instead. For example, if you detect that an instruction always
2579 produces a constant value (for example through constant folding), you can
2580 replace all uses of the instruction with the constant like this:</p>
2581
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2582<div class="doc_code">
2583<pre>
2584Inst-&gt;replaceAllUsesWith(ConstVal);
2585</pre>
2586</div>
2587
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2588</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2589
2590</div>
2591
2592<!-- ======================================================================= -->
2593<div class="doc_subsection">
2594 <a name="User">The <tt>User</tt> class</a>
2595</div>
2596
2597<div class="doc_text">
2598
2599<p>
2600<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]2601doxygen info: <a href="/doxygen/classllvm_1_1User.html">User Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2602Superclass: <a href="#Value"><tt>Value</tt></a></p>
2603
2604<p>The <tt>User</tt> class is the common base class of all LLVM nodes that may
2605refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
2606that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
2607referring to. The <tt>User</tt> class itself is a subclass of
2608<tt>Value</tt>.</p>
2609
2610<p>The operands of a <tt>User</tt> point directly to the LLVM <a
2611href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
2612Single Assignment (SSA) form, there can only be one definition referred to,
2613allowing this direct connection. This connection provides the use-def
2614information in LLVM.</p>
2615
2616</div>
2617
2618<!-- _______________________________________________________________________ -->
2619<div class="doc_subsubsection">
2620 <a name="m_User">Important Public Members of the <tt>User</tt> class</a>
2621</div>
2622
2623<div class="doc_text">
2624
2625<p>The <tt>User</tt> class exposes the operand list in two ways: through
2626an index access interface and through an iterator based interface.</p>
2627
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2628<ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2629 <li><tt>Value *getOperand(unsigned i)</tt><br>
2630 <tt>unsigned getNumOperands()</tt>
2631 <p> These two methods expose the operands of the <tt>User</tt> in a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2632convenient form for direct access.</p></li>
2633
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2634 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand
2635list<br>
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]2636 <tt>op_iterator op_begin()</tt> - Get an iterator to the start of
2637the operand list.<br>
2638 <tt>op_iterator op_end()</tt> - Get an iterator to the end of the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2639operand list.
2640 <p> Together, these methods make up the iterator based interface to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2641the operands of a <tt>User</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2642</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2643
2644</div>
2645
2646<!-- ======================================================================= -->
2647<div class="doc_subsection">
2648 <a name="Instruction">The <tt>Instruction</tt> class</a>
2649</div>
2650
2651<div class="doc_text">
2652
2653<p><tt>#include "</tt><tt><a
2654href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt><br>
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]2655doxygen info: <a href="/doxygen/classllvm_1_1Instruction.html">Instruction Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2656Superclasses: <a href="#User"><tt>User</tt></a>, <a
2657href="#Value"><tt>Value</tt></a></p>
2658
2659<p>The <tt>Instruction</tt> class is the common base class for all LLVM
2660instructions. It provides only a few methods, but is a very commonly used
2661class. The primary data tracked by the <tt>Instruction</tt> class itself is the
2662opcode (instruction type) and the parent <a
2663href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
2664into. To represent a specific type of instruction, one of many subclasses of
2665<tt>Instruction</tt> are used.</p>
2666
2667<p> Because the <tt>Instruction</tt> class subclasses the <a
2668href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
2669way as for other <a href="#User"><tt>User</tt></a>s (with the
2670<tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
2671<tt>op_begin()</tt>/<tt>op_end()</tt> methods).</p> <p> An important file for
2672the <tt>Instruction</tt> class is the <tt>llvm/Instruction.def</tt> file. This
2673file contains some meta-data about the various different types of instructions
2674in LLVM. It describes the enum values that are used as opcodes (for example
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]2675<tt>Instruction::Add</tt> and <tt>Instruction::ICmp</tt>), as well as the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2676concrete sub-classes of <tt>Instruction</tt> that implement the instruction (for
2677example <tt><a href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]2678href="#CmpInst">CmpInst</a></tt>). Unfortunately, the use of macros in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2679this file confuses doxygen, so these enum values don't show up correctly in the
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]2680<a href="/doxygen/classllvm_1_1Instruction.html">doxygen output</a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2681
2682</div>
2683
2684<!-- _______________________________________________________________________ -->
2685<div class="doc_subsubsection">
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]2686 <a name="s_Instruction">Important Subclasses of the <tt>Instruction</tt>
2687 class</a>
2688</div>
2689<div class="doc_text">
2690 <ul>
2691 <li><tt><a name="BinaryOperator">BinaryOperator</a></tt>
2692 <p>This subclasses represents all two operand instructions whose operands
2693 must be the same type, except for the comparison instructions.</p></li>
2694 <li><tt><a name="CastInst">CastInst</a></tt>
2695 <p>This subclass is the parent of the 12 casting instructions. It provides
2696 common operations on cast instructions.</p>
2697 <li><tt><a name="CmpInst">CmpInst</a></tt>
2698 <p>This subclass respresents the two comparison instructions,
2699 <a href="LangRef.html#i_icmp">ICmpInst</a> (integer opreands), and
2700 <a href="LangRef.html#i_fcmp">FCmpInst</a> (floating point operands).</p>
2701 <li><tt><a name="TerminatorInst">TerminatorInst</a></tt>
2702 <p>This subclass is the parent of all terminator instructions (those which
2703 can terminate a block).</p>
2704 </ul>
2705 </div>
2706
2707<!-- _______________________________________________________________________ -->
2708<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2709 <a name="m_Instruction">Important Public Members of the <tt>Instruction</tt>
2710 class</a>
2711</div>
2712
2713<div class="doc_text">
2714
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2715<ul>
2716 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2717 <p>Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that
2718this <tt>Instruction</tt> is embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2719 <li><tt>bool mayWriteToMemory()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2720 <p>Returns true if the instruction writes to memory, i.e. it is a
2721 <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]2722 <li><tt>unsigned getOpcode()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2723 <p>Returns the opcode for the <tt>Instruction</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2724 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2725 <p>Returns another instance of the specified instruction, identical
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2726in all ways to the original except that the instruction has no parent
2727(ie it's not embedded into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>),
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2728and it has no name</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2729</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2730
2731</div>
2732
2733<!-- ======================================================================= -->
2734<div class="doc_subsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2735 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2736</div>
2737
2738<div class="doc_text">
2739
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2740<p>Constant represents a base class for different types of constants. It
2741is subclassed by ConstantInt, ConstantArray, etc. for representing
2742the various types of Constants. <a href="#GlobalValue">GlobalValue</a> is also
2743a subclass, which represents the address of a global variable or function.
2744</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2745
2746</div>
2747
2748<!-- _______________________________________________________________________ -->
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2749<div class="doc_subsubsection">Important Subclasses of Constant </div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2750<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2751<ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2752 <li>ConstantInt : This subclass of Constant represents an integer constant of
2753 any width.
2754 <ul>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]2755 <li><tt>const APInt&amp; getValue() const</tt>: Returns the underlying
2756 value of this constant, an APInt value.</li>
2757 <li><tt>int64_t getSExtValue() const</tt>: Converts the underlying APInt
2758 value to an int64_t via sign extension. If the value (not the bit width)
2759 of the APInt is too large to fit in an int64_t, an assertion will result.
2760 For this reason, use of this method is discouraged.</li>
2761 <li><tt>uint64_t getZExtValue() const</tt>: Converts the underlying APInt
2762 value to a uint64_t via zero extension. IF the value (not the bit width)
2763 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]2764 For this reason, use of this method is discouraged.</li>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]2765 <li><tt>static ConstantInt* get(const APInt&amp; Val)</tt>: Returns the
2766 ConstantInt object that represents the value provided by <tt>Val</tt>.
2767 The type is implied as the IntegerType that corresponds to the bit width
2768 of <tt>Val</tt>.</li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2769 <li><tt>static ConstantInt* get(const Type *Ty, uint64_t Val)</tt>:
2770 Returns the ConstantInt object that represents the value provided by
2771 <tt>Val</tt> for integer type <tt>Ty</tt>.</li>
2772 </ul>
2773 </li>
2774 <li>ConstantFP : This class represents a floating point constant.
2775 <ul>
2776 <li><tt>double getValue() const</tt>: Returns the underlying value of
2777 this constant. </li>
2778 </ul>
2779 </li>
2780 <li>ConstantArray : This represents a constant array.
2781 <ul>
2782 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
2783 a vector of component constants that makeup this array. </li>
2784 </ul>
2785 </li>
2786 <li>ConstantStruct : This represents a constant struct.
2787 <ul>
2788 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
2789 a vector of component constants that makeup this array. </li>
2790 </ul>
2791 </li>
2792 <li>GlobalValue : This represents either a global variable or a function. In
2793 either case, the value is a constant fixed address (after linking).
2794 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2795</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2796</div>
2797
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2798
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2799<!-- ======================================================================= -->
2800<div class="doc_subsection">
2801 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
2802</div>
2803
2804<div class="doc_text">
2805
2806<p><tt>#include "<a
2807href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt><br>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]2808doxygen info: <a href="/doxygen/classllvm_1_1GlobalValue.html">GlobalValue
2809Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]2810Superclasses: <a href="#Constant"><tt>Constant</tt></a>,
2811<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]2812
2813<p>Global values (<a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
2814href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
2815visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
2816Because they are visible at global scope, they are also subject to linking with
2817other globals defined in different translation units. To control the linking
2818process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
2819<tt>GlobalValue</tt>s know whether they have internal or external linkage, as
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]2820defined by the <tt>LinkageTypes</tt> enumeration.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2821
2822<p>If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
2823<tt>static</tt> in C), it is not visible to code outside the current translation
2824unit, and does not participate in linking. If it has external linkage, it is
2825visible to external code, and does participate in linking. In addition to
2826linkage information, <tt>GlobalValue</tt>s keep track of which <a
2827href="#Module"><tt>Module</tt></a> they are currently part of.</p>
2828
2829<p>Because <tt>GlobalValue</tt>s are memory objects, they are always referred to
2830by their <b>address</b>. As such, the <a href="#Type"><tt>Type</tt></a> of a
2831global is always a pointer to its contents. It is important to remember this
2832when using the <tt>GetElementPtrInst</tt> instruction because this pointer must
2833be dereferenced first. For example, if you have a <tt>GlobalVariable</tt> (a
2834subclass 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]2835i32]</tt>, then the <tt>GlobalVariable</tt> is a pointer to that array. Although
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2836the address of the first element of this array and the value of the
2837<tt>GlobalVariable</tt> are the same, they have different types. The
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2838<tt>GlobalVariable</tt>'s type is <tt>[24 x i32]</tt>. The first element's type
2839is <tt>i32.</tt> Because of this, accessing a global value requires you to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2840dereference the pointer with <tt>GetElementPtrInst</tt> first, then its elements
2841can be accessed. This is explained in the <a href="LangRef.html#globalvars">LLVM
2842Language Reference Manual</a>.</p>
2843
2844</div>
2845
2846<!-- _______________________________________________________________________ -->
2847<div class="doc_subsubsection">
2848 <a name="m_GlobalValue">Important Public Members of the <tt>GlobalValue</tt>
2849 class</a>
2850</div>
2851
2852<div class="doc_text">
2853
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2854<ul>
2855 <li><tt>bool hasInternalLinkage() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]2856 <tt>bool hasExternalLinkage() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2857 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt>
2858 <p> These methods manipulate the linkage characteristics of the <tt>GlobalValue</tt>.</p>
2859 <p> </p>
2860 </li>
2861 <li><tt><a href="#Module">Module</a> *getParent()</tt>
2862 <p> This returns the <a href="#Module"><tt>Module</tt></a> that the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2863GlobalValue is currently embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2864</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2865
2866</div>
2867
2868<!-- ======================================================================= -->
2869<div class="doc_subsection">
2870 <a name="Function">The <tt>Function</tt> class</a>
2871</div>
2872
2873<div class="doc_text">
2874
2875<p><tt>#include "<a
2876href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt><br> doxygen
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]2877info: <a href="/doxygen/classllvm_1_1Function.html">Function Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]2878Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
2879<a href="#Constant"><tt>Constant</tt></a>,
2880<a href="#User"><tt>User</tt></a>,
2881<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2882
2883<p>The <tt>Function</tt> class represents a single procedure in LLVM. It is
2884actually one of the more complex classes in the LLVM heirarchy because it must
2885keep track of a large amount of data. The <tt>Function</tt> class keeps track
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]2886of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal
2887<a href="#Argument"><tt>Argument</tt></a>s, and a
2888<a href="#SymbolTable"><tt>SymbolTable</tt></a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2889
2890<p>The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most
2891commonly used part of <tt>Function</tt> objects. The list imposes an implicit
2892ordering of the blocks in the function, which indicate how the code will be
2893layed out by the backend. Additionally, the first <a
2894href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
2895<tt>Function</tt>. It is not legal in LLVM to explicitly branch to this initial
2896block. There are no implicit exit nodes, and in fact there may be multiple exit
2897nodes from a single <tt>Function</tt>. If the <a
2898href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
2899the <tt>Function</tt> is actually a function declaration: the actual body of the
2900function hasn't been linked in yet.</p>
2901
2902<p>In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
2903<tt>Function</tt> class also keeps track of the list of formal <a
2904href="#Argument"><tt>Argument</tt></a>s that the function receives. This
2905container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
2906nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
2907the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.</p>
2908
2909<p>The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used
2910LLVM feature that is only used when you have to look up a value by name. Aside
2911from that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used
2912internally to make sure that there are not conflicts between the names of <a
2913href="#Instruction"><tt>Instruction</tt></a>s, <a
2914href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
2915href="#Argument"><tt>Argument</tt></a>s in the function body.</p>
2916
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]2917<p>Note that <tt>Function</tt> is a <a href="#GlobalValue">GlobalValue</a>
2918and therefore also a <a href="#Constant">Constant</a>. The value of the function
2919is its address (after linking) which is guaranteed to be constant.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2920</div>
2921
2922<!-- _______________________________________________________________________ -->
2923<div class="doc_subsubsection">
2924 <a name="m_Function">Important Public Members of the <tt>Function</tt>
2925 class</a>
2926</div>
2927
2928<div class="doc_text">
2929
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2930<ul>
2931 <li><tt>Function(const </tt><tt><a href="#FunctionType">FunctionType</a>
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]2932 *Ty, LinkageTypes Linkage, const std::string &amp;N = "", Module* Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2933
2934 <p>Constructor used when you need to create new <tt>Function</tt>s to add
2935 the the program. The constructor must specify the type of the function to
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]2936 create and what type of linkage the function should have. The <a
2937 href="#FunctionType"><tt>FunctionType</tt></a> argument
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2938 specifies the formal arguments and return value for the function. The same
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]2939 <a href="#FunctionType"><tt>FunctionType</tt></a> value can be used to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2940 create multiple functions. The <tt>Parent</tt> argument specifies the Module
2941 in which the function is defined. If this argument is provided, the function
2942 will automatically be inserted into that module's list of
2943 functions.</p></li>
2944
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2945 <li><tt>bool isExternal()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2946
2947 <p>Return whether or not the <tt>Function</tt> has a body defined. If the
2948 function is "external", it does not have a body, and thus must be resolved
2949 by linking with a function defined in a different translation unit.</p></li>
2950
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2951 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]2952 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2953
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]2954 <tt>begin()</tt>, <tt>end()</tt>
2955 <tt>size()</tt>, <tt>empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2956
2957 <p>These are forwarding methods that make it easy to access the contents of
2958 a <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
2959 list.</p></li>
2960
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2961 <li><tt>Function::BasicBlockListType &amp;getBasicBlockList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2962
2963 <p>Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This
2964 is necessary to use when you need to update the list or perform a complex
2965 action that doesn't have a forwarding method.</p></li>
2966
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]2967 <li><tt>Function::arg_iterator</tt> - Typedef for the argument list
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2968iterator<br>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]2969 <tt>Function::const_arg_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2970
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]2971 <tt>arg_begin()</tt>, <tt>arg_end()</tt>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]2972 <tt>arg_size()</tt>, <tt>arg_empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2973
2974 <p>These are forwarding methods that make it easy to access the contents of
2975 a <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a>
2976 list.</p></li>
2977
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2978 <li><tt>Function::ArgumentListType &amp;getArgumentList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2979
2980 <p>Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
2981 necessary to use when you need to update the list or perform a complex
2982 action that doesn't have a forwarding method.</p></li>
2983
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2984 <li><tt><a href="#BasicBlock">BasicBlock</a> &amp;getEntryBlock()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2985
2986 <p>Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
2987 function. Because the entry block for the function is always the first
2988 block, this returns the first block of the <tt>Function</tt>.</p></li>
2989
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2990 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
2991 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2992
2993 <p>This traverses the <a href="#Type"><tt>Type</tt></a> of the
2994 <tt>Function</tt> and returns the return type of the function, or the <a
2995 href="#FunctionType"><tt>FunctionType</tt></a> of the actual
2996 function.</p></li>
2997
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2998 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2999
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3000 <p> Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3001 for this <tt>Function</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3002</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3003
3004</div>
3005
3006<!-- ======================================================================= -->
3007<div class="doc_subsection">
3008 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
3009</div>
3010
3011<div class="doc_text">
3012
3013<p><tt>#include "<a
3014href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt>
3015<br>
Tanya Lattnera3da7772004-06-22 08:02:25 +0000[diff] [blame]3016doxygen info: <a href="/doxygen/classllvm_1_1GlobalVariable.html">GlobalVariable
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3017 Class</a><br>
3018Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
3019<a href="#Constant"><tt>Constant</tt></a>,
3020<a href="#User"><tt>User</tt></a>,
3021<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3022
3023<p>Global variables are represented with the (suprise suprise)
3024<tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are also
3025subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such are
3026always referenced by their address (global values must live in memory, so their
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3027"name" refers to their constant address). See
3028<a href="#GlobalValue"><tt>GlobalValue</tt></a> for more on this. Global
3029variables may have an initial value (which must be a
3030<a href="#Constant"><tt>Constant</tt></a>), and if they have an initializer,
3031they may be marked as "constant" themselves (indicating that their contents
3032never change at runtime).</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3033</div>
3034
3035<!-- _______________________________________________________________________ -->
3036<div class="doc_subsubsection">
3037 <a name="m_GlobalVariable">Important Public Members of the
3038 <tt>GlobalVariable</tt> class</a>
3039</div>
3040
3041<div class="doc_text">
3042
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3043<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3044 <li><tt>GlobalVariable(const </tt><tt><a href="#Type">Type</a> *Ty, bool
3045 isConstant, LinkageTypes&amp; Linkage, <a href="#Constant">Constant</a>
3046 *Initializer = 0, const std::string &amp;Name = "", Module* Parent = 0)</tt>
3047
3048 <p>Create a new global variable of the specified type. If
3049 <tt>isConstant</tt> is true then the global variable will be marked as
3050 unchanging for the program. The Linkage parameter specifies the type of
3051 linkage (internal, external, weak, linkonce, appending) for the variable. If
3052 the linkage is InternalLinkage, WeakLinkage, or LinkOnceLinkage,&nbsp; then
3053 the resultant global variable will have internal linkage. AppendingLinkage
3054 concatenates together all instances (in different translation units) of the
3055 variable into a single variable but is only applicable to arrays. &nbsp;See
3056 the <a href="LangRef.html#modulestructure">LLVM Language Reference</a> for
3057 further details on linkage types. Optionally an initializer, a name, and the
3058 module to put the variable into may be specified for the global variable as
3059 well.</p></li>
3060
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3061 <li><tt>bool isConstant() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3062
3063 <p>Returns true if this is a global variable that is known not to
3064 be modified at runtime.</p></li>
3065
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3066 <li><tt>bool hasInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3067
3068 <p>Returns true if this <tt>GlobalVariable</tt> has an intializer.</p></li>
3069
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3070 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3071
3072 <p>Returns the intial value for a <tt>GlobalVariable</tt>. It is not legal
3073 to call this method if there is no initializer.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3074</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3075
3076</div>
3077
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3078
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3079<!-- ======================================================================= -->
3080<div class="doc_subsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3081 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3082</div>
3083
3084<div class="doc_text">
3085
3086<p><tt>#include "<a
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3087href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt><br>
3088doxygen info: <a href="/doxygen/structllvm_1_1BasicBlock.html">BasicBlock
3089Class</a><br>
3090Superclass: <a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3091
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3092<p>This class represents a single entry multiple exit section of the code,
3093commonly known as a basic block by the compiler community. The
3094<tt>BasicBlock</tt> class maintains a list of <a
3095href="#Instruction"><tt>Instruction</tt></a>s, which form the body of the block.
3096Matching the language definition, the last element of this list of instructions
3097is always a terminator instruction (a subclass of the <a
3098href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).</p>
3099
3100<p>In addition to tracking the list of instructions that make up the block, the
3101<tt>BasicBlock</tt> class also keeps track of the <a
3102href="#Function"><tt>Function</tt></a> that it is embedded into.</p>
3103
3104<p>Note that <tt>BasicBlock</tt>s themselves are <a
3105href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
3106like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
3107<tt>label</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3108
3109</div>
3110
3111<!-- _______________________________________________________________________ -->
3112<div class="doc_subsubsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3113 <a name="m_BasicBlock">Important Public Members of the <tt>BasicBlock</tt>
3114 class</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3115</div>
3116
3117<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3118<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3119
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3120<li><tt>BasicBlock(const std::string &amp;Name = "", </tt><tt><a
3121 href="#Function">Function</a> *Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3122
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3123<p>The <tt>BasicBlock</tt> constructor is used to create new basic blocks for
3124insertion into a function. The constructor optionally takes a name for the new
3125block, and a <a href="#Function"><tt>Function</tt></a> to insert it into. If
3126the <tt>Parent</tt> parameter is specified, the new <tt>BasicBlock</tt> is
3127automatically inserted at the end of the specified <a
3128href="#Function"><tt>Function</tt></a>, if not specified, the BasicBlock must be
3129manually inserted into the <a href="#Function"><tt>Function</tt></a>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3130
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3131<li><tt>BasicBlock::iterator</tt> - Typedef for instruction list iterator<br>
3132<tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
3133<tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
3134<tt>size()</tt>, <tt>empty()</tt>
3135STL-style functions for accessing the instruction list.
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3136
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3137<p>These methods and typedefs are forwarding functions that have the same
3138semantics as the standard library methods of the same names. These methods
3139expose the underlying instruction list of a basic block in a way that is easy to
3140manipulate. To get the full complement of container operations (including
3141operations to update the list), you must use the <tt>getInstList()</tt>
3142method.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3143
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3144<li><tt>BasicBlock::InstListType &amp;getInstList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3145
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3146<p>This method is used to get access to the underlying container that actually
3147holds the Instructions. This method must be used when there isn't a forwarding
3148function in the <tt>BasicBlock</tt> class for the operation that you would like
3149to perform. Because there are no forwarding functions for "updating"
3150operations, you need to use this if you want to update the contents of a
3151<tt>BasicBlock</tt>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3152
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3153<li><tt><a href="#Function">Function</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3154
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3155<p> Returns a pointer to <a href="#Function"><tt>Function</tt></a> the block is
3156embedded into, or a null pointer if it is homeless.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3157
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3158<li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3159
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3160<p> Returns a pointer to the terminator instruction that appears at the end of
3161the <tt>BasicBlock</tt>. If there is no terminator instruction, or if the last
3162instruction in the block is not a terminator, then a null pointer is
3163returned.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3164
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3165</ul>
3166
3167</div>
3168
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3169
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3170<!-- ======================================================================= -->
3171<div class="doc_subsection">
3172 <a name="Argument">The <tt>Argument</tt> class</a>
3173</div>
3174
3175<div class="doc_text">
3176
3177<p>This subclass of Value defines the interface for incoming formal
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]3178arguments to a function. A Function maintains a list of its formal
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3179arguments. An argument has a pointer to the parent Function.</p>
3180
3181</div>
3182
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3183<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3184<hr>
3185<address>
3186 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
3187 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
3188 <a href="http://validator.w3.org/check/referer"><img
3189 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
3190
3191 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
3192 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]3193 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3194 Last modified: $Date$
3195</address>
3196
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3197</body>
3198</html>