blob: 1d5ad0e82c0d8319fa5dea47da7e8f80c5f5ee46 [file] [log] [blame]
Misha Brukman13fd15c2004-01-15 00:14:41 +00001<!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 +00003<html>
4<head>
5 <title>LLVM Programmer's Manual</title>
Misha Brukman13fd15c2004-01-15 00:14:41 +00006 <link rel="stylesheet" href="llvm.css" type="text/css">
Chris Lattner261efe92003-11-25 01:02:51 +00007</head>
Misha Brukman13fd15c2004-01-15 00:14:41 +00008<body>
9
10<div class="doc_title">
11 LLVM Programmer's Manual
12</div>
13
Chris Lattner9355b472002-09-06 02:50:58 +000014<ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +000015 <li><a href="#introduction">Introduction</a></li>
Chris Lattner9355b472002-09-06 02:50:58 +000016 <li><a href="#general">General Information</a>
Chris Lattner261efe92003-11-25 01:02:51 +000017 <ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +000018 <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 +000023
Reid Spencerfe8f4ff2004-11-01 09:02:53 +000024-->
Chris Lattner84b7f8d2003-08-01 22:20:59 +000025 </ul>
Chris Lattner261efe92003-11-25 01:02:51 +000026 </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 +000031 <li><a href="#DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt>
Chris Lattner261efe92003-11-25 01:02:51 +000032option</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 +000038 <li><a href="#Statistic">The <tt>Statistic</tt> class &amp; <tt>-stats</tt>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +000039option</a></li>
40<!--
41 <li>The <tt>InstVisitor</tt> template
42 <li>The general graph API
43-->
Chris Lattnerf623a082005-10-17 01:36:23 +000044 <li><a href="#ViewGraph">Viewing graphs while debugging code</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000045 </ul>
46 </li>
Chris Lattner098129a2007-02-03 03:04:03 +000047 <li><a href="#datastructure">Picking the Right Data Structure for a Task</a>
48 <ul>
Chris Lattner74c4ca12007-02-03 07:59:07 +000049 <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 +000058 <li><a href="#dss_other">Other Sequential Container Options</a></li>
Chris Lattner098129a2007-02-03 03:04:03 +000059 </ul></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +000060 <li><a href="#ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
61 <ul>
62 <li><a href="#dss_sortedvectorset">A sorted 'vector'</a></li>
63 <li><a href="#dss_smallset">"llvm/ADT/SmallSet.h"</a></li>
64 <li><a href="#dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a></li>
65 <li><a href="#dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a></li>
66 <li><a href="#dss_set">&lt;set&gt;</a></li>
67 <li><a href="#dss_setvector">"llvm/ADT/SetVector.h"</a></li>
Chris Lattnerc5722432007-02-03 19:49:31 +000068 <li><a href="#dss_uniquevector">"llvm/ADT/UniqueVector.h"</a></li>
69 <li><a href="#dss_otherset">Other Set-Like ContainerOptions</a></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +000070 </ul></li>
Chris Lattnerf3692522007-02-03 19:51:56 +000071 <li><a href="#ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
72 <ul>
73 <li><a href="#dss_sortedvectormap">A sorted 'vector'</a></li>
74 <li><a href="#dss_cstringmap">"llvm/ADT/CStringMap.h"</a></li>
75 <li><a href="#dss_indexedmap">"llvm/ADT/IndexedMap.h"</a></li>
76 <li><a href="#dss_densemap">"llvm/ADT/DenseMap.h"</a></li>
77 <li><a href="#dss_map">&lt;map&gt;</a></li>
78 <li><a href="#dss_othermap">Other Map-Like Container Options</a></li>
79 </ul></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +000080 </ul>
Chris Lattner098129a2007-02-03 03:04:03 +000081 </li>
Chris Lattnerae7f7592002-09-06 18:31:18 +000082 <li><a href="#common">Helpful Hints for Common Operations</a>
Chris Lattnerae7f7592002-09-06 18:31:18 +000083 <ul>
Chris Lattner261efe92003-11-25 01:02:51 +000084 <li><a href="#inspection">Basic Inspection and Traversal Routines</a>
85 <ul>
86 <li><a href="#iterate_function">Iterating over the <tt>BasicBlock</tt>s
87in a <tt>Function</tt></a> </li>
88 <li><a href="#iterate_basicblock">Iterating over the <tt>Instruction</tt>s
89in a <tt>BasicBlock</tt></a> </li>
90 <li><a href="#iterate_institer">Iterating over the <tt>Instruction</tt>s
91in a <tt>Function</tt></a> </li>
92 <li><a href="#iterate_convert">Turning an iterator into a
93class pointer</a> </li>
94 <li><a href="#iterate_complex">Finding call sites: a more
95complex example</a> </li>
96 <li><a href="#calls_and_invokes">Treating calls and invokes
97the same way</a> </li>
98 <li><a href="#iterate_chains">Iterating over def-use &amp;
99use-def chains</a> </li>
100 </ul>
101 </li>
102 <li><a href="#simplechanges">Making simple changes</a>
103 <ul>
104 <li><a href="#schanges_creating">Creating and inserting new
105 <tt>Instruction</tt>s</a> </li>
106 <li><a href="#schanges_deleting">Deleting <tt>Instruction</tt>s</a> </li>
107 <li><a href="#schanges_replacing">Replacing an <tt>Instruction</tt>
108with another <tt>Value</tt></a> </li>
109 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000110 </li>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000111<!--
112 <li>Working with the Control Flow Graph
113 <ul>
114 <li>Accessing predecessors and successors of a <tt>BasicBlock</tt>
115 <li>
116 <li>
117 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000118-->
Chris Lattner261efe92003-11-25 01:02:51 +0000119 </ul>
120 </li>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000121
122 <li><a href="#advanced">Advanced Topics</a>
123 <ul>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000124 <li><a href="#TypeResolve">LLVM Type Resolution</a>
125 <ul>
126 <li><a href="#BuildRecType">Basic Recursive Type Construction</a></li>
127 <li><a href="#refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a></li>
128 <li><a href="#PATypeHolder">The PATypeHolder Class</a></li>
129 <li><a href="#AbstractTypeUser">The AbstractTypeUser Class</a></li>
130 </ul></li>
131
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000132 <li><a href="#SymbolTable">The <tt>SymbolTable</tt> class </a></li>
133 </ul></li>
134
Joel Stanley9b96c442002-09-06 21:55:13 +0000135 <li><a href="#coreclasses">The Core LLVM Class Hierarchy Reference</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000136 <ul>
Reid Spencer303c4b42007-01-12 17:26:25 +0000137 <li><a href="#Type">The <tt>Type</tt> class</a> </li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000138 <li><a href="#Module">The <tt>Module</tt> class</a></li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000139 <li><a href="#Value">The <tt>Value</tt> class</a>
Chris Lattner2b78d962007-02-03 20:02:25 +0000140 <ul>
141 <li><a href="#User">The <tt>User</tt> class</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000142 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000143 <li><a href="#Instruction">The <tt>Instruction</tt> class</a></li>
144 <li><a href="#Constant">The <tt>Constant</tt> class</a>
145 <ul>
146 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000147 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000148 <li><a href="#Function">The <tt>Function</tt> class</a></li>
149 <li><a href="#GlobalVariable">The <tt>GlobalVariable</tt> class</a></li>
150 </ul>
151 </li>
152 </ul>
153 </li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000154 </ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000155 </li>
156 <li><a href="#BasicBlock">The <tt>BasicBlock</tt> class</a></li>
157 <li><a href="#Argument">The <tt>Argument</tt> class</a></li>
158 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000159 </li>
160 </ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000161 </li>
Chris Lattner9355b472002-09-06 02:50:58 +0000162</ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000163
Chris Lattner69bf8a92004-05-23 21:06:58 +0000164<div class="doc_author">
165 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
Chris Lattner94c43592004-05-26 16:52:55 +0000166 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>,
167 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a>, and
168 <a href="mailto:rspencer@x10sys.com">Reid Spencer</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000169</div>
170
Chris Lattner9355b472002-09-06 02:50:58 +0000171<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000172<div class="doc_section">
173 <a name="introduction">Introduction </a>
174</div>
Chris Lattner9355b472002-09-06 02:50:58 +0000175<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000176
177<div class="doc_text">
178
179<p>This document is meant to highlight some of the important classes and
Chris Lattner261efe92003-11-25 01:02:51 +0000180interfaces available in the LLVM source-base. This manual is not
181intended to explain what LLVM is, how it works, and what LLVM code looks
182like. It assumes that you know the basics of LLVM and are interested
183in writing transformations or otherwise analyzing or manipulating the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000184code.</p>
185
186<p>This document should get you oriented so that you can find your
Chris Lattner261efe92003-11-25 01:02:51 +0000187way in the continuously growing source code that makes up the LLVM
188infrastructure. Note that this manual is not intended to serve as a
189replacement for reading the source code, so if you think there should be
190a method in one of these classes to do something, but it's not listed,
191check the source. Links to the <a href="/doxygen/">doxygen</a> sources
192are provided to make this as easy as possible.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000193
194<p>The first section of this document describes general information that is
195useful to know when working in the LLVM infrastructure, and the second describes
196the Core LLVM classes. In the future this manual will be extended with
197information describing how to use extension libraries, such as dominator
198information, CFG traversal routines, and useful utilities like the <tt><a
199href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.</p>
200
201</div>
202
Chris Lattner9355b472002-09-06 02:50:58 +0000203<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000204<div class="doc_section">
205 <a name="general">General Information</a>
206</div>
207<!-- *********************************************************************** -->
208
209<div class="doc_text">
210
211<p>This section contains general information that is useful if you are working
212in the LLVM source-base, but that isn't specific to any particular API.</p>
213
214</div>
215
216<!-- ======================================================================= -->
217<div class="doc_subsection">
218 <a name="stl">The C++ Standard Template Library</a>
219</div>
220
221<div class="doc_text">
222
223<p>LLVM makes heavy use of the C++ Standard Template Library (STL),
Chris Lattner261efe92003-11-25 01:02:51 +0000224perhaps much more than you are used to, or have seen before. Because of
225this, you might want to do a little background reading in the
226techniques used and capabilities of the library. There are many good
227pages that discuss the STL, and several books on the subject that you
Misha Brukman13fd15c2004-01-15 00:14:41 +0000228can get, so it will not be discussed in this document.</p>
229
230<p>Here are some useful links:</p>
231
232<ol>
233
234<li><a href="http://www.dinkumware.com/refxcpp.html">Dinkumware C++ Library
235reference</a> - an excellent reference for the STL and other parts of the
236standard C++ library.</li>
237
238<li><a href="http://www.tempest-sw.com/cpp/">C++ In a Nutshell</a> - This is an
Tanya Lattner09cf73c2004-06-22 04:24:55 +0000239O'Reilly book in the making. It has a decent
240Standard Library
241Reference that rivals Dinkumware's, and is unfortunately no longer free since the book has been
Misha Brukman13fd15c2004-01-15 00:14:41 +0000242published.</li>
243
244<li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
245Questions</a></li>
246
247<li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
248Contains a useful <a
249href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
250STL</a>.</li>
251
252<li><a href="http://www.research.att.com/%7Ebs/C++.html">Bjarne Stroustrup's C++
253Page</a></li>
254
Tanya Lattner79445ba2004-12-08 18:34:56 +0000255<li><a href="http://64.78.49.204/">
Reid Spencer096603a2004-05-26 08:41:35 +0000256Bruce Eckel's Thinking in C++, 2nd ed. Volume 2 Revision 4.0 (even better, get
257the book).</a></li>
258
Misha Brukman13fd15c2004-01-15 00:14:41 +0000259</ol>
260
261<p>You are also encouraged to take a look at the <a
262href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
263to write maintainable code more than where to put your curly braces.</p>
264
265</div>
266
267<!-- ======================================================================= -->
268<div class="doc_subsection">
269 <a name="stl">Other useful references</a>
270</div>
271
272<div class="doc_text">
273
Misha Brukman13fd15c2004-01-15 00:14:41 +0000274<ol>
275<li><a href="http://www.psc.edu/%7Esemke/cvs_branches.html">CVS
Chris Lattner261efe92003-11-25 01:02:51 +0000276Branch and Tag Primer</a></li>
Misha Brukmana0f71e42004-06-18 18:39:00 +0000277<li><a href="http://www.fortran-2000.com/ArnaudRecipes/sharedlib.html">Using
278static and shared libraries across platforms</a></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000279</ol>
280
281</div>
282
Chris Lattner9355b472002-09-06 02:50:58 +0000283<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000284<div class="doc_section">
285 <a name="apis">Important and useful LLVM APIs</a>
286</div>
287<!-- *********************************************************************** -->
288
289<div class="doc_text">
290
291<p>Here we highlight some LLVM APIs that are generally useful and good to
292know about when writing transformations.</p>
293
294</div>
295
296<!-- ======================================================================= -->
297<div class="doc_subsection">
Misha Brukman2c122ce2005-11-01 21:12:49 +0000298 <a name="isa">The <tt>isa&lt;&gt;</tt>, <tt>cast&lt;&gt;</tt> and
299 <tt>dyn_cast&lt;&gt;</tt> templates</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000300</div>
301
302<div class="doc_text">
303
304<p>The LLVM source-base makes extensive use of a custom form of RTTI.
Chris Lattner261efe92003-11-25 01:02:51 +0000305These templates have many similarities to the C++ <tt>dynamic_cast&lt;&gt;</tt>
306operator, but they don't have some drawbacks (primarily stemming from
307the fact that <tt>dynamic_cast&lt;&gt;</tt> only works on classes that
308have a v-table). Because they are used so often, you must know what they
309do and how they work. All of these templates are defined in the <a
Chris Lattner695b78b2005-04-26 22:56:16 +0000310 href="/doxygen/Casting_8h-source.html"><tt>llvm/Support/Casting.h</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000311file (note that you very rarely have to include this file directly).</p>
312
313<dl>
314 <dt><tt>isa&lt;&gt;</tt>: </dt>
315
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000316 <dd><p>The <tt>isa&lt;&gt;</tt> operator works exactly like the Java
Misha Brukman13fd15c2004-01-15 00:14:41 +0000317 "<tt>instanceof</tt>" operator. It returns true or false depending on whether
318 a reference or pointer points to an instance of the specified class. This can
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000319 be very useful for constraint checking of various sorts (example below).</p>
320 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000321
322 <dt><tt>cast&lt;&gt;</tt>: </dt>
323
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000324 <dd><p>The <tt>cast&lt;&gt;</tt> operator is a "checked cast" operation. It
Misha Brukman13fd15c2004-01-15 00:14:41 +0000325 converts a pointer or reference from a base class to a derived cast, causing
326 an assertion failure if it is not really an instance of the right type. This
327 should be used in cases where you have some information that makes you believe
328 that something is of the right type. An example of the <tt>isa&lt;&gt;</tt>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000329 and <tt>cast&lt;&gt;</tt> template is:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000330
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000331<div class="doc_code">
332<pre>
333static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
334 if (isa&lt;<a href="#Constant">Constant</a>&gt;(V) || isa&lt;<a href="#Argument">Argument</a>&gt;(V) || isa&lt;<a href="#GlobalValue">GlobalValue</a>&gt;(V))
335 return true;
Chris Lattner69bf8a92004-05-23 21:06:58 +0000336
Bill Wendling82e2eea2006-10-11 18:00:22 +0000337 // <i>Otherwise, it must be an instruction...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000338 return !L-&gt;contains(cast&lt;<a href="#Instruction">Instruction</a>&gt;(V)-&gt;getParent());
339}
340</pre>
341</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000342
343 <p>Note that you should <b>not</b> use an <tt>isa&lt;&gt;</tt> test followed
344 by a <tt>cast&lt;&gt;</tt>, for that use the <tt>dyn_cast&lt;&gt;</tt>
345 operator.</p>
346
347 </dd>
348
349 <dt><tt>dyn_cast&lt;&gt;</tt>:</dt>
350
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000351 <dd><p>The <tt>dyn_cast&lt;&gt;</tt> operator is a "checking cast" operation.
352 It checks to see if the operand is of the specified type, and if so, returns a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000353 pointer to it (this operator does not work with references). If the operand is
354 not of the correct type, a null pointer is returned. Thus, this works very
Misha Brukman2c122ce2005-11-01 21:12:49 +0000355 much like the <tt>dynamic_cast&lt;&gt;</tt> operator in C++, and should be
356 used in the same circumstances. Typically, the <tt>dyn_cast&lt;&gt;</tt>
357 operator is used in an <tt>if</tt> statement or some other flow control
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000358 statement like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000359
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000360<div class="doc_code">
361<pre>
362if (<a href="#AllocationInst">AllocationInst</a> *AI = dyn_cast&lt;<a href="#AllocationInst">AllocationInst</a>&gt;(Val)) {
Bill Wendling82e2eea2006-10-11 18:00:22 +0000363 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000364}
365</pre>
366</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000367
Misha Brukman2c122ce2005-11-01 21:12:49 +0000368 <p>This form of the <tt>if</tt> statement effectively combines together a call
369 to <tt>isa&lt;&gt;</tt> and a call to <tt>cast&lt;&gt;</tt> into one
370 statement, which is very convenient.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000371
Misha Brukman2c122ce2005-11-01 21:12:49 +0000372 <p>Note that the <tt>dyn_cast&lt;&gt;</tt> operator, like C++'s
373 <tt>dynamic_cast&lt;&gt;</tt> or Java's <tt>instanceof</tt> operator, can be
374 abused. In particular, you should not use big chained <tt>if/then/else</tt>
375 blocks to check for lots of different variants of classes. If you find
376 yourself wanting to do this, it is much cleaner and more efficient to use the
377 <tt>InstVisitor</tt> class to dispatch over the instruction type directly.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000378
Misha Brukman2c122ce2005-11-01 21:12:49 +0000379 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000380
Misha Brukman2c122ce2005-11-01 21:12:49 +0000381 <dt><tt>cast_or_null&lt;&gt;</tt>: </dt>
382
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000383 <dd><p>The <tt>cast_or_null&lt;&gt;</tt> operator works just like the
Misha Brukman2c122ce2005-11-01 21:12:49 +0000384 <tt>cast&lt;&gt;</tt> operator, except that it allows for a null pointer as an
385 argument (which it then propagates). This can sometimes be useful, allowing
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000386 you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000387
Misha Brukman2c122ce2005-11-01 21:12:49 +0000388 <dt><tt>dyn_cast_or_null&lt;&gt;</tt>: </dt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000389
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000390 <dd><p>The <tt>dyn_cast_or_null&lt;&gt;</tt> operator works just like the
Misha Brukman2c122ce2005-11-01 21:12:49 +0000391 <tt>dyn_cast&lt;&gt;</tt> operator, except that it allows for a null pointer
392 as an argument (which it then propagates). This can sometimes be useful,
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000393 allowing you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000394
Misha Brukman2c122ce2005-11-01 21:12:49 +0000395</dl>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000396
397<p>These five templates can be used with any classes, whether they have a
398v-table or not. To add support for these templates, you simply need to add
399<tt>classof</tt> static methods to the class you are interested casting
400to. Describing this is currently outside the scope of this document, but there
401are lots of examples in the LLVM source base.</p>
402
403</div>
404
405<!-- ======================================================================= -->
406<div class="doc_subsection">
Misha Brukman2c122ce2005-11-01 21:12:49 +0000407 <a name="DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt> option</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000408</div>
409
410<div class="doc_text">
411
412<p>Often when working on your pass you will put a bunch of debugging printouts
413and other code into your pass. After you get it working, you want to remove
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000414it, but you may need it again in the future (to work out new bugs that you run
Misha Brukman13fd15c2004-01-15 00:14:41 +0000415across).</p>
416
417<p> Naturally, because of this, you don't want to delete the debug printouts,
418but you don't want them to always be noisy. A standard compromise is to comment
419them out, allowing you to enable them if you need them in the future.</p>
420
Chris Lattner695b78b2005-04-26 22:56:16 +0000421<p>The "<tt><a href="/doxygen/Debug_8h-source.html">llvm/Support/Debug.h</a></tt>"
Misha Brukman13fd15c2004-01-15 00:14:41 +0000422file provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to
423this problem. Basically, you can put arbitrary code into the argument of the
424<tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
425tool) is run with the '<tt>-debug</tt>' command line argument:</p>
426
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000427<div class="doc_code">
428<pre>
Bill Wendling832171c2006-12-07 20:04:42 +0000429DOUT &lt;&lt; "I am here!\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000430</pre>
431</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000432
433<p>Then you can run your pass like this:</p>
434
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000435<div class="doc_code">
436<pre>
437$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000438<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000439$ opt &lt; a.bc &gt; /dev/null -mypass -debug
440I am here!
441</pre>
442</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000443
444<p>Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you
445to not have to create "yet another" command line option for the debug output for
446your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
447so they do not cause a performance impact at all (for the same reason, they
448should also not contain side-effects!).</p>
449
450<p>One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
451enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
452"<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
453program hasn't been started yet, you can always just run it with
454<tt>-debug</tt>.</p>
455
456</div>
457
458<!-- _______________________________________________________________________ -->
459<div class="doc_subsubsection">
Chris Lattnerc9151082005-04-26 22:57:07 +0000460 <a name="DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt> and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000461 the <tt>-debug-only</tt> option</a>
462</div>
463
464<div class="doc_text">
465
466<p>Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
467just turns on <b>too much</b> information (such as when working on the code
468generator). If you want to enable debug information with more fine-grained
469control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
470option as follows:</p>
471
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000472<div class="doc_code">
473<pre>
Bill Wendling832171c2006-12-07 20:04:42 +0000474DOUT &lt;&lt; "No debug type\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000475#undef DEBUG_TYPE
476#define DEBUG_TYPE "foo"
Bill Wendling832171c2006-12-07 20:04:42 +0000477DOUT &lt;&lt; "'foo' debug type\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000478#undef DEBUG_TYPE
479#define DEBUG_TYPE "bar"
Bill Wendling832171c2006-12-07 20:04:42 +0000480DOUT &lt;&lt; "'bar' debug type\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000481#undef DEBUG_TYPE
482#define DEBUG_TYPE ""
Bill Wendling832171c2006-12-07 20:04:42 +0000483DOUT &lt;&lt; "No debug type (2)\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000484</pre>
485</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000486
487<p>Then you can run your pass like this:</p>
488
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000489<div class="doc_code">
490<pre>
491$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000492<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000493$ opt &lt; a.bc &gt; /dev/null -mypass -debug
494No debug type
495'foo' debug type
496'bar' debug type
497No debug type (2)
498$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=foo
499'foo' debug type
500$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=bar
501'bar' debug type
502</pre>
503</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000504
505<p>Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of
506a file, to specify the debug type for the entire module (if you do this before
Chris Lattner695b78b2005-04-26 22:56:16 +0000507you <tt>#include "llvm/Support/Debug.h"</tt>, you don't have to insert the ugly
Misha Brukman13fd15c2004-01-15 00:14:41 +0000508<tt>#undef</tt>'s). Also, you should use names more meaningful than "foo" and
509"bar", because there is no system in place to ensure that names do not
510conflict. If two different modules use the same string, they will all be turned
511on when the name is specified. This allows, for example, all debug information
512for instruction scheduling to be enabled with <tt>-debug-type=InstrSched</tt>,
Chris Lattner261efe92003-11-25 01:02:51 +0000513even if the source lives in multiple files.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000514
515</div>
516
517<!-- ======================================================================= -->
518<div class="doc_subsection">
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000519 <a name="Statistic">The <tt>Statistic</tt> class &amp; <tt>-stats</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000520 option</a>
521</div>
522
523<div class="doc_text">
524
525<p>The "<tt><a
Chris Lattner695b78b2005-04-26 22:56:16 +0000526href="/doxygen/Statistic_8h-source.html">llvm/ADT/Statistic.h</a></tt>" file
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000527provides a class named <tt>Statistic</tt> that is used as a unified way to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000528keep track of what the LLVM compiler is doing and how effective various
529optimizations are. It is useful to see what optimizations are contributing to
530making a particular program run faster.</p>
531
532<p>Often you may run your pass on some big program, and you're interested to see
533how many times it makes a certain transformation. Although you can do this with
534hand inspection, or some ad-hoc method, this is a real pain and not very useful
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000535for big programs. Using the <tt>Statistic</tt> class makes it very easy to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000536keep track of this information, and the calculated information is presented in a
537uniform manner with the rest of the passes being executed.</p>
538
539<p>There are many examples of <tt>Statistic</tt> uses, but the basics of using
540it are as follows:</p>
541
542<ol>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000543 <li><p>Define your statistic like this:</p>
544
545<div class="doc_code">
546<pre>
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000547#define <a href="#DEBUG_TYPE">DEBUG_TYPE</a> "mypassname" <i>// This goes before any #includes.</i>
548STATISTIC(NumXForms, "The # of times I did stuff");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000549</pre>
550</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000551
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000552 <p>The <tt>STATISTIC</tt> macro defines a static variable, whose name is
553 specified by the first argument. The pass name is taken from the DEBUG_TYPE
554 macro, and the description is taken from the second argument. The variable
Reid Spencer06565dc2007-01-12 17:11:23 +0000555 defined ("NumXForms" in this case) acts like an unsigned integer.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000556
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000557 <li><p>Whenever you make a transformation, bump the counter:</p>
558
559<div class="doc_code">
560<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000561++NumXForms; // <i>I did stuff!</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000562</pre>
563</div>
564
Chris Lattner261efe92003-11-25 01:02:51 +0000565 </li>
566 </ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000567
568 <p>That's all you have to do. To get '<tt>opt</tt>' to print out the
569 statistics gathered, use the '<tt>-stats</tt>' option:</p>
570
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000571<div class="doc_code">
572<pre>
573$ opt -stats -mypassname &lt; program.bc &gt; /dev/null
Bill Wendling82e2eea2006-10-11 18:00:22 +0000574<i>... statistics output ...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000575</pre>
576</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000577
Chris Lattner261efe92003-11-25 01:02:51 +0000578 <p> When running <tt>gccas</tt> on a C file from the SPEC benchmark
579suite, it gives a report that looks like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000580
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000581<div class="doc_code">
582<pre>
583 7646 bytecodewriter - Number of normal instructions
584 725 bytecodewriter - Number of oversized instructions
585 129996 bytecodewriter - Number of bytecode bytes written
586 2817 raise - Number of insts DCEd or constprop'd
587 3213 raise - Number of cast-of-self removed
588 5046 raise - Number of expression trees converted
589 75 raise - Number of other getelementptr's formed
590 138 raise - Number of load/store peepholes
591 42 deadtypeelim - Number of unused typenames removed from symtab
592 392 funcresolve - Number of varargs functions resolved
593 27 globaldce - Number of global variables removed
594 2 adce - Number of basic blocks removed
595 134 cee - Number of branches revectored
596 49 cee - Number of setcc instruction eliminated
597 532 gcse - Number of loads removed
598 2919 gcse - Number of instructions removed
599 86 indvars - Number of canonical indvars added
600 87 indvars - Number of aux indvars removed
601 25 instcombine - Number of dead inst eliminate
602 434 instcombine - Number of insts combined
603 248 licm - Number of load insts hoisted
604 1298 licm - Number of insts hoisted to a loop pre-header
605 3 licm - Number of insts hoisted to multiple loop preds (bad, no loop pre-header)
606 75 mem2reg - Number of alloca's promoted
607 1444 cfgsimplify - Number of blocks simplified
608</pre>
609</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000610
611<p>Obviously, with so many optimizations, having a unified framework for this
612stuff is very nice. Making your pass fit well into the framework makes it more
613maintainable and useful.</p>
614
615</div>
616
Chris Lattnerf623a082005-10-17 01:36:23 +0000617<!-- ======================================================================= -->
618<div class="doc_subsection">
619 <a name="ViewGraph">Viewing graphs while debugging code</a>
620</div>
621
622<div class="doc_text">
623
624<p>Several of the important data structures in LLVM are graphs: for example
625CFGs made out of LLVM <a href="#BasicBlock">BasicBlock</a>s, CFGs made out of
626LLVM <a href="CodeGenerator.html#machinebasicblock">MachineBasicBlock</a>s, and
627<a href="CodeGenerator.html#selectiondag_intro">Instruction Selection
628DAGs</a>. In many cases, while debugging various parts of the compiler, it is
629nice to instantly visualize these graphs.</p>
630
631<p>LLVM provides several callbacks that are available in a debug build to do
632exactly that. If you call the <tt>Function::viewCFG()</tt> method, for example,
633the current LLVM tool will pop up a window containing the CFG for the function
634where each basic block is a node in the graph, and each node contains the
635instructions in the block. Similarly, there also exists
636<tt>Function::viewCFGOnly()</tt> (does not include the instructions), the
637<tt>MachineFunction::viewCFG()</tt> and <tt>MachineFunction::viewCFGOnly()</tt>,
638and the <tt>SelectionDAG::viewGraph()</tt> methods. Within GDB, for example,
Jim Laskey543a0ee2006-10-02 12:28:07 +0000639you can usually use something like <tt>call DAG.viewGraph()</tt> to pop
Chris Lattnerf623a082005-10-17 01:36:23 +0000640up a window. Alternatively, you can sprinkle calls to these functions in your
641code in places you want to debug.</p>
642
643<p>Getting this to work requires a small amount of configuration. On Unix
644systems with X11, install the <a href="http://www.graphviz.org">graphviz</a>
645toolkit, and make sure 'dot' and 'gv' are in your path. If you are running on
646Mac OS/X, download and install the Mac OS/X <a
647href="http://www.pixelglow.com/graphviz/">Graphviz program</a>, and add
Reid Spencer128a7a72007-02-03 21:06:43 +0000648<tt>/Applications/Graphviz.app/Contents/MacOS/</tt> (or wherever you install
Chris Lattnerf623a082005-10-17 01:36:23 +0000649it) to your path. Once in your system and path are set up, rerun the LLVM
650configure script and rebuild LLVM to enable this functionality.</p>
651
Jim Laskey543a0ee2006-10-02 12:28:07 +0000652<p><tt>SelectionDAG</tt> has been extended to make it easier to locate
653<i>interesting</i> nodes in large complex graphs. From gdb, if you
654<tt>call DAG.setGraphColor(<i>node</i>, "<i>color</i>")</tt>, then the
Reid Spencer128a7a72007-02-03 21:06:43 +0000655next <tt>call DAG.viewGraph()</tt> would highlight the node in the
Jim Laskey543a0ee2006-10-02 12:28:07 +0000656specified color (choices of colors can be found at <a
Chris Lattner302da1e2007-02-03 03:05:57 +0000657href="http://www.graphviz.org/doc/info/colors.html">colors</a>.) More
Jim Laskey543a0ee2006-10-02 12:28:07 +0000658complex node attributes can be provided with <tt>call
659DAG.setGraphAttrs(<i>node</i>, "<i>attributes</i>")</tt> (choices can be
660found at <a href="http://www.graphviz.org/doc/info/attrs.html">Graph
661Attributes</a>.) If you want to restart and clear all the current graph
662attributes, then you can <tt>call DAG.clearGraphAttrs()</tt>. </p>
663
Chris Lattnerf623a082005-10-17 01:36:23 +0000664</div>
665
Chris Lattner098129a2007-02-03 03:04:03 +0000666<!-- *********************************************************************** -->
667<div class="doc_section">
668 <a name="datastructure">Picking the Right Data Structure for a Task</a>
669</div>
670<!-- *********************************************************************** -->
671
672<div class="doc_text">
673
Reid Spencer128a7a72007-02-03 21:06:43 +0000674<p>LLVM has a plethora of data structures in the <tt>llvm/ADT/</tt> directory,
675 and we commonly use STL data structures. This section describes the trade-offs
Chris Lattner098129a2007-02-03 03:04:03 +0000676 you should consider when you pick one.</p>
677
678<p>
679The first step is a choose your own adventure: do you want a sequential
680container, a set-like container, or a map-like container? The most important
681thing when choosing a container is the algorithmic properties of how you plan to
682access the container. Based on that, you should use:</p>
683
684<ul>
Reid Spencer128a7a72007-02-03 21:06:43 +0000685<li>a <a href="#ds_map">map-like</a> container if you need efficient look-up
Chris Lattner098129a2007-02-03 03:04:03 +0000686 of an value based on another value. Map-like containers also support
687 efficient queries for containment (whether a key is in the map). Map-like
688 containers generally do not support efficient reverse mapping (values to
689 keys). If you need that, use two maps. Some map-like containers also
690 support efficient iteration through the keys in sorted order. Map-like
691 containers are the most expensive sort, only use them if you need one of
692 these capabilities.</li>
693
694<li>a <a href="#ds_set">set-like</a> container if you need to put a bunch of
695 stuff into a container that automatically eliminates duplicates. Some
696 set-like containers support efficient iteration through the elements in
697 sorted order. Set-like containers are more expensive than sequential
698 containers.
699</li>
700
701<li>a <a href="#ds_sequential">sequential</a> container provides
702 the most efficient way to add elements and keeps track of the order they are
703 added to the collection. They permit duplicates and support efficient
Reid Spencer128a7a72007-02-03 21:06:43 +0000704 iteration, but do not support efficient look-up based on a key.
Chris Lattner098129a2007-02-03 03:04:03 +0000705</li>
706
707</ul>
708
709<p>
Reid Spencer128a7a72007-02-03 21:06:43 +0000710Once the proper category of container is determined, you can fine tune the
Chris Lattner098129a2007-02-03 03:04:03 +0000711memory use, constant factors, and cache behaviors of access by intelligently
Reid Spencer128a7a72007-02-03 21:06:43 +0000712picking a member of the category. Note that constant factors and cache behavior
Chris Lattner098129a2007-02-03 03:04:03 +0000713can be a big deal. If you have a vector that usually only contains a few
714elements (but could contain many), for example, it's much better to use
715<a href="#dss_smallvector">SmallVector</a> than <a href="#dss_vector">vector</a>
716. Doing so avoids (relatively) expensive malloc/free calls, which dwarf the
717cost of adding the elements to the container. </p>
718
719</div>
720
721<!-- ======================================================================= -->
722<div class="doc_subsection">
723 <a name="ds_sequential">Sequential Containers (std::vector, std::list, etc)</a>
724</div>
725
726<div class="doc_text">
727There are a variety of sequential containers available for you, based on your
728needs. Pick the first in this section that will do what you want.
729</div>
730
731<!-- _______________________________________________________________________ -->
732<div class="doc_subsubsection">
733 <a name="dss_fixedarrays">Fixed Size Arrays</a>
734</div>
735
736<div class="doc_text">
737<p>Fixed size arrays are very simple and very fast. They are good if you know
738exactly how many elements you have, or you have a (low) upper bound on how many
739you have.</p>
740</div>
741
742<!-- _______________________________________________________________________ -->
743<div class="doc_subsubsection">
744 <a name="dss_heaparrays">Heap Allocated Arrays</a>
745</div>
746
747<div class="doc_text">
748<p>Heap allocated arrays (new[] + delete[]) are also simple. They are good if
749the number of elements is variable, if you know how many elements you will need
750before the array is allocated, and if the array is usually large (if not,
751consider a <a href="#dss_smallvector">SmallVector</a>). The cost of a heap
752allocated array is the cost of the new/delete (aka malloc/free). Also note that
753if you are allocating an array of a type with a constructor, the constructor and
Reid Spencer128a7a72007-02-03 21:06:43 +0000754destructors will be run for every element in the array (re-sizable vectors only
Chris Lattner098129a2007-02-03 03:04:03 +0000755construct those elements actually used).</p>
756</div>
757
758<!-- _______________________________________________________________________ -->
759<div class="doc_subsubsection">
760 <a name="dss_smallvector">"llvm/ADT/SmallVector.h"</a>
761</div>
762
763<div class="doc_text">
764<p><tt>SmallVector&lt;Type, N&gt;</tt> is a simple class that looks and smells
765just like <tt>vector&lt;Type&gt;</tt>:
766it supports efficient iteration, lays out elements in memory order (so you can
767do pointer arithmetic between elements), supports efficient push_back/pop_back
768operations, supports efficient random access to its elements, etc.</p>
769
770<p>The advantage of SmallVector is that it allocates space for
771some number of elements (N) <b>in the object itself</b>. Because of this, if
772the SmallVector is dynamically smaller than N, no malloc is performed. This can
773be a big win in cases where the malloc/free call is far more expensive than the
774code that fiddles around with the elements.</p>
775
776<p>This is good for vectors that are "usually small" (e.g. the number of
777predecessors/successors of a block is usually less than 8). On the other hand,
778this makes the size of the SmallVector itself large, so you don't want to
779allocate lots of them (doing so will waste a lot of space). As such,
780SmallVectors are most useful when on the stack.</p>
781
782<p>SmallVector also provides a nice portable and efficient replacement for
783<tt>alloca</tt>.</p>
784
785</div>
786
787<!-- _______________________________________________________________________ -->
788<div class="doc_subsubsection">
789 <a name="dss_vector">&lt;vector&gt;</a>
790</div>
791
792<div class="doc_text">
793<p>
794std::vector is well loved and respected. It is useful when SmallVector isn't:
795when the size of the vector is often large (thus the small optimization will
796rarely be a benefit) or if you will be allocating many instances of the vector
797itself (which would waste space for elements that aren't in the container).
798vector is also useful when interfacing with code that expects vectors :).
799</p>
800</div>
801
802<!-- _______________________________________________________________________ -->
803<div class="doc_subsubsection">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000804 <a name="dss_deque">&lt;deque&gt;</a>
805</div>
806
807<div class="doc_text">
808<p>std::deque is, in some senses, a generalized version of std::vector. Like
809std::vector, it provides constant time random access and other similar
810properties, but it also provides efficient access to the front of the list. It
811does not guarantee continuity of elements within memory.</p>
812
813<p>In exchange for this extra flexibility, std::deque has significantly higher
814constant factor costs than std::vector. If possible, use std::vector or
815something cheaper.</p>
816</div>
817
818<!-- _______________________________________________________________________ -->
819<div class="doc_subsubsection">
Chris Lattner098129a2007-02-03 03:04:03 +0000820 <a name="dss_list">&lt;list&gt;</a>
821</div>
822
823<div class="doc_text">
824<p>std::list is an extremely inefficient class that is rarely useful.
825It performs a heap allocation for every element inserted into it, thus having an
826extremely high constant factor, particularly for small data types. std::list
827also only supports bidirectional iteration, not random access iteration.</p>
828
829<p>In exchange for this high cost, std::list supports efficient access to both
830ends of the list (like std::deque, but unlike std::vector or SmallVector). In
831addition, the iterator invalidation characteristics of std::list are stronger
832than that of a vector class: inserting or removing an element into the list does
833not invalidate iterator or pointers to other elements in the list.</p>
834</div>
835
836<!-- _______________________________________________________________________ -->
837<div class="doc_subsubsection">
838 <a name="dss_ilist">llvm/ADT/ilist</a>
839</div>
840
841<div class="doc_text">
842<p><tt>ilist&lt;T&gt;</tt> implements an 'intrusive' doubly-linked list. It is
843intrusive, because it requires the element to store and provide access to the
844prev/next pointers for the list.</p>
845
846<p>ilist has the same drawbacks as std::list, and additionally requires an
847ilist_traits implementation for the element type, but it provides some novel
848characteristics. In particular, it can efficiently store polymorphic objects,
849the traits class is informed when an element is inserted or removed from the
850list, and ilists are guaranteed to support a constant-time splice operation.
851</p>
852
853<p>These properties are exactly what we want for things like Instructions and
854basic blocks, which is why these are implemented with ilists.</p>
855</div>
856
857<!-- _______________________________________________________________________ -->
858<div class="doc_subsubsection">
Chris Lattnerc5722432007-02-03 19:49:31 +0000859 <a name="dss_other">Other Sequential Container options</a>
Chris Lattner098129a2007-02-03 03:04:03 +0000860</div>
861
862<div class="doc_text">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000863<p>Other STL containers are available, such as std::string.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000864
865<p>There are also various STL adapter classes such as std::queue,
866std::priority_queue, std::stack, etc. These provide simplified access to an
867underlying container but don't affect the cost of the container itself.</p>
868
869</div>
870
871
872<!-- ======================================================================= -->
873<div class="doc_subsection">
874 <a name="ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
875</div>
876
877<div class="doc_text">
878
Chris Lattner74c4ca12007-02-03 07:59:07 +0000879<p>Set-like containers are useful when you need to canonicalize multiple values
880into a single representation. There are several different choices for how to do
881this, providing various trade-offs.</p>
882
883</div>
884
885
886<!-- _______________________________________________________________________ -->
887<div class="doc_subsubsection">
888 <a name="dss_sortedvectorset">A sorted 'vector'</a>
889</div>
890
891<div class="doc_text">
892
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000893<p>If you intend to insert a lot of elements, then do a lot of queries, a
894great approach is to use a vector (or other sequential container) with
Chris Lattner74c4ca12007-02-03 07:59:07 +0000895std::sort+std::unique to remove duplicates. This approach works really well if
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000896your usage pattern has these two distinct phases (insert then query), and can be
897coupled with a good choice of <a href="#ds_sequential">sequential container</a>.
898</p>
899
900<p>
901This combination provides the several nice properties: the result data is
902contiguous in memory (good for cache locality), has few allocations, is easy to
903address (iterators in the final vector are just indices or pointers), and can be
904efficiently queried with a standard binary or radix search.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000905
906</div>
907
908<!-- _______________________________________________________________________ -->
909<div class="doc_subsubsection">
910 <a name="dss_smallset">"llvm/ADT/SmallSet.h"</a>
911</div>
912
913<div class="doc_text">
914
Reid Spencer128a7a72007-02-03 21:06:43 +0000915<p>If you have a set-like data structure that is usually small and whose elements
Chris Lattner4ddfac12007-02-03 07:59:51 +0000916are reasonably small, a <tt>SmallSet&lt;Type, N&gt;</tt> is a good choice. This set
Chris Lattner74c4ca12007-02-03 07:59:07 +0000917has space for N elements in place (thus, if the set is dynamically smaller than
Chris Lattner14868db2007-02-03 08:20:15 +0000918N, no malloc traffic is required) and accesses them with a simple linear search.
919When the set grows beyond 'N' elements, it allocates a more expensive representation that
Chris Lattner74c4ca12007-02-03 07:59:07 +0000920guarantees efficient access (for most types, it falls back to std::set, but for
Chris Lattner14868db2007-02-03 08:20:15 +0000921pointers it uses something far better, <a
Chris Lattner74c4ca12007-02-03 07:59:07 +0000922href="#dss_smallptrset">SmallPtrSet</a>).</p>
923
924<p>The magic of this class is that it handles small sets extremely efficiently,
925but gracefully handles extremely large sets without loss of efficiency. The
926drawback is that the interface is quite small: it supports insertion, queries
927and erasing, but does not support iteration.</p>
928
929</div>
930
931<!-- _______________________________________________________________________ -->
932<div class="doc_subsubsection">
933 <a name="dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a>
934</div>
935
936<div class="doc_text">
937
938<p>SmallPtrSet has all the advantages of SmallSet (and a SmallSet of pointers is
Reid Spencer128a7a72007-02-03 21:06:43 +0000939transparently implemented with a SmallPtrSet), but also supports iterators. If
Chris Lattner14868db2007-02-03 08:20:15 +0000940more than 'N' insertions are performed, a single quadratically
Chris Lattner74c4ca12007-02-03 07:59:07 +0000941probed hash table is allocated and grows as needed, providing extremely
942efficient access (constant time insertion/deleting/queries with low constant
943factors) and is very stingy with malloc traffic.</p>
944
945<p>Note that, unlike std::set, the iterators of SmallPtrSet are invalidated
946whenever an insertion occurs. Also, the values visited by the iterators are not
947visited in sorted order.</p>
948
949</div>
950
951<!-- _______________________________________________________________________ -->
952<div class="doc_subsubsection">
953 <a name="dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a>
954</div>
955
956<div class="doc_text">
957
Chris Lattner098129a2007-02-03 03:04:03 +0000958<p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000959FoldingSet is an aggregate class that is really good at uniquing
960expensive-to-create or polymorphic objects. It is a combination of a chained
961hash table with intrusive links (uniqued objects are required to inherit from
Chris Lattner14868db2007-02-03 08:20:15 +0000962FoldingSetNode) that uses <a href="#dss_smallvector">SmallVector</a> as part of
963its ID process.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000964
Chris Lattner14868db2007-02-03 08:20:15 +0000965<p>Consider a case where you want to implement a "getOrCreateFoo" method for
Chris Lattner74c4ca12007-02-03 07:59:07 +0000966a complex object (for example, a node in the code generator). The client has a
967description of *what* it wants to generate (it knows the opcode and all the
968operands), but we don't want to 'new' a node, then try inserting it into a set
Chris Lattner14868db2007-02-03 08:20:15 +0000969only to find out it already exists, at which point we would have to delete it
970and return the node that already exists.
Chris Lattner098129a2007-02-03 03:04:03 +0000971</p>
972
Chris Lattner74c4ca12007-02-03 07:59:07 +0000973<p>To support this style of client, FoldingSet perform a query with a
974FoldingSetNodeID (which wraps SmallVector) that can be used to describe the
975element that we want to query for. The query either returns the element
976matching the ID or it returns an opaque ID that indicates where insertion should
Chris Lattner14868db2007-02-03 08:20:15 +0000977take place. Construction of the ID usually does not require heap traffic.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000978
979<p>Because FoldingSet uses intrusive links, it can support polymorphic objects
980in the set (for example, you can have SDNode instances mixed with LoadSDNodes).
981Because the elements are individually allocated, pointers to the elements are
982stable: inserting or removing elements does not invalidate any pointers to other
983elements.
984</p>
985
986</div>
987
988<!-- _______________________________________________________________________ -->
989<div class="doc_subsubsection">
990 <a name="dss_set">&lt;set&gt;</a>
991</div>
992
993<div class="doc_text">
994
Chris Lattnerc5722432007-02-03 19:49:31 +0000995<p><tt>std::set</tt> is a reasonable all-around set class, which is decent at
996many things but great at nothing. std::set allocates memory for each element
Chris Lattner74c4ca12007-02-03 07:59:07 +0000997inserted (thus it is very malloc intensive) and typically stores three pointers
Chris Lattner14868db2007-02-03 08:20:15 +0000998per element in the set (thus adding a large amount of per-element space
999overhead). It offers guaranteed log(n) performance, which is not particularly
Chris Lattnerc5722432007-02-03 19:49:31 +00001000fast from a complexity standpoint (particularly if the elements of the set are
1001expensive to compare, like strings), and has extremely high constant factors for
1002lookup, insertion and removal.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +00001003
Chris Lattner14868db2007-02-03 08:20:15 +00001004<p>The advantages of std::set are that its iterators are stable (deleting or
Chris Lattner74c4ca12007-02-03 07:59:07 +00001005inserting an element from the set does not affect iterators or pointers to other
1006elements) and that iteration over the set is guaranteed to be in sorted order.
1007If the elements in the set are large, then the relative overhead of the pointers
1008and malloc traffic is not a big deal, but if the elements of the set are small,
1009std::set is almost never a good choice.</p>
1010
1011</div>
1012
1013<!-- _______________________________________________________________________ -->
1014<div class="doc_subsubsection">
1015 <a name="dss_setvector">"llvm/ADT/SetVector.h"</a>
1016</div>
1017
1018<div class="doc_text">
Chris Lattneredca3c52007-02-04 00:00:26 +00001019<p>LLVM's SetVector&lt;Type&gt; is an adapter class that combines your choice of
1020a set-like container along with a <a href="#ds_sequential">Sequential
1021Container</a>. The important property
Chris Lattner74c4ca12007-02-03 07:59:07 +00001022that this provides is efficient insertion with uniquing (duplicate elements are
1023ignored) with iteration support. It implements this by inserting elements into
1024both a set-like container and the sequential container, using the set-like
1025container for uniquing and the sequential container for iteration.
1026</p>
1027
1028<p>The difference between SetVector and other sets is that the order of
1029iteration is guaranteed to match the order of insertion into the SetVector.
1030This property is really important for things like sets of pointers. Because
1031pointer values are non-deterministic (e.g. vary across runs of the program on
Chris Lattneredca3c52007-02-04 00:00:26 +00001032different machines), iterating over the pointers in the set will
Chris Lattner74c4ca12007-02-03 07:59:07 +00001033not be in a well-defined order.</p>
1034
1035<p>
1036The drawback of SetVector is that it requires twice as much space as a normal
1037set and has the sum of constant factors from the set-like container and the
1038sequential container that it uses. Use it *only* if you need to iterate over
1039the elements in a deterministic order. SetVector is also expensive to delete
Chris Lattneredca3c52007-02-04 00:00:26 +00001040elements out of (linear time), unless you use it's "pop_back" method, which is
1041faster.
Chris Lattner74c4ca12007-02-03 07:59:07 +00001042</p>
1043
Chris Lattneredca3c52007-02-04 00:00:26 +00001044<p>SetVector is an adapter class that defaults to using std::vector and std::set
1045for the underlying containers, so it is quite expensive. However,
1046<tt>"llvm/ADT/SetVector.h"</tt> also provides a SmallSetVector class, which
1047defaults to using a SmallVector and SmallSet of a specified size. If you use
1048this, and if your sets are dynamically smaller than N, you will save a lot of
1049heap traffic.</p>
1050
Chris Lattner74c4ca12007-02-03 07:59:07 +00001051</div>
1052
1053<!-- _______________________________________________________________________ -->
1054<div class="doc_subsubsection">
Chris Lattnerc5722432007-02-03 19:49:31 +00001055 <a name="dss_uniquevector">"llvm/ADT/UniqueVector.h"</a>
1056</div>
1057
1058<div class="doc_text">
1059
1060<p>
1061UniqueVector is similar to <a href="#dss_setvector">SetVector</a>, but it
1062retains a unique ID for each element inserted into the set. It internally
1063contains a map and a vector, and it assigns a unique ID for each value inserted
1064into the set.</p>
1065
1066<p>UniqueVector is very expensive: its cost is the sum of the cost of
1067maintaining both the map and vector, it has high complexity, high constant
1068factors, and produces a lot of malloc traffic. It should be avoided.</p>
1069
1070</div>
1071
1072
1073<!-- _______________________________________________________________________ -->
1074<div class="doc_subsubsection">
1075 <a name="dss_otherset">Other Set-Like Container Options</a>
Chris Lattner74c4ca12007-02-03 07:59:07 +00001076</div>
1077
1078<div class="doc_text">
1079
1080<p>
1081The STL provides several other options, such as std::multiset and the various
Chris Lattnerc5722432007-02-03 19:49:31 +00001082"hash_set" like containers (whether from C++ TR1 or from the SGI library).</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +00001083
1084<p>std::multiset is useful if you're not interested in elimination of
Chris Lattner14868db2007-02-03 08:20:15 +00001085duplicates, but has all the drawbacks of std::set. A sorted vector (where you
1086don't delete duplicate entries) or some other approach is almost always
1087better.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +00001088
1089<p>The various hash_set implementations (exposed portably by
Chris Lattner14868db2007-02-03 08:20:15 +00001090"llvm/ADT/hash_set") is a simple chained hashtable. This algorithm is as malloc
1091intensive as std::set (performing an allocation for each element inserted,
Chris Lattner74c4ca12007-02-03 07:59:07 +00001092thus having really high constant factors) but (usually) provides O(1)
1093insertion/deletion of elements. This can be useful if your elements are large
Chris Lattner14868db2007-02-03 08:20:15 +00001094(thus making the constant-factor cost relatively low) or if comparisons are
1095expensive. Element iteration does not visit elements in a useful order.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +00001096
Chris Lattner098129a2007-02-03 03:04:03 +00001097</div>
1098
1099<!-- ======================================================================= -->
1100<div class="doc_subsection">
1101 <a name="ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
1102</div>
1103
1104<div class="doc_text">
Chris Lattnerc5722432007-02-03 19:49:31 +00001105Map-like containers are useful when you want to associate data to a key. As
1106usual, there are a lot of different ways to do this. :)
1107</div>
1108
1109<!-- _______________________________________________________________________ -->
1110<div class="doc_subsubsection">
1111 <a name="dss_sortedvectormap">A sorted 'vector'</a>
1112</div>
1113
1114<div class="doc_text">
1115
1116<p>
1117If your usage pattern follows a strict insert-then-query approach, you can
1118trivially use the same approach as <a href="#dss_sortedvectorset">sorted vectors
1119for set-like containers</a>. The only difference is that your query function
1120(which uses std::lower_bound to get efficient log(n) lookup) should only compare
1121the key, not both the key and value. This yields the same advantages as sorted
1122vectors for sets.
1123</p>
1124</div>
1125
1126<!-- _______________________________________________________________________ -->
1127<div class="doc_subsubsection">
1128 <a name="dss_cstringmap">"llvm/ADT/CStringMap.h"</a>
1129</div>
1130
1131<div class="doc_text">
1132
1133<p>
1134Strings are commonly used as keys in maps, and they are difficult to support
1135efficiently: they are variable length, inefficient to hash and compare when
1136long, expensive to copy, etc. CStringMap is a specialized container designed to
1137cope with these issues. It supports mapping an arbitrary range of bytes that
Chris Lattnere3683e12007-02-03 22:04:27 +00001138does not have an embedded nul character in it ("C strings") to an arbitrary
Chris Lattnerc5722432007-02-03 19:49:31 +00001139other object.</p>
1140
1141<p>The CStringMap implementation uses a quadratically-probed hash table, where
1142the buckets store a pointer to the heap allocated entries (and some other
1143stuff). The entries in the map must be heap allocated because the strings are
1144variable length. The string data (key) and the element object (value) are
1145stored in the same allocation with the string data immediately after the element
1146object. This container guarantees the "<tt>(char*)(&amp;Value+1)</tt>" points
1147to the key string for a value.</p>
1148
1149<p>The CStringMap is very fast for several reasons: quadratic probing is very
1150cache efficient for lookups, the hash value of strings in buckets is not
1151recomputed when lookup up an element, CStringMap rarely has to touch the
1152memory for unrelated objects when looking up a value (even when hash collisions
1153happen), hash table growth does not recompute the hash values for strings
1154already in the table, and each pair in the map is store in a single allocation
1155(the string data is stored in the same allocation as the Value of a pair).</p>
1156
1157<p>CStringMap also provides query methods that take byte ranges, so it only ever
1158copies a string if a value is inserted into the table.</p>
1159</div>
1160
1161<!-- _______________________________________________________________________ -->
1162<div class="doc_subsubsection">
1163 <a name="dss_indexedmap">"llvm/ADT/IndexedMap.h"</a>
1164</div>
1165
1166<div class="doc_text">
1167<p>
1168IndexedMap is a specialized container for mapping small dense integers (or
1169values that can be mapped to small dense integers) to some other type. It is
1170internally implemented as a vector with a mapping function that maps the keys to
1171the dense integer range.
1172</p>
1173
1174<p>
1175This is useful for cases like virtual registers in the LLVM code generator: they
1176have a dense mapping that is offset by a compile-time constant (the first
1177virtual register ID).</p>
1178
1179</div>
1180
1181<!-- _______________________________________________________________________ -->
1182<div class="doc_subsubsection">
1183 <a name="dss_densemap">"llvm/ADT/DenseMap.h"</a>
1184</div>
1185
1186<div class="doc_text">
1187
1188<p>
1189DenseMap is a simple quadratically probed hash table. It excels at supporting
1190small keys and values: it uses a single allocation to hold all of the pairs that
1191are currently inserted in the map. DenseMap is a great way to map pointers to
1192pointers, or map other small types to each other.
1193</p>
1194
1195<p>
1196There are several aspects of DenseMap that you should be aware of, however. The
1197iterators in a densemap are invalidated whenever an insertion occurs, unlike
1198map. Also, because DenseMap allocates space for a large number of key/value
Chris Lattnera4a264d2007-02-03 20:17:53 +00001199pairs (it starts with 64 by default), it will waste a lot of space if your keys
1200or values are large. Finally, you must implement a partial specialization of
Chris Lattnerc5722432007-02-03 19:49:31 +00001201DenseMapKeyInfo for the key that you want, if it isn't already supported. This
1202is required to tell DenseMap about two special marker values (which can never be
Chris Lattnera4a264d2007-02-03 20:17:53 +00001203inserted into the map) that it needs internally.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +00001204
1205</div>
1206
1207<!-- _______________________________________________________________________ -->
1208<div class="doc_subsubsection">
1209 <a name="dss_map">&lt;map&gt;</a>
1210</div>
1211
1212<div class="doc_text">
1213
1214<p>
1215std::map has similar characteristics to <a href="#dss_set">std::set</a>: it uses
1216a single allocation per pair inserted into the map, it offers log(n) lookup with
1217an extremely large constant factor, imposes a space penalty of 3 pointers per
1218pair in the map, etc.</p>
1219
1220<p>std::map is most useful when your keys or values are very large, if you need
1221to iterate over the collection in sorted order, or if you need stable iterators
1222into the map (i.e. they don't get invalidated if an insertion or deletion of
1223another element takes place).</p>
1224
1225</div>
1226
1227<!-- _______________________________________________________________________ -->
1228<div class="doc_subsubsection">
1229 <a name="dss_othermap">Other Map-Like Container Options</a>
1230</div>
1231
1232<div class="doc_text">
1233
1234<p>
1235The STL provides several other options, such as std::multimap and the various
1236"hash_map" like containers (whether from C++ TR1 or from the SGI library).</p>
1237
1238<p>std::multimap is useful if you want to map a key to multiple values, but has
1239all the drawbacks of std::map. A sorted vector or some other approach is almost
1240always better.</p>
1241
1242<p>The various hash_map implementations (exposed portably by
1243"llvm/ADT/hash_map") are simple chained hash tables. This algorithm is as
1244malloc intensive as std::map (performing an allocation for each element
1245inserted, thus having really high constant factors) but (usually) provides O(1)
1246insertion/deletion of elements. This can be useful if your elements are large
1247(thus making the constant-factor cost relatively low) or if comparisons are
1248expensive. Element iteration does not visit elements in a useful order.</p>
1249
Chris Lattner098129a2007-02-03 03:04:03 +00001250</div>
1251
Chris Lattnerf623a082005-10-17 01:36:23 +00001252
Misha Brukman13fd15c2004-01-15 00:14:41 +00001253<!-- *********************************************************************** -->
1254<div class="doc_section">
1255 <a name="common">Helpful Hints for Common Operations</a>
1256</div>
1257<!-- *********************************************************************** -->
1258
1259<div class="doc_text">
1260
1261<p>This section describes how to perform some very simple transformations of
1262LLVM code. This is meant to give examples of common idioms used, showing the
1263practical side of LLVM transformations. <p> Because this is a "how-to" section,
1264you should also read about the main classes that you will be working with. The
1265<a href="#coreclasses">Core LLVM Class Hierarchy Reference</a> contains details
1266and descriptions of the main classes that you should know about.</p>
1267
1268</div>
1269
1270<!-- NOTE: this section should be heavy on example code -->
1271<!-- ======================================================================= -->
1272<div class="doc_subsection">
1273 <a name="inspection">Basic Inspection and Traversal Routines</a>
1274</div>
1275
1276<div class="doc_text">
1277
1278<p>The LLVM compiler infrastructure have many different data structures that may
1279be traversed. Following the example of the C++ standard template library, the
1280techniques used to traverse these various data structures are all basically the
1281same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
1282method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
1283function returns an iterator pointing to one past the last valid element of the
1284sequence, and there is some <tt>XXXiterator</tt> data type that is common
1285between the two operations.</p>
1286
1287<p>Because the pattern for iteration is common across many different aspects of
1288the program representation, the standard template library algorithms may be used
1289on them, and it is easier to remember how to iterate. First we show a few common
1290examples of the data structures that need to be traversed. Other data
1291structures are traversed in very similar ways.</p>
1292
1293</div>
1294
1295<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +00001296<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +00001297 <a name="iterate_function">Iterating over the </a><a
1298 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
1299 href="#Function"><tt>Function</tt></a>
1300</div>
1301
1302<div class="doc_text">
1303
1304<p>It's quite common to have a <tt>Function</tt> instance that you'd like to
1305transform in some way; in particular, you'd like to manipulate its
1306<tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over all of
1307the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>. The following is
1308an example that prints the name of a <tt>BasicBlock</tt> and the number of
1309<tt>Instruction</tt>s it contains:</p>
1310
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001311<div class="doc_code">
1312<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +00001313// <i>func is a pointer to a Function instance</i>
1314for (Function::iterator i = func-&gt;begin(), e = func-&gt;end(); i != e; ++i)
1315 // <i>Print out the name of the basic block if it has one, and then the</i>
1316 // <i>number of instructions that it contains</i>
Bill Wendling832171c2006-12-07 20:04:42 +00001317 llvm::cerr &lt;&lt; "Basic block (name=" &lt;&lt; i-&gt;getName() &lt;&lt; ") has "
1318 &lt;&lt; i-&gt;size() &lt;&lt; " instructions.\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001319</pre>
1320</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001321
1322<p>Note that i can be used as if it were a pointer for the purposes of
Joel Stanley9b96c442002-09-06 21:55:13 +00001323invoking member functions of the <tt>Instruction</tt> class. This is
1324because the indirection operator is overloaded for the iterator
Chris Lattner7496ec52003-08-05 22:54:23 +00001325classes. In the above code, the expression <tt>i-&gt;size()</tt> is
Misha Brukman13fd15c2004-01-15 00:14:41 +00001326exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.</p>
1327
1328</div>
1329
1330<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +00001331<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +00001332 <a name="iterate_basicblock">Iterating over the </a><a
1333 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1334 href="#BasicBlock"><tt>BasicBlock</tt></a>
1335</div>
1336
1337<div class="doc_text">
1338
1339<p>Just like when dealing with <tt>BasicBlock</tt>s in <tt>Function</tt>s, it's
1340easy to iterate over the individual instructions that make up
1341<tt>BasicBlock</tt>s. Here's a code snippet that prints out each instruction in
1342a <tt>BasicBlock</tt>:</p>
1343
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001344<div class="doc_code">
Chris Lattner55c04612005-03-06 06:00:13 +00001345<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +00001346// <i>blk is a pointer to a BasicBlock instance</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001347for (BasicBlock::iterator i = blk-&gt;begin(), e = blk-&gt;end(); i != e; ++i)
Bill Wendling82e2eea2006-10-11 18:00:22 +00001348 // <i>The next statement works since operator&lt;&lt;(ostream&amp;,...)</i>
1349 // <i>is overloaded for Instruction&amp;</i>
Bill Wendling832171c2006-12-07 20:04:42 +00001350 llvm::cerr &lt;&lt; *i &lt;&lt; "\n";
Chris Lattner55c04612005-03-06 06:00:13 +00001351</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001352</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001353
1354<p>However, this isn't really the best way to print out the contents of a
1355<tt>BasicBlock</tt>! Since the ostream operators are overloaded for virtually
1356anything you'll care about, you could have just invoked the print routine on the
Bill Wendling832171c2006-12-07 20:04:42 +00001357basic block itself: <tt>llvm::cerr &lt;&lt; *blk &lt;&lt; "\n";</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001358
1359</div>
1360
1361<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +00001362<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +00001363 <a name="iterate_institer">Iterating over the </a><a
1364 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1365 href="#Function"><tt>Function</tt></a>
1366</div>
1367
1368<div class="doc_text">
1369
1370<p>If you're finding that you commonly iterate over a <tt>Function</tt>'s
1371<tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s <tt>Instruction</tt>s,
1372<tt>InstIterator</tt> should be used instead. You'll need to include <a
1373href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>,
1374and then instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
Chris Lattner69bf8a92004-05-23 21:06:58 +00001375small example that shows how to dump all instructions in a function to the standard error stream:<p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001376
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001377<div class="doc_code">
1378<pre>
1379#include "<a href="/doxygen/InstIterator_8h-source.html">llvm/Support/InstIterator.h</a>"
1380
Reid Spencer128a7a72007-02-03 21:06:43 +00001381// <i>F is a pointer to a Function instance</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001382for (inst_iterator i = inst_begin(F), e = inst_end(F); i != e; ++i)
Bill Wendling832171c2006-12-07 20:04:42 +00001383 llvm::cerr &lt;&lt; *i &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001384</pre>
1385</div>
1386
1387<p>Easy, isn't it? You can also use <tt>InstIterator</tt>s to fill a
Reid Spencer128a7a72007-02-03 21:06:43 +00001388work list with its initial contents. For example, if you wanted to
1389initialize a work list to contain all instructions in a <tt>Function</tt>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001390F, all you would need to do is something like:</p>
1391
1392<div class="doc_code">
1393<pre>
1394std::set&lt;Instruction*&gt; worklist;
1395worklist.insert(inst_begin(F), inst_end(F));
1396</pre>
1397</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001398
1399<p>The STL set <tt>worklist</tt> would now contain all instructions in the
1400<tt>Function</tt> pointed to by F.</p>
1401
1402</div>
1403
1404<!-- _______________________________________________________________________ -->
1405<div class="doc_subsubsection">
1406 <a name="iterate_convert">Turning an iterator into a class pointer (and
1407 vice-versa)</a>
1408</div>
1409
1410<div class="doc_text">
1411
1412<p>Sometimes, it'll be useful to grab a reference (or pointer) to a class
Joel Stanley9b96c442002-09-06 21:55:13 +00001413instance when all you've got at hand is an iterator. Well, extracting
Chris Lattner69bf8a92004-05-23 21:06:58 +00001414a reference or a pointer from an iterator is very straight-forward.
Chris Lattner261efe92003-11-25 01:02:51 +00001415Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and <tt>j</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001416is a <tt>BasicBlock::const_iterator</tt>:</p>
1417
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001418<div class="doc_code">
1419<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +00001420Instruction&amp; inst = *i; // <i>Grab reference to instruction reference</i>
1421Instruction* pinst = &amp;*i; // <i>Grab pointer to instruction reference</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001422const Instruction&amp; inst = *j;
1423</pre>
1424</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001425
1426<p>However, the iterators you'll be working with in the LLVM framework are
1427special: they will automatically convert to a ptr-to-instance type whenever they
1428need to. Instead of dereferencing the iterator and then taking the address of
1429the result, you can simply assign the iterator to the proper pointer type and
1430you get the dereference and address-of operation as a result of the assignment
1431(behind the scenes, this is a result of overloading casting mechanisms). Thus
1432the last line of the last example,</p>
1433
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001434<div class="doc_code">
1435<pre>
1436Instruction* pinst = &amp;*i;
1437</pre>
1438</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001439
1440<p>is semantically equivalent to</p>
1441
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001442<div class="doc_code">
1443<pre>
1444Instruction* pinst = i;
1445</pre>
1446</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001447
Chris Lattner69bf8a92004-05-23 21:06:58 +00001448<p>It's also possible to turn a class pointer into the corresponding iterator,
1449and this is a constant time operation (very efficient). The following code
1450snippet illustrates use of the conversion constructors provided by LLVM
1451iterators. By using these, you can explicitly grab the iterator of something
1452without actually obtaining it via iteration over some structure:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001453
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001454<div class="doc_code">
1455<pre>
1456void printNextInstruction(Instruction* inst) {
1457 BasicBlock::iterator it(inst);
Bill Wendling82e2eea2006-10-11 18:00:22 +00001458 ++it; // <i>After this line, it refers to the instruction after *inst</i>
Bill Wendling832171c2006-12-07 20:04:42 +00001459 if (it != inst-&gt;getParent()-&gt;end()) llvm::cerr &lt;&lt; *it &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001460}
1461</pre>
1462</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001463
Misha Brukman13fd15c2004-01-15 00:14:41 +00001464</div>
1465
1466<!--_______________________________________________________________________-->
1467<div class="doc_subsubsection">
1468 <a name="iterate_complex">Finding call sites: a slightly more complex
1469 example</a>
1470</div>
1471
1472<div class="doc_text">
1473
1474<p>Say that you're writing a FunctionPass and would like to count all the
1475locations in the entire module (that is, across every <tt>Function</tt>) where a
1476certain function (i.e., some <tt>Function</tt>*) is already in scope. As you'll
1477learn later, you may want to use an <tt>InstVisitor</tt> to accomplish this in a
Chris Lattner69bf8a92004-05-23 21:06:58 +00001478much more straight-forward manner, but this example will allow us to explore how
Reid Spencer128a7a72007-02-03 21:06:43 +00001479you'd do it if you didn't have <tt>InstVisitor</tt> around. In pseudo-code, this
Misha Brukman13fd15c2004-01-15 00:14:41 +00001480is what we want to do:</p>
1481
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001482<div class="doc_code">
1483<pre>
1484initialize callCounter to zero
1485for each Function f in the Module
1486 for each BasicBlock b in f
1487 for each Instruction i in b
1488 if (i is a CallInst and calls the given function)
1489 increment callCounter
1490</pre>
1491</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001492
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001493<p>And the actual code is (remember, because we're writing a
Misha Brukman13fd15c2004-01-15 00:14:41 +00001494<tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply has to
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001495override the <tt>runOnFunction</tt> method):</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001496
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001497<div class="doc_code">
1498<pre>
1499Function* targetFunc = ...;
1500
1501class OurFunctionPass : public FunctionPass {
1502 public:
1503 OurFunctionPass(): callCounter(0) { }
1504
1505 virtual runOnFunction(Function&amp; F) {
1506 for (Function::iterator b = F.begin(), be = F.end(); b != be; ++b) {
1507 for (BasicBlock::iterator i = b-&gt;begin(); ie = b-&gt;end(); i != ie; ++i) {
1508 if (<a href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a>&lt;<a
1509 href="#CallInst">CallInst</a>&gt;(&amp;*i)) {
Bill Wendling82e2eea2006-10-11 18:00:22 +00001510 // <i>We know we've encountered a call instruction, so we</i>
1511 // <i>need to determine if it's a call to the</i>
1512 // <i>function pointed to by m_func or not</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001513
1514 if (callInst-&gt;getCalledFunction() == targetFunc)
1515 ++callCounter;
1516 }
1517 }
1518 }
Bill Wendling82e2eea2006-10-11 18:00:22 +00001519 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001520
1521 private:
1522 unsigned callCounter;
1523};
1524</pre>
1525</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001526
1527</div>
1528
Brian Gaekef1972c62003-11-07 19:25:45 +00001529<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +00001530<div class="doc_subsubsection">
1531 <a name="calls_and_invokes">Treating calls and invokes the same way</a>
1532</div>
1533
1534<div class="doc_text">
1535
1536<p>You may have noticed that the previous example was a bit oversimplified in
1537that it did not deal with call sites generated by 'invoke' instructions. In
1538this, and in other situations, you may find that you want to treat
1539<tt>CallInst</tt>s and <tt>InvokeInst</tt>s the same way, even though their
1540most-specific common base class is <tt>Instruction</tt>, which includes lots of
1541less closely-related things. For these cases, LLVM provides a handy wrapper
1542class called <a
Reid Spencer05fe4b02006-03-14 05:39:39 +00001543href="http://llvm.org/doxygen/classllvm_1_1CallSite.html"><tt>CallSite</tt></a>.
Chris Lattner69bf8a92004-05-23 21:06:58 +00001544It is essentially a wrapper around an <tt>Instruction</tt> pointer, with some
1545methods that provide functionality common to <tt>CallInst</tt>s and
Misha Brukman13fd15c2004-01-15 00:14:41 +00001546<tt>InvokeInst</tt>s.</p>
1547
Chris Lattner69bf8a92004-05-23 21:06:58 +00001548<p>This class has "value semantics": it should be passed by value, not by
1549reference and it should not be dynamically allocated or deallocated using
1550<tt>operator new</tt> or <tt>operator delete</tt>. It is efficiently copyable,
1551assignable and constructable, with costs equivalents to that of a bare pointer.
1552If you look at its definition, it has only a single pointer member.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001553
1554</div>
1555
Chris Lattner1a3105b2002-09-09 05:49:39 +00001556<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +00001557<div class="doc_subsubsection">
1558 <a name="iterate_chains">Iterating over def-use &amp; use-def chains</a>
1559</div>
1560
1561<div class="doc_text">
1562
1563<p>Frequently, we might have an instance of the <a
Chris Lattner00815172007-01-04 22:01:45 +00001564href="/doxygen/classllvm_1_1Value.html">Value Class</a> and we want to
Misha Brukman384047f2004-06-03 23:29:12 +00001565determine which <tt>User</tt>s use the <tt>Value</tt>. The list of all
1566<tt>User</tt>s of a particular <tt>Value</tt> is called a <i>def-use</i> chain.
1567For example, let's say we have a <tt>Function*</tt> named <tt>F</tt> to a
1568particular function <tt>foo</tt>. Finding all of the instructions that
1569<i>use</i> <tt>foo</tt> is as simple as iterating over the <i>def-use</i> chain
1570of <tt>F</tt>:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001571
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001572<div class="doc_code">
1573<pre>
1574Function* F = ...;
1575
Bill Wendling82e2eea2006-10-11 18:00:22 +00001576for (Value::use_iterator i = F-&gt;use_begin(), e = F-&gt;use_end(); i != e; ++i)
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001577 if (Instruction *Inst = dyn_cast&lt;Instruction&gt;(*i)) {
Bill Wendling832171c2006-12-07 20:04:42 +00001578 llvm::cerr &lt;&lt; "F is used in instruction:\n";
1579 llvm::cerr &lt;&lt; *Inst &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001580 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001581</pre>
1582</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001583
1584<p>Alternately, it's common to have an instance of the <a
Misha Brukman384047f2004-06-03 23:29:12 +00001585href="/doxygen/classllvm_1_1User.html">User Class</a> and need to know what
Misha Brukman13fd15c2004-01-15 00:14:41 +00001586<tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used by a
1587<tt>User</tt> is known as a <i>use-def</i> chain. Instances of class
1588<tt>Instruction</tt> are common <tt>User</tt>s, so we might want to iterate over
1589all of the values that a particular instruction uses (that is, the operands of
1590the particular <tt>Instruction</tt>):</p>
1591
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001592<div class="doc_code">
1593<pre>
1594Instruction* pi = ...;
1595
1596for (User::op_iterator i = pi-&gt;op_begin(), e = pi-&gt;op_end(); i != e; ++i) {
1597 Value* v = *i;
Bill Wendling82e2eea2006-10-11 18:00:22 +00001598 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001599}
1600</pre>
1601</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001602
Chris Lattner1a3105b2002-09-09 05:49:39 +00001603<!--
1604 def-use chains ("finding all users of"): Value::use_begin/use_end
1605 use-def chains ("finding all values used"): User::op_begin/op_end [op=operand]
Misha Brukman13fd15c2004-01-15 00:14:41 +00001606-->
1607
1608</div>
1609
1610<!-- ======================================================================= -->
1611<div class="doc_subsection">
1612 <a name="simplechanges">Making simple changes</a>
1613</div>
1614
1615<div class="doc_text">
1616
1617<p>There are some primitive transformation operations present in the LLVM
Joel Stanley753eb712002-09-11 22:32:24 +00001618infrastructure that are worth knowing about. When performing
Chris Lattner261efe92003-11-25 01:02:51 +00001619transformations, it's fairly common to manipulate the contents of basic
1620blocks. This section describes some of the common methods for doing so
Misha Brukman13fd15c2004-01-15 00:14:41 +00001621and gives example code.</p>
1622
1623</div>
1624
Chris Lattner261efe92003-11-25 01:02:51 +00001625<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +00001626<div class="doc_subsubsection">
1627 <a name="schanges_creating">Creating and inserting new
1628 <tt>Instruction</tt>s</a>
1629</div>
1630
1631<div class="doc_text">
1632
1633<p><i>Instantiating Instructions</i></p>
1634
Chris Lattner69bf8a92004-05-23 21:06:58 +00001635<p>Creation of <tt>Instruction</tt>s is straight-forward: simply call the
Misha Brukman13fd15c2004-01-15 00:14:41 +00001636constructor for the kind of instruction to instantiate and provide the necessary
1637parameters. For example, an <tt>AllocaInst</tt> only <i>requires</i> a
1638(const-ptr-to) <tt>Type</tt>. Thus:</p>
1639
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001640<div class="doc_code">
1641<pre>
1642AllocaInst* ai = new AllocaInst(Type::IntTy);
1643</pre>
1644</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001645
1646<p>will create an <tt>AllocaInst</tt> instance that represents the allocation of
Reid Spencer128a7a72007-02-03 21:06:43 +00001647one integer in the current stack frame, at run time. Each <tt>Instruction</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001648subclass is likely to have varying default parameters which change the semantics
1649of the instruction, so refer to the <a
Misha Brukman31ca1de2004-06-03 23:35:54 +00001650href="/doxygen/classllvm_1_1Instruction.html">doxygen documentation for the subclass of
Misha Brukman13fd15c2004-01-15 00:14:41 +00001651Instruction</a> that you're interested in instantiating.</p>
1652
1653<p><i>Naming values</i></p>
1654
1655<p>It is very useful to name the values of instructions when you're able to, as
1656this facilitates the debugging of your transformations. If you end up looking
1657at generated LLVM machine code, you definitely want to have logical names
1658associated with the results of instructions! By supplying a value for the
1659<tt>Name</tt> (default) parameter of the <tt>Instruction</tt> constructor, you
1660associate a logical name with the result of the instruction's execution at
Reid Spencer128a7a72007-02-03 21:06:43 +00001661run time. For example, say that I'm writing a transformation that dynamically
Misha Brukman13fd15c2004-01-15 00:14:41 +00001662allocates space for an integer on the stack, and that integer is going to be
1663used as some kind of index by some other code. To accomplish this, I place an
1664<tt>AllocaInst</tt> at the first point in the first <tt>BasicBlock</tt> of some
1665<tt>Function</tt>, and I'm intending to use it within the same
1666<tt>Function</tt>. I might do:</p>
1667
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001668<div class="doc_code">
1669<pre>
1670AllocaInst* pa = new AllocaInst(Type::IntTy, 0, "indexLoc");
1671</pre>
1672</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001673
1674<p>where <tt>indexLoc</tt> is now the logical name of the instruction's
Reid Spencer128a7a72007-02-03 21:06:43 +00001675execution value, which is a pointer to an integer on the run time stack.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001676
1677<p><i>Inserting instructions</i></p>
1678
1679<p>There are essentially two ways to insert an <tt>Instruction</tt>
1680into an existing sequence of instructions that form a <tt>BasicBlock</tt>:</p>
1681
Joel Stanley9dd1ad62002-09-18 03:17:23 +00001682<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001683 <li>Insertion into an explicit instruction list
1684
1685 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within that
1686 <tt>BasicBlock</tt>, and a newly-created instruction we wish to insert
1687 before <tt>*pi</tt>, we do the following: </p>
1688
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001689<div class="doc_code">
1690<pre>
1691BasicBlock *pb = ...;
1692Instruction *pi = ...;
1693Instruction *newInst = new Instruction(...);
1694
Bill Wendling82e2eea2006-10-11 18:00:22 +00001695pb-&gt;getInstList().insert(pi, newInst); // <i>Inserts newInst before pi in pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001696</pre>
1697</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +00001698
1699 <p>Appending to the end of a <tt>BasicBlock</tt> is so common that
1700 the <tt>Instruction</tt> class and <tt>Instruction</tt>-derived
1701 classes provide constructors which take a pointer to a
1702 <tt>BasicBlock</tt> to be appended to. For example code that
1703 looked like: </p>
1704
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001705<div class="doc_code">
1706<pre>
1707BasicBlock *pb = ...;
1708Instruction *newInst = new Instruction(...);
1709
Bill Wendling82e2eea2006-10-11 18:00:22 +00001710pb-&gt;getInstList().push_back(newInst); // <i>Appends newInst to pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001711</pre>
1712</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +00001713
1714 <p>becomes: </p>
1715
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001716<div class="doc_code">
1717<pre>
1718BasicBlock *pb = ...;
1719Instruction *newInst = new Instruction(..., pb);
1720</pre>
1721</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +00001722
1723 <p>which is much cleaner, especially if you are creating
1724 long instruction streams.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001725
1726 <li>Insertion into an implicit instruction list
1727
1728 <p><tt>Instruction</tt> instances that are already in <tt>BasicBlock</tt>s
1729 are implicitly associated with an existing instruction list: the instruction
1730 list of the enclosing basic block. Thus, we could have accomplished the same
1731 thing as the above code without being given a <tt>BasicBlock</tt> by doing:
1732 </p>
1733
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001734<div class="doc_code">
1735<pre>
1736Instruction *pi = ...;
1737Instruction *newInst = new Instruction(...);
1738
1739pi-&gt;getParent()-&gt;getInstList().insert(pi, newInst);
1740</pre>
1741</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001742
1743 <p>In fact, this sequence of steps occurs so frequently that the
1744 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes provide
1745 constructors which take (as a default parameter) a pointer to an
1746 <tt>Instruction</tt> which the newly-created <tt>Instruction</tt> should
1747 precede. That is, <tt>Instruction</tt> constructors are capable of
1748 inserting the newly-created instance into the <tt>BasicBlock</tt> of a
1749 provided instruction, immediately before that instruction. Using an
1750 <tt>Instruction</tt> constructor with a <tt>insertBefore</tt> (default)
1751 parameter, the above code becomes:</p>
1752
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001753<div class="doc_code">
1754<pre>
1755Instruction* pi = ...;
1756Instruction* newInst = new Instruction(..., pi);
1757</pre>
1758</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001759
1760 <p>which is much cleaner, especially if you're creating a lot of
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001761 instructions and adding them to <tt>BasicBlock</tt>s.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001762</ul>
1763
1764</div>
1765
1766<!--_______________________________________________________________________-->
1767<div class="doc_subsubsection">
1768 <a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a>
1769</div>
1770
1771<div class="doc_text">
1772
1773<p>Deleting an instruction from an existing sequence of instructions that form a
Chris Lattner69bf8a92004-05-23 21:06:58 +00001774<a href="#BasicBlock"><tt>BasicBlock</tt></a> is very straight-forward. First,
Misha Brukman13fd15c2004-01-15 00:14:41 +00001775you must have a pointer to the instruction that you wish to delete. Second, you
1776need to obtain the pointer to that instruction's basic block. You use the
1777pointer to the basic block to get its list of instructions and then use the
1778erase function to remove your instruction. For example:</p>
1779
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001780<div class="doc_code">
1781<pre>
1782<a href="#Instruction">Instruction</a> *I = .. ;
1783<a href="#BasicBlock">BasicBlock</a> *BB = I-&gt;getParent();
1784
1785BB-&gt;getInstList().erase(I);
1786</pre>
1787</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001788
1789</div>
1790
1791<!--_______________________________________________________________________-->
1792<div class="doc_subsubsection">
1793 <a name="schanges_replacing">Replacing an <tt>Instruction</tt> with another
1794 <tt>Value</tt></a>
1795</div>
1796
1797<div class="doc_text">
1798
1799<p><i>Replacing individual instructions</i></p>
1800
1801<p>Including "<a href="/doxygen/BasicBlockUtils_8h-source.html">llvm/Transforms/Utils/BasicBlockUtils.h</a>"
Chris Lattner261efe92003-11-25 01:02:51 +00001802permits use of two very useful replace functions: <tt>ReplaceInstWithValue</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001803and <tt>ReplaceInstWithInst</tt>.</p>
1804
Chris Lattner261efe92003-11-25 01:02:51 +00001805<h4><a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a></h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001806
Chris Lattner261efe92003-11-25 01:02:51 +00001807<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001808 <li><tt>ReplaceInstWithValue</tt>
1809
1810 <p>This function replaces all uses (within a basic block) of a given
1811 instruction with a value, and then removes the original instruction. The
1812 following example illustrates the replacement of the result of a particular
Chris Lattner58360822005-01-17 00:12:04 +00001813 <tt>AllocaInst</tt> that allocates memory for a single integer with a null
Misha Brukman13fd15c2004-01-15 00:14:41 +00001814 pointer to an integer.</p>
1815
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001816<div class="doc_code">
1817<pre>
1818AllocaInst* instToReplace = ...;
1819BasicBlock::iterator ii(instToReplace);
1820
1821ReplaceInstWithValue(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
1822 Constant::getNullValue(PointerType::get(Type::IntTy)));
1823</pre></div></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001824
1825 <li><tt>ReplaceInstWithInst</tt>
1826
1827 <p>This function replaces a particular instruction with another
1828 instruction. The following example illustrates the replacement of one
1829 <tt>AllocaInst</tt> with another.</p>
1830
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001831<div class="doc_code">
1832<pre>
1833AllocaInst* instToReplace = ...;
1834BasicBlock::iterator ii(instToReplace);
1835
1836ReplaceInstWithInst(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
1837 new AllocaInst(Type::IntTy, 0, "ptrToReplacedInt"));
1838</pre></div></li>
Chris Lattner261efe92003-11-25 01:02:51 +00001839</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001840
1841<p><i>Replacing multiple uses of <tt>User</tt>s and <tt>Value</tt>s</i></p>
1842
1843<p>You can use <tt>Value::replaceAllUsesWith</tt> and
1844<tt>User::replaceUsesOfWith</tt> to change more than one use at a time. See the
Chris Lattner00815172007-01-04 22:01:45 +00001845doxygen documentation for the <a href="/doxygen/classllvm_1_1Value.html">Value Class</a>
Misha Brukman384047f2004-06-03 23:29:12 +00001846and <a href="/doxygen/classllvm_1_1User.html">User Class</a>, respectively, for more
Misha Brukman13fd15c2004-01-15 00:14:41 +00001847information.</p>
1848
1849<!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
1850include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
1851ReplaceInstWithValue, ReplaceInstWithInst -->
1852
1853</div>
1854
Chris Lattner9355b472002-09-06 02:50:58 +00001855<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +00001856<div class="doc_section">
Chris Lattnerd9d6e102005-04-23 16:10:52 +00001857 <a name="advanced">Advanced Topics</a>
1858</div>
1859<!-- *********************************************************************** -->
1860
1861<div class="doc_text">
Chris Lattnerf1b200b2005-04-23 17:27:36 +00001862<p>
1863This section describes some of the advanced or obscure API's that most clients
1864do not need to be aware of. These API's tend manage the inner workings of the
1865LLVM system, and only need to be accessed in unusual circumstances.
1866</p>
1867</div>
Chris Lattnerd9d6e102005-04-23 16:10:52 +00001868
Chris Lattnerf1b200b2005-04-23 17:27:36 +00001869<!-- ======================================================================= -->
1870<div class="doc_subsection">
1871 <a name="TypeResolve">LLVM Type Resolution</a>
1872</div>
Chris Lattnerd9d6e102005-04-23 16:10:52 +00001873
Chris Lattnerf1b200b2005-04-23 17:27:36 +00001874<div class="doc_text">
1875
1876<p>
1877The LLVM type system has a very simple goal: allow clients to compare types for
1878structural equality with a simple pointer comparison (aka a shallow compare).
1879This goal makes clients much simpler and faster, and is used throughout the LLVM
1880system.
1881</p>
1882
1883<p>
1884Unfortunately achieving this goal is not a simple matter. In particular,
1885recursive types and late resolution of opaque types makes the situation very
1886difficult to handle. Fortunately, for the most part, our implementation makes
1887most clients able to be completely unaware of the nasty internal details. The
1888primary case where clients are exposed to the inner workings of it are when
1889building a recursive type. In addition to this case, the LLVM bytecode reader,
1890assembly parser, and linker also have to be aware of the inner workings of this
1891system.
1892</p>
1893
Chris Lattner0f876db2005-04-25 15:47:57 +00001894<p>
1895For our purposes below, we need three concepts. First, an "Opaque Type" is
1896exactly as defined in the <a href="LangRef.html#t_opaque">language
1897reference</a>. Second an "Abstract Type" is any type which includes an
Reid Spencer06565dc2007-01-12 17:11:23 +00001898opaque type as part of its type graph (for example "<tt>{ opaque, i32 }</tt>").
1899Third, a concrete type is a type that is not an abstract type (e.g. "<tt>{ i32,
Chris Lattner0f876db2005-04-25 15:47:57 +00001900float }</tt>").
1901</p>
1902
Chris Lattnerf1b200b2005-04-23 17:27:36 +00001903</div>
1904
1905<!-- ______________________________________________________________________ -->
1906<div class="doc_subsubsection">
1907 <a name="BuildRecType">Basic Recursive Type Construction</a>
1908</div>
1909
1910<div class="doc_text">
1911
1912<p>
1913Because the most common question is "how do I build a recursive type with LLVM",
1914we answer it now and explain it as we go. Here we include enough to cause this
1915to be emitted to an output .ll file:
1916</p>
1917
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001918<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +00001919<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +00001920%mylist = type { %mylist*, i32 }
Chris Lattnerf1b200b2005-04-23 17:27:36 +00001921</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001922</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +00001923
1924<p>
1925To build this, use the following LLVM APIs:
1926</p>
1927
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001928<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +00001929<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +00001930// <i>Create the initial outer struct</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001931<a href="#PATypeHolder">PATypeHolder</a> StructTy = OpaqueType::get();
1932std::vector&lt;const Type*&gt; Elts;
1933Elts.push_back(PointerType::get(StructTy));
1934Elts.push_back(Type::IntTy);
1935StructType *NewSTy = StructType::get(Elts);
Chris Lattnerf1b200b2005-04-23 17:27:36 +00001936
Reid Spencer06565dc2007-01-12 17:11:23 +00001937// <i>At this point, NewSTy = "{ opaque*, i32 }". Tell VMCore that</i>
Bill Wendling82e2eea2006-10-11 18:00:22 +00001938// <i>the struct and the opaque type are actually the same.</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001939cast&lt;OpaqueType&gt;(StructTy.get())-&gt;<a href="#refineAbstractTypeTo">refineAbstractTypeTo</a>(NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +00001940
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001941// <i>NewSTy is potentially invalidated, but StructTy (a <a href="#PATypeHolder">PATypeHolder</a>) is</i>
Bill Wendling82e2eea2006-10-11 18:00:22 +00001942// <i>kept up-to-date</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001943NewSTy = cast&lt;StructType&gt;(StructTy.get());
Chris Lattnerf1b200b2005-04-23 17:27:36 +00001944
Bill Wendling82e2eea2006-10-11 18:00:22 +00001945// <i>Add a name for the type to the module symbol table (optional)</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001946MyModule-&gt;addTypeName("mylist", NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +00001947</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00001948</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +00001949
1950<p>
1951This code shows the basic approach used to build recursive types: build a
1952non-recursive type using 'opaque', then use type unification to close the cycle.
1953The type unification step is performed by the <tt><a
Chris Lattneraff26d12007-02-03 03:06:52 +00001954href="#refineAbstractTypeTo">refineAbstractTypeTo</a></tt> method, which is
Chris Lattnerf1b200b2005-04-23 17:27:36 +00001955described next. After that, we describe the <a
1956href="#PATypeHolder">PATypeHolder class</a>.
1957</p>
1958
1959</div>
1960
1961<!-- ______________________________________________________________________ -->
1962<div class="doc_subsubsection">
1963 <a name="refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a>
1964</div>
1965
1966<div class="doc_text">
1967<p>
1968The <tt>refineAbstractTypeTo</tt> method starts the type unification process.
1969While this method is actually a member of the DerivedType class, it is most
1970often used on OpaqueType instances. Type unification is actually a recursive
1971process. After unification, types can become structurally isomorphic to
1972existing types, and all duplicates are deleted (to preserve pointer equality).
1973</p>
1974
1975<p>
1976In the example above, the OpaqueType object is definitely deleted.
Reid Spencer06565dc2007-01-12 17:11:23 +00001977Additionally, if there is an "{ \2*, i32}" type already created in the system,
Chris Lattnerf1b200b2005-04-23 17:27:36 +00001978the pointer and struct type created are <b>also</b> deleted. Obviously whenever
1979a type is deleted, any "Type*" pointers in the program are invalidated. As
1980such, it is safest to avoid having <i>any</i> "Type*" pointers to abstract types
1981live across a call to <tt>refineAbstractTypeTo</tt> (note that non-abstract
1982types can never move or be deleted). To deal with this, the <a
1983href="#PATypeHolder">PATypeHolder</a> class is used to maintain a stable
1984reference to a possibly refined type, and the <a
1985href="#AbstractTypeUser">AbstractTypeUser</a> class is used to update more
1986complex datastructures.
1987</p>
1988
1989</div>
1990
1991<!-- ______________________________________________________________________ -->
1992<div class="doc_subsubsection">
1993 <a name="PATypeHolder">The PATypeHolder Class</a>
1994</div>
1995
1996<div class="doc_text">
1997<p>
1998PATypeHolder is a form of a "smart pointer" for Type objects. When VMCore
1999happily goes about nuking types that become isomorphic to existing types, it
2000automatically updates all PATypeHolder objects to point to the new type. In the
2001example above, this allows the code to maintain a pointer to the resultant
2002resolved recursive type, even though the Type*'s are potentially invalidated.
2003</p>
2004
2005<p>
2006PATypeHolder is an extremely light-weight object that uses a lazy union-find
2007implementation to update pointers. For example the pointer from a Value to its
2008Type is maintained by PATypeHolder objects.
2009</p>
2010
2011</div>
2012
2013<!-- ______________________________________________________________________ -->
2014<div class="doc_subsubsection">
2015 <a name="AbstractTypeUser">The AbstractTypeUser Class</a>
2016</div>
2017
2018<div class="doc_text">
2019
2020<p>
2021Some data structures need more to perform more complex updates when types get
2022resolved. The <a href="#SymbolTable">SymbolTable</a> class, for example, needs
2023move and potentially merge type planes in its representation when a pointer
2024changes.</p>
2025
2026<p>
2027To support this, a class can derive from the AbstractTypeUser class. This class
2028allows it to get callbacks when certain types are resolved. To register to get
2029callbacks for a particular type, the DerivedType::{add/remove}AbstractTypeUser
Chris Lattner0f876db2005-04-25 15:47:57 +00002030methods can be called on a type. Note that these methods only work for <i>
Reid Spencer06565dc2007-01-12 17:11:23 +00002031 abstract</i> types. Concrete types (those that do not include any opaque
2032objects) can never be refined.
Chris Lattnerf1b200b2005-04-23 17:27:36 +00002033</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +00002034</div>
2035
2036
2037<!-- ======================================================================= -->
2038<div class="doc_subsection">
2039 <a name="SymbolTable">The <tt>SymbolTable</tt> class</a>
2040</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +00002041
Chris Lattnerd9d6e102005-04-23 16:10:52 +00002042<div class="doc_text">
2043<p>This class provides a symbol table that the <a
2044href="#Function"><tt>Function</tt></a> and <a href="#Module">
2045<tt>Module</tt></a> classes use for naming definitions. The symbol table can
Reid Spencera6362242007-01-07 00:41:39 +00002046provide a name for any <a href="#Value"><tt>Value</tt></a>.
2047<tt>SymbolTable</tt> is an abstract data type. It hides the data it contains
2048and provides access to it through a controlled interface.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +00002049
Reid Spencera6362242007-01-07 00:41:39 +00002050<p>Note that the <tt>SymbolTable</tt> class should not be directly accessed
2051by most clients. It should only be used when iteration over the symbol table
2052names themselves are required, which is very special purpose. Note that not
2053all LLVM
Chris Lattnerd9d6e102005-04-23 16:10:52 +00002054<a href="#Value">Value</a>s have names, and those without names (i.e. they have
2055an empty name) do not exist in the symbol table.
2056</p>
2057
2058<p>To use the <tt>SymbolTable</tt> well, you need to understand the
2059structure of the information it holds. The class contains two
2060<tt>std::map</tt> objects. The first, <tt>pmap</tt>, is a map of
2061<tt>Type*</tt> to maps of name (<tt>std::string</tt>) to <tt>Value*</tt>.
Reid Spencera6362242007-01-07 00:41:39 +00002062Thus, Values are stored in two-dimensions and accessed by <tt>Type</tt> and
2063name.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +00002064
2065<p>The interface of this class provides three basic types of operations:
2066<ol>
2067 <li><em>Accessors</em>. Accessors provide read-only access to information
2068 such as finding a value for a name with the
2069 <a href="#SymbolTable_lookup">lookup</a> method.</li>
2070 <li><em>Mutators</em>. Mutators allow the user to add information to the
2071 <tt>SymbolTable</tt> with methods like
2072 <a href="#SymbolTable_insert"><tt>insert</tt></a>.</li>
2073 <li><em>Iterators</em>. Iterators allow the user to traverse the content
2074 of the symbol table in well defined ways, such as the method
Reid Spencera6362242007-01-07 00:41:39 +00002075 <a href="#SymbolTable_plane_begin"><tt>plane_begin</tt></a>.</li>
Chris Lattnerd9d6e102005-04-23 16:10:52 +00002076</ol>
2077
2078<h3>Accessors</h3>
2079<dl>
2080 <dt><tt>Value* lookup(const Type* Ty, const std::string&amp; name) const</tt>:
2081 </dt>
2082 <dd>The <tt>lookup</tt> method searches the type plane given by the
2083 <tt>Ty</tt> parameter for a <tt>Value</tt> with the provided <tt>name</tt>.
2084 If a suitable <tt>Value</tt> is not found, null is returned.</dd>
2085
Chris Lattnerd9d6e102005-04-23 16:10:52 +00002086 <dt><tt>bool isEmpty() const</tt>:</dt>
2087 <dd>This function returns true if both the value and types maps are
2088 empty</dd>
2089</dl>
2090
2091<h3>Mutators</h3>
2092<dl>
2093 <dt><tt>void insert(Value *Val)</tt>:</dt>
2094 <dd>This method adds the provided value to the symbol table. The Value must
2095 have both a name and a type which are extracted and used to place the value
2096 in the correct type plane under the value's name.</dd>
2097
2098 <dt><tt>void insert(const std::string&amp; Name, Value *Val)</tt>:</dt>
2099 <dd> Inserts a constant or type into the symbol table with the specified
2100 name. There can be a many to one mapping between names and constants
2101 or types.</dd>
2102
Chris Lattnerd9d6e102005-04-23 16:10:52 +00002103 <dt><tt>void remove(Value* Val)</tt>:</dt>
2104 <dd> This method removes a named value from the symbol table. The
2105 type and name of the Value are extracted from \p N and used to
2106 lookup the Value in the correct type plane. If the Value is
2107 not in the symbol table, this method silently ignores the
2108 request.</dd>
2109
Chris Lattnerd9d6e102005-04-23 16:10:52 +00002110 <dt><tt>Value* remove(const std::string&amp; Name, Value *Val)</tt>:</dt>
2111 <dd> Remove a constant or type with the specified name from the
2112 symbol table.</dd>
2113
Reid Spencera6362242007-01-07 00:41:39 +00002114 <dt><tt>Value *remove(const value_iterator&amp; It)</tt>:</dt>
Chris Lattnerd9d6e102005-04-23 16:10:52 +00002115 <dd> Removes a specific value from the symbol table.
2116 Returns the removed value.</dd>
2117
2118 <dt><tt>bool strip()</tt>:</dt>
2119 <dd> This method will strip the symbol table of its names leaving
2120 the type and values. </dd>
2121
2122 <dt><tt>void clear()</tt>:</dt>
2123 <dd>Empty the symbol table completely.</dd>
2124</dl>
2125
2126<h3>Iteration</h3>
2127<p>The following functions describe three types of iterators you can obtain
2128the beginning or end of the sequence for both const and non-const. It is
2129important to keep track of the different kinds of iterators. There are
2130three idioms worth pointing out:</p>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00002131
Chris Lattnerd9d6e102005-04-23 16:10:52 +00002132<table>
2133 <tr><th>Units</th><th>Iterator</th><th>Idiom</th></tr>
2134 <tr>
2135 <td align="left">Planes Of name/Value maps</td><td>PI</td>
2136 <td align="left"><pre><tt>
2137for (SymbolTable::plane_const_iterator PI = ST.plane_begin(),
2138 PE = ST.plane_end(); PI != PE; ++PI ) {
Bill Wendling82e2eea2006-10-11 18:00:22 +00002139 PI-&gt;first // <i>This is the Type* of the plane</i>
2140 PI-&gt;second // <i>This is the SymbolTable::ValueMap of name/Value pairs</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00002141}
Chris Lattnerd9d6e102005-04-23 16:10:52 +00002142 </tt></pre></td>
2143 </tr>
2144 <tr>
Chris Lattnerd9d6e102005-04-23 16:10:52 +00002145 <td align="left">name/Value pairs in a plane</td><td>VI</td>
2146 <td align="left"><pre><tt>
2147for (SymbolTable::value_const_iterator VI = ST.value_begin(SomeType),
Bill Wendling3cd5ca62006-10-11 06:30:10 +00002148 VE = ST.value_end(SomeType); VI != VE; ++VI ) {
Bill Wendling82e2eea2006-10-11 18:00:22 +00002149 VI-&gt;first // <i>This is the name of the Value</i>
2150 VI-&gt;second // <i>This is the Value* value associated with the name</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +00002151}
Chris Lattnerd9d6e102005-04-23 16:10:52 +00002152 </tt></pre></td>
2153 </tr>
2154</table>
2155
2156<p>Using the recommended iterator names and idioms will help you avoid
2157making mistakes. Of particular note, make sure that whenever you use
2158value_begin(SomeType) that you always compare the resulting iterator
2159with value_end(SomeType) not value_end(SomeOtherType) or else you
2160will loop infinitely.</p>
2161
2162<dl>
2163
2164 <dt><tt>plane_iterator plane_begin()</tt>:</dt>
2165 <dd>Get an iterator that starts at the beginning of the type planes.
2166 The iterator will iterate over the Type/ValueMap pairs in the
2167 type planes. </dd>
2168
2169 <dt><tt>plane_const_iterator plane_begin() const</tt>:</dt>
2170 <dd>Get a const_iterator that starts at the beginning of the type
2171 planes. The iterator will iterate over the Type/ValueMap pairs
2172 in the type planes. </dd>
2173
2174 <dt><tt>plane_iterator plane_end()</tt>:</dt>
2175 <dd>Get an iterator at the end of the type planes. This serves as
2176 the marker for end of iteration over the type planes.</dd>
2177
2178 <dt><tt>plane_const_iterator plane_end() const</tt>:</dt>
2179 <dd>Get a const_iterator at the end of the type planes. This serves as
2180 the marker for end of iteration over the type planes.</dd>
2181
2182 <dt><tt>value_iterator value_begin(const Type *Typ)</tt>:</dt>
2183 <dd>Get an iterator that starts at the beginning of a type plane.
2184 The iterator will iterate over the name/value pairs in the type plane.
2185 Note: The type plane must already exist before using this.</dd>
2186
2187 <dt><tt>value_const_iterator value_begin(const Type *Typ) const</tt>:</dt>
2188 <dd>Get a const_iterator that starts at the beginning of a type plane.
2189 The iterator will iterate over the name/value pairs in the type plane.
2190 Note: The type plane must already exist before using this.</dd>
2191
2192 <dt><tt>value_iterator value_end(const Type *Typ)</tt>:</dt>
2193 <dd>Get an iterator to the end of a type plane. This serves as the marker
2194 for end of iteration of the type plane.
2195 Note: The type plane must already exist before using this.</dd>
2196
2197 <dt><tt>value_const_iterator value_end(const Type *Typ) const</tt>:</dt>
2198 <dd>Get a const_iterator to the end of a type plane. This serves as the
2199 marker for end of iteration of the type plane.
2200 Note: the type plane must already exist before using this.</dd>
2201
Chris Lattnerd9d6e102005-04-23 16:10:52 +00002202 <dt><tt>plane_const_iterator find(const Type* Typ ) const</tt>:</dt>
2203 <dd>This method returns a plane_const_iterator for iteration over
2204 the type planes starting at a specific plane, given by \p Ty.</dd>
2205
2206 <dt><tt>plane_iterator find( const Type* Typ </tt>:</dt>
2207 <dd>This method returns a plane_iterator for iteration over the
2208 type planes starting at a specific plane, given by \p Ty.</dd>
2209
2210</dl>
2211</div>
2212
2213
2214
2215<!-- *********************************************************************** -->
2216<div class="doc_section">
Misha Brukman13fd15c2004-01-15 00:14:41 +00002217 <a name="coreclasses">The Core LLVM Class Hierarchy Reference </a>
2218</div>
2219<!-- *********************************************************************** -->
2220
2221<div class="doc_text">
Reid Spencer303c4b42007-01-12 17:26:25 +00002222<p><tt>#include "<a href="/doxygen/Type_8h-source.html">llvm/Type.h</a>"</tt>
2223<br>doxygen info: <a href="/doxygen/classllvm_1_1Type.html">Type Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002224
2225<p>The Core LLVM classes are the primary means of representing the program
Chris Lattner261efe92003-11-25 01:02:51 +00002226being inspected or transformed. The core LLVM classes are defined in
2227header files in the <tt>include/llvm/</tt> directory, and implemented in
Misha Brukman13fd15c2004-01-15 00:14:41 +00002228the <tt>lib/VMCore</tt> directory.</p>
2229
2230</div>
2231
2232<!-- ======================================================================= -->
2233<div class="doc_subsection">
Reid Spencer303c4b42007-01-12 17:26:25 +00002234 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
2235</div>
2236
2237<div class="doc_text">
2238
2239 <p><tt>Type</tt> is a superclass of all type classes. Every <tt>Value</tt> has
2240 a <tt>Type</tt>. <tt>Type</tt> cannot be instantiated directly but only
2241 through its subclasses. Certain primitive types (<tt>VoidType</tt>,
2242 <tt>LabelType</tt>, <tt>FloatType</tt> and <tt>DoubleType</tt>) have hidden
2243 subclasses. They are hidden because they offer no useful functionality beyond
2244 what the <tt>Type</tt> class offers except to distinguish themselves from
2245 other subclasses of <tt>Type</tt>.</p>
2246 <p>All other types are subclasses of <tt>DerivedType</tt>. Types can be
2247 named, but this is not a requirement. There exists exactly
2248 one instance of a given shape at any one time. This allows type equality to
2249 be performed with address equality of the Type Instance. That is, given two
2250 <tt>Type*</tt> values, the types are identical if the pointers are identical.
2251 </p>
2252</div>
2253
2254<!-- _______________________________________________________________________ -->
2255<div class="doc_subsubsection">
2256 <a name="m_Value">Important Public Methods</a>
2257</div>
2258
2259<div class="doc_text">
2260
2261<ul>
Chris Lattner8f79df32007-01-15 01:55:32 +00002262 <li><tt>bool isInteger() const</tt>: Returns true for any integer type.</li>
Reid Spencer303c4b42007-01-12 17:26:25 +00002263
2264 <li><tt>bool isFloatingPoint()</tt>: Return true if this is one of the two
2265 floating point types.</li>
2266
2267 <li><tt>bool isAbstract()</tt>: Return true if the type is abstract (contains
2268 an OpaqueType anywhere in its definition).</li>
2269
2270 <li><tt>bool isSized()</tt>: Return true if the type has known size. Things
2271 that don't have a size are abstract types, labels and void.</li>
2272
2273</ul>
2274</div>
2275
2276<!-- _______________________________________________________________________ -->
2277<div class="doc_subsubsection">
2278 <a name="m_Value">Important Derived Types</a>
2279</div>
2280<div class="doc_text">
2281<dl>
2282 <dt><tt>IntegerType</tt></dt>
2283 <dd>Subclass of DerivedType that represents integer types of any bit width.
2284 Any bit width between <tt>IntegerType::MIN_INT_BITS</tt> (1) and
2285 <tt>IntegerType::MAX_INT_BITS</tt> (~8 million) can be represented.
2286 <ul>
2287 <li><tt>static const IntegerType* get(unsigned NumBits)</tt>: get an integer
2288 type of a specific bit width.</li>
2289 <li><tt>unsigned getBitWidth() const</tt>: Get the bit width of an integer
2290 type.</li>
2291 </ul>
2292 </dd>
2293 <dt><tt>SequentialType</tt></dt>
2294 <dd>This is subclassed by ArrayType and PointerType
2295 <ul>
2296 <li><tt>const Type * getElementType() const</tt>: Returns the type of each
2297 of the elements in the sequential type. </li>
2298 </ul>
2299 </dd>
2300 <dt><tt>ArrayType</tt></dt>
2301 <dd>This is a subclass of SequentialType and defines the interface for array
2302 types.
2303 <ul>
2304 <li><tt>unsigned getNumElements() const</tt>: Returns the number of
2305 elements in the array. </li>
2306 </ul>
2307 </dd>
2308 <dt><tt>PointerType</tt></dt>
Chris Lattner302da1e2007-02-03 03:05:57 +00002309 <dd>Subclass of SequentialType for pointer types.</dd>
Reid Spencer303c4b42007-01-12 17:26:25 +00002310 <dt><tt>PackedType</tt></dt>
2311 <dd>Subclass of SequentialType for packed (vector) types. A
2312 packed type is similar to an ArrayType but is distinguished because it is
2313 a first class type wherease ArrayType is not. Packed types are used for
2314 vector operations and are usually small vectors of of an integer or floating
2315 point type.</dd>
2316 <dt><tt>StructType</tt></dt>
2317 <dd>Subclass of DerivedTypes for struct types.</dd>
2318 <dt><tt>FunctionType</tt></dt>
2319 <dd>Subclass of DerivedTypes for function types.
2320 <ul>
2321 <li><tt>bool isVarArg() const</tt>: Returns true if its a vararg
2322 function</li>
2323 <li><tt> const Type * getReturnType() const</tt>: Returns the
2324 return type of the function.</li>
2325 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns
2326 the type of the ith parameter.</li>
2327 <li><tt> const unsigned getNumParams() const</tt>: Returns the
2328 number of formal parameters.</li>
2329 </ul>
2330 </dd>
2331 <dt><tt>OpaqueType</tt></dt>
2332 <dd>Sublcass of DerivedType for abstract types. This class
2333 defines no content and is used as a placeholder for some other type. Note
2334 that OpaqueType is used (temporarily) during type resolution for forward
2335 references of types. Once the referenced type is resolved, the OpaqueType
2336 is replaced with the actual type. OpaqueType can also be used for data
2337 abstraction. At link time opaque types can be resolved to actual types
2338 of the same name.</dd>
2339</dl>
2340</div>
2341
Chris Lattner2b78d962007-02-03 20:02:25 +00002342
2343
2344<!-- ======================================================================= -->
2345<div class="doc_subsection">
2346 <a name="Module">The <tt>Module</tt> class</a>
2347</div>
2348
2349<div class="doc_text">
2350
2351<p><tt>#include "<a
2352href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt><br> doxygen info:
2353<a href="/doxygen/classllvm_1_1Module.html">Module Class</a></p>
2354
2355<p>The <tt>Module</tt> class represents the top level structure present in LLVM
2356programs. An LLVM module is effectively either a translation unit of the
2357original program or a combination of several translation units merged by the
2358linker. The <tt>Module</tt> class keeps track of a list of <a
2359href="#Function"><tt>Function</tt></a>s, a list of <a
2360href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
2361href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
2362helpful member functions that try to make common operations easy.</p>
2363
2364</div>
2365
2366<!-- _______________________________________________________________________ -->
2367<div class="doc_subsubsection">
2368 <a name="m_Module">Important Public Members of the <tt>Module</tt> class</a>
2369</div>
2370
2371<div class="doc_text">
2372
2373<ul>
2374 <li><tt>Module::Module(std::string name = "")</tt></li>
2375</ul>
2376
2377<p>Constructing a <a href="#Module">Module</a> is easy. You can optionally
2378provide a name for it (probably based on the name of the translation unit).</p>
2379
2380<ul>
2381 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
2382 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
2383
2384 <tt>begin()</tt>, <tt>end()</tt>
2385 <tt>size()</tt>, <tt>empty()</tt>
2386
2387 <p>These are forwarding methods that make it easy to access the contents of
2388 a <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
2389 list.</p></li>
2390
2391 <li><tt>Module::FunctionListType &amp;getFunctionList()</tt>
2392
2393 <p> Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
2394 necessary to use when you need to update the list or perform a complex
2395 action that doesn't have a forwarding method.</p>
2396
2397 <p><!-- Global Variable --></p></li>
2398</ul>
2399
2400<hr>
2401
2402<ul>
2403 <li><tt>Module::global_iterator</tt> - Typedef for global variable list iterator<br>
2404
2405 <tt>Module::const_global_iterator</tt> - Typedef for const_iterator.<br>
2406
2407 <tt>global_begin()</tt>, <tt>global_end()</tt>
2408 <tt>global_size()</tt>, <tt>global_empty()</tt>
2409
2410 <p> These are forwarding methods that make it easy to access the contents of
2411 a <tt>Module</tt> object's <a
2412 href="#GlobalVariable"><tt>GlobalVariable</tt></a> list.</p></li>
2413
2414 <li><tt>Module::GlobalListType &amp;getGlobalList()</tt>
2415
2416 <p>Returns the list of <a
2417 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s. This is necessary to
2418 use when you need to update the list or perform a complex action that
2419 doesn't have a forwarding method.</p>
2420
2421 <p><!-- Symbol table stuff --> </p></li>
2422</ul>
2423
2424<hr>
2425
2426<ul>
2427 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
2428
2429 <p>Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
2430 for this <tt>Module</tt>.</p>
2431
2432 <p><!-- Convenience methods --></p></li>
2433</ul>
2434
2435<hr>
2436
2437<ul>
2438 <li><tt><a href="#Function">Function</a> *getFunction(const std::string
2439 &amp;Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt>
2440
2441 <p>Look up the specified function in the <tt>Module</tt> <a
2442 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
2443 <tt>null</tt>.</p></li>
2444
2445 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const
2446 std::string &amp;Name, const <a href="#FunctionType">FunctionType</a> *T)</tt>
2447
2448 <p>Look up the specified function in the <tt>Module</tt> <a
2449 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
2450 external declaration for the function and return it.</p></li>
2451
2452 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt>
2453
2454 <p>If there is at least one entry in the <a
2455 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
2456 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
2457 string.</p></li>
2458
2459 <li><tt>bool addTypeName(const std::string &amp;Name, const <a
2460 href="#Type">Type</a> *Ty)</tt>
2461
2462 <p>Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
2463 mapping <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this
2464 name, true is returned and the <a
2465 href="#SymbolTable"><tt>SymbolTable</tt></a> is not modified.</p></li>
2466</ul>
2467
2468</div>
2469
2470
Reid Spencer303c4b42007-01-12 17:26:25 +00002471<!-- ======================================================================= -->
2472<div class="doc_subsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +00002473 <a name="Value">The <tt>Value</tt> class</a>
2474</div>
2475
Chris Lattner2b78d962007-02-03 20:02:25 +00002476<div class="doc_text">
Misha Brukman13fd15c2004-01-15 00:14:41 +00002477
2478<p><tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt>
2479<br>
Chris Lattner00815172007-01-04 22:01:45 +00002480doxygen info: <a href="/doxygen/classllvm_1_1Value.html">Value Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002481
2482<p>The <tt>Value</tt> class is the most important class in the LLVM Source
2483base. It represents a typed value that may be used (among other things) as an
2484operand to an instruction. There are many different types of <tt>Value</tt>s,
2485such as <a href="#Constant"><tt>Constant</tt></a>s,<a
2486href="#Argument"><tt>Argument</tt></a>s. Even <a
2487href="#Instruction"><tt>Instruction</tt></a>s and <a
2488href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.</p>
2489
2490<p>A particular <tt>Value</tt> may be used many times in the LLVM representation
2491for a program. For example, an incoming argument to a function (represented
2492with an instance of the <a href="#Argument">Argument</a> class) is "used" by
2493every instruction in the function that references the argument. To keep track
2494of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
2495href="#User"><tt>User</tt></a>s that is using it (the <a
2496href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
2497graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
2498def-use information in the program, and is accessible through the <tt>use_</tt>*
2499methods, shown below.</p>
2500
2501<p>Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed,
2502and this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
2503method. In addition, all LLVM values can be named. The "name" of the
2504<tt>Value</tt> is a symbolic string printed in the LLVM code:</p>
2505
Bill Wendling3cd5ca62006-10-11 06:30:10 +00002506<div class="doc_code">
2507<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +00002508%<b>foo</b> = add i32 1, 2
Bill Wendling3cd5ca62006-10-11 06:30:10 +00002509</pre>
2510</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002511
2512<p><a name="#nameWarning">The name of this instruction is "foo".</a> <b>NOTE</b>
2513that the name of any value may be missing (an empty string), so names should
2514<b>ONLY</b> be used for debugging (making the source code easier to read,
2515debugging printouts), they should not be used to keep track of values or map
2516between them. For this purpose, use a <tt>std::map</tt> of pointers to the
2517<tt>Value</tt> itself instead.</p>
2518
2519<p>One important aspect of LLVM is that there is no distinction between an SSA
2520variable and the operation that produces it. Because of this, any reference to
2521the value produced by an instruction (or the value available as an incoming
Chris Lattnerd5fc4fc2004-03-18 14:58:55 +00002522argument, for example) is represented as a direct pointer to the instance of
2523the class that
Misha Brukman13fd15c2004-01-15 00:14:41 +00002524represents this value. Although this may take some getting used to, it
2525simplifies the representation and makes it easier to manipulate.</p>
2526
2527</div>
2528
2529<!-- _______________________________________________________________________ -->
2530<div class="doc_subsubsection">
2531 <a name="m_Value">Important Public Members of the <tt>Value</tt> class</a>
2532</div>
2533
2534<div class="doc_text">
2535
Chris Lattner261efe92003-11-25 01:02:51 +00002536<ul>
2537 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the
2538use-list<br>
2539 <tt>Value::use_const_iterator</tt> - Typedef for const_iterator over
2540the use-list<br>
2541 <tt>unsigned use_size()</tt> - Returns the number of users of the
2542value.<br>
Chris Lattner9355b472002-09-06 02:50:58 +00002543 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
Chris Lattner261efe92003-11-25 01:02:51 +00002544 <tt>use_iterator use_begin()</tt> - Get an iterator to the start of
2545the use-list.<br>
2546 <tt>use_iterator use_end()</tt> - Get an iterator to the end of the
2547use-list.<br>
2548 <tt><a href="#User">User</a> *use_back()</tt> - Returns the last
2549element in the list.
2550 <p> These methods are the interface to access the def-use
2551information in LLVM. As with all other iterators in LLVM, the naming
2552conventions follow the conventions defined by the <a href="#stl">STL</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002553 </li>
2554 <li><tt><a href="#Type">Type</a> *getType() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002555 <p>This method returns the Type of the Value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002556 </li>
2557 <li><tt>bool hasName() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +00002558 <tt>std::string getName() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +00002559 <tt>void setName(const std::string &amp;Name)</tt>
2560 <p> This family of methods is used to access and assign a name to a <tt>Value</tt>,
2561be aware of the <a href="#nameWarning">precaution above</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002562 </li>
2563 <li><tt>void replaceAllUsesWith(Value *V)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002564
2565 <p>This method traverses the use list of a <tt>Value</tt> changing all <a
2566 href="#User"><tt>User</tt>s</a> of the current value to refer to
2567 "<tt>V</tt>" instead. For example, if you detect that an instruction always
2568 produces a constant value (for example through constant folding), you can
2569 replace all uses of the instruction with the constant like this:</p>
2570
Bill Wendling3cd5ca62006-10-11 06:30:10 +00002571<div class="doc_code">
2572<pre>
2573Inst-&gt;replaceAllUsesWith(ConstVal);
2574</pre>
2575</div>
2576
Chris Lattner261efe92003-11-25 01:02:51 +00002577</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002578
2579</div>
2580
2581<!-- ======================================================================= -->
2582<div class="doc_subsection">
2583 <a name="User">The <tt>User</tt> class</a>
2584</div>
2585
2586<div class="doc_text">
2587
2588<p>
2589<tt>#include "<a href="/doxygen/User_8h-source.html">llvm/User.h</a>"</tt><br>
Misha Brukman384047f2004-06-03 23:29:12 +00002590doxygen info: <a href="/doxygen/classllvm_1_1User.html">User Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002591Superclass: <a href="#Value"><tt>Value</tt></a></p>
2592
2593<p>The <tt>User</tt> class is the common base class of all LLVM nodes that may
2594refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
2595that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
2596referring to. The <tt>User</tt> class itself is a subclass of
2597<tt>Value</tt>.</p>
2598
2599<p>The operands of a <tt>User</tt> point directly to the LLVM <a
2600href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
2601Single Assignment (SSA) form, there can only be one definition referred to,
2602allowing this direct connection. This connection provides the use-def
2603information in LLVM.</p>
2604
2605</div>
2606
2607<!-- _______________________________________________________________________ -->
2608<div class="doc_subsubsection">
2609 <a name="m_User">Important Public Members of the <tt>User</tt> class</a>
2610</div>
2611
2612<div class="doc_text">
2613
2614<p>The <tt>User</tt> class exposes the operand list in two ways: through
2615an index access interface and through an iterator based interface.</p>
2616
Chris Lattner261efe92003-11-25 01:02:51 +00002617<ul>
Chris Lattner261efe92003-11-25 01:02:51 +00002618 <li><tt>Value *getOperand(unsigned i)</tt><br>
2619 <tt>unsigned getNumOperands()</tt>
2620 <p> These two methods expose the operands of the <tt>User</tt> in a
Misha Brukman13fd15c2004-01-15 00:14:41 +00002621convenient form for direct access.</p></li>
2622
Chris Lattner261efe92003-11-25 01:02:51 +00002623 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand
2624list<br>
Chris Lattner58360822005-01-17 00:12:04 +00002625 <tt>op_iterator op_begin()</tt> - Get an iterator to the start of
2626the operand list.<br>
2627 <tt>op_iterator op_end()</tt> - Get an iterator to the end of the
Chris Lattner261efe92003-11-25 01:02:51 +00002628operand list.
2629 <p> Together, these methods make up the iterator based interface to
Misha Brukman13fd15c2004-01-15 00:14:41 +00002630the operands of a <tt>User</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00002631</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002632
2633</div>
2634
2635<!-- ======================================================================= -->
2636<div class="doc_subsection">
2637 <a name="Instruction">The <tt>Instruction</tt> class</a>
2638</div>
2639
2640<div class="doc_text">
2641
2642<p><tt>#include "</tt><tt><a
2643href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt><br>
Misha Brukman31ca1de2004-06-03 23:35:54 +00002644doxygen info: <a href="/doxygen/classllvm_1_1Instruction.html">Instruction Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002645Superclasses: <a href="#User"><tt>User</tt></a>, <a
2646href="#Value"><tt>Value</tt></a></p>
2647
2648<p>The <tt>Instruction</tt> class is the common base class for all LLVM
2649instructions. It provides only a few methods, but is a very commonly used
2650class. The primary data tracked by the <tt>Instruction</tt> class itself is the
2651opcode (instruction type) and the parent <a
2652href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
2653into. To represent a specific type of instruction, one of many subclasses of
2654<tt>Instruction</tt> are used.</p>
2655
2656<p> Because the <tt>Instruction</tt> class subclasses the <a
2657href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
2658way as for other <a href="#User"><tt>User</tt></a>s (with the
2659<tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
2660<tt>op_begin()</tt>/<tt>op_end()</tt> methods).</p> <p> An important file for
2661the <tt>Instruction</tt> class is the <tt>llvm/Instruction.def</tt> file. This
2662file contains some meta-data about the various different types of instructions
2663in LLVM. It describes the enum values that are used as opcodes (for example
Reid Spencerc92d25d2006-12-19 19:47:19 +00002664<tt>Instruction::Add</tt> and <tt>Instruction::ICmp</tt>), as well as the
Misha Brukman13fd15c2004-01-15 00:14:41 +00002665concrete sub-classes of <tt>Instruction</tt> that implement the instruction (for
2666example <tt><a href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
Reid Spencerc92d25d2006-12-19 19:47:19 +00002667href="#CmpInst">CmpInst</a></tt>). Unfortunately, the use of macros in
Misha Brukman13fd15c2004-01-15 00:14:41 +00002668this file confuses doxygen, so these enum values don't show up correctly in the
Misha Brukman31ca1de2004-06-03 23:35:54 +00002669<a href="/doxygen/classllvm_1_1Instruction.html">doxygen output</a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002670
2671</div>
2672
2673<!-- _______________________________________________________________________ -->
2674<div class="doc_subsubsection">
Reid Spencerc92d25d2006-12-19 19:47:19 +00002675 <a name="s_Instruction">Important Subclasses of the <tt>Instruction</tt>
2676 class</a>
2677</div>
2678<div class="doc_text">
2679 <ul>
2680 <li><tt><a name="BinaryOperator">BinaryOperator</a></tt>
2681 <p>This subclasses represents all two operand instructions whose operands
2682 must be the same type, except for the comparison instructions.</p></li>
2683 <li><tt><a name="CastInst">CastInst</a></tt>
2684 <p>This subclass is the parent of the 12 casting instructions. It provides
2685 common operations on cast instructions.</p>
2686 <li><tt><a name="CmpInst">CmpInst</a></tt>
2687 <p>This subclass respresents the two comparison instructions,
2688 <a href="LangRef.html#i_icmp">ICmpInst</a> (integer opreands), and
2689 <a href="LangRef.html#i_fcmp">FCmpInst</a> (floating point operands).</p>
2690 <li><tt><a name="TerminatorInst">TerminatorInst</a></tt>
2691 <p>This subclass is the parent of all terminator instructions (those which
2692 can terminate a block).</p>
2693 </ul>
2694 </div>
2695
2696<!-- _______________________________________________________________________ -->
2697<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +00002698 <a name="m_Instruction">Important Public Members of the <tt>Instruction</tt>
2699 class</a>
2700</div>
2701
2702<div class="doc_text">
2703
Chris Lattner261efe92003-11-25 01:02:51 +00002704<ul>
2705 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002706 <p>Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that
2707this <tt>Instruction</tt> is embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00002708 <li><tt>bool mayWriteToMemory()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002709 <p>Returns true if the instruction writes to memory, i.e. it is a
2710 <tt>call</tt>,<tt>free</tt>,<tt>invoke</tt>, or <tt>store</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00002711 <li><tt>unsigned getOpcode()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002712 <p>Returns the opcode for the <tt>Instruction</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00002713 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002714 <p>Returns another instance of the specified instruction, identical
Chris Lattner261efe92003-11-25 01:02:51 +00002715in all ways to the original except that the instruction has no parent
2716(ie it's not embedded into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>),
Misha Brukman13fd15c2004-01-15 00:14:41 +00002717and it has no name</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00002718</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002719
2720</div>
2721
2722<!-- ======================================================================= -->
2723<div class="doc_subsection">
Chris Lattner2b78d962007-02-03 20:02:25 +00002724 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002725</div>
2726
2727<div class="doc_text">
2728
Chris Lattner2b78d962007-02-03 20:02:25 +00002729<p>Constant represents a base class for different types of constants. It
2730is subclassed by ConstantInt, ConstantArray, etc. for representing
2731the various types of Constants. <a href="#GlobalValue">GlobalValue</a> is also
2732a subclass, which represents the address of a global variable or function.
2733</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002734
2735</div>
2736
2737<!-- _______________________________________________________________________ -->
Chris Lattner2b78d962007-02-03 20:02:25 +00002738<div class="doc_subsubsection">Important Subclasses of Constant </div>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002739<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +00002740<ul>
Chris Lattner2b78d962007-02-03 20:02:25 +00002741 <li>ConstantInt : This subclass of Constant represents an integer constant of
2742 any width.
2743 <ul>
2744 <li><tt>int64_t getSExtValue() const</tt>: Returns the underlying value of
2745 this constant as a sign extended signed integer value.</li>
2746 <li><tt>uint64_t getZExtValue() const</tt>: Returns the underlying value
2747 of this constant as a zero extended unsigned integer value.</li>
2748 <li><tt>static ConstantInt* get(const Type *Ty, uint64_t Val)</tt>:
2749 Returns the ConstantInt object that represents the value provided by
2750 <tt>Val</tt> for integer type <tt>Ty</tt>.</li>
2751 </ul>
2752 </li>
2753 <li>ConstantFP : This class represents a floating point constant.
2754 <ul>
2755 <li><tt>double getValue() const</tt>: Returns the underlying value of
2756 this constant. </li>
2757 </ul>
2758 </li>
2759 <li>ConstantArray : This represents a constant array.
2760 <ul>
2761 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
2762 a vector of component constants that makeup this array. </li>
2763 </ul>
2764 </li>
2765 <li>ConstantStruct : This represents a constant struct.
2766 <ul>
2767 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
2768 a vector of component constants that makeup this array. </li>
2769 </ul>
2770 </li>
2771 <li>GlobalValue : This represents either a global variable or a function. In
2772 either case, the value is a constant fixed address (after linking).
2773 </li>
Chris Lattner261efe92003-11-25 01:02:51 +00002774</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002775</div>
2776
Chris Lattner2b78d962007-02-03 20:02:25 +00002777
Misha Brukman13fd15c2004-01-15 00:14:41 +00002778<!-- ======================================================================= -->
2779<div class="doc_subsection">
2780 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
2781</div>
2782
2783<div class="doc_text">
2784
2785<p><tt>#include "<a
2786href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt><br>
Misha Brukman384047f2004-06-03 23:29:12 +00002787doxygen info: <a href="/doxygen/classllvm_1_1GlobalValue.html">GlobalValue
2788Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +00002789Superclasses: <a href="#Constant"><tt>Constant</tt></a>,
2790<a href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002791
2792<p>Global values (<a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
2793href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
2794visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
2795Because they are visible at global scope, they are also subject to linking with
2796other globals defined in different translation units. To control the linking
2797process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
2798<tt>GlobalValue</tt>s know whether they have internal or external linkage, as
Reid Spencer8b2da7a2004-07-18 13:10:31 +00002799defined by the <tt>LinkageTypes</tt> enumeration.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002800
2801<p>If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
2802<tt>static</tt> in C), it is not visible to code outside the current translation
2803unit, and does not participate in linking. If it has external linkage, it is
2804visible to external code, and does participate in linking. In addition to
2805linkage information, <tt>GlobalValue</tt>s keep track of which <a
2806href="#Module"><tt>Module</tt></a> they are currently part of.</p>
2807
2808<p>Because <tt>GlobalValue</tt>s are memory objects, they are always referred to
2809by their <b>address</b>. As such, the <a href="#Type"><tt>Type</tt></a> of a
2810global is always a pointer to its contents. It is important to remember this
2811when using the <tt>GetElementPtrInst</tt> instruction because this pointer must
2812be dereferenced first. For example, if you have a <tt>GlobalVariable</tt> (a
2813subclass of <tt>GlobalValue)</tt> that is an array of 24 ints, type <tt>[24 x
Reid Spencer06565dc2007-01-12 17:11:23 +00002814i32]</tt>, then the <tt>GlobalVariable</tt> is a pointer to that array. Although
Misha Brukman13fd15c2004-01-15 00:14:41 +00002815the address of the first element of this array and the value of the
2816<tt>GlobalVariable</tt> are the same, they have different types. The
Reid Spencer06565dc2007-01-12 17:11:23 +00002817<tt>GlobalVariable</tt>'s type is <tt>[24 x i32]</tt>. The first element's type
2818is <tt>i32.</tt> Because of this, accessing a global value requires you to
Misha Brukman13fd15c2004-01-15 00:14:41 +00002819dereference the pointer with <tt>GetElementPtrInst</tt> first, then its elements
2820can be accessed. This is explained in the <a href="LangRef.html#globalvars">LLVM
2821Language Reference Manual</a>.</p>
2822
2823</div>
2824
2825<!-- _______________________________________________________________________ -->
2826<div class="doc_subsubsection">
2827 <a name="m_GlobalValue">Important Public Members of the <tt>GlobalValue</tt>
2828 class</a>
2829</div>
2830
2831<div class="doc_text">
2832
Chris Lattner261efe92003-11-25 01:02:51 +00002833<ul>
2834 <li><tt>bool hasInternalLinkage() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +00002835 <tt>bool hasExternalLinkage() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +00002836 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt>
2837 <p> These methods manipulate the linkage characteristics of the <tt>GlobalValue</tt>.</p>
2838 <p> </p>
2839 </li>
2840 <li><tt><a href="#Module">Module</a> *getParent()</tt>
2841 <p> This returns the <a href="#Module"><tt>Module</tt></a> that the
Misha Brukman13fd15c2004-01-15 00:14:41 +00002842GlobalValue is currently embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00002843</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002844
2845</div>
2846
2847<!-- ======================================================================= -->
2848<div class="doc_subsection">
2849 <a name="Function">The <tt>Function</tt> class</a>
2850</div>
2851
2852<div class="doc_text">
2853
2854<p><tt>#include "<a
2855href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt><br> doxygen
Misha Brukman31ca1de2004-06-03 23:35:54 +00002856info: <a href="/doxygen/classllvm_1_1Function.html">Function Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +00002857Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
2858<a href="#Constant"><tt>Constant</tt></a>,
2859<a href="#User"><tt>User</tt></a>,
2860<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002861
2862<p>The <tt>Function</tt> class represents a single procedure in LLVM. It is
2863actually one of the more complex classes in the LLVM heirarchy because it must
2864keep track of a large amount of data. The <tt>Function</tt> class keeps track
Reid Spencerbe5e85e2006-04-14 14:11:48 +00002865of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal
2866<a href="#Argument"><tt>Argument</tt></a>s, and a
2867<a href="#SymbolTable"><tt>SymbolTable</tt></a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002868
2869<p>The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most
2870commonly used part of <tt>Function</tt> objects. The list imposes an implicit
2871ordering of the blocks in the function, which indicate how the code will be
2872layed out by the backend. Additionally, the first <a
2873href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
2874<tt>Function</tt>. It is not legal in LLVM to explicitly branch to this initial
2875block. There are no implicit exit nodes, and in fact there may be multiple exit
2876nodes from a single <tt>Function</tt>. If the <a
2877href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
2878the <tt>Function</tt> is actually a function declaration: the actual body of the
2879function hasn't been linked in yet.</p>
2880
2881<p>In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
2882<tt>Function</tt> class also keeps track of the list of formal <a
2883href="#Argument"><tt>Argument</tt></a>s that the function receives. This
2884container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
2885nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
2886the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.</p>
2887
2888<p>The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used
2889LLVM feature that is only used when you have to look up a value by name. Aside
2890from that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used
2891internally to make sure that there are not conflicts between the names of <a
2892href="#Instruction"><tt>Instruction</tt></a>s, <a
2893href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
2894href="#Argument"><tt>Argument</tt></a>s in the function body.</p>
2895
Reid Spencer8b2da7a2004-07-18 13:10:31 +00002896<p>Note that <tt>Function</tt> is a <a href="#GlobalValue">GlobalValue</a>
2897and therefore also a <a href="#Constant">Constant</a>. The value of the function
2898is its address (after linking) which is guaranteed to be constant.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002899</div>
2900
2901<!-- _______________________________________________________________________ -->
2902<div class="doc_subsubsection">
2903 <a name="m_Function">Important Public Members of the <tt>Function</tt>
2904 class</a>
2905</div>
2906
2907<div class="doc_text">
2908
Chris Lattner261efe92003-11-25 01:02:51 +00002909<ul>
2910 <li><tt>Function(const </tt><tt><a href="#FunctionType">FunctionType</a>
Chris Lattnerac479e52004-08-04 05:10:48 +00002911 *Ty, LinkageTypes Linkage, const std::string &amp;N = "", Module* Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002912
2913 <p>Constructor used when you need to create new <tt>Function</tt>s to add
2914 the the program. The constructor must specify the type of the function to
Chris Lattnerac479e52004-08-04 05:10:48 +00002915 create and what type of linkage the function should have. The <a
2916 href="#FunctionType"><tt>FunctionType</tt></a> argument
Misha Brukman13fd15c2004-01-15 00:14:41 +00002917 specifies the formal arguments and return value for the function. The same
2918 <a href="#FunctionTypel"><tt>FunctionType</tt></a> value can be used to
2919 create multiple functions. The <tt>Parent</tt> argument specifies the Module
2920 in which the function is defined. If this argument is provided, the function
2921 will automatically be inserted into that module's list of
2922 functions.</p></li>
2923
Chris Lattner261efe92003-11-25 01:02:51 +00002924 <li><tt>bool isExternal()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002925
2926 <p>Return whether or not the <tt>Function</tt> has a body defined. If the
2927 function is "external", it does not have a body, and thus must be resolved
2928 by linking with a function defined in a different translation unit.</p></li>
2929
Chris Lattner261efe92003-11-25 01:02:51 +00002930 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
Chris Lattner9355b472002-09-06 02:50:58 +00002931 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002932
Chris Lattner77d69242005-03-15 05:19:20 +00002933 <tt>begin()</tt>, <tt>end()</tt>
2934 <tt>size()</tt>, <tt>empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002935
2936 <p>These are forwarding methods that make it easy to access the contents of
2937 a <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
2938 list.</p></li>
2939
Chris Lattner261efe92003-11-25 01:02:51 +00002940 <li><tt>Function::BasicBlockListType &amp;getBasicBlockList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002941
2942 <p>Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This
2943 is necessary to use when you need to update the list or perform a complex
2944 action that doesn't have a forwarding method.</p></li>
2945
Chris Lattner89cc2652005-03-15 04:48:32 +00002946 <li><tt>Function::arg_iterator</tt> - Typedef for the argument list
Chris Lattner261efe92003-11-25 01:02:51 +00002947iterator<br>
Chris Lattner89cc2652005-03-15 04:48:32 +00002948 <tt>Function::const_arg_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002949
Chris Lattner77d69242005-03-15 05:19:20 +00002950 <tt>arg_begin()</tt>, <tt>arg_end()</tt>
Chris Lattner89cc2652005-03-15 04:48:32 +00002951 <tt>arg_size()</tt>, <tt>arg_empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002952
2953 <p>These are forwarding methods that make it easy to access the contents of
2954 a <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a>
2955 list.</p></li>
2956
Chris Lattner261efe92003-11-25 01:02:51 +00002957 <li><tt>Function::ArgumentListType &amp;getArgumentList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002958
2959 <p>Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
2960 necessary to use when you need to update the list or perform a complex
2961 action that doesn't have a forwarding method.</p></li>
2962
Chris Lattner261efe92003-11-25 01:02:51 +00002963 <li><tt><a href="#BasicBlock">BasicBlock</a> &amp;getEntryBlock()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002964
2965 <p>Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
2966 function. Because the entry block for the function is always the first
2967 block, this returns the first block of the <tt>Function</tt>.</p></li>
2968
Chris Lattner261efe92003-11-25 01:02:51 +00002969 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
2970 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002971
2972 <p>This traverses the <a href="#Type"><tt>Type</tt></a> of the
2973 <tt>Function</tt> and returns the return type of the function, or the <a
2974 href="#FunctionType"><tt>FunctionType</tt></a> of the actual
2975 function.</p></li>
2976
Chris Lattner261efe92003-11-25 01:02:51 +00002977 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002978
Chris Lattner261efe92003-11-25 01:02:51 +00002979 <p> Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002980 for this <tt>Function</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00002981</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00002982
2983</div>
2984
2985<!-- ======================================================================= -->
2986<div class="doc_subsection">
2987 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
2988</div>
2989
2990<div class="doc_text">
2991
2992<p><tt>#include "<a
2993href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt>
2994<br>
Tanya Lattnera3da7772004-06-22 08:02:25 +00002995doxygen info: <a href="/doxygen/classllvm_1_1GlobalVariable.html">GlobalVariable
Reid Spencerbe5e85e2006-04-14 14:11:48 +00002996 Class</a><br>
2997Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
2998<a href="#Constant"><tt>Constant</tt></a>,
2999<a href="#User"><tt>User</tt></a>,
3000<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003001
3002<p>Global variables are represented with the (suprise suprise)
3003<tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are also
3004subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such are
3005always referenced by their address (global values must live in memory, so their
Reid Spencerbe5e85e2006-04-14 14:11:48 +00003006"name" refers to their constant address). See
3007<a href="#GlobalValue"><tt>GlobalValue</tt></a> for more on this. Global
3008variables may have an initial value (which must be a
3009<a href="#Constant"><tt>Constant</tt></a>), and if they have an initializer,
3010they may be marked as "constant" themselves (indicating that their contents
3011never change at runtime).</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003012</div>
3013
3014<!-- _______________________________________________________________________ -->
3015<div class="doc_subsubsection">
3016 <a name="m_GlobalVariable">Important Public Members of the
3017 <tt>GlobalVariable</tt> class</a>
3018</div>
3019
3020<div class="doc_text">
3021
Chris Lattner261efe92003-11-25 01:02:51 +00003022<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003023 <li><tt>GlobalVariable(const </tt><tt><a href="#Type">Type</a> *Ty, bool
3024 isConstant, LinkageTypes&amp; Linkage, <a href="#Constant">Constant</a>
3025 *Initializer = 0, const std::string &amp;Name = "", Module* Parent = 0)</tt>
3026
3027 <p>Create a new global variable of the specified type. If
3028 <tt>isConstant</tt> is true then the global variable will be marked as
3029 unchanging for the program. The Linkage parameter specifies the type of
3030 linkage (internal, external, weak, linkonce, appending) for the variable. If
3031 the linkage is InternalLinkage, WeakLinkage, or LinkOnceLinkage,&nbsp; then
3032 the resultant global variable will have internal linkage. AppendingLinkage
3033 concatenates together all instances (in different translation units) of the
3034 variable into a single variable but is only applicable to arrays. &nbsp;See
3035 the <a href="LangRef.html#modulestructure">LLVM Language Reference</a> for
3036 further details on linkage types. Optionally an initializer, a name, and the
3037 module to put the variable into may be specified for the global variable as
3038 well.</p></li>
3039
Chris Lattner261efe92003-11-25 01:02:51 +00003040 <li><tt>bool isConstant() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003041
3042 <p>Returns true if this is a global variable that is known not to
3043 be modified at runtime.</p></li>
3044
Chris Lattner261efe92003-11-25 01:02:51 +00003045 <li><tt>bool hasInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003046
3047 <p>Returns true if this <tt>GlobalVariable</tt> has an intializer.</p></li>
3048
Chris Lattner261efe92003-11-25 01:02:51 +00003049 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003050
3051 <p>Returns the intial value for a <tt>GlobalVariable</tt>. It is not legal
3052 to call this method if there is no initializer.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00003053</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003054
3055</div>
3056
Chris Lattner2b78d962007-02-03 20:02:25 +00003057
Misha Brukman13fd15c2004-01-15 00:14:41 +00003058<!-- ======================================================================= -->
3059<div class="doc_subsection">
Chris Lattner2b78d962007-02-03 20:02:25 +00003060 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003061</div>
3062
3063<div class="doc_text">
3064
3065<p><tt>#include "<a
Chris Lattner2b78d962007-02-03 20:02:25 +00003066href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt><br>
3067doxygen info: <a href="/doxygen/structllvm_1_1BasicBlock.html">BasicBlock
3068Class</a><br>
3069Superclass: <a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003070
Chris Lattner2b78d962007-02-03 20:02:25 +00003071<p>This class represents a single entry multiple exit section of the code,
3072commonly known as a basic block by the compiler community. The
3073<tt>BasicBlock</tt> class maintains a list of <a
3074href="#Instruction"><tt>Instruction</tt></a>s, which form the body of the block.
3075Matching the language definition, the last element of this list of instructions
3076is always a terminator instruction (a subclass of the <a
3077href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).</p>
3078
3079<p>In addition to tracking the list of instructions that make up the block, the
3080<tt>BasicBlock</tt> class also keeps track of the <a
3081href="#Function"><tt>Function</tt></a> that it is embedded into.</p>
3082
3083<p>Note that <tt>BasicBlock</tt>s themselves are <a
3084href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
3085like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
3086<tt>label</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003087
3088</div>
3089
3090<!-- _______________________________________________________________________ -->
3091<div class="doc_subsubsection">
Chris Lattner2b78d962007-02-03 20:02:25 +00003092 <a name="m_BasicBlock">Important Public Members of the <tt>BasicBlock</tt>
3093 class</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003094</div>
3095
3096<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +00003097<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003098
Chris Lattner2b78d962007-02-03 20:02:25 +00003099<li><tt>BasicBlock(const std::string &amp;Name = "", </tt><tt><a
3100 href="#Function">Function</a> *Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003101
Chris Lattner2b78d962007-02-03 20:02:25 +00003102<p>The <tt>BasicBlock</tt> constructor is used to create new basic blocks for
3103insertion into a function. The constructor optionally takes a name for the new
3104block, and a <a href="#Function"><tt>Function</tt></a> to insert it into. If
3105the <tt>Parent</tt> parameter is specified, the new <tt>BasicBlock</tt> is
3106automatically inserted at the end of the specified <a
3107href="#Function"><tt>Function</tt></a>, if not specified, the BasicBlock must be
3108manually inserted into the <a href="#Function"><tt>Function</tt></a>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003109
Chris Lattner2b78d962007-02-03 20:02:25 +00003110<li><tt>BasicBlock::iterator</tt> - Typedef for instruction list iterator<br>
3111<tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
3112<tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
3113<tt>size()</tt>, <tt>empty()</tt>
3114STL-style functions for accessing the instruction list.
Misha Brukman13fd15c2004-01-15 00:14:41 +00003115
Chris Lattner2b78d962007-02-03 20:02:25 +00003116<p>These methods and typedefs are forwarding functions that have the same
3117semantics as the standard library methods of the same names. These methods
3118expose the underlying instruction list of a basic block in a way that is easy to
3119manipulate. To get the full complement of container operations (including
3120operations to update the list), you must use the <tt>getInstList()</tt>
3121method.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003122
Chris Lattner2b78d962007-02-03 20:02:25 +00003123<li><tt>BasicBlock::InstListType &amp;getInstList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003124
Chris Lattner2b78d962007-02-03 20:02:25 +00003125<p>This method is used to get access to the underlying container that actually
3126holds the Instructions. This method must be used when there isn't a forwarding
3127function in the <tt>BasicBlock</tt> class for the operation that you would like
3128to perform. Because there are no forwarding functions for "updating"
3129operations, you need to use this if you want to update the contents of a
3130<tt>BasicBlock</tt>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003131
Chris Lattner2b78d962007-02-03 20:02:25 +00003132<li><tt><a href="#Function">Function</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003133
Chris Lattner2b78d962007-02-03 20:02:25 +00003134<p> Returns a pointer to <a href="#Function"><tt>Function</tt></a> the block is
3135embedded into, or a null pointer if it is homeless.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003136
Chris Lattner2b78d962007-02-03 20:02:25 +00003137<li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003138
Chris Lattner2b78d962007-02-03 20:02:25 +00003139<p> Returns a pointer to the terminator instruction that appears at the end of
3140the <tt>BasicBlock</tt>. If there is no terminator instruction, or if the last
3141instruction in the block is not a terminator, then a null pointer is
3142returned.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003143
Misha Brukman13fd15c2004-01-15 00:14:41 +00003144</ul>
3145
3146</div>
3147
Misha Brukman13fd15c2004-01-15 00:14:41 +00003148
Misha Brukman13fd15c2004-01-15 00:14:41 +00003149<!-- ======================================================================= -->
3150<div class="doc_subsection">
3151 <a name="Argument">The <tt>Argument</tt> class</a>
3152</div>
3153
3154<div class="doc_text">
3155
3156<p>This subclass of Value defines the interface for incoming formal
Chris Lattner58360822005-01-17 00:12:04 +00003157arguments to a function. A Function maintains a list of its formal
Misha Brukman13fd15c2004-01-15 00:14:41 +00003158arguments. An argument has a pointer to the parent Function.</p>
3159
3160</div>
3161
Chris Lattner9355b472002-09-06 02:50:58 +00003162<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +00003163<hr>
3164<address>
3165 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
3166 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
3167 <a href="http://validator.w3.org/check/referer"><img
3168 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
3169
3170 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
3171 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +00003172 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +00003173 Last modified: $Date$
3174</address>
3175
Chris Lattner261efe92003-11-25 01:02:51 +00003176</body>
3177</html>