Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / llvm / 59fec6a53234df91d6b66d5826ef1f8ce60af2fa / . / docs / ProgrammersManual.html
blob: bce592e9529fa64b67f767a6037694d3586a038c [file] [log] [blame]
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3<html>
4<head>
Owen Anderson5e8c50e2009-06-16 17:40:28 +0000[diff] [blame]5 <meta http-equiv="Content-type" content="text/html;charset=UTF-8">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]6 <title>LLVM Programmer's Manual</title>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]7 <link rel="stylesheet" href="llvm.css" type="text/css">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]8</head>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]9<body>
10
11<div class="doc_title">
12 LLVM Programmer's Manual
13</div>
14
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]15<ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]16 <li><a href="#introduction">Introduction</a></li>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]17 <li><a href="#general">General Information</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]18 <ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]19 <li><a href="#stl">The C++ Standard Template Library</a></li>
20<!--
21 <li>The <tt>-time-passes</tt> option</li>
22 <li>How to use the LLVM Makefile system</li>
23 <li>How to write a regression test</li>
Chris Lattner61db4652004-12-08 19:05:44 +0000[diff] [blame]24
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]25-->
Chris Lattner84b7f8d2003-08-01 22:20:59 +0000[diff] [blame]26 </ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]27 </li>
28 <li><a href="#apis">Important and useful LLVM APIs</a>
29 <ul>
30 <li><a href="#isa">The <tt>isa&lt;&gt;</tt>, <tt>cast&lt;&gt;</tt>
31and <tt>dyn_cast&lt;&gt;</tt> templates</a> </li>
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]32 <li><a href="#DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]33option</a>
34 <ul>
35 <li><a href="#DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt>
36and the <tt>-debug-only</tt> option</a> </li>
37 </ul>
38 </li>
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]39 <li><a href="#Statistic">The <tt>Statistic</tt> class &amp; <tt>-stats</tt>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]40option</a></li>
41<!--
42 <li>The <tt>InstVisitor</tt> template
43 <li>The general graph API
44-->
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]45 <li><a href="#ViewGraph">Viewing graphs while debugging code</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]46 </ul>
47 </li>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]48 <li><a href="#datastructure">Picking the Right Data Structure for a Task</a>
49 <ul>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]50 <li><a href="#ds_sequential">Sequential Containers (std::vector, std::list, etc)</a>
51 <ul>
52 <li><a href="#dss_fixedarrays">Fixed Size Arrays</a></li>
53 <li><a href="#dss_heaparrays">Heap Allocated Arrays</a></li>
54 <li><a href="#dss_smallvector">"llvm/ADT/SmallVector.h"</a></li>
55 <li><a href="#dss_vector">&lt;vector&gt;</a></li>
56 <li><a href="#dss_deque">&lt;deque&gt;</a></li>
57 <li><a href="#dss_list">&lt;list&gt;</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]58 <li><a href="#dss_ilist">llvm/ADT/ilist.h</a></li>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]59 <li><a href="#dss_other">Other Sequential Container Options</a></li>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]60 </ul></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]61 <li><a href="#ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
62 <ul>
63 <li><a href="#dss_sortedvectorset">A sorted 'vector'</a></li>
64 <li><a href="#dss_smallset">"llvm/ADT/SmallSet.h"</a></li>
65 <li><a href="#dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a></li>
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]66 <li><a href="#dss_denseset">"llvm/ADT/DenseSet.h"</a></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]67 <li><a href="#dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a></li>
68 <li><a href="#dss_set">&lt;set&gt;</a></li>
69 <li><a href="#dss_setvector">"llvm/ADT/SetVector.h"</a></li>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]70 <li><a href="#dss_uniquevector">"llvm/ADT/UniqueVector.h"</a></li>
71 <li><a href="#dss_otherset">Other Set-Like ContainerOptions</a></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]72 </ul></li>
Chris Lattnerf3692522007-02-03 19:51:56 +0000[diff] [blame]73 <li><a href="#ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
74 <ul>
75 <li><a href="#dss_sortedvectormap">A sorted 'vector'</a></li>
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]76 <li><a href="#dss_stringmap">"llvm/ADT/StringMap.h"</a></li>
Chris Lattnerf3692522007-02-03 19:51:56 +0000[diff] [blame]77 <li><a href="#dss_indexedmap">"llvm/ADT/IndexedMap.h"</a></li>
78 <li><a href="#dss_densemap">"llvm/ADT/DenseMap.h"</a></li>
79 <li><a href="#dss_map">&lt;map&gt;</a></li>
80 <li><a href="#dss_othermap">Other Map-Like Container Options</a></li>
81 </ul></li>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]82 <li><a href="#ds_bit">BitVector-like containers</a>
83 <ul>
84 <li><a href="#dss_bitvector">A dense bitvector</a></li>
85 <li><a href="#dss_sparsebitvector">A sparse bitvector</a></li>
86 </ul></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]87 </ul>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]88 </li>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]89 <li><a href="#common">Helpful Hints for Common Operations</a>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]90 <ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]91 <li><a href="#inspection">Basic Inspection and Traversal Routines</a>
92 <ul>
93 <li><a href="#iterate_function">Iterating over the <tt>BasicBlock</tt>s
94in a <tt>Function</tt></a> </li>
95 <li><a href="#iterate_basicblock">Iterating over the <tt>Instruction</tt>s
96in a <tt>BasicBlock</tt></a> </li>
97 <li><a href="#iterate_institer">Iterating over the <tt>Instruction</tt>s
98in a <tt>Function</tt></a> </li>
99 <li><a href="#iterate_convert">Turning an iterator into a
100class pointer</a> </li>
101 <li><a href="#iterate_complex">Finding call sites: a more
102complex example</a> </li>
103 <li><a href="#calls_and_invokes">Treating calls and invokes
104the same way</a> </li>
105 <li><a href="#iterate_chains">Iterating over def-use &amp;
106use-def chains</a> </li>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]107 <li><a href="#iterate_preds">Iterating over predecessors &amp;
108successors of blocks</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]109 </ul>
110 </li>
111 <li><a href="#simplechanges">Making simple changes</a>
112 <ul>
113 <li><a href="#schanges_creating">Creating and inserting new
114 <tt>Instruction</tt>s</a> </li>
115 <li><a href="#schanges_deleting">Deleting <tt>Instruction</tt>s</a> </li>
116 <li><a href="#schanges_replacing">Replacing an <tt>Instruction</tt>
117with another <tt>Value</tt></a> </li>
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]118 <li><a href="#schanges_deletingGV">Deleting <tt>GlobalVariable</tt>s</a> </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]119 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]120 </li>
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]121 <li><a href="#create_types">How to Create Types</a></li>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]122<!--
123 <li>Working with the Control Flow Graph
124 <ul>
125 <li>Accessing predecessors and successors of a <tt>BasicBlock</tt>
126 <li>
127 <li>
128 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]129-->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]130 </ul>
131 </li>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]132
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]133 <li><a href="#threading">Threads and LLVM</a>
134 <ul>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]135 <li><a href="#startmultithreaded">Entering and Exiting Multithreaded Mode
136 </a></li>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]137 <li><a href="#shutdown">Ending execution with <tt>llvm_shutdown()</tt></a></li>
138 <li><a href="#managedstatic">Lazy initialization with <tt>ManagedStatic</tt></a></li>
139 </ul>
140 </li>
141
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]142 <li><a href="#advanced">Advanced Topics</a>
143 <ul>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]144 <li><a href="#TypeResolve">LLVM Type Resolution</a>
145 <ul>
146 <li><a href="#BuildRecType">Basic Recursive Type Construction</a></li>
147 <li><a href="#refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a></li>
148 <li><a href="#PATypeHolder">The PATypeHolder Class</a></li>
149 <li><a href="#AbstractTypeUser">The AbstractTypeUser Class</a></li>
150 </ul></li>
151
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]152 <li><a href="#SymbolTable">The <tt>ValueSymbolTable</tt> and <tt>TypeSymbolTable</tt> classes</a></li>
153 <li><a href="#UserLayout">The <tt>User</tt> and owned <tt>Use</tt> classes' memory layout</a></li>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]154 </ul></li>
155
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]156 <li><a href="#coreclasses">The Core LLVM Class Hierarchy Reference</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]157 <ul>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]158 <li><a href="#Type">The <tt>Type</tt> class</a> </li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]159 <li><a href="#Module">The <tt>Module</tt> class</a></li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]160 <li><a href="#Value">The <tt>Value</tt> class</a>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]161 <ul>
162 <li><a href="#User">The <tt>User</tt> class</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]163 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]164 <li><a href="#Instruction">The <tt>Instruction</tt> class</a></li>
165 <li><a href="#Constant">The <tt>Constant</tt> class</a>
166 <ul>
167 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]168 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]169 <li><a href="#Function">The <tt>Function</tt> class</a></li>
170 <li><a href="#GlobalVariable">The <tt>GlobalVariable</tt> class</a></li>
171 </ul>
172 </li>
173 </ul>
174 </li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]175 </ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]176 </li>
177 <li><a href="#BasicBlock">The <tt>BasicBlock</tt> class</a></li>
178 <li><a href="#Argument">The <tt>Argument</tt> class</a></li>
179 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]180 </li>
181 </ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]182 </li>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]183</ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]184
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]185<div class="doc_author">
186 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
Chris Lattner94c43592004-05-26 16:52:55 +0000[diff] [blame]187 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>,
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]188 <a href="mailto:ggreif@gmail.com">Gabor Greif</a>,
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]189 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a>,
190 <a href="mailto:rspencer@x10sys.com">Reid Spencer</a> and
191 <a href="mailto:owen@apple.com">Owen Anderson</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]192</div>
193
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]194<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]195<div class="doc_section">
196 <a name="introduction">Introduction </a>
197</div>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]198<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]199
200<div class="doc_text">
201
202<p>This document is meant to highlight some of the important classes and
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]203interfaces available in the LLVM source-base. This manual is not
204intended to explain what LLVM is, how it works, and what LLVM code looks
205like. It assumes that you know the basics of LLVM and are interested
206in writing transformations or otherwise analyzing or manipulating the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]207code.</p>
208
209<p>This document should get you oriented so that you can find your
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]210way in the continuously growing source code that makes up the LLVM
211infrastructure. Note that this manual is not intended to serve as a
212replacement for reading the source code, so if you think there should be
213a method in one of these classes to do something, but it's not listed,
214check the source. Links to the <a href="/doxygen/">doxygen</a> sources
215are provided to make this as easy as possible.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]216
217<p>The first section of this document describes general information that is
218useful to know when working in the LLVM infrastructure, and the second describes
219the Core LLVM classes. In the future this manual will be extended with
220information describing how to use extension libraries, such as dominator
221information, CFG traversal routines, and useful utilities like the <tt><a
222href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.</p>
223
224</div>
225
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]226<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]227<div class="doc_section">
228 <a name="general">General Information</a>
229</div>
230<!-- *********************************************************************** -->
231
232<div class="doc_text">
233
234<p>This section contains general information that is useful if you are working
235in the LLVM source-base, but that isn't specific to any particular API.</p>
236
237</div>
238
239<!-- ======================================================================= -->
240<div class="doc_subsection">
241 <a name="stl">The C++ Standard Template Library</a>
242</div>
243
244<div class="doc_text">
245
246<p>LLVM makes heavy use of the C++ Standard Template Library (STL),
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]247perhaps much more than you are used to, or have seen before. Because of
248this, you might want to do a little background reading in the
249techniques used and capabilities of the library. There are many good
250pages that discuss the STL, and several books on the subject that you
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]251can get, so it will not be discussed in this document.</p>
252
253<p>Here are some useful links:</p>
254
255<ol>
256
257<li><a href="http://www.dinkumware.com/refxcpp.html">Dinkumware C++ Library
258reference</a> - an excellent reference for the STL and other parts of the
259standard C++ library.</li>
260
261<li><a href="http://www.tempest-sw.com/cpp/">C++ In a Nutshell</a> - This is an
Gabor Greif0cbcabe2009-03-12 09:47:03 +0000[diff] [blame]262O'Reilly book in the making. It has a decent Standard Library
263Reference that rivals Dinkumware's, and is unfortunately no longer free since the
264book has been published.</li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]265
266<li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
267Questions</a></li>
268
269<li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
270Contains a useful <a
271href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
272STL</a>.</li>
273
274<li><a href="http://www.research.att.com/%7Ebs/C++.html">Bjarne Stroustrup's C++
275Page</a></li>
276
Tanya Lattner79445ba2004-12-08 18:34:56 +0000[diff] [blame]277<li><a href="http://64.78.49.204/">
Reid Spencer096603a2004-05-26 08:41:35 +0000[diff] [blame]278Bruce Eckel's Thinking in C++, 2nd ed. Volume 2 Revision 4.0 (even better, get
279the book).</a></li>
280
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]281</ol>
282
283<p>You are also encouraged to take a look at the <a
284href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
285to write maintainable code more than where to put your curly braces.</p>
286
287</div>
288
289<!-- ======================================================================= -->
290<div class="doc_subsection">
291 <a name="stl">Other useful references</a>
292</div>
293
294<div class="doc_text">
295
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]296<ol>
297<li><a href="http://www.psc.edu/%7Esemke/cvs_branches.html">CVS
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]298Branch and Tag Primer</a></li>
Misha Brukmana0f71e42004-06-18 18:39:00 +0000[diff] [blame]299<li><a href="http://www.fortran-2000.com/ArnaudRecipes/sharedlib.html">Using
300static and shared libraries across platforms</a></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]301</ol>
302
303</div>
304
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]305<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]306<div class="doc_section">
307 <a name="apis">Important and useful LLVM APIs</a>
308</div>
309<!-- *********************************************************************** -->
310
311<div class="doc_text">
312
313<p>Here we highlight some LLVM APIs that are generally useful and good to
314know about when writing transformations.</p>
315
316</div>
317
318<!-- ======================================================================= -->
319<div class="doc_subsection">
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]320 <a name="isa">The <tt>isa&lt;&gt;</tt>, <tt>cast&lt;&gt;</tt> and
321 <tt>dyn_cast&lt;&gt;</tt> templates</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]322</div>
323
324<div class="doc_text">
325
326<p>The LLVM source-base makes extensive use of a custom form of RTTI.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]327These templates have many similarities to the C++ <tt>dynamic_cast&lt;&gt;</tt>
328operator, but they don't have some drawbacks (primarily stemming from
329the fact that <tt>dynamic_cast&lt;&gt;</tt> only works on classes that
330have a v-table). Because they are used so often, you must know what they
331do and how they work. All of these templates are defined in the <a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]332 href="/doxygen/Casting_8h-source.html"><tt>llvm/Support/Casting.h</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]333file (note that you very rarely have to include this file directly).</p>
334
335<dl>
336 <dt><tt>isa&lt;&gt;</tt>: </dt>
337
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]338 <dd><p>The <tt>isa&lt;&gt;</tt> operator works exactly like the Java
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]339 "<tt>instanceof</tt>" operator. It returns true or false depending on whether
340 a reference or pointer points to an instance of the specified class. This can
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]341 be very useful for constraint checking of various sorts (example below).</p>
342 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]343
344 <dt><tt>cast&lt;&gt;</tt>: </dt>
345
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]346 <dd><p>The <tt>cast&lt;&gt;</tt> operator is a "checked cast" operation. It
Chris Lattner28e6ff52008-06-20 05:03:17 +0000[diff] [blame]347 converts a pointer or reference from a base class to a derived class, causing
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]348 an assertion failure if it is not really an instance of the right type. This
349 should be used in cases where you have some information that makes you believe
350 that something is of the right type. An example of the <tt>isa&lt;&gt;</tt>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]351 and <tt>cast&lt;&gt;</tt> template is:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]352
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]353<div class="doc_code">
354<pre>
355static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
356 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))
357 return true;
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]358
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]359 // <i>Otherwise, it must be an instruction...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]360 return !L-&gt;contains(cast&lt;<a href="#Instruction">Instruction</a>&gt;(V)-&gt;getParent());
361}
362</pre>
363</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]364
365 <p>Note that you should <b>not</b> use an <tt>isa&lt;&gt;</tt> test followed
366 by a <tt>cast&lt;&gt;</tt>, for that use the <tt>dyn_cast&lt;&gt;</tt>
367 operator.</p>
368
369 </dd>
370
371 <dt><tt>dyn_cast&lt;&gt;</tt>:</dt>
372
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]373 <dd><p>The <tt>dyn_cast&lt;&gt;</tt> operator is a "checking cast" operation.
374 It checks to see if the operand is of the specified type, and if so, returns a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]375 pointer to it (this operator does not work with references). If the operand is
376 not of the correct type, a null pointer is returned. Thus, this works very
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]377 much like the <tt>dynamic_cast&lt;&gt;</tt> operator in C++, and should be
378 used in the same circumstances. Typically, the <tt>dyn_cast&lt;&gt;</tt>
379 operator is used in an <tt>if</tt> statement or some other flow control
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]380 statement like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]381
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]382<div class="doc_code">
383<pre>
384if (<a href="#AllocationInst">AllocationInst</a> *AI = dyn_cast&lt;<a href="#AllocationInst">AllocationInst</a>&gt;(Val)) {
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]385 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]386}
387</pre>
388</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]389
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]390 <p>This form of the <tt>if</tt> statement effectively combines together a call
391 to <tt>isa&lt;&gt;</tt> and a call to <tt>cast&lt;&gt;</tt> into one
392 statement, which is very convenient.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]393
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]394 <p>Note that the <tt>dyn_cast&lt;&gt;</tt> operator, like C++'s
395 <tt>dynamic_cast&lt;&gt;</tt> or Java's <tt>instanceof</tt> operator, can be
396 abused. In particular, you should not use big chained <tt>if/then/else</tt>
397 blocks to check for lots of different variants of classes. If you find
398 yourself wanting to do this, it is much cleaner and more efficient to use the
399 <tt>InstVisitor</tt> class to dispatch over the instruction type directly.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]400
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]401 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]402
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]403 <dt><tt>cast_or_null&lt;&gt;</tt>: </dt>
404
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]405 <dd><p>The <tt>cast_or_null&lt;&gt;</tt> operator works just like the
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]406 <tt>cast&lt;&gt;</tt> operator, except that it allows for a null pointer as an
407 argument (which it then propagates). This can sometimes be useful, allowing
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]408 you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]409
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]410 <dt><tt>dyn_cast_or_null&lt;&gt;</tt>: </dt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]411
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]412 <dd><p>The <tt>dyn_cast_or_null&lt;&gt;</tt> operator works just like the
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]413 <tt>dyn_cast&lt;&gt;</tt> operator, except that it allows for a null pointer
414 as an argument (which it then propagates). This can sometimes be useful,
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]415 allowing you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]416
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]417</dl>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]418
419<p>These five templates can be used with any classes, whether they have a
420v-table or not. To add support for these templates, you simply need to add
421<tt>classof</tt> static methods to the class you are interested casting
422to. Describing this is currently outside the scope of this document, but there
423are lots of examples in the LLVM source base.</p>
424
425</div>
426
427<!-- ======================================================================= -->
428<div class="doc_subsection">
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]429 <a name="DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt> option</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]430</div>
431
432<div class="doc_text">
433
434<p>Often when working on your pass you will put a bunch of debugging printouts
435and other code into your pass. After you get it working, you want to remove
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]436it, but you may need it again in the future (to work out new bugs that you run
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]437across).</p>
438
439<p> Naturally, because of this, you don't want to delete the debug printouts,
440but you don't want them to always be noisy. A standard compromise is to comment
441them out, allowing you to enable them if you need them in the future.</p>
442
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]443<p>The "<tt><a href="/doxygen/Debug_8h-source.html">llvm/Support/Debug.h</a></tt>"
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]444file provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to
445this problem. Basically, you can put arbitrary code into the argument of the
446<tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
447tool) is run with the '<tt>-debug</tt>' command line argument:</p>
448
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]449<div class="doc_code">
450<pre>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]451DOUT &lt;&lt; "I am here!\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]452</pre>
453</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]454
455<p>Then you can run your pass like this:</p>
456
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]457<div class="doc_code">
458<pre>
459$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]460<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]461$ opt &lt; a.bc &gt; /dev/null -mypass -debug
462I am here!
463</pre>
464</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]465
466<p>Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you
467to not have to create "yet another" command line option for the debug output for
468your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
469so they do not cause a performance impact at all (for the same reason, they
470should also not contain side-effects!).</p>
471
472<p>One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
473enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
474"<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
475program hasn't been started yet, you can always just run it with
476<tt>-debug</tt>.</p>
477
478</div>
479
480<!-- _______________________________________________________________________ -->
481<div class="doc_subsubsection">
Chris Lattnerc9151082005-04-26 22:57:07 +0000[diff] [blame]482 <a name="DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt> and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]483 the <tt>-debug-only</tt> option</a>
484</div>
485
486<div class="doc_text">
487
488<p>Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
489just turns on <b>too much</b> information (such as when working on the code
490generator). If you want to enable debug information with more fine-grained
491control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
492option as follows:</p>
493
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]494<div class="doc_code">
495<pre>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]496DOUT &lt;&lt; "No debug type\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]497#undef DEBUG_TYPE
498#define DEBUG_TYPE "foo"
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]499DOUT &lt;&lt; "'foo' debug type\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]500#undef DEBUG_TYPE
501#define DEBUG_TYPE "bar"
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]502DOUT &lt;&lt; "'bar' debug type\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]503#undef DEBUG_TYPE
504#define DEBUG_TYPE ""
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]505DOUT &lt;&lt; "No debug type (2)\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]506</pre>
507</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]508
509<p>Then you can run your pass like this:</p>
510
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]511<div class="doc_code">
512<pre>
513$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]514<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]515$ opt &lt; a.bc &gt; /dev/null -mypass -debug
516No debug type
517'foo' debug type
518'bar' debug type
519No debug type (2)
520$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=foo
521'foo' debug type
522$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=bar
523'bar' debug type
524</pre>
525</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]526
527<p>Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of
528a file, to specify the debug type for the entire module (if you do this before
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]529you <tt>#include "llvm/Support/Debug.h"</tt>, you don't have to insert the ugly
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]530<tt>#undef</tt>'s). Also, you should use names more meaningful than "foo" and
531"bar", because there is no system in place to ensure that names do not
532conflict. If two different modules use the same string, they will all be turned
533on when the name is specified. This allows, for example, all debug information
534for instruction scheduling to be enabled with <tt>-debug-type=InstrSched</tt>,
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]535even if the source lives in multiple files.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]536
537</div>
538
539<!-- ======================================================================= -->
540<div class="doc_subsection">
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]541 <a name="Statistic">The <tt>Statistic</tt> class &amp; <tt>-stats</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]542 option</a>
543</div>
544
545<div class="doc_text">
546
547<p>The "<tt><a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]548href="/doxygen/Statistic_8h-source.html">llvm/ADT/Statistic.h</a></tt>" file
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]549provides a class named <tt>Statistic</tt> that is used as a unified way to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]550keep track of what the LLVM compiler is doing and how effective various
551optimizations are. It is useful to see what optimizations are contributing to
552making a particular program run faster.</p>
553
554<p>Often you may run your pass on some big program, and you're interested to see
555how many times it makes a certain transformation. Although you can do this with
556hand inspection, or some ad-hoc method, this is a real pain and not very useful
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]557for big programs. Using the <tt>Statistic</tt> class makes it very easy to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]558keep track of this information, and the calculated information is presented in a
559uniform manner with the rest of the passes being executed.</p>
560
561<p>There are many examples of <tt>Statistic</tt> uses, but the basics of using
562it are as follows:</p>
563
564<ol>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]565 <li><p>Define your statistic like this:</p>
566
567<div class="doc_code">
568<pre>
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]569#define <a href="#DEBUG_TYPE">DEBUG_TYPE</a> "mypassname" <i>// This goes before any #includes.</i>
570STATISTIC(NumXForms, "The # of times I did stuff");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]571</pre>
572</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]573
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]574 <p>The <tt>STATISTIC</tt> macro defines a static variable, whose name is
575 specified by the first argument. The pass name is taken from the DEBUG_TYPE
576 macro, and the description is taken from the second argument. The variable
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]577 defined ("NumXForms" in this case) acts like an unsigned integer.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]578
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]579 <li><p>Whenever you make a transformation, bump the counter:</p>
580
581<div class="doc_code">
582<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]583++NumXForms; // <i>I did stuff!</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]584</pre>
585</div>
586
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]587 </li>
588 </ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]589
590 <p>That's all you have to do. To get '<tt>opt</tt>' to print out the
591 statistics gathered, use the '<tt>-stats</tt>' option:</p>
592
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]593<div class="doc_code">
594<pre>
595$ opt -stats -mypassname &lt; program.bc &gt; /dev/null
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]596<i>... statistics output ...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]597</pre>
598</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]599
Reid Spencer6b6c73e2007-02-09 16:00:28 +0000[diff] [blame]600 <p> When running <tt>opt</tt> on a C file from the SPEC benchmark
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]601suite, it gives a report that looks like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]602
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]603<div class="doc_code">
604<pre>
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]605 7646 bitcodewriter - Number of normal instructions
606 725 bitcodewriter - Number of oversized instructions
607 129996 bitcodewriter - Number of bitcode bytes written
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]608 2817 raise - Number of insts DCEd or constprop'd
609 3213 raise - Number of cast-of-self removed
610 5046 raise - Number of expression trees converted
611 75 raise - Number of other getelementptr's formed
612 138 raise - Number of load/store peepholes
613 42 deadtypeelim - Number of unused typenames removed from symtab
614 392 funcresolve - Number of varargs functions resolved
615 27 globaldce - Number of global variables removed
616 2 adce - Number of basic blocks removed
617 134 cee - Number of branches revectored
618 49 cee - Number of setcc instruction eliminated
619 532 gcse - Number of loads removed
620 2919 gcse - Number of instructions removed
621 86 indvars - Number of canonical indvars added
622 87 indvars - Number of aux indvars removed
623 25 instcombine - Number of dead inst eliminate
624 434 instcombine - Number of insts combined
625 248 licm - Number of load insts hoisted
626 1298 licm - Number of insts hoisted to a loop pre-header
627 3 licm - Number of insts hoisted to multiple loop preds (bad, no loop pre-header)
628 75 mem2reg - Number of alloca's promoted
629 1444 cfgsimplify - Number of blocks simplified
630</pre>
631</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]632
633<p>Obviously, with so many optimizations, having a unified framework for this
634stuff is very nice. Making your pass fit well into the framework makes it more
635maintainable and useful.</p>
636
637</div>
638
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]639<!-- ======================================================================= -->
640<div class="doc_subsection">
641 <a name="ViewGraph">Viewing graphs while debugging code</a>
642</div>
643
644<div class="doc_text">
645
646<p>Several of the important data structures in LLVM are graphs: for example
647CFGs made out of LLVM <a href="#BasicBlock">BasicBlock</a>s, CFGs made out of
648LLVM <a href="CodeGenerator.html#machinebasicblock">MachineBasicBlock</a>s, and
649<a href="CodeGenerator.html#selectiondag_intro">Instruction Selection
650DAGs</a>. In many cases, while debugging various parts of the compiler, it is
651nice to instantly visualize these graphs.</p>
652
653<p>LLVM provides several callbacks that are available in a debug build to do
654exactly that. If you call the <tt>Function::viewCFG()</tt> method, for example,
655the current LLVM tool will pop up a window containing the CFG for the function
656where each basic block is a node in the graph, and each node contains the
657instructions in the block. Similarly, there also exists
658<tt>Function::viewCFGOnly()</tt> (does not include the instructions), the
659<tt>MachineFunction::viewCFG()</tt> and <tt>MachineFunction::viewCFGOnly()</tt>,
660and the <tt>SelectionDAG::viewGraph()</tt> methods. Within GDB, for example,
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]661you can usually use something like <tt>call DAG.viewGraph()</tt> to pop
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]662up a window. Alternatively, you can sprinkle calls to these functions in your
663code in places you want to debug.</p>
664
665<p>Getting this to work requires a small amount of configuration. On Unix
666systems with X11, install the <a href="http://www.graphviz.org">graphviz</a>
667toolkit, and make sure 'dot' and 'gv' are in your path. If you are running on
668Mac OS/X, download and install the Mac OS/X <a
669href="http://www.pixelglow.com/graphviz/">Graphviz program</a>, and add
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]670<tt>/Applications/Graphviz.app/Contents/MacOS/</tt> (or wherever you install
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]671it) to your path. Once in your system and path are set up, rerun the LLVM
672configure script and rebuild LLVM to enable this functionality.</p>
673
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]674<p><tt>SelectionDAG</tt> has been extended to make it easier to locate
675<i>interesting</i> nodes in large complex graphs. From gdb, if you
676<tt>call DAG.setGraphColor(<i>node</i>, "<i>color</i>")</tt>, then the
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]677next <tt>call DAG.viewGraph()</tt> would highlight the node in the
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]678specified color (choices of colors can be found at <a
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]679href="http://www.graphviz.org/doc/info/colors.html">colors</a>.) More
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]680complex node attributes can be provided with <tt>call
681DAG.setGraphAttrs(<i>node</i>, "<i>attributes</i>")</tt> (choices can be
682found at <a href="http://www.graphviz.org/doc/info/attrs.html">Graph
683Attributes</a>.) If you want to restart and clear all the current graph
684attributes, then you can <tt>call DAG.clearGraphAttrs()</tt>. </p>
685
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]686</div>
687
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]688<!-- *********************************************************************** -->
689<div class="doc_section">
690 <a name="datastructure">Picking the Right Data Structure for a Task</a>
691</div>
692<!-- *********************************************************************** -->
693
694<div class="doc_text">
695
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]696<p>LLVM has a plethora of data structures in the <tt>llvm/ADT/</tt> directory,
697 and we commonly use STL data structures. This section describes the trade-offs
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]698 you should consider when you pick one.</p>
699
700<p>
701The first step is a choose your own adventure: do you want a sequential
702container, a set-like container, or a map-like container? The most important
703thing when choosing a container is the algorithmic properties of how you plan to
704access the container. Based on that, you should use:</p>
705
706<ul>
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]707<li>a <a href="#ds_map">map-like</a> container if you need efficient look-up
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]708 of an value based on another value. Map-like containers also support
709 efficient queries for containment (whether a key is in the map). Map-like
710 containers generally do not support efficient reverse mapping (values to
711 keys). If you need that, use two maps. Some map-like containers also
712 support efficient iteration through the keys in sorted order. Map-like
713 containers are the most expensive sort, only use them if you need one of
714 these capabilities.</li>
715
716<li>a <a href="#ds_set">set-like</a> container if you need to put a bunch of
717 stuff into a container that automatically eliminates duplicates. Some
718 set-like containers support efficient iteration through the elements in
719 sorted order. Set-like containers are more expensive than sequential
720 containers.
721</li>
722
723<li>a <a href="#ds_sequential">sequential</a> container provides
724 the most efficient way to add elements and keeps track of the order they are
725 added to the collection. They permit duplicates and support efficient
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]726 iteration, but do not support efficient look-up based on a key.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]727</li>
728
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]729<li>a <a href="#ds_bit">bit</a> container provides an efficient way to store and
730 perform set operations on sets of numeric id's, while automatically
731 eliminating duplicates. Bit containers require a maximum of 1 bit for each
732 identifier you want to store.
733</li>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]734</ul>
735
736<p>
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]737Once the proper category of container is determined, you can fine tune the
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]738memory use, constant factors, and cache behaviors of access by intelligently
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]739picking a member of the category. Note that constant factors and cache behavior
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]740can be a big deal. If you have a vector that usually only contains a few
741elements (but could contain many), for example, it's much better to use
742<a href="#dss_smallvector">SmallVector</a> than <a href="#dss_vector">vector</a>
743. Doing so avoids (relatively) expensive malloc/free calls, which dwarf the
744cost of adding the elements to the container. </p>
745
746</div>
747
748<!-- ======================================================================= -->
749<div class="doc_subsection">
750 <a name="ds_sequential">Sequential Containers (std::vector, std::list, etc)</a>
751</div>
752
753<div class="doc_text">
754There are a variety of sequential containers available for you, based on your
755needs. Pick the first in this section that will do what you want.
756</div>
757
758<!-- _______________________________________________________________________ -->
759<div class="doc_subsubsection">
760 <a name="dss_fixedarrays">Fixed Size Arrays</a>
761</div>
762
763<div class="doc_text">
764<p>Fixed size arrays are very simple and very fast. They are good if you know
765exactly how many elements you have, or you have a (low) upper bound on how many
766you have.</p>
767</div>
768
769<!-- _______________________________________________________________________ -->
770<div class="doc_subsubsection">
771 <a name="dss_heaparrays">Heap Allocated Arrays</a>
772</div>
773
774<div class="doc_text">
775<p>Heap allocated arrays (new[] + delete[]) are also simple. They are good if
776the number of elements is variable, if you know how many elements you will need
777before the array is allocated, and if the array is usually large (if not,
778consider a <a href="#dss_smallvector">SmallVector</a>). The cost of a heap
779allocated array is the cost of the new/delete (aka malloc/free). Also note that
780if you are allocating an array of a type with a constructor, the constructor and
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]781destructors will be run for every element in the array (re-sizable vectors only
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]782construct those elements actually used).</p>
783</div>
784
785<!-- _______________________________________________________________________ -->
786<div class="doc_subsubsection">
787 <a name="dss_smallvector">"llvm/ADT/SmallVector.h"</a>
788</div>
789
790<div class="doc_text">
791<p><tt>SmallVector&lt;Type, N&gt;</tt> is a simple class that looks and smells
792just like <tt>vector&lt;Type&gt;</tt>:
793it supports efficient iteration, lays out elements in memory order (so you can
794do pointer arithmetic between elements), supports efficient push_back/pop_back
795operations, supports efficient random access to its elements, etc.</p>
796
797<p>The advantage of SmallVector is that it allocates space for
798some number of elements (N) <b>in the object itself</b>. Because of this, if
799the SmallVector is dynamically smaller than N, no malloc is performed. This can
800be a big win in cases where the malloc/free call is far more expensive than the
801code that fiddles around with the elements.</p>
802
803<p>This is good for vectors that are "usually small" (e.g. the number of
804predecessors/successors of a block is usually less than 8). On the other hand,
805this makes the size of the SmallVector itself large, so you don't want to
806allocate lots of them (doing so will waste a lot of space). As such,
807SmallVectors are most useful when on the stack.</p>
808
809<p>SmallVector also provides a nice portable and efficient replacement for
810<tt>alloca</tt>.</p>
811
812</div>
813
814<!-- _______________________________________________________________________ -->
815<div class="doc_subsubsection">
816 <a name="dss_vector">&lt;vector&gt;</a>
817</div>
818
819<div class="doc_text">
820<p>
821std::vector is well loved and respected. It is useful when SmallVector isn't:
822when the size of the vector is often large (thus the small optimization will
823rarely be a benefit) or if you will be allocating many instances of the vector
824itself (which would waste space for elements that aren't in the container).
825vector is also useful when interfacing with code that expects vectors :).
826</p>
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]827
828<p>One worthwhile note about std::vector: avoid code like this:</p>
829
830<div class="doc_code">
831<pre>
832for ( ... ) {
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]833 std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]834 use V;
835}
836</pre>
837</div>
838
839<p>Instead, write this as:</p>
840
841<div class="doc_code">
842<pre>
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]843std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]844for ( ... ) {
845 use V;
846 V.clear();
847}
848</pre>
849</div>
850
851<p>Doing so will save (at least) one heap allocation and free per iteration of
852the loop.</p>
853
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]854</div>
855
856<!-- _______________________________________________________________________ -->
857<div class="doc_subsubsection">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]858 <a name="dss_deque">&lt;deque&gt;</a>
859</div>
860
861<div class="doc_text">
862<p>std::deque is, in some senses, a generalized version of std::vector. Like
863std::vector, it provides constant time random access and other similar
864properties, but it also provides efficient access to the front of the list. It
865does not guarantee continuity of elements within memory.</p>
866
867<p>In exchange for this extra flexibility, std::deque has significantly higher
868constant factor costs than std::vector. If possible, use std::vector or
869something cheaper.</p>
870</div>
871
872<!-- _______________________________________________________________________ -->
873<div class="doc_subsubsection">
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]874 <a name="dss_list">&lt;list&gt;</a>
875</div>
876
877<div class="doc_text">
878<p>std::list is an extremely inefficient class that is rarely useful.
879It performs a heap allocation for every element inserted into it, thus having an
880extremely high constant factor, particularly for small data types. std::list
881also only supports bidirectional iteration, not random access iteration.</p>
882
883<p>In exchange for this high cost, std::list supports efficient access to both
884ends of the list (like std::deque, but unlike std::vector or SmallVector). In
885addition, the iterator invalidation characteristics of std::list are stronger
886than that of a vector class: inserting or removing an element into the list does
887not invalidate iterator or pointers to other elements in the list.</p>
888</div>
889
890<!-- _______________________________________________________________________ -->
891<div class="doc_subsubsection">
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]892 <a name="dss_ilist">llvm/ADT/ilist.h</a>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]893</div>
894
895<div class="doc_text">
896<p><tt>ilist&lt;T&gt;</tt> implements an 'intrusive' doubly-linked list. It is
897intrusive, because it requires the element to store and provide access to the
898prev/next pointers for the list.</p>
899
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]900<p><tt>ilist</tt> has the same drawbacks as <tt>std::list</tt>, and additionally
901requires an <tt>ilist_traits</tt> implementation for the element type, but it
902provides some novel characteristics. In particular, it can efficiently store
903polymorphic objects, the traits class is informed when an element is inserted or
Gabor Greif0cbcabe2009-03-12 09:47:03 +0000[diff] [blame]904removed from the list, and <tt>ilist</tt>s are guaranteed to support a
905constant-time splice operation.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]906
Gabor Greif0cbcabe2009-03-12 09:47:03 +0000[diff] [blame]907<p>These properties are exactly what we want for things like
908<tt>Instruction</tt>s and basic blocks, which is why these are implemented with
909<tt>ilist</tt>s.</p>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]910
911Related classes of interest are explained in the following subsections:
912 <ul>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]913 <li><a href="#dss_ilist_traits">ilist_traits</a></li>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]914 <li><a href="#dss_iplist">iplist</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]915 <li><a href="#dss_ilist_node">llvm/ADT/ilist_node.h</a></li>
Gabor Greif6a65f422009-03-12 10:30:31 +0000[diff] [blame]916 <li><a href="#dss_ilist_sentinel">Sentinels</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]917 </ul>
918</div>
919
920<!-- _______________________________________________________________________ -->
921<div class="doc_subsubsection">
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]922 <a name="dss_ilist_traits">ilist_traits</a>
923</div>
924
925<div class="doc_text">
926<p><tt>ilist_traits&lt;T&gt;</tt> is <tt>ilist&lt;T&gt;</tt>'s customization
927mechanism. <tt>iplist&lt;T&gt;</tt> (and consequently <tt>ilist&lt;T&gt;</tt>)
928publicly derive from this traits class.</p>
929</div>
930
931<!-- _______________________________________________________________________ -->
932<div class="doc_subsubsection">
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]933 <a name="dss_iplist">iplist</a>
934</div>
935
936<div class="doc_text">
937<p><tt>iplist&lt;T&gt;</tt> is <tt>ilist&lt;T&gt;</tt>'s base and as such
Gabor Greif0cbcabe2009-03-12 09:47:03 +0000[diff] [blame]938supports a slightly narrower interface. Notably, inserters from
939<tt>T&amp;</tt> are absent.</p>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]940
941<p><tt>ilist_traits&lt;T&gt;</tt> is a public base of this class and can be
942used for a wide variety of customizations.</p>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]943</div>
944
945<!-- _______________________________________________________________________ -->
946<div class="doc_subsubsection">
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]947 <a name="dss_ilist_node">llvm/ADT/ilist_node.h</a>
948</div>
949
950<div class="doc_text">
951<p><tt>ilist_node&lt;T&gt;</tt> implements a the forward and backward links
952that are expected by the <tt>ilist&lt;T&gt;</tt> (and analogous containers)
953in the default manner.</p>
954
955<p><tt>ilist_node&lt;T&gt;</tt>s are meant to be embedded in the node type
Gabor Greif0cbcabe2009-03-12 09:47:03 +0000[diff] [blame]956<tt>T</tt>, usually <tt>T</tt> publicly derives from
957<tt>ilist_node&lt;T&gt;</tt>.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]958</div>
959
960<!-- _______________________________________________________________________ -->
961<div class="doc_subsubsection">
Gabor Greif6a65f422009-03-12 10:30:31 +0000[diff] [blame]962 <a name="dss_ilist_sentinel">Sentinels</a>
963</div>
964
965<div class="doc_text">
966<p><tt>ilist</tt>s have another speciality that must be considered. To be a good
967citizen in the C++ ecosystem, it needs to support the standard container
968operations, such as <tt>begin</tt> and <tt>end</tt> iterators, etc. Also, the
969<tt>operator--</tt> must work correctly on the <tt>end</tt> iterator in the
970case of non-empty <tt>ilist</tt>s.</p>
971
972<p>The only sensible solution to this problem is to allocate a so-called
973<i>sentinel</i> along with the intrusive list, which serves as the <tt>end</tt>
974iterator, providing the back-link to the last element. However conforming to the
975C++ convention it is illegal to <tt>operator++</tt> beyond the sentinel and it
976also must not be dereferenced.</p>
977
978<p>These constraints allow for some implementation freedom to the <tt>ilist</tt>
979how to allocate and store the sentinel. The corresponding policy is dictated
980by <tt>ilist_traits&lt;T&gt;</tt>. By default a <tt>T</tt> gets heap-allocated
981whenever the need for a sentinel arises.</p>
982
983<p>While the default policy is sufficient in most cases, it may break down when
984<tt>T</tt> does not provide a default constructor. Also, in the case of many
985instances of <tt>ilist</tt>s, the memory overhead of the associated sentinels
986is wasted. To alleviate the situation with numerous and voluminous
987<tt>T</tt>-sentinels, sometimes a trick is employed, leading to <i>ghostly
988sentinels</i>.</p>
989
990<p>Ghostly sentinels are obtained by specially-crafted <tt>ilist_traits&lt;T&gt;</tt>
991which superpose the sentinel with the <tt>ilist</tt> instance in memory. Pointer
992arithmetic is used to obtain the sentinel, which is relative to the
993<tt>ilist</tt>'s <tt>this</tt> pointer. The <tt>ilist</tt> is augmented by an
994extra pointer, which serves as the back-link of the sentinel. This is the only
995field in the ghostly sentinel which can be legally accessed.</p>
996</div>
997
998<!-- _______________________________________________________________________ -->
999<div class="doc_subsubsection">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1000 <a name="dss_other">Other Sequential Container options</a>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1001</div>
1002
1003<div class="doc_text">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1004<p>Other STL containers are available, such as std::string.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1005
1006<p>There are also various STL adapter classes such as std::queue,
1007std::priority_queue, std::stack, etc. These provide simplified access to an
1008underlying container but don't affect the cost of the container itself.</p>
1009
1010</div>
1011
1012
1013<!-- ======================================================================= -->
1014<div class="doc_subsection">
1015 <a name="ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
1016</div>
1017
1018<div class="doc_text">
1019
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1020<p>Set-like containers are useful when you need to canonicalize multiple values
1021into a single representation. There are several different choices for how to do
1022this, providing various trade-offs.</p>
1023
1024</div>
1025
1026
1027<!-- _______________________________________________________________________ -->
1028<div class="doc_subsubsection">
1029 <a name="dss_sortedvectorset">A sorted 'vector'</a>
1030</div>
1031
1032<div class="doc_text">
1033
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]1034<p>If you intend to insert a lot of elements, then do a lot of queries, a
1035great approach is to use a vector (or other sequential container) with
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1036std::sort+std::unique to remove duplicates. This approach works really well if
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]1037your usage pattern has these two distinct phases (insert then query), and can be
1038coupled with a good choice of <a href="#ds_sequential">sequential container</a>.
1039</p>
1040
1041<p>
1042This combination provides the several nice properties: the result data is
1043contiguous in memory (good for cache locality), has few allocations, is easy to
1044address (iterators in the final vector are just indices or pointers), and can be
1045efficiently queried with a standard binary or radix search.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1046
1047</div>
1048
1049<!-- _______________________________________________________________________ -->
1050<div class="doc_subsubsection">
1051 <a name="dss_smallset">"llvm/ADT/SmallSet.h"</a>
1052</div>
1053
1054<div class="doc_text">
1055
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1056<p>If you have a set-like data structure that is usually small and whose elements
Chris Lattner4ddfac12007-02-03 07:59:51 +0000[diff] [blame]1057are reasonably small, a <tt>SmallSet&lt;Type, N&gt;</tt> is a good choice. This set
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1058has space for N elements in place (thus, if the set is dynamically smaller than
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1059N, no malloc traffic is required) and accesses them with a simple linear search.
1060When the set grows beyond 'N' elements, it allocates a more expensive representation that
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1061guarantees efficient access (for most types, it falls back to std::set, but for
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1062pointers it uses something far better, <a
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1063href="#dss_smallptrset">SmallPtrSet</a>).</p>
1064
1065<p>The magic of this class is that it handles small sets extremely efficiently,
1066but gracefully handles extremely large sets without loss of efficiency. The
1067drawback is that the interface is quite small: it supports insertion, queries
1068and erasing, but does not support iteration.</p>
1069
1070</div>
1071
1072<!-- _______________________________________________________________________ -->
1073<div class="doc_subsubsection">
1074 <a name="dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a>
1075</div>
1076
1077<div class="doc_text">
1078
1079<p>SmallPtrSet has all the advantages of SmallSet (and a SmallSet of pointers is
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1080transparently implemented with a SmallPtrSet), but also supports iterators. If
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1081more than 'N' insertions are performed, a single quadratically
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1082probed hash table is allocated and grows as needed, providing extremely
1083efficient access (constant time insertion/deleting/queries with low constant
1084factors) and is very stingy with malloc traffic.</p>
1085
1086<p>Note that, unlike std::set, the iterators of SmallPtrSet are invalidated
1087whenever an insertion occurs. Also, the values visited by the iterators are not
1088visited in sorted order.</p>
1089
1090</div>
1091
1092<!-- _______________________________________________________________________ -->
1093<div class="doc_subsubsection">
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]1094 <a name="dss_denseset">"llvm/ADT/DenseSet.h"</a>
1095</div>
1096
1097<div class="doc_text">
1098
1099<p>
1100DenseSet is a simple quadratically probed hash table. It excels at supporting
1101small values: it uses a single allocation to hold all of the pairs that
1102are currently inserted in the set. DenseSet is a great way to unique small
1103values that are not simple pointers (use <a
1104href="#dss_smallptrset">SmallPtrSet</a> for pointers). Note that DenseSet has
1105the same requirements for the value type that <a
1106href="#dss_densemap">DenseMap</a> has.
1107</p>
1108
1109</div>
1110
1111<!-- _______________________________________________________________________ -->
1112<div class="doc_subsubsection">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1113 <a name="dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a>
1114</div>
1115
1116<div class="doc_text">
1117
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1118<p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1119FoldingSet is an aggregate class that is really good at uniquing
1120expensive-to-create or polymorphic objects. It is a combination of a chained
1121hash table with intrusive links (uniqued objects are required to inherit from
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1122FoldingSetNode) that uses <a href="#dss_smallvector">SmallVector</a> as part of
1123its ID process.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1124
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1125<p>Consider a case where you want to implement a "getOrCreateFoo" method for
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1126a complex object (for example, a node in the code generator). The client has a
1127description of *what* it wants to generate (it knows the opcode and all the
1128operands), but we don't want to 'new' a node, then try inserting it into a set
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1129only to find out it already exists, at which point we would have to delete it
1130and return the node that already exists.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1131</p>
1132
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1133<p>To support this style of client, FoldingSet perform a query with a
1134FoldingSetNodeID (which wraps SmallVector) that can be used to describe the
1135element that we want to query for. The query either returns the element
1136matching the ID or it returns an opaque ID that indicates where insertion should
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1137take place. Construction of the ID usually does not require heap traffic.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1138
1139<p>Because FoldingSet uses intrusive links, it can support polymorphic objects
1140in the set (for example, you can have SDNode instances mixed with LoadSDNodes).
1141Because the elements are individually allocated, pointers to the elements are
1142stable: inserting or removing elements does not invalidate any pointers to other
1143elements.
1144</p>
1145
1146</div>
1147
1148<!-- _______________________________________________________________________ -->
1149<div class="doc_subsubsection">
1150 <a name="dss_set">&lt;set&gt;</a>
1151</div>
1152
1153<div class="doc_text">
1154
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1155<p><tt>std::set</tt> is a reasonable all-around set class, which is decent at
1156many things but great at nothing. std::set allocates memory for each element
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1157inserted (thus it is very malloc intensive) and typically stores three pointers
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1158per element in the set (thus adding a large amount of per-element space
1159overhead). It offers guaranteed log(n) performance, which is not particularly
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1160fast from a complexity standpoint (particularly if the elements of the set are
1161expensive to compare, like strings), and has extremely high constant factors for
1162lookup, insertion and removal.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1163
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1164<p>The advantages of std::set are that its iterators are stable (deleting or
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1165inserting an element from the set does not affect iterators or pointers to other
1166elements) and that iteration over the set is guaranteed to be in sorted order.
1167If the elements in the set are large, then the relative overhead of the pointers
1168and malloc traffic is not a big deal, but if the elements of the set are small,
1169std::set is almost never a good choice.</p>
1170
1171</div>
1172
1173<!-- _______________________________________________________________________ -->
1174<div class="doc_subsubsection">
1175 <a name="dss_setvector">"llvm/ADT/SetVector.h"</a>
1176</div>
1177
1178<div class="doc_text">
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1179<p>LLVM's SetVector&lt;Type&gt; is an adapter class that combines your choice of
1180a set-like container along with a <a href="#ds_sequential">Sequential
1181Container</a>. The important property
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1182that this provides is efficient insertion with uniquing (duplicate elements are
1183ignored) with iteration support. It implements this by inserting elements into
1184both a set-like container and the sequential container, using the set-like
1185container for uniquing and the sequential container for iteration.
1186</p>
1187
1188<p>The difference between SetVector and other sets is that the order of
1189iteration is guaranteed to match the order of insertion into the SetVector.
1190This property is really important for things like sets of pointers. Because
1191pointer values are non-deterministic (e.g. vary across runs of the program on
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1192different machines), iterating over the pointers in the set will
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1193not be in a well-defined order.</p>
1194
1195<p>
1196The drawback of SetVector is that it requires twice as much space as a normal
1197set and has the sum of constant factors from the set-like container and the
1198sequential container that it uses. Use it *only* if you need to iterate over
1199the elements in a deterministic order. SetVector is also expensive to delete
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1200elements out of (linear time), unless you use it's "pop_back" method, which is
1201faster.
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1202</p>
1203
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1204<p>SetVector is an adapter class that defaults to using std::vector and std::set
1205for the underlying containers, so it is quite expensive. However,
1206<tt>"llvm/ADT/SetVector.h"</tt> also provides a SmallSetVector class, which
1207defaults to using a SmallVector and SmallSet of a specified size. If you use
1208this, and if your sets are dynamically smaller than N, you will save a lot of
1209heap traffic.</p>
1210
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1211</div>
1212
1213<!-- _______________________________________________________________________ -->
1214<div class="doc_subsubsection">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1215 <a name="dss_uniquevector">"llvm/ADT/UniqueVector.h"</a>
1216</div>
1217
1218<div class="doc_text">
1219
1220<p>
1221UniqueVector is similar to <a href="#dss_setvector">SetVector</a>, but it
1222retains a unique ID for each element inserted into the set. It internally
1223contains a map and a vector, and it assigns a unique ID for each value inserted
1224into the set.</p>
1225
1226<p>UniqueVector is very expensive: its cost is the sum of the cost of
1227maintaining both the map and vector, it has high complexity, high constant
1228factors, and produces a lot of malloc traffic. It should be avoided.</p>
1229
1230</div>
1231
1232
1233<!-- _______________________________________________________________________ -->
1234<div class="doc_subsubsection">
1235 <a name="dss_otherset">Other Set-Like Container Options</a>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1236</div>
1237
1238<div class="doc_text">
1239
1240<p>
1241The STL provides several other options, such as std::multiset and the various
Chris Lattnerf1a30822009-03-09 05:20:45 +0000[diff] [blame]1242"hash_set" like containers (whether from C++ TR1 or from the SGI library). We
1243never use hash_set and unordered_set because they are generally very expensive
1244(each insertion requires a malloc) and very non-portable.
1245</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1246
1247<p>std::multiset is useful if you're not interested in elimination of
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1248duplicates, but has all the drawbacks of std::set. A sorted vector (where you
1249don't delete duplicate entries) or some other approach is almost always
1250better.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1251
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1252</div>
1253
1254<!-- ======================================================================= -->
1255<div class="doc_subsection">
1256 <a name="ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
1257</div>
1258
1259<div class="doc_text">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1260Map-like containers are useful when you want to associate data to a key. As
1261usual, there are a lot of different ways to do this. :)
1262</div>
1263
1264<!-- _______________________________________________________________________ -->
1265<div class="doc_subsubsection">
1266 <a name="dss_sortedvectormap">A sorted 'vector'</a>
1267</div>
1268
1269<div class="doc_text">
1270
1271<p>
1272If your usage pattern follows a strict insert-then-query approach, you can
1273trivially use the same approach as <a href="#dss_sortedvectorset">sorted vectors
1274for set-like containers</a>. The only difference is that your query function
1275(which uses std::lower_bound to get efficient log(n) lookup) should only compare
1276the key, not both the key and value. This yields the same advantages as sorted
1277vectors for sets.
1278</p>
1279</div>
1280
1281<!-- _______________________________________________________________________ -->
1282<div class="doc_subsubsection">
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1283 <a name="dss_stringmap">"llvm/ADT/StringMap.h"</a>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1284</div>
1285
1286<div class="doc_text">
1287
1288<p>
1289Strings are commonly used as keys in maps, and they are difficult to support
1290efficiently: they are variable length, inefficient to hash and compare when
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1291long, expensive to copy, etc. StringMap is a specialized container designed to
1292cope with these issues. It supports mapping an arbitrary range of bytes to an
1293arbitrary other object.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1294
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1295<p>The StringMap implementation uses a quadratically-probed hash table, where
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1296the buckets store a pointer to the heap allocated entries (and some other
1297stuff). The entries in the map must be heap allocated because the strings are
1298variable length. The string data (key) and the element object (value) are
1299stored in the same allocation with the string data immediately after the element
1300object. This container guarantees the "<tt>(char*)(&amp;Value+1)</tt>" points
1301to the key string for a value.</p>
1302
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1303<p>The StringMap is very fast for several reasons: quadratic probing is very
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1304cache efficient for lookups, the hash value of strings in buckets is not
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1305recomputed when lookup up an element, StringMap rarely has to touch the
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1306memory for unrelated objects when looking up a value (even when hash collisions
1307happen), hash table growth does not recompute the hash values for strings
1308already in the table, and each pair in the map is store in a single allocation
1309(the string data is stored in the same allocation as the Value of a pair).</p>
1310
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1311<p>StringMap also provides query methods that take byte ranges, so it only ever
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1312copies a string if a value is inserted into the table.</p>
1313</div>
1314
1315<!-- _______________________________________________________________________ -->
1316<div class="doc_subsubsection">
1317 <a name="dss_indexedmap">"llvm/ADT/IndexedMap.h"</a>
1318</div>
1319
1320<div class="doc_text">
1321<p>
1322IndexedMap is a specialized container for mapping small dense integers (or
1323values that can be mapped to small dense integers) to some other type. It is
1324internally implemented as a vector with a mapping function that maps the keys to
1325the dense integer range.
1326</p>
1327
1328<p>
1329This is useful for cases like virtual registers in the LLVM code generator: they
1330have a dense mapping that is offset by a compile-time constant (the first
1331virtual register ID).</p>
1332
1333</div>
1334
1335<!-- _______________________________________________________________________ -->
1336<div class="doc_subsubsection">
1337 <a name="dss_densemap">"llvm/ADT/DenseMap.h"</a>
1338</div>
1339
1340<div class="doc_text">
1341
1342<p>
1343DenseMap is a simple quadratically probed hash table. It excels at supporting
1344small keys and values: it uses a single allocation to hold all of the pairs that
1345are currently inserted in the map. DenseMap is a great way to map pointers to
1346pointers, or map other small types to each other.
1347</p>
1348
1349<p>
1350There are several aspects of DenseMap that you should be aware of, however. The
1351iterators in a densemap are invalidated whenever an insertion occurs, unlike
1352map. Also, because DenseMap allocates space for a large number of key/value
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1353pairs (it starts with 64 by default), it will waste a lot of space if your keys
1354or values are large. Finally, you must implement a partial specialization of
Chris Lattner76c1b972007-09-17 18:34:04 +0000[diff] [blame]1355DenseMapInfo for the key that you want, if it isn't already supported. This
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1356is required to tell DenseMap about two special marker values (which can never be
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1357inserted into the map) that it needs internally.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1358
1359</div>
1360
1361<!-- _______________________________________________________________________ -->
1362<div class="doc_subsubsection">
1363 <a name="dss_map">&lt;map&gt;</a>
1364</div>
1365
1366<div class="doc_text">
1367
1368<p>
1369std::map has similar characteristics to <a href="#dss_set">std::set</a>: it uses
1370a single allocation per pair inserted into the map, it offers log(n) lookup with
1371an extremely large constant factor, imposes a space penalty of 3 pointers per
1372pair in the map, etc.</p>
1373
1374<p>std::map is most useful when your keys or values are very large, if you need
1375to iterate over the collection in sorted order, or if you need stable iterators
1376into the map (i.e. they don't get invalidated if an insertion or deletion of
1377another element takes place).</p>
1378
1379</div>
1380
1381<!-- _______________________________________________________________________ -->
1382<div class="doc_subsubsection">
1383 <a name="dss_othermap">Other Map-Like Container Options</a>
1384</div>
1385
1386<div class="doc_text">
1387
1388<p>
1389The STL provides several other options, such as std::multimap and the various
Chris Lattnerf1a30822009-03-09 05:20:45 +0000[diff] [blame]1390"hash_map" like containers (whether from C++ TR1 or from the SGI library). We
1391never use hash_set and unordered_set because they are generally very expensive
1392(each insertion requires a malloc) and very non-portable.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1393
1394<p>std::multimap is useful if you want to map a key to multiple values, but has
1395all the drawbacks of std::map. A sorted vector or some other approach is almost
1396always better.</p>
1397
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1398</div>
1399
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1400<!-- ======================================================================= -->
1401<div class="doc_subsection">
1402 <a name="ds_bit">Bit storage containers (BitVector, SparseBitVector)</a>
1403</div>
1404
1405<div class="doc_text">
Chris Lattner7086ce72007-09-25 22:37:50 +0000[diff] [blame]1406<p>Unlike the other containers, there are only two bit storage containers, and
1407choosing when to use each is relatively straightforward.</p>
1408
1409<p>One additional option is
1410<tt>std::vector&lt;bool&gt;</tt>: we discourage its use for two reasons 1) the
1411implementation in many common compilers (e.g. commonly available versions of
1412GCC) is extremely inefficient and 2) the C++ standards committee is likely to
1413deprecate this container and/or change it significantly somehow. In any case,
1414please don't use it.</p>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1415</div>
1416
1417<!-- _______________________________________________________________________ -->
1418<div class="doc_subsubsection">
1419 <a name="dss_bitvector">BitVector</a>
1420</div>
1421
1422<div class="doc_text">
1423<p> The BitVector container provides a fixed size set of bits for manipulation.
1424It supports individual bit setting/testing, as well as set operations. The set
1425operations take time O(size of bitvector), but operations are performed one word
1426at a time, instead of one bit at a time. This makes the BitVector very fast for
1427set operations compared to other containers. Use the BitVector when you expect
1428the number of set bits to be high (IE a dense set).
1429</p>
1430</div>
1431
1432<!-- _______________________________________________________________________ -->
1433<div class="doc_subsubsection">
1434 <a name="dss_sparsebitvector">SparseBitVector</a>
1435</div>
1436
1437<div class="doc_text">
1438<p> The SparseBitVector container is much like BitVector, with one major
1439difference: Only the bits that are set, are stored. This makes the
1440SparseBitVector much more space efficient than BitVector when the set is sparse,
1441as well as making set operations O(number of set bits) instead of O(size of
1442universe). The downside to the SparseBitVector is that setting and testing of random bits is O(N), and on large SparseBitVectors, this can be slower than BitVector. In our implementation, setting or testing bits in sorted order
1443(either forwards or reverse) is O(1) worst case. Testing and setting bits within 128 bits (depends on size) of the current bit is also O(1). As a general statement, testing/setting bits in a SparseBitVector is O(distance away from last set bit).
1444</p>
1445</div>
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]1446
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1447<!-- *********************************************************************** -->
1448<div class="doc_section">
1449 <a name="common">Helpful Hints for Common Operations</a>
1450</div>
1451<!-- *********************************************************************** -->
1452
1453<div class="doc_text">
1454
1455<p>This section describes how to perform some very simple transformations of
1456LLVM code. This is meant to give examples of common idioms used, showing the
1457practical side of LLVM transformations. <p> Because this is a "how-to" section,
1458you should also read about the main classes that you will be working with. The
1459<a href="#coreclasses">Core LLVM Class Hierarchy Reference</a> contains details
1460and descriptions of the main classes that you should know about.</p>
1461
1462</div>
1463
1464<!-- NOTE: this section should be heavy on example code -->
1465<!-- ======================================================================= -->
1466<div class="doc_subsection">
1467 <a name="inspection">Basic Inspection and Traversal Routines</a>
1468</div>
1469
1470<div class="doc_text">
1471
1472<p>The LLVM compiler infrastructure have many different data structures that may
1473be traversed. Following the example of the C++ standard template library, the
1474techniques used to traverse these various data structures are all basically the
1475same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
1476method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
1477function returns an iterator pointing to one past the last valid element of the
1478sequence, and there is some <tt>XXXiterator</tt> data type that is common
1479between the two operations.</p>
1480
1481<p>Because the pattern for iteration is common across many different aspects of
1482the program representation, the standard template library algorithms may be used
1483on them, and it is easier to remember how to iterate. First we show a few common
1484examples of the data structures that need to be traversed. Other data
1485structures are traversed in very similar ways.</p>
1486
1487</div>
1488
1489<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1490<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1491 <a name="iterate_function">Iterating over the </a><a
1492 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
1493 href="#Function"><tt>Function</tt></a>
1494</div>
1495
1496<div class="doc_text">
1497
1498<p>It's quite common to have a <tt>Function</tt> instance that you'd like to
1499transform in some way; in particular, you'd like to manipulate its
1500<tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over all of
1501the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>. The following is
1502an example that prints the name of a <tt>BasicBlock</tt> and the number of
1503<tt>Instruction</tt>s it contains:</p>
1504
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1505<div class="doc_code">
1506<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1507// <i>func is a pointer to a Function instance</i>
1508for (Function::iterator i = func-&gt;begin(), e = func-&gt;end(); i != e; ++i)
1509 // <i>Print out the name of the basic block if it has one, and then the</i>
1510 // <i>number of instructions that it contains</i>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1511 llvm::cerr &lt;&lt; "Basic block (name=" &lt;&lt; i-&gt;getName() &lt;&lt; ") has "
1512 &lt;&lt; i-&gt;size() &lt;&lt; " instructions.\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1513</pre>
1514</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1515
1516<p>Note that i can be used as if it were a pointer for the purposes of
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]1517invoking member functions of the <tt>Instruction</tt> class. This is
1518because the indirection operator is overloaded for the iterator
Chris Lattner7496ec52003-08-05 22:54:23 +0000[diff] [blame]1519classes. In the above code, the expression <tt>i-&gt;size()</tt> is
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1520exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.</p>
1521
1522</div>
1523
1524<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1525<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1526 <a name="iterate_basicblock">Iterating over the </a><a
1527 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1528 href="#BasicBlock"><tt>BasicBlock</tt></a>
1529</div>
1530
1531<div class="doc_text">
1532
1533<p>Just like when dealing with <tt>BasicBlock</tt>s in <tt>Function</tt>s, it's
1534easy to iterate over the individual instructions that make up
1535<tt>BasicBlock</tt>s. Here's a code snippet that prints out each instruction in
1536a <tt>BasicBlock</tt>:</p>
1537
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1538<div class="doc_code">
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1539<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1540// <i>blk is a pointer to a BasicBlock instance</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1541for (BasicBlock::iterator i = blk-&gt;begin(), e = blk-&gt;end(); i != e; ++i)
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1542 // <i>The next statement works since operator&lt;&lt;(ostream&amp;,...)</i>
1543 // <i>is overloaded for Instruction&amp;</i>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1544 llvm::cerr &lt;&lt; *i &lt;&lt; "\n";
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1545</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1546</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1547
1548<p>However, this isn't really the best way to print out the contents of a
1549<tt>BasicBlock</tt>! Since the ostream operators are overloaded for virtually
1550anything you'll care about, you could have just invoked the print routine on the
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1551basic block itself: <tt>llvm::cerr &lt;&lt; *blk &lt;&lt; "\n";</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1552
1553</div>
1554
1555<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1556<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1557 <a name="iterate_institer">Iterating over the </a><a
1558 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1559 href="#Function"><tt>Function</tt></a>
1560</div>
1561
1562<div class="doc_text">
1563
1564<p>If you're finding that you commonly iterate over a <tt>Function</tt>'s
1565<tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s <tt>Instruction</tt>s,
1566<tt>InstIterator</tt> should be used instead. You'll need to include <a
1567href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>,
1568and then instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1569small example that shows how to dump all instructions in a function to the standard error stream:<p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1570
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1571<div class="doc_code">
1572<pre>
1573#include "<a href="/doxygen/InstIterator_8h-source.html">llvm/Support/InstIterator.h</a>"
1574
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1575// <i>F is a pointer to a Function instance</i>
Chris Lattnerda021aa2008-06-04 18:20:42 +0000[diff] [blame]1576for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
1577 llvm::cerr &lt;&lt; *I &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1578</pre>
1579</div>
1580
1581<p>Easy, isn't it? You can also use <tt>InstIterator</tt>s to fill a
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1582work list with its initial contents. For example, if you wanted to
1583initialize a work list to contain all instructions in a <tt>Function</tt>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1584F, all you would need to do is something like:</p>
1585
1586<div class="doc_code">
1587<pre>
1588std::set&lt;Instruction*&gt; worklist;
Chris Lattnerda021aa2008-06-04 18:20:42 +0000[diff] [blame]1589// or better yet, SmallPtrSet&lt;Instruction*, 64&gt; worklist;
1590
1591for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
1592 worklist.insert(&amp;*I);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1593</pre>
1594</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1595
1596<p>The STL set <tt>worklist</tt> would now contain all instructions in the
1597<tt>Function</tt> pointed to by F.</p>
1598
1599</div>
1600
1601<!-- _______________________________________________________________________ -->
1602<div class="doc_subsubsection">
1603 <a name="iterate_convert">Turning an iterator into a class pointer (and
1604 vice-versa)</a>
1605</div>
1606
1607<div class="doc_text">
1608
1609<p>Sometimes, it'll be useful to grab a reference (or pointer) to a class
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]1610instance when all you've got at hand is an iterator. Well, extracting
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1611a reference or a pointer from an iterator is very straight-forward.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1612Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and <tt>j</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1613is a <tt>BasicBlock::const_iterator</tt>:</p>
1614
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1615<div class="doc_code">
1616<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1617Instruction&amp; inst = *i; // <i>Grab reference to instruction reference</i>
1618Instruction* pinst = &amp;*i; // <i>Grab pointer to instruction reference</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1619const Instruction&amp; inst = *j;
1620</pre>
1621</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1622
1623<p>However, the iterators you'll be working with in the LLVM framework are
1624special: they will automatically convert to a ptr-to-instance type whenever they
1625need to. Instead of dereferencing the iterator and then taking the address of
1626the result, you can simply assign the iterator to the proper pointer type and
1627you get the dereference and address-of operation as a result of the assignment
1628(behind the scenes, this is a result of overloading casting mechanisms). Thus
1629the last line of the last example,</p>
1630
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1631<div class="doc_code">
1632<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1633Instruction *pinst = &amp;*i;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1634</pre>
1635</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1636
1637<p>is semantically equivalent to</p>
1638
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1639<div class="doc_code">
1640<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1641Instruction *pinst = i;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1642</pre>
1643</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1644
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1645<p>It's also possible to turn a class pointer into the corresponding iterator,
1646and this is a constant time operation (very efficient). The following code
1647snippet illustrates use of the conversion constructors provided by LLVM
1648iterators. By using these, you can explicitly grab the iterator of something
1649without actually obtaining it via iteration over some structure:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1650
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1651<div class="doc_code">
1652<pre>
1653void printNextInstruction(Instruction* inst) {
1654 BasicBlock::iterator it(inst);
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1655 ++it; // <i>After this line, it refers to the instruction after *inst</i>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1656 if (it != inst-&gt;getParent()-&gt;end()) llvm::cerr &lt;&lt; *it &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1657}
1658</pre>
1659</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1660
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1661</div>
1662
1663<!--_______________________________________________________________________-->
1664<div class="doc_subsubsection">
1665 <a name="iterate_complex">Finding call sites: a slightly more complex
1666 example</a>
1667</div>
1668
1669<div class="doc_text">
1670
1671<p>Say that you're writing a FunctionPass and would like to count all the
1672locations in the entire module (that is, across every <tt>Function</tt>) where a
1673certain function (i.e., some <tt>Function</tt>*) is already in scope. As you'll
1674learn later, you may want to use an <tt>InstVisitor</tt> to accomplish this in a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1675much more straight-forward manner, but this example will allow us to explore how
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1676you'd do it if you didn't have <tt>InstVisitor</tt> around. In pseudo-code, this
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1677is what we want to do:</p>
1678
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1679<div class="doc_code">
1680<pre>
1681initialize callCounter to zero
1682for each Function f in the Module
1683 for each BasicBlock b in f
1684 for each Instruction i in b
1685 if (i is a CallInst and calls the given function)
1686 increment callCounter
1687</pre>
1688</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1689
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1690<p>And the actual code is (remember, because we're writing a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1691<tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply has to
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1692override the <tt>runOnFunction</tt> method):</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1693
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1694<div class="doc_code">
1695<pre>
1696Function* targetFunc = ...;
1697
1698class OurFunctionPass : public FunctionPass {
1699 public:
1700 OurFunctionPass(): callCounter(0) { }
1701
1702 virtual runOnFunction(Function&amp; F) {
1703 for (Function::iterator b = F.begin(), be = F.end(); b != be; ++b) {
Eric Christopher203e71d2008-11-08 08:20:49 +0000[diff] [blame]1704 for (BasicBlock::iterator i = b-&gt;begin(), ie = b-&gt;end(); i != ie; ++i) {
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1705 if (<a href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a>&lt;<a
1706 href="#CallInst">CallInst</a>&gt;(&amp;*i)) {
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1707 // <i>We know we've encountered a call instruction, so we</i>
1708 // <i>need to determine if it's a call to the</i>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1709 // <i>function pointed to by m_func or not.</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1710 if (callInst-&gt;getCalledFunction() == targetFunc)
1711 ++callCounter;
1712 }
1713 }
1714 }
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1715 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1716
1717 private:
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1718 unsigned callCounter;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1719};
1720</pre>
1721</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1722
1723</div>
1724
Brian Gaekef1972c62003-11-07 19:25:45 +0000[diff] [blame]1725<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1726<div class="doc_subsubsection">
1727 <a name="calls_and_invokes">Treating calls and invokes the same way</a>
1728</div>
1729
1730<div class="doc_text">
1731
1732<p>You may have noticed that the previous example was a bit oversimplified in
1733that it did not deal with call sites generated by 'invoke' instructions. In
1734this, and in other situations, you may find that you want to treat
1735<tt>CallInst</tt>s and <tt>InvokeInst</tt>s the same way, even though their
1736most-specific common base class is <tt>Instruction</tt>, which includes lots of
1737less closely-related things. For these cases, LLVM provides a handy wrapper
1738class called <a
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]1739href="http://llvm.org/doxygen/classllvm_1_1CallSite.html"><tt>CallSite</tt></a>.
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1740It is essentially a wrapper around an <tt>Instruction</tt> pointer, with some
1741methods that provide functionality common to <tt>CallInst</tt>s and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1742<tt>InvokeInst</tt>s.</p>
1743
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1744<p>This class has "value semantics": it should be passed by value, not by
1745reference and it should not be dynamically allocated or deallocated using
1746<tt>operator new</tt> or <tt>operator delete</tt>. It is efficiently copyable,
1747assignable and constructable, with costs equivalents to that of a bare pointer.
1748If you look at its definition, it has only a single pointer member.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1749
1750</div>
1751
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]1752<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1753<div class="doc_subsubsection">
1754 <a name="iterate_chains">Iterating over def-use &amp; use-def chains</a>
1755</div>
1756
1757<div class="doc_text">
1758
1759<p>Frequently, we might have an instance of the <a
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]1760href="/doxygen/classllvm_1_1Value.html">Value Class</a> and we want to
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]1761determine which <tt>User</tt>s use the <tt>Value</tt>. The list of all
1762<tt>User</tt>s of a particular <tt>Value</tt> is called a <i>def-use</i> chain.
1763For example, let's say we have a <tt>Function*</tt> named <tt>F</tt> to a
1764particular function <tt>foo</tt>. Finding all of the instructions that
1765<i>use</i> <tt>foo</tt> is as simple as iterating over the <i>def-use</i> chain
1766of <tt>F</tt>:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1767
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1768<div class="doc_code">
1769<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1770Function *F = ...;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1771
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1772for (Value::use_iterator i = F-&gt;use_begin(), e = F-&gt;use_end(); i != e; ++i)
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1773 if (Instruction *Inst = dyn_cast&lt;Instruction&gt;(*i)) {
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1774 llvm::cerr &lt;&lt; "F is used in instruction:\n";
1775 llvm::cerr &lt;&lt; *Inst &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1776 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1777</pre>
1778</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1779
1780<p>Alternately, it's common to have an instance of the <a
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]1781href="/doxygen/classllvm_1_1User.html">User Class</a> and need to know what
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1782<tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used by a
1783<tt>User</tt> is known as a <i>use-def</i> chain. Instances of class
1784<tt>Instruction</tt> are common <tt>User</tt>s, so we might want to iterate over
1785all of the values that a particular instruction uses (that is, the operands of
1786the particular <tt>Instruction</tt>):</p>
1787
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1788<div class="doc_code">
1789<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1790Instruction *pi = ...;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1791
1792for (User::op_iterator i = pi-&gt;op_begin(), e = pi-&gt;op_end(); i != e; ++i) {
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1793 Value *v = *i;
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1794 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1795}
1796</pre>
1797</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1798
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]1799<!--
1800 def-use chains ("finding all users of"): Value::use_begin/use_end
1801 use-def chains ("finding all values used"): User::op_begin/op_end [op=operand]
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1802-->
1803
1804</div>
1805
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1806<!--_______________________________________________________________________-->
1807<div class="doc_subsubsection">
1808 <a name="iterate_preds">Iterating over predecessors &amp;
1809successors of blocks</a>
1810</div>
1811
1812<div class="doc_text">
1813
1814<p>Iterating over the predecessors and successors of a block is quite easy
1815with the routines defined in <tt>"llvm/Support/CFG.h"</tt>. Just use code like
1816this to iterate over all predecessors of BB:</p>
1817
1818<div class="doc_code">
1819<pre>
1820#include "llvm/Support/CFG.h"
1821BasicBlock *BB = ...;
1822
1823for (pred_iterator PI = pred_begin(BB), E = pred_end(BB); PI != E; ++PI) {
1824 BasicBlock *Pred = *PI;
1825 // <i>...</i>
1826}
1827</pre>
1828</div>
1829
1830<p>Similarly, to iterate over successors use
1831succ_iterator/succ_begin/succ_end.</p>
1832
1833</div>
1834
1835
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1836<!-- ======================================================================= -->
1837<div class="doc_subsection">
1838 <a name="simplechanges">Making simple changes</a>
1839</div>
1840
1841<div class="doc_text">
1842
1843<p>There are some primitive transformation operations present in the LLVM
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]1844infrastructure that are worth knowing about. When performing
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1845transformations, it's fairly common to manipulate the contents of basic
1846blocks. This section describes some of the common methods for doing so
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1847and gives example code.</p>
1848
1849</div>
1850
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1851<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1852<div class="doc_subsubsection">
1853 <a name="schanges_creating">Creating and inserting new
1854 <tt>Instruction</tt>s</a>
1855</div>
1856
1857<div class="doc_text">
1858
1859<p><i>Instantiating Instructions</i></p>
1860
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1861<p>Creation of <tt>Instruction</tt>s is straight-forward: simply call the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1862constructor for the kind of instruction to instantiate and provide the necessary
1863parameters. For example, an <tt>AllocaInst</tt> only <i>requires</i> a
1864(const-ptr-to) <tt>Type</tt>. Thus:</p>
1865
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1866<div class="doc_code">
1867<pre>
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]1868AllocaInst* ai = new AllocaInst(Type::Int32Ty);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1869</pre>
1870</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1871
1872<p>will create an <tt>AllocaInst</tt> instance that represents the allocation of
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1873one integer in the current stack frame, at run time. Each <tt>Instruction</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1874subclass is likely to have varying default parameters which change the semantics
1875of the instruction, so refer to the <a
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]1876href="/doxygen/classllvm_1_1Instruction.html">doxygen documentation for the subclass of
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1877Instruction</a> that you're interested in instantiating.</p>
1878
1879<p><i>Naming values</i></p>
1880
1881<p>It is very useful to name the values of instructions when you're able to, as
1882this facilitates the debugging of your transformations. If you end up looking
1883at generated LLVM machine code, you definitely want to have logical names
1884associated with the results of instructions! By supplying a value for the
1885<tt>Name</tt> (default) parameter of the <tt>Instruction</tt> constructor, you
1886associate a logical name with the result of the instruction's execution at
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1887run time. For example, say that I'm writing a transformation that dynamically
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1888allocates space for an integer on the stack, and that integer is going to be
1889used as some kind of index by some other code. To accomplish this, I place an
1890<tt>AllocaInst</tt> at the first point in the first <tt>BasicBlock</tt> of some
1891<tt>Function</tt>, and I'm intending to use it within the same
1892<tt>Function</tt>. I might do:</p>
1893
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1894<div class="doc_code">
1895<pre>
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]1896AllocaInst* pa = new AllocaInst(Type::Int32Ty, 0, "indexLoc");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1897</pre>
1898</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1899
1900<p>where <tt>indexLoc</tt> is now the logical name of the instruction's
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1901execution value, which is a pointer to an integer on the run time stack.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1902
1903<p><i>Inserting instructions</i></p>
1904
1905<p>There are essentially two ways to insert an <tt>Instruction</tt>
1906into an existing sequence of instructions that form a <tt>BasicBlock</tt>:</p>
1907
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]1908<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1909 <li>Insertion into an explicit instruction list
1910
1911 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within that
1912 <tt>BasicBlock</tt>, and a newly-created instruction we wish to insert
1913 before <tt>*pi</tt>, we do the following: </p>
1914
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1915<div class="doc_code">
1916<pre>
1917BasicBlock *pb = ...;
1918Instruction *pi = ...;
1919Instruction *newInst = new Instruction(...);
1920
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1921pb-&gt;getInstList().insert(pi, newInst); // <i>Inserts newInst before pi in pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1922</pre>
1923</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]1924
1925 <p>Appending to the end of a <tt>BasicBlock</tt> is so common that
1926 the <tt>Instruction</tt> class and <tt>Instruction</tt>-derived
1927 classes provide constructors which take a pointer to a
1928 <tt>BasicBlock</tt> to be appended to. For example code that
1929 looked like: </p>
1930
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1931<div class="doc_code">
1932<pre>
1933BasicBlock *pb = ...;
1934Instruction *newInst = new Instruction(...);
1935
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1936pb-&gt;getInstList().push_back(newInst); // <i>Appends newInst to pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1937</pre>
1938</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]1939
1940 <p>becomes: </p>
1941
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1942<div class="doc_code">
1943<pre>
1944BasicBlock *pb = ...;
1945Instruction *newInst = new Instruction(..., pb);
1946</pre>
1947</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]1948
1949 <p>which is much cleaner, especially if you are creating
1950 long instruction streams.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1951
1952 <li>Insertion into an implicit instruction list
1953
1954 <p><tt>Instruction</tt> instances that are already in <tt>BasicBlock</tt>s
1955 are implicitly associated with an existing instruction list: the instruction
1956 list of the enclosing basic block. Thus, we could have accomplished the same
1957 thing as the above code without being given a <tt>BasicBlock</tt> by doing:
1958 </p>
1959
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1960<div class="doc_code">
1961<pre>
1962Instruction *pi = ...;
1963Instruction *newInst = new Instruction(...);
1964
1965pi-&gt;getParent()-&gt;getInstList().insert(pi, newInst);
1966</pre>
1967</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1968
1969 <p>In fact, this sequence of steps occurs so frequently that the
1970 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes provide
1971 constructors which take (as a default parameter) a pointer to an
1972 <tt>Instruction</tt> which the newly-created <tt>Instruction</tt> should
1973 precede. That is, <tt>Instruction</tt> constructors are capable of
1974 inserting the newly-created instance into the <tt>BasicBlock</tt> of a
1975 provided instruction, immediately before that instruction. Using an
1976 <tt>Instruction</tt> constructor with a <tt>insertBefore</tt> (default)
1977 parameter, the above code becomes:</p>
1978
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1979<div class="doc_code">
1980<pre>
1981Instruction* pi = ...;
1982Instruction* newInst = new Instruction(..., pi);
1983</pre>
1984</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1985
1986 <p>which is much cleaner, especially if you're creating a lot of
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1987 instructions and adding them to <tt>BasicBlock</tt>s.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1988</ul>
1989
1990</div>
1991
1992<!--_______________________________________________________________________-->
1993<div class="doc_subsubsection">
1994 <a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a>
1995</div>
1996
1997<div class="doc_text">
1998
1999<p>Deleting an instruction from an existing sequence of instructions that form a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]2000<a href="#BasicBlock"><tt>BasicBlock</tt></a> is very straight-forward. First,
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2001you must have a pointer to the instruction that you wish to delete. Second, you
2002need to obtain the pointer to that instruction's basic block. You use the
2003pointer to the basic block to get its list of instructions and then use the
2004erase function to remove your instruction. For example:</p>
2005
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2006<div class="doc_code">
2007<pre>
2008<a href="#Instruction">Instruction</a> *I = .. ;
Chris Lattner9f8ec252008-02-15 22:57:17 +0000[diff] [blame]2009I-&gt;eraseFromParent();
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2010</pre>
2011</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2012
2013</div>
2014
2015<!--_______________________________________________________________________-->
2016<div class="doc_subsubsection">
2017 <a name="schanges_replacing">Replacing an <tt>Instruction</tt> with another
2018 <tt>Value</tt></a>
2019</div>
2020
2021<div class="doc_text">
2022
2023<p><i>Replacing individual instructions</i></p>
2024
2025<p>Including "<a href="/doxygen/BasicBlockUtils_8h-source.html">llvm/Transforms/Utils/BasicBlockUtils.h</a>"
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2026permits use of two very useful replace functions: <tt>ReplaceInstWithValue</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2027and <tt>ReplaceInstWithInst</tt>.</p>
2028
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2029<h4><a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a></h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2030
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2031<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2032 <li><tt>ReplaceInstWithValue</tt>
2033
Nick Lewyckyb6d1f392008-09-15 06:31:52 +0000[diff] [blame]2034 <p>This function replaces all uses of a given instruction with a value,
2035 and then removes the original instruction. The following example
2036 illustrates the replacement of the result of a particular
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]2037 <tt>AllocaInst</tt> that allocates memory for a single integer with a null
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2038 pointer to an integer.</p>
2039
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2040<div class="doc_code">
2041<pre>
2042AllocaInst* instToReplace = ...;
2043BasicBlock::iterator ii(instToReplace);
2044
2045ReplaceInstWithValue(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
Daniel Dunbar58c2ac02008-10-03 22:17:25 +0000[diff] [blame]2046 Constant::getNullValue(PointerType::getUnqual(Type::Int32Ty)));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2047</pre></div></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2048
2049 <li><tt>ReplaceInstWithInst</tt>
2050
2051 <p>This function replaces a particular instruction with another
Nick Lewyckyb6d1f392008-09-15 06:31:52 +0000[diff] [blame]2052 instruction, inserting the new instruction into the basic block at the
2053 location where the old instruction was, and replacing any uses of the old
2054 instruction with the new instruction. The following example illustrates
2055 the replacement of one <tt>AllocaInst</tt> with another.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2056
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2057<div class="doc_code">
2058<pre>
2059AllocaInst* instToReplace = ...;
2060BasicBlock::iterator ii(instToReplace);
2061
2062ReplaceInstWithInst(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2063 new AllocaInst(Type::Int32Ty, 0, "ptrToReplacedInt"));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2064</pre></div></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2065</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2066
2067<p><i>Replacing multiple uses of <tt>User</tt>s and <tt>Value</tt>s</i></p>
2068
2069<p>You can use <tt>Value::replaceAllUsesWith</tt> and
2070<tt>User::replaceUsesOfWith</tt> to change more than one use at a time. See the
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]2071doxygen documentation for the <a href="/doxygen/classllvm_1_1Value.html">Value Class</a>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]2072and <a href="/doxygen/classllvm_1_1User.html">User Class</a>, respectively, for more
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2073information.</p>
2074
2075<!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
2076include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
2077ReplaceInstWithValue, ReplaceInstWithInst -->
2078
2079</div>
2080
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2081<!--_______________________________________________________________________-->
2082<div class="doc_subsubsection">
2083 <a name="schanges_deletingGV">Deleting <tt>GlobalVariable</tt>s</a>
2084</div>
2085
2086<div class="doc_text">
2087
Tanya Lattnerc5dfcdb2007-06-20 20:46:37 +0000[diff] [blame]2088<p>Deleting a global variable from a module is just as easy as deleting an
2089Instruction. First, you must have a pointer to the global variable that you wish
2090 to delete. You use this pointer to erase it from its parent, the module.
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2091 For example:</p>
2092
2093<div class="doc_code">
2094<pre>
2095<a href="#GlobalVariable">GlobalVariable</a> *GV = .. ;
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2096
Tanya Lattnerc5dfcdb2007-06-20 20:46:37 +0000[diff] [blame]2097GV-&gt;eraseFromParent();
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2098</pre>
2099</div>
2100
2101</div>
2102
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2103<!-- ======================================================================= -->
2104<div class="doc_subsection">
2105 <a name="create_types">How to Create Types</a>
2106</div>
2107
2108<div class="doc_text">
2109
2110<p>In generating IR, you may need some complex types. If you know these types
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2111statically, you can use <tt>TypeBuilder&lt;...&gt;::get()</tt>, defined
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2112in <tt>llvm/Support/TypeBuilder.h</tt>, to retrieve them. <tt>TypeBuilder</tt>
2113has two forms depending on whether you're building types for cross-compilation
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2114or native library use. <tt>TypeBuilder&lt;T, true&gt;</tt> requires
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2115that <tt>T</tt> be independent of the host environment, meaning that it's built
2116out of types from
2117the <a href="/doxygen/namespacellvm_1_1types.html"><tt>llvm::types</tt></a>
2118namespace and pointers, functions, arrays, etc. built of
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2119those. <tt>TypeBuilder&lt;T, false&gt;</tt> additionally allows native C types
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2120whose size may depend on the host compiler. For example,</p>
2121
2122<div class="doc_code">
2123<pre>
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2124FunctionType *ft = TypeBuilder&lt;types::i&lt;8&gt;(types::i&lt;32&gt;*), true&gt;::get();
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2125</pre>
2126</div>
2127
2128<p>is easier to read and write than the equivalent</p>
2129
2130<div class="doc_code">
2131<pre>
Owen Anderson5e8c50e2009-06-16 17:40:28 +0000[diff] [blame]2132std::vector&lt;const Type*&gt; params;
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2133params.push_back(PointerType::getUnqual(Type::Int32Ty));
2134FunctionType *ft = FunctionType::get(Type::Int8Ty, params, false);
2135</pre>
2136</div>
2137
2138<p>See the <a href="/doxygen/TypeBuilder_8h-source.html#l00001">class
2139comment</a> for more details.</p>
2140
2141</div>
2142
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]2143<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2144<div class="doc_section">
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2145 <a name="threading">Threads and LLVM</a>
2146</div>
2147<!-- *********************************************************************** -->
2148
2149<div class="doc_text">
2150<p>
2151This section describes the interaction of the LLVM APIs with multithreading,
2152both on the part of client applications, and in the JIT, in the hosted
2153application.
2154</p>
2155
2156<p>
2157Note that LLVM's support for multithreading is still relatively young. Up
2158through version 2.5, the execution of threaded hosted applications was
2159supported, but not threaded client access to the APIs. While this use case is
2160now supported, clients <em>must</em> adhere to the guidelines specified below to
2161ensure proper operation in multithreaded mode.
2162</p>
2163
2164<p>
2165Note that, on Unix-like platforms, LLVM requires the presence of GCC's atomic
2166intrinsics in order to support threaded operation. If you need a
2167multhreading-capable LLVM on a platform without a suitably modern system
2168compiler, consider compiling LLVM and LLVM-GCC in single-threaded mode, and
2169using the resultant compiler to build a copy of LLVM with multithreading
2170support.
2171</p>
2172</div>
2173
2174<!-- ======================================================================= -->
2175<div class="doc_subsection">
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2176 <a name="startmultithreaded">Entering and Exiting Multithreaded Mode</a>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2177</div>
2178
2179<div class="doc_text">
2180
2181<p>
2182In order to properly protect its internal data structures while avoiding
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2183excessive locking overhead in the single-threaded case, the LLVM must intialize
2184certain data structures necessary to provide guards around its internals. To do
2185so, the client program must invoke <tt>llvm_start_multithreaded()</tt> before
2186making any concurrent LLVM API calls. To subsequently tear down these
2187structures, use the <tt>llvm_stop_multithreaded()</tt> call. You can also use
2188the <tt>llvm_is_multithreaded()</tt> call to check the status of multithreaded
2189mode.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2190</p>
2191
2192<p>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2193Note that both of these calls must be made <em>in isolation</em>. That is to
2194say that no other LLVM API calls may be executing at any time during the
2195execution of <tt>llvm_start_multithreaded()</tt> or <tt>llvm_stop_multithreaded
2196</tt>. It's is the client's responsibility to enforce this isolation.
2197</p>
2198
2199<p>
2200The return value of <tt>llvm_start_multithreaded()</tt> indicates the success or
2201failure of the initialization. Failure typically indicates that your copy of
2202LLVM was built without multithreading support, typically because GCC atomic
2203intrinsics were not found in your system compiler. In this case, the LLVM API
2204will not be safe for concurrent calls. However, it <em>will</em> be safe for
2205hosting threaded applications in the JIT, though care must be taken to ensure
2206that side exits and the like do not accidentally result in concurrent LLVM API
2207calls.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2208</p>
2209</div>
2210
2211<!-- ======================================================================= -->
2212<div class="doc_subsection">
2213 <a name="shutdown">Ending Execution with <tt>llvm_shutdown()</tt></a>
2214</div>
2215
2216<div class="doc_text">
2217<p>
2218When you are done using the LLVM APIs, you should call <tt>llvm_shutdown()</tt>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2219to deallocate memory used for internal structures. This will also invoke
2220<tt>llvm_stop_multithreaded()</tt> if LLVM is operating in multithreaded mode.
2221As such, <tt>llvm_shutdown()</tt> requires the same isolation guarantees as
2222<tt>llvm_stop_multithreaded()</tt>.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2223</p>
2224
2225<p>
2226Note that, if you use scope-based shutdown, you can use the
2227<tt>llvm_shutdown_obj</tt> class, which calls <tt>llvm_shutdown()</tt> in its
2228destructor.
2229</div>
2230
2231<!-- ======================================================================= -->
2232<div class="doc_subsection">
2233 <a name="managedstatic">Lazy Initialization with <tt>ManagedStatic</tt></a>
2234</div>
2235
2236<div class="doc_text">
2237<p>
2238<tt>ManagedStatic</tt> is a utility class in LLVM used to implement static
2239initialization of static resources, such as the global type tables. Before the
2240invocation of <tt>llvm_shutdown()</tt>, it implements a simple lazy
2241initialization scheme. Once <tt>llvm_start_multithreaded()</tt> returns,
2242however, it uses double-checked locking to implement thread-safe lazy
2243initialization.
2244</p>
2245
2246<p>
2247Note that, because no other threads are allowed to issue LLVM API calls before
2248<tt>llvm_start_multithreaded()</tt> returns, it is possible to have
2249<tt>ManagedStatic</tt>s of <tt>llvm::sys::Mutex</tt>s.
2250</p>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2251
2252<p>
2253The <tt>llvm_acquire_global_lock()</tt> and <tt>llvm_release_global_lock</tt>
2254APIs provide access to the global lock used to implement the double-checked
2255locking for lazy initialization. These should only be used internally to LLVM,
2256and only if you know what you're doing!
2257</p>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2258</div>
2259
2260<!-- *********************************************************************** -->
2261<div class="doc_section">
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2262 <a name="advanced">Advanced Topics</a>
2263</div>
2264<!-- *********************************************************************** -->
2265
2266<div class="doc_text">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2267<p>
2268This section describes some of the advanced or obscure API's that most clients
2269do not need to be aware of. These API's tend manage the inner workings of the
2270LLVM system, and only need to be accessed in unusual circumstances.
2271</p>
2272</div>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2273
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2274<!-- ======================================================================= -->
2275<div class="doc_subsection">
2276 <a name="TypeResolve">LLVM Type Resolution</a>
2277</div>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2278
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2279<div class="doc_text">
2280
2281<p>
2282The LLVM type system has a very simple goal: allow clients to compare types for
2283structural equality with a simple pointer comparison (aka a shallow compare).
2284This goal makes clients much simpler and faster, and is used throughout the LLVM
2285system.
2286</p>
2287
2288<p>
2289Unfortunately achieving this goal is not a simple matter. In particular,
2290recursive types and late resolution of opaque types makes the situation very
2291difficult to handle. Fortunately, for the most part, our implementation makes
2292most clients able to be completely unaware of the nasty internal details. The
2293primary case where clients are exposed to the inner workings of it are when
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]2294building a recursive type. In addition to this case, the LLVM bitcode reader,
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2295assembly parser, and linker also have to be aware of the inner workings of this
2296system.
2297</p>
2298
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]2299<p>
2300For our purposes below, we need three concepts. First, an "Opaque Type" is
2301exactly as defined in the <a href="LangRef.html#t_opaque">language
2302reference</a>. Second an "Abstract Type" is any type which includes an
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2303opaque type as part of its type graph (for example "<tt>{ opaque, i32 }</tt>").
2304Third, a concrete type is a type that is not an abstract type (e.g. "<tt>{ i32,
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]2305float }</tt>").
2306</p>
2307
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2308</div>
2309
2310<!-- ______________________________________________________________________ -->
2311<div class="doc_subsubsection">
2312 <a name="BuildRecType">Basic Recursive Type Construction</a>
2313</div>
2314
2315<div class="doc_text">
2316
2317<p>
2318Because the most common question is "how do I build a recursive type with LLVM",
2319we answer it now and explain it as we go. Here we include enough to cause this
2320to be emitted to an output .ll file:
2321</p>
2322
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2323<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2324<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2325%mylist = type { %mylist*, i32 }
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2326</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2327</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2328
2329<p>
2330To build this, use the following LLVM APIs:
2331</p>
2332
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2333<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2334<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2335// <i>Create the initial outer struct</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2336<a href="#PATypeHolder">PATypeHolder</a> StructTy = OpaqueType::get();
2337std::vector&lt;const Type*&gt; Elts;
Daniel Dunbar58c2ac02008-10-03 22:17:25 +0000[diff] [blame]2338Elts.push_back(PointerType::getUnqual(StructTy));
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2339Elts.push_back(Type::Int32Ty);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2340StructType *NewSTy = StructType::get(Elts);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2341
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2342// <i>At this point, NewSTy = "{ opaque*, i32 }". Tell VMCore that</i>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2343// <i>the struct and the opaque type are actually the same.</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2344cast&lt;OpaqueType&gt;(StructTy.get())-&gt;<a href="#refineAbstractTypeTo">refineAbstractTypeTo</a>(NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2345
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2346// <i>NewSTy is potentially invalidated, but StructTy (a <a href="#PATypeHolder">PATypeHolder</a>) is</i>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2347// <i>kept up-to-date</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2348NewSTy = cast&lt;StructType&gt;(StructTy.get());
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2349
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2350// <i>Add a name for the type to the module symbol table (optional)</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2351MyModule-&gt;addTypeName("mylist", NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2352</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2353</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2354
2355<p>
2356This code shows the basic approach used to build recursive types: build a
2357non-recursive type using 'opaque', then use type unification to close the cycle.
2358The type unification step is performed by the <tt><a
Chris Lattneraff26d12007-02-03 03:06:52 +0000[diff] [blame]2359href="#refineAbstractTypeTo">refineAbstractTypeTo</a></tt> method, which is
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2360described next. After that, we describe the <a
2361href="#PATypeHolder">PATypeHolder class</a>.
2362</p>
2363
2364</div>
2365
2366<!-- ______________________________________________________________________ -->
2367<div class="doc_subsubsection">
2368 <a name="refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a>
2369</div>
2370
2371<div class="doc_text">
2372<p>
2373The <tt>refineAbstractTypeTo</tt> method starts the type unification process.
2374While this method is actually a member of the DerivedType class, it is most
2375often used on OpaqueType instances. Type unification is actually a recursive
2376process. After unification, types can become structurally isomorphic to
2377existing types, and all duplicates are deleted (to preserve pointer equality).
2378</p>
2379
2380<p>
2381In the example above, the OpaqueType object is definitely deleted.
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2382Additionally, if there is an "{ \2*, i32}" type already created in the system,
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2383the pointer and struct type created are <b>also</b> deleted. Obviously whenever
2384a type is deleted, any "Type*" pointers in the program are invalidated. As
2385such, it is safest to avoid having <i>any</i> "Type*" pointers to abstract types
2386live across a call to <tt>refineAbstractTypeTo</tt> (note that non-abstract
2387types can never move or be deleted). To deal with this, the <a
2388href="#PATypeHolder">PATypeHolder</a> class is used to maintain a stable
2389reference to a possibly refined type, and the <a
2390href="#AbstractTypeUser">AbstractTypeUser</a> class is used to update more
2391complex datastructures.
2392</p>
2393
2394</div>
2395
2396<!-- ______________________________________________________________________ -->
2397<div class="doc_subsubsection">
2398 <a name="PATypeHolder">The PATypeHolder Class</a>
2399</div>
2400
2401<div class="doc_text">
2402<p>
2403PATypeHolder is a form of a "smart pointer" for Type objects. When VMCore
2404happily goes about nuking types that become isomorphic to existing types, it
2405automatically updates all PATypeHolder objects to point to the new type. In the
2406example above, this allows the code to maintain a pointer to the resultant
2407resolved recursive type, even though the Type*'s are potentially invalidated.
2408</p>
2409
2410<p>
2411PATypeHolder is an extremely light-weight object that uses a lazy union-find
2412implementation to update pointers. For example the pointer from a Value to its
2413Type is maintained by PATypeHolder objects.
2414</p>
2415
2416</div>
2417
2418<!-- ______________________________________________________________________ -->
2419<div class="doc_subsubsection">
2420 <a name="AbstractTypeUser">The AbstractTypeUser Class</a>
2421</div>
2422
2423<div class="doc_text">
2424
2425<p>
2426Some data structures need more to perform more complex updates when types get
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2427resolved. To support this, a class can derive from the AbstractTypeUser class.
2428This class
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2429allows it to get callbacks when certain types are resolved. To register to get
2430callbacks for a particular type, the DerivedType::{add/remove}AbstractTypeUser
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]2431methods can be called on a type. Note that these methods only work for <i>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2432 abstract</i> types. Concrete types (those that do not include any opaque
2433objects) can never be refined.
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2434</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2435</div>
2436
2437
2438<!-- ======================================================================= -->
2439<div class="doc_subsection">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2440 <a name="SymbolTable">The <tt>ValueSymbolTable</tt> and
2441 <tt>TypeSymbolTable</tt> classes</a>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2442</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2443
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2444<div class="doc_text">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2445<p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1ValueSymbolTable.html">
2446ValueSymbolTable</a></tt> class provides a symbol table that the <a
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2447href="#Function"><tt>Function</tt></a> and <a href="#Module">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2448<tt>Module</tt></a> classes use for naming value definitions. The symbol table
2449can provide a name for any <a href="#Value"><tt>Value</tt></a>.
2450The <tt><a href="http://llvm.org/doxygen/classllvm_1_1TypeSymbolTable.html">
2451TypeSymbolTable</a></tt> class is used by the <tt>Module</tt> class to store
2452names for types.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2453
Reid Spencera6362242007-01-07 00:41:39 +0000[diff] [blame]2454<p>Note that the <tt>SymbolTable</tt> class should not be directly accessed
2455by most clients. It should only be used when iteration over the symbol table
2456names themselves are required, which is very special purpose. Note that not
2457all LLVM
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2458<tt><a href="#Value">Value</a></tt>s have names, and those without names (i.e. they have
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2459an empty name) do not exist in the symbol table.
2460</p>
2461
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2462<p>These symbol tables support iteration over the values/types in the symbol
2463table with <tt>begin/end/iterator</tt> and supports querying to see if a
2464specific name is in the symbol table (with <tt>lookup</tt>). The
2465<tt>ValueSymbolTable</tt> class exposes no public mutator methods, instead,
2466simply call <tt>setName</tt> on a value, which will autoinsert it into the
2467appropriate symbol table. For types, use the Module::addTypeName method to
2468insert entries into the symbol table.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2469
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2470</div>
2471
2472
2473
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2474<!-- ======================================================================= -->
2475<div class="doc_subsection">
2476 <a name="UserLayout">The <tt>User</tt> and owned <tt>Use</tt> classes' memory layout</a>
2477</div>
2478
2479<div class="doc_text">
2480<p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1User.html">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2481User</a></tt> class provides a basis for expressing the ownership of <tt>User</tt>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2482towards other <tt><a href="http://llvm.org/doxygen/classllvm_1_1Value.html">
2483Value</a></tt>s. The <tt><a href="http://llvm.org/doxygen/classllvm_1_1Use.html">
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2484Use</a></tt> helper class is employed to do the bookkeeping and to facilitate <i>O(1)</i>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2485addition and removal.</p>
2486
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2487<!-- ______________________________________________________________________ -->
2488<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2489 <a name="Use2User">Interaction and relationship between <tt>User</tt> and <tt>Use</tt> objects</a>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2490</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2491
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2492<div class="doc_text">
2493<p>
2494A subclass of <tt>User</tt> can choose between incorporating its <tt>Use</tt> objects
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2495or refer to them out-of-line by means of a pointer. A mixed variant
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2496(some <tt>Use</tt>s inline others hung off) is impractical and breaks the invariant
2497that the <tt>Use</tt> objects belonging to the same <tt>User</tt> form a contiguous array.
2498</p>
2499</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2500
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2501<p>
2502We have 2 different layouts in the <tt>User</tt> (sub)classes:
2503<ul>
2504<li><p>Layout a)
2505The <tt>Use</tt> object(s) are inside (resp. at fixed offset) of the <tt>User</tt>
2506object and there are a fixed number of them.</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2507
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2508<li><p>Layout b)
2509The <tt>Use</tt> object(s) are referenced by a pointer to an
2510array from the <tt>User</tt> object and there may be a variable
2511number of them.</p>
2512</ul>
2513<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2514As of v2.4 each layout still possesses a direct pointer to the
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2515start of the array of <tt>Use</tt>s. Though not mandatory for layout a),
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2516we stick to this redundancy for the sake of simplicity.
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2517The <tt>User</tt> object also stores the number of <tt>Use</tt> objects it
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2518has. (Theoretically this information can also be calculated
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2519given the scheme presented below.)</p>
2520<p>
2521Special forms of allocation operators (<tt>operator new</tt>)
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2522enforce the following memory layouts:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2523
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2524<ul>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2525<li><p>Layout a) is modelled by prepending the <tt>User</tt> object by the <tt>Use[]</tt> array.</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2526
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2527<pre>
2528...---.---.---.---.-------...
2529 | P | P | P | P | User
2530'''---'---'---'---'-------'''
2531</pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2532
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2533<li><p>Layout b) is modelled by pointing at the <tt>Use[]</tt> array.</p>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2534<pre>
2535.-------...
2536| User
2537'-------'''
2538 |
2539 v
2540 .---.---.---.---...
2541 | P | P | P | P |
2542 '---'---'---'---'''
2543</pre>
2544</ul>
2545<i>(In the above figures '<tt>P</tt>' stands for the <tt>Use**</tt> that
2546 is stored in each <tt>Use</tt> object in the member <tt>Use::Prev</tt>)</i>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2547
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2548<!-- ______________________________________________________________________ -->
2549<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2550 <a name="Waymarking">The waymarking algorithm</a>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2551</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2552
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2553<div class="doc_text">
2554<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2555Since the <tt>Use</tt> objects are deprived of the direct (back)pointer to
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2556their <tt>User</tt> objects, there must be a fast and exact method to
2557recover it. This is accomplished by the following scheme:</p>
2558</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2559
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2560A bit-encoding in the 2 LSBits (least significant bits) of the <tt>Use::Prev</tt> allows to find the
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2561start of the <tt>User</tt> object:
2562<ul>
2563<li><tt>00</tt> &mdash;&gt; binary digit 0</li>
2564<li><tt>01</tt> &mdash;&gt; binary digit 1</li>
2565<li><tt>10</tt> &mdash;&gt; stop and calculate (<tt>s</tt>)</li>
2566<li><tt>11</tt> &mdash;&gt; full stop (<tt>S</tt>)</li>
2567</ul>
2568<p>
2569Given a <tt>Use*</tt>, all we have to do is to walk till we get
2570a stop and we either have a <tt>User</tt> immediately behind or
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2571we have to walk to the next stop picking up digits
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2572and calculating the offset:</p>
2573<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2574.---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.----------------
2575| 1 | s | 1 | 0 | 1 | 0 | s | 1 | 1 | 0 | s | 1 | 1 | s | 1 | S | User (or User*)
2576'---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'----------------
2577 |+15 |+10 |+6 |+3 |+1
2578 | | | | |__>
2579 | | | |__________>
2580 | | |______________________>
2581 | |______________________________________>
2582 |__________________________________________________________>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2583</pre>
2584<p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2585Only the significant number of bits need to be stored between the
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2586stops, so that the <i>worst case is 20 memory accesses</i> when there are
25871000 <tt>Use</tt> objects associated with a <tt>User</tt>.</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2588
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2589<!-- ______________________________________________________________________ -->
2590<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2591 <a name="ReferenceImpl">Reference implementation</a>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2592</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2593
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2594<div class="doc_text">
2595<p>
2596The following literate Haskell fragment demonstrates the concept:</p>
2597</div>
2598
2599<div class="doc_code">
2600<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2601> import Test.QuickCheck
2602>
2603> digits :: Int -> [Char] -> [Char]
2604> digits 0 acc = '0' : acc
2605> digits 1 acc = '1' : acc
2606> digits n acc = digits (n `div` 2) $ digits (n `mod` 2) acc
2607>
2608> dist :: Int -> [Char] -> [Char]
2609> dist 0 [] = ['S']
2610> dist 0 acc = acc
2611> dist 1 acc = let r = dist 0 acc in 's' : digits (length r) r
2612> dist n acc = dist (n - 1) $ dist 1 acc
2613>
2614> takeLast n ss = reverse $ take n $ reverse ss
2615>
2616> test = takeLast 40 $ dist 20 []
2617>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2618</pre>
2619</div>
2620<p>
2621Printing &lt;test&gt; gives: <tt>"1s100000s11010s10100s1111s1010s110s11s1S"</tt></p>
2622<p>
2623The reverse algorithm computes the length of the string just by examining
2624a certain prefix:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2625
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2626<div class="doc_code">
2627<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2628> pref :: [Char] -> Int
2629> pref "S" = 1
2630> pref ('s':'1':rest) = decode 2 1 rest
2631> pref (_:rest) = 1 + pref rest
2632>
2633> decode walk acc ('0':rest) = decode (walk + 1) (acc * 2) rest
2634> decode walk acc ('1':rest) = decode (walk + 1) (acc * 2 + 1) rest
2635> decode walk acc _ = walk + acc
2636>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2637</pre>
2638</div>
2639<p>
2640Now, as expected, printing &lt;pref test&gt; gives <tt>40</tt>.</p>
2641<p>
2642We can <i>quickCheck</i> this with following property:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2643
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2644<div class="doc_code">
2645<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2646> testcase = dist 2000 []
2647> testcaseLength = length testcase
2648>
2649> identityProp n = n > 0 && n <= testcaseLength ==> length arr == pref arr
2650> where arr = takeLast n testcase
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2651>
2652</pre>
2653</div>
2654<p>
2655As expected &lt;quickCheck identityProp&gt; gives:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2656
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2657<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2658*Main> quickCheck identityProp
2659OK, passed 100 tests.
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2660</pre>
2661<p>
2662Let's be a bit more exhaustive:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2663
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2664<div class="doc_code">
2665<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2666>
2667> deepCheck p = check (defaultConfig { configMaxTest = 500 }) p
2668>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2669</pre>
2670</div>
2671<p>
2672And here is the result of &lt;deepCheck identityProp&gt;:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2673
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2674<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2675*Main> deepCheck identityProp
2676OK, passed 500 tests.
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2677</pre>
2678
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2679<!-- ______________________________________________________________________ -->
2680<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2681 <a name="Tagging">Tagging considerations</a>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2682</div>
2683
2684<p>
2685To maintain the invariant that the 2 LSBits of each <tt>Use**</tt> in <tt>Use</tt>
2686never change after being set up, setters of <tt>Use::Prev</tt> must re-tag the
2687new <tt>Use**</tt> on every modification. Accordingly getters must strip the
2688tag bits.</p>
2689<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2690For layout b) instead of the <tt>User</tt> we find a pointer (<tt>User*</tt> with LSBit set).
2691Following this pointer brings us to the <tt>User</tt>. A portable trick ensures
2692that the first bytes of <tt>User</tt> (if interpreted as a pointer) never has
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2693the LSBit set. (Portability is relying on the fact that all known compilers place the
2694<tt>vptr</tt> in the first word of the instances.)</p>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2695
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2696</div>
2697
2698 <!-- *********************************************************************** -->
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2699<div class="doc_section">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2700 <a name="coreclasses">The Core LLVM Class Hierarchy Reference </a>
2701</div>
2702<!-- *********************************************************************** -->
2703
2704<div class="doc_text">
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2705<p><tt>#include "<a href="/doxygen/Type_8h-source.html">llvm/Type.h</a>"</tt>
2706<br>doxygen info: <a href="/doxygen/classllvm_1_1Type.html">Type Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2707
2708<p>The Core LLVM classes are the primary means of representing the program
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2709being inspected or transformed. The core LLVM classes are defined in
2710header files in the <tt>include/llvm/</tt> directory, and implemented in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2711the <tt>lib/VMCore</tt> directory.</p>
2712
2713</div>
2714
2715<!-- ======================================================================= -->
2716<div class="doc_subsection">
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2717 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
2718</div>
2719
2720<div class="doc_text">
2721
2722 <p><tt>Type</tt> is a superclass of all type classes. Every <tt>Value</tt> has
2723 a <tt>Type</tt>. <tt>Type</tt> cannot be instantiated directly but only
2724 through its subclasses. Certain primitive types (<tt>VoidType</tt>,
2725 <tt>LabelType</tt>, <tt>FloatType</tt> and <tt>DoubleType</tt>) have hidden
2726 subclasses. They are hidden because they offer no useful functionality beyond
2727 what the <tt>Type</tt> class offers except to distinguish themselves from
2728 other subclasses of <tt>Type</tt>.</p>
2729 <p>All other types are subclasses of <tt>DerivedType</tt>. Types can be
2730 named, but this is not a requirement. There exists exactly
2731 one instance of a given shape at any one time. This allows type equality to
2732 be performed with address equality of the Type Instance. That is, given two
2733 <tt>Type*</tt> values, the types are identical if the pointers are identical.
2734 </p>
2735</div>
2736
2737<!-- _______________________________________________________________________ -->
2738<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2739 <a name="m_Type">Important Public Methods</a>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2740</div>
2741
2742<div class="doc_text">
2743
2744<ul>
Chris Lattner8f79df32007-01-15 01:55:32 +0000[diff] [blame]2745 <li><tt>bool isInteger() const</tt>: Returns true for any integer type.</li>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2746
2747 <li><tt>bool isFloatingPoint()</tt>: Return true if this is one of the two
2748 floating point types.</li>
2749
2750 <li><tt>bool isAbstract()</tt>: Return true if the type is abstract (contains
2751 an OpaqueType anywhere in its definition).</li>
2752
2753 <li><tt>bool isSized()</tt>: Return true if the type has known size. Things
2754 that don't have a size are abstract types, labels and void.</li>
2755
2756</ul>
2757</div>
2758
2759<!-- _______________________________________________________________________ -->
2760<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2761 <a name="derivedtypes">Important Derived Types</a>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2762</div>
2763<div class="doc_text">
2764<dl>
2765 <dt><tt>IntegerType</tt></dt>
2766 <dd>Subclass of DerivedType that represents integer types of any bit width.
2767 Any bit width between <tt>IntegerType::MIN_INT_BITS</tt> (1) and
2768 <tt>IntegerType::MAX_INT_BITS</tt> (~8 million) can be represented.
2769 <ul>
2770 <li><tt>static const IntegerType* get(unsigned NumBits)</tt>: get an integer
2771 type of a specific bit width.</li>
2772 <li><tt>unsigned getBitWidth() const</tt>: Get the bit width of an integer
2773 type.</li>
2774 </ul>
2775 </dd>
2776 <dt><tt>SequentialType</tt></dt>
2777 <dd>This is subclassed by ArrayType and PointerType
2778 <ul>
2779 <li><tt>const Type * getElementType() const</tt>: Returns the type of each
2780 of the elements in the sequential type. </li>
2781 </ul>
2782 </dd>
2783 <dt><tt>ArrayType</tt></dt>
2784 <dd>This is a subclass of SequentialType and defines the interface for array
2785 types.
2786 <ul>
2787 <li><tt>unsigned getNumElements() const</tt>: Returns the number of
2788 elements in the array. </li>
2789 </ul>
2790 </dd>
2791 <dt><tt>PointerType</tt></dt>
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]2792 <dd>Subclass of SequentialType for pointer types.</dd>
Reid Spencer9d6565a2007-02-15 02:26:10 +0000[diff] [blame]2793 <dt><tt>VectorType</tt></dt>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2794 <dd>Subclass of SequentialType for vector types. A
2795 vector type is similar to an ArrayType but is distinguished because it is
2796 a first class type wherease ArrayType is not. Vector types are used for
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2797 vector operations and are usually small vectors of of an integer or floating
2798 point type.</dd>
2799 <dt><tt>StructType</tt></dt>
2800 <dd>Subclass of DerivedTypes for struct types.</dd>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]2801 <dt><tt><a name="FunctionType">FunctionType</a></tt></dt>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2802 <dd>Subclass of DerivedTypes for function types.
2803 <ul>
2804 <li><tt>bool isVarArg() const</tt>: Returns true if its a vararg
2805 function</li>
2806 <li><tt> const Type * getReturnType() const</tt>: Returns the
2807 return type of the function.</li>
2808 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns
2809 the type of the ith parameter.</li>
2810 <li><tt> const unsigned getNumParams() const</tt>: Returns the
2811 number of formal parameters.</li>
2812 </ul>
2813 </dd>
2814 <dt><tt>OpaqueType</tt></dt>
2815 <dd>Sublcass of DerivedType for abstract types. This class
2816 defines no content and is used as a placeholder for some other type. Note
2817 that OpaqueType is used (temporarily) during type resolution for forward
2818 references of types. Once the referenced type is resolved, the OpaqueType
2819 is replaced with the actual type. OpaqueType can also be used for data
2820 abstraction. At link time opaque types can be resolved to actual types
2821 of the same name.</dd>
2822</dl>
2823</div>
2824
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2825
2826
2827<!-- ======================================================================= -->
2828<div class="doc_subsection">
2829 <a name="Module">The <tt>Module</tt> class</a>
2830</div>
2831
2832<div class="doc_text">
2833
2834<p><tt>#include "<a
2835href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt><br> doxygen info:
2836<a href="/doxygen/classllvm_1_1Module.html">Module Class</a></p>
2837
2838<p>The <tt>Module</tt> class represents the top level structure present in LLVM
2839programs. An LLVM module is effectively either a translation unit of the
2840original program or a combination of several translation units merged by the
2841linker. The <tt>Module</tt> class keeps track of a list of <a
2842href="#Function"><tt>Function</tt></a>s, a list of <a
2843href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
2844href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
2845helpful member functions that try to make common operations easy.</p>
2846
2847</div>
2848
2849<!-- _______________________________________________________________________ -->
2850<div class="doc_subsubsection">
2851 <a name="m_Module">Important Public Members of the <tt>Module</tt> class</a>
2852</div>
2853
2854<div class="doc_text">
2855
2856<ul>
2857 <li><tt>Module::Module(std::string name = "")</tt></li>
2858</ul>
2859
2860<p>Constructing a <a href="#Module">Module</a> is easy. You can optionally
2861provide a name for it (probably based on the name of the translation unit).</p>
2862
2863<ul>
2864 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
2865 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
2866
2867 <tt>begin()</tt>, <tt>end()</tt>
2868 <tt>size()</tt>, <tt>empty()</tt>
2869
2870 <p>These are forwarding methods that make it easy to access the contents of
2871 a <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
2872 list.</p></li>
2873
2874 <li><tt>Module::FunctionListType &amp;getFunctionList()</tt>
2875
2876 <p> Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
2877 necessary to use when you need to update the list or perform a complex
2878 action that doesn't have a forwarding method.</p>
2879
2880 <p><!-- Global Variable --></p></li>
2881</ul>
2882
2883<hr>
2884
2885<ul>
2886 <li><tt>Module::global_iterator</tt> - Typedef for global variable list iterator<br>
2887
2888 <tt>Module::const_global_iterator</tt> - Typedef for const_iterator.<br>
2889
2890 <tt>global_begin()</tt>, <tt>global_end()</tt>
2891 <tt>global_size()</tt>, <tt>global_empty()</tt>
2892
2893 <p> These are forwarding methods that make it easy to access the contents of
2894 a <tt>Module</tt> object's <a
2895 href="#GlobalVariable"><tt>GlobalVariable</tt></a> list.</p></li>
2896
2897 <li><tt>Module::GlobalListType &amp;getGlobalList()</tt>
2898
2899 <p>Returns the list of <a
2900 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s. This is necessary to
2901 use when you need to update the list or perform a complex action that
2902 doesn't have a forwarding method.</p>
2903
2904 <p><!-- Symbol table stuff --> </p></li>
2905</ul>
2906
2907<hr>
2908
2909<ul>
2910 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
2911
2912 <p>Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
2913 for this <tt>Module</tt>.</p>
2914
2915 <p><!-- Convenience methods --></p></li>
2916</ul>
2917
2918<hr>
2919
2920<ul>
2921 <li><tt><a href="#Function">Function</a> *getFunction(const std::string
2922 &amp;Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt>
2923
2924 <p>Look up the specified function in the <tt>Module</tt> <a
2925 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
2926 <tt>null</tt>.</p></li>
2927
2928 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const
2929 std::string &amp;Name, const <a href="#FunctionType">FunctionType</a> *T)</tt>
2930
2931 <p>Look up the specified function in the <tt>Module</tt> <a
2932 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
2933 external declaration for the function and return it.</p></li>
2934
2935 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt>
2936
2937 <p>If there is at least one entry in the <a
2938 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
2939 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
2940 string.</p></li>
2941
2942 <li><tt>bool addTypeName(const std::string &amp;Name, const <a
2943 href="#Type">Type</a> *Ty)</tt>
2944
2945 <p>Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
2946 mapping <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this
2947 name, true is returned and the <a
2948 href="#SymbolTable"><tt>SymbolTable</tt></a> is not modified.</p></li>
2949</ul>
2950
2951</div>
2952
2953
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2954<!-- ======================================================================= -->
2955<div class="doc_subsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2956 <a name="Value">The <tt>Value</tt> class</a>
2957</div>
2958
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2959<div class="doc_text">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2960
2961<p><tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt>
2962<br>
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]2963doxygen info: <a href="/doxygen/classllvm_1_1Value.html">Value Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2964
2965<p>The <tt>Value</tt> class is the most important class in the LLVM Source
2966base. It represents a typed value that may be used (among other things) as an
2967operand to an instruction. There are many different types of <tt>Value</tt>s,
2968such as <a href="#Constant"><tt>Constant</tt></a>s,<a
2969href="#Argument"><tt>Argument</tt></a>s. Even <a
2970href="#Instruction"><tt>Instruction</tt></a>s and <a
2971href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.</p>
2972
2973<p>A particular <tt>Value</tt> may be used many times in the LLVM representation
2974for a program. For example, an incoming argument to a function (represented
2975with an instance of the <a href="#Argument">Argument</a> class) is "used" by
2976every instruction in the function that references the argument. To keep track
2977of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
2978href="#User"><tt>User</tt></a>s that is using it (the <a
2979href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
2980graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
2981def-use information in the program, and is accessible through the <tt>use_</tt>*
2982methods, shown below.</p>
2983
2984<p>Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed,
2985and this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
2986method. In addition, all LLVM values can be named. The "name" of the
2987<tt>Value</tt> is a symbolic string printed in the LLVM code:</p>
2988
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2989<div class="doc_code">
2990<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2991%<b>foo</b> = add i32 1, 2
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2992</pre>
2993</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2994
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]2995<p><a name="nameWarning">The name of this instruction is "foo".</a> <b>NOTE</b>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2996that the name of any value may be missing (an empty string), so names should
2997<b>ONLY</b> be used for debugging (making the source code easier to read,
2998debugging printouts), they should not be used to keep track of values or map
2999between them. For this purpose, use a <tt>std::map</tt> of pointers to the
3000<tt>Value</tt> itself instead.</p>
3001
3002<p>One important aspect of LLVM is that there is no distinction between an SSA
3003variable and the operation that produces it. Because of this, any reference to
3004the value produced by an instruction (or the value available as an incoming
Chris Lattnerd5fc4fc2004-03-18 14:58:55 +0000[diff] [blame]3005argument, for example) is represented as a direct pointer to the instance of
3006the class that
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3007represents this value. Although this may take some getting used to, it
3008simplifies the representation and makes it easier to manipulate.</p>
3009
3010</div>
3011
3012<!-- _______________________________________________________________________ -->
3013<div class="doc_subsubsection">
3014 <a name="m_Value">Important Public Members of the <tt>Value</tt> class</a>
3015</div>
3016
3017<div class="doc_text">
3018
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3019<ul>
3020 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the
3021use-list<br>
3022 <tt>Value::use_const_iterator</tt> - Typedef for const_iterator over
3023the use-list<br>
3024 <tt>unsigned use_size()</tt> - Returns the number of users of the
3025value.<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3026 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3027 <tt>use_iterator use_begin()</tt> - Get an iterator to the start of
3028the use-list.<br>
3029 <tt>use_iterator use_end()</tt> - Get an iterator to the end of the
3030use-list.<br>
3031 <tt><a href="#User">User</a> *use_back()</tt> - Returns the last
3032element in the list.
3033 <p> These methods are the interface to access the def-use
3034information in LLVM. As with all other iterators in LLVM, the naming
3035conventions follow the conventions defined by the <a href="#stl">STL</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3036 </li>
3037 <li><tt><a href="#Type">Type</a> *getType() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3038 <p>This method returns the Type of the Value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3039 </li>
3040 <li><tt>bool hasName() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3041 <tt>std::string getName() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3042 <tt>void setName(const std::string &amp;Name)</tt>
3043 <p> This family of methods is used to access and assign a name to a <tt>Value</tt>,
3044be aware of the <a href="#nameWarning">precaution above</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3045 </li>
3046 <li><tt>void replaceAllUsesWith(Value *V)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3047
3048 <p>This method traverses the use list of a <tt>Value</tt> changing all <a
3049 href="#User"><tt>User</tt>s</a> of the current value to refer to
3050 "<tt>V</tt>" instead. For example, if you detect that an instruction always
3051 produces a constant value (for example through constant folding), you can
3052 replace all uses of the instruction with the constant like this:</p>
3053
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3054<div class="doc_code">
3055<pre>
3056Inst-&gt;replaceAllUsesWith(ConstVal);
3057</pre>
3058</div>
3059
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3060</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3061
3062</div>
3063
3064<!-- ======================================================================= -->
3065<div class="doc_subsection">
3066 <a name="User">The <tt>User</tt> class</a>
3067</div>
3068
3069<div class="doc_text">
3070
3071<p>
3072<tt>#include "<a href="/doxygen/User_8h-source.html">llvm/User.h</a>"</tt><br>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]3073doxygen info: <a href="/doxygen/classllvm_1_1User.html">User Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3074Superclass: <a href="#Value"><tt>Value</tt></a></p>
3075
3076<p>The <tt>User</tt> class is the common base class of all LLVM nodes that may
3077refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
3078that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
3079referring to. The <tt>User</tt> class itself is a subclass of
3080<tt>Value</tt>.</p>
3081
3082<p>The operands of a <tt>User</tt> point directly to the LLVM <a
3083href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
3084Single Assignment (SSA) form, there can only be one definition referred to,
3085allowing this direct connection. This connection provides the use-def
3086information in LLVM.</p>
3087
3088</div>
3089
3090<!-- _______________________________________________________________________ -->
3091<div class="doc_subsubsection">
3092 <a name="m_User">Important Public Members of the <tt>User</tt> class</a>
3093</div>
3094
3095<div class="doc_text">
3096
3097<p>The <tt>User</tt> class exposes the operand list in two ways: through
3098an index access interface and through an iterator based interface.</p>
3099
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3100<ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3101 <li><tt>Value *getOperand(unsigned i)</tt><br>
3102 <tt>unsigned getNumOperands()</tt>
3103 <p> These two methods expose the operands of the <tt>User</tt> in a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3104convenient form for direct access.</p></li>
3105
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3106 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand
3107list<br>
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]3108 <tt>op_iterator op_begin()</tt> - Get an iterator to the start of
3109the operand list.<br>
3110 <tt>op_iterator op_end()</tt> - Get an iterator to the end of the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3111operand list.
3112 <p> Together, these methods make up the iterator based interface to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3113the operands of a <tt>User</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3114</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3115
3116</div>
3117
3118<!-- ======================================================================= -->
3119<div class="doc_subsection">
3120 <a name="Instruction">The <tt>Instruction</tt> class</a>
3121</div>
3122
3123<div class="doc_text">
3124
3125<p><tt>#include "</tt><tt><a
3126href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt><br>
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3127doxygen info: <a href="/doxygen/classllvm_1_1Instruction.html">Instruction Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3128Superclasses: <a href="#User"><tt>User</tt></a>, <a
3129href="#Value"><tt>Value</tt></a></p>
3130
3131<p>The <tt>Instruction</tt> class is the common base class for all LLVM
3132instructions. It provides only a few methods, but is a very commonly used
3133class. The primary data tracked by the <tt>Instruction</tt> class itself is the
3134opcode (instruction type) and the parent <a
3135href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
3136into. To represent a specific type of instruction, one of many subclasses of
3137<tt>Instruction</tt> are used.</p>
3138
3139<p> Because the <tt>Instruction</tt> class subclasses the <a
3140href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
3141way as for other <a href="#User"><tt>User</tt></a>s (with the
3142<tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
3143<tt>op_begin()</tt>/<tt>op_end()</tt> methods).</p> <p> An important file for
3144the <tt>Instruction</tt> class is the <tt>llvm/Instruction.def</tt> file. This
3145file contains some meta-data about the various different types of instructions
3146in LLVM. It describes the enum values that are used as opcodes (for example
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3147<tt>Instruction::Add</tt> and <tt>Instruction::ICmp</tt>), as well as the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3148concrete sub-classes of <tt>Instruction</tt> that implement the instruction (for
3149example <tt><a href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3150href="#CmpInst">CmpInst</a></tt>). Unfortunately, the use of macros in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3151this file confuses doxygen, so these enum values don't show up correctly in the
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3152<a href="/doxygen/classllvm_1_1Instruction.html">doxygen output</a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3153
3154</div>
3155
3156<!-- _______________________________________________________________________ -->
3157<div class="doc_subsubsection">
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3158 <a name="s_Instruction">Important Subclasses of the <tt>Instruction</tt>
3159 class</a>
3160</div>
3161<div class="doc_text">
3162 <ul>
3163 <li><tt><a name="BinaryOperator">BinaryOperator</a></tt>
3164 <p>This subclasses represents all two operand instructions whose operands
3165 must be the same type, except for the comparison instructions.</p></li>
3166 <li><tt><a name="CastInst">CastInst</a></tt>
3167 <p>This subclass is the parent of the 12 casting instructions. It provides
3168 common operations on cast instructions.</p>
3169 <li><tt><a name="CmpInst">CmpInst</a></tt>
3170 <p>This subclass respresents the two comparison instructions,
3171 <a href="LangRef.html#i_icmp">ICmpInst</a> (integer opreands), and
3172 <a href="LangRef.html#i_fcmp">FCmpInst</a> (floating point operands).</p>
3173 <li><tt><a name="TerminatorInst">TerminatorInst</a></tt>
3174 <p>This subclass is the parent of all terminator instructions (those which
3175 can terminate a block).</p>
3176 </ul>
3177 </div>
3178
3179<!-- _______________________________________________________________________ -->
3180<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3181 <a name="m_Instruction">Important Public Members of the <tt>Instruction</tt>
3182 class</a>
3183</div>
3184
3185<div class="doc_text">
3186
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3187<ul>
3188 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3189 <p>Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that
3190this <tt>Instruction</tt> is embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3191 <li><tt>bool mayWriteToMemory()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3192 <p>Returns true if the instruction writes to memory, i.e. it is a
3193 <tt>call</tt>,<tt>free</tt>,<tt>invoke</tt>, or <tt>store</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3194 <li><tt>unsigned getOpcode()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3195 <p>Returns the opcode for the <tt>Instruction</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3196 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3197 <p>Returns another instance of the specified instruction, identical
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3198in all ways to the original except that the instruction has no parent
3199(ie it's not embedded into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>),
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3200and it has no name</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3201</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3202
3203</div>
3204
3205<!-- ======================================================================= -->
3206<div class="doc_subsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3207 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3208</div>
3209
3210<div class="doc_text">
3211
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3212<p>Constant represents a base class for different types of constants. It
3213is subclassed by ConstantInt, ConstantArray, etc. for representing
3214the various types of Constants. <a href="#GlobalValue">GlobalValue</a> is also
3215a subclass, which represents the address of a global variable or function.
3216</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3217
3218</div>
3219
3220<!-- _______________________________________________________________________ -->
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3221<div class="doc_subsubsection">Important Subclasses of Constant </div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3222<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3223<ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3224 <li>ConstantInt : This subclass of Constant represents an integer constant of
3225 any width.
3226 <ul>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]3227 <li><tt>const APInt&amp; getValue() const</tt>: Returns the underlying
3228 value of this constant, an APInt value.</li>
3229 <li><tt>int64_t getSExtValue() const</tt>: Converts the underlying APInt
3230 value to an int64_t via sign extension. If the value (not the bit width)
3231 of the APInt is too large to fit in an int64_t, an assertion will result.
3232 For this reason, use of this method is discouraged.</li>
3233 <li><tt>uint64_t getZExtValue() const</tt>: Converts the underlying APInt
3234 value to a uint64_t via zero extension. IF the value (not the bit width)
3235 of the APInt is too large to fit in a uint64_t, an assertion will result.
Reid Spencer4474d872007-03-02 01:31:31 +0000[diff] [blame]3236 For this reason, use of this method is discouraged.</li>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]3237 <li><tt>static ConstantInt* get(const APInt&amp; Val)</tt>: Returns the
3238 ConstantInt object that represents the value provided by <tt>Val</tt>.
3239 The type is implied as the IntegerType that corresponds to the bit width
3240 of <tt>Val</tt>.</li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3241 <li><tt>static ConstantInt* get(const Type *Ty, uint64_t Val)</tt>:
3242 Returns the ConstantInt object that represents the value provided by
3243 <tt>Val</tt> for integer type <tt>Ty</tt>.</li>
3244 </ul>
3245 </li>
3246 <li>ConstantFP : This class represents a floating point constant.
3247 <ul>
3248 <li><tt>double getValue() const</tt>: Returns the underlying value of
3249 this constant. </li>
3250 </ul>
3251 </li>
3252 <li>ConstantArray : This represents a constant array.
3253 <ul>
3254 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
3255 a vector of component constants that makeup this array. </li>
3256 </ul>
3257 </li>
3258 <li>ConstantStruct : This represents a constant struct.
3259 <ul>
3260 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
3261 a vector of component constants that makeup this array. </li>
3262 </ul>
3263 </li>
3264 <li>GlobalValue : This represents either a global variable or a function. In
3265 either case, the value is a constant fixed address (after linking).
3266 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3267</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3268</div>
3269
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3270
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3271<!-- ======================================================================= -->
3272<div class="doc_subsection">
3273 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
3274</div>
3275
3276<div class="doc_text">
3277
3278<p><tt>#include "<a
3279href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt><br>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]3280doxygen info: <a href="/doxygen/classllvm_1_1GlobalValue.html">GlobalValue
3281Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3282Superclasses: <a href="#Constant"><tt>Constant</tt></a>,
3283<a href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3284
3285<p>Global values (<a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
3286href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
3287visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
3288Because they are visible at global scope, they are also subject to linking with
3289other globals defined in different translation units. To control the linking
3290process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
3291<tt>GlobalValue</tt>s know whether they have internal or external linkage, as
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]3292defined by the <tt>LinkageTypes</tt> enumeration.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3293
3294<p>If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
3295<tt>static</tt> in C), it is not visible to code outside the current translation
3296unit, and does not participate in linking. If it has external linkage, it is
3297visible to external code, and does participate in linking. In addition to
3298linkage information, <tt>GlobalValue</tt>s keep track of which <a
3299href="#Module"><tt>Module</tt></a> they are currently part of.</p>
3300
3301<p>Because <tt>GlobalValue</tt>s are memory objects, they are always referred to
3302by their <b>address</b>. As such, the <a href="#Type"><tt>Type</tt></a> of a
3303global is always a pointer to its contents. It is important to remember this
3304when using the <tt>GetElementPtrInst</tt> instruction because this pointer must
3305be dereferenced first. For example, if you have a <tt>GlobalVariable</tt> (a
3306subclass of <tt>GlobalValue)</tt> that is an array of 24 ints, type <tt>[24 x
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]3307i32]</tt>, then the <tt>GlobalVariable</tt> is a pointer to that array. Although
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3308the address of the first element of this array and the value of the
3309<tt>GlobalVariable</tt> are the same, they have different types. The
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]3310<tt>GlobalVariable</tt>'s type is <tt>[24 x i32]</tt>. The first element's type
3311is <tt>i32.</tt> Because of this, accessing a global value requires you to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3312dereference the pointer with <tt>GetElementPtrInst</tt> first, then its elements
3313can be accessed. This is explained in the <a href="LangRef.html#globalvars">LLVM
3314Language Reference Manual</a>.</p>
3315
3316</div>
3317
3318<!-- _______________________________________________________________________ -->
3319<div class="doc_subsubsection">
3320 <a name="m_GlobalValue">Important Public Members of the <tt>GlobalValue</tt>
3321 class</a>
3322</div>
3323
3324<div class="doc_text">
3325
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3326<ul>
3327 <li><tt>bool hasInternalLinkage() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3328 <tt>bool hasExternalLinkage() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3329 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt>
3330 <p> These methods manipulate the linkage characteristics of the <tt>GlobalValue</tt>.</p>
3331 <p> </p>
3332 </li>
3333 <li><tt><a href="#Module">Module</a> *getParent()</tt>
3334 <p> This returns the <a href="#Module"><tt>Module</tt></a> that the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3335GlobalValue is currently embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3336</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3337
3338</div>
3339
3340<!-- ======================================================================= -->
3341<div class="doc_subsection">
3342 <a name="Function">The <tt>Function</tt> class</a>
3343</div>
3344
3345<div class="doc_text">
3346
3347<p><tt>#include "<a
3348href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt><br> doxygen
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3349info: <a href="/doxygen/classllvm_1_1Function.html">Function Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3350Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
3351<a href="#Constant"><tt>Constant</tt></a>,
3352<a href="#User"><tt>User</tt></a>,
3353<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3354
3355<p>The <tt>Function</tt> class represents a single procedure in LLVM. It is
3356actually one of the more complex classes in the LLVM heirarchy because it must
3357keep track of a large amount of data. The <tt>Function</tt> class keeps track
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3358of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal
3359<a href="#Argument"><tt>Argument</tt></a>s, and a
3360<a href="#SymbolTable"><tt>SymbolTable</tt></a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3361
3362<p>The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most
3363commonly used part of <tt>Function</tt> objects. The list imposes an implicit
3364ordering of the blocks in the function, which indicate how the code will be
3365layed out by the backend. Additionally, the first <a
3366href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
3367<tt>Function</tt>. It is not legal in LLVM to explicitly branch to this initial
3368block. There are no implicit exit nodes, and in fact there may be multiple exit
3369nodes from a single <tt>Function</tt>. If the <a
3370href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
3371the <tt>Function</tt> is actually a function declaration: the actual body of the
3372function hasn't been linked in yet.</p>
3373
3374<p>In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
3375<tt>Function</tt> class also keeps track of the list of formal <a
3376href="#Argument"><tt>Argument</tt></a>s that the function receives. This
3377container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
3378nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
3379the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.</p>
3380
3381<p>The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used
3382LLVM feature that is only used when you have to look up a value by name. Aside
3383from that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used
3384internally to make sure that there are not conflicts between the names of <a
3385href="#Instruction"><tt>Instruction</tt></a>s, <a
3386href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
3387href="#Argument"><tt>Argument</tt></a>s in the function body.</p>
3388
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]3389<p>Note that <tt>Function</tt> is a <a href="#GlobalValue">GlobalValue</a>
3390and therefore also a <a href="#Constant">Constant</a>. The value of the function
3391is its address (after linking) which is guaranteed to be constant.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3392</div>
3393
3394<!-- _______________________________________________________________________ -->
3395<div class="doc_subsubsection">
3396 <a name="m_Function">Important Public Members of the <tt>Function</tt>
3397 class</a>
3398</div>
3399
3400<div class="doc_text">
3401
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3402<ul>
3403 <li><tt>Function(const </tt><tt><a href="#FunctionType">FunctionType</a>
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]3404 *Ty, LinkageTypes Linkage, const std::string &amp;N = "", Module* Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3405
3406 <p>Constructor used when you need to create new <tt>Function</tt>s to add
3407 the the program. The constructor must specify the type of the function to
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]3408 create and what type of linkage the function should have. The <a
3409 href="#FunctionType"><tt>FunctionType</tt></a> argument
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3410 specifies the formal arguments and return value for the function. The same
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3411 <a href="#FunctionType"><tt>FunctionType</tt></a> value can be used to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3412 create multiple functions. The <tt>Parent</tt> argument specifies the Module
3413 in which the function is defined. If this argument is provided, the function
3414 will automatically be inserted into that module's list of
3415 functions.</p></li>
3416
Chris Lattner62810e32008-11-25 18:34:50 +0000[diff] [blame]3417 <li><tt>bool isDeclaration()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3418
3419 <p>Return whether or not the <tt>Function</tt> has a body defined. If the
3420 function is "external", it does not have a body, and thus must be resolved
3421 by linking with a function defined in a different translation unit.</p></li>
3422
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3423 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3424 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3425
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]3426 <tt>begin()</tt>, <tt>end()</tt>
3427 <tt>size()</tt>, <tt>empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3428
3429 <p>These are forwarding methods that make it easy to access the contents of
3430 a <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
3431 list.</p></li>
3432
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3433 <li><tt>Function::BasicBlockListType &amp;getBasicBlockList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3434
3435 <p>Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This
3436 is necessary to use when you need to update the list or perform a complex
3437 action that doesn't have a forwarding method.</p></li>
3438
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3439 <li><tt>Function::arg_iterator</tt> - Typedef for the argument list
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3440iterator<br>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3441 <tt>Function::const_arg_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3442
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]3443 <tt>arg_begin()</tt>, <tt>arg_end()</tt>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3444 <tt>arg_size()</tt>, <tt>arg_empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3445
3446 <p>These are forwarding methods that make it easy to access the contents of
3447 a <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a>
3448 list.</p></li>
3449
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3450 <li><tt>Function::ArgumentListType &amp;getArgumentList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3451
3452 <p>Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
3453 necessary to use when you need to update the list or perform a complex
3454 action that doesn't have a forwarding method.</p></li>
3455
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3456 <li><tt><a href="#BasicBlock">BasicBlock</a> &amp;getEntryBlock()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3457
3458 <p>Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
3459 function. Because the entry block for the function is always the first
3460 block, this returns the first block of the <tt>Function</tt>.</p></li>
3461
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3462 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
3463 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3464
3465 <p>This traverses the <a href="#Type"><tt>Type</tt></a> of the
3466 <tt>Function</tt> and returns the return type of the function, or the <a
3467 href="#FunctionType"><tt>FunctionType</tt></a> of the actual
3468 function.</p></li>
3469
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3470 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3471
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3472 <p> Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3473 for this <tt>Function</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3474</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3475
3476</div>
3477
3478<!-- ======================================================================= -->
3479<div class="doc_subsection">
3480 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
3481</div>
3482
3483<div class="doc_text">
3484
3485<p><tt>#include "<a
3486href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt>
3487<br>
Tanya Lattnera3da7772004-06-22 08:02:25 +0000[diff] [blame]3488doxygen info: <a href="/doxygen/classllvm_1_1GlobalVariable.html">GlobalVariable
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3489 Class</a><br>
3490Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
3491<a href="#Constant"><tt>Constant</tt></a>,
3492<a href="#User"><tt>User</tt></a>,
3493<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3494
3495<p>Global variables are represented with the (suprise suprise)
3496<tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are also
3497subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such are
3498always referenced by their address (global values must live in memory, so their
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3499"name" refers to their constant address). See
3500<a href="#GlobalValue"><tt>GlobalValue</tt></a> for more on this. Global
3501variables may have an initial value (which must be a
3502<a href="#Constant"><tt>Constant</tt></a>), and if they have an initializer,
3503they may be marked as "constant" themselves (indicating that their contents
3504never change at runtime).</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3505</div>
3506
3507<!-- _______________________________________________________________________ -->
3508<div class="doc_subsubsection">
3509 <a name="m_GlobalVariable">Important Public Members of the
3510 <tt>GlobalVariable</tt> class</a>
3511</div>
3512
3513<div class="doc_text">
3514
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3515<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3516 <li><tt>GlobalVariable(const </tt><tt><a href="#Type">Type</a> *Ty, bool
3517 isConstant, LinkageTypes&amp; Linkage, <a href="#Constant">Constant</a>
3518 *Initializer = 0, const std::string &amp;Name = "", Module* Parent = 0)</tt>
3519
3520 <p>Create a new global variable of the specified type. If
3521 <tt>isConstant</tt> is true then the global variable will be marked as
3522 unchanging for the program. The Linkage parameter specifies the type of
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]3523 linkage (internal, external, weak, linkonce, appending) for the variable.
3524 If the linkage is InternalLinkage, WeakAnyLinkage, WeakODRLinkage,
3525 LinkOnceAnyLinkage or LinkOnceODRLinkage,&nbsp; then the resultant
3526 global variable will have internal linkage. AppendingLinkage concatenates
3527 together all instances (in different translation units) of the variable
3528 into a single variable but is only applicable to arrays. &nbsp;See
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3529 the <a href="LangRef.html#modulestructure">LLVM Language Reference</a> for
3530 further details on linkage types. Optionally an initializer, a name, and the
3531 module to put the variable into may be specified for the global variable as
3532 well.</p></li>
3533
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3534 <li><tt>bool isConstant() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3535
3536 <p>Returns true if this is a global variable that is known not to
3537 be modified at runtime.</p></li>
3538
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3539 <li><tt>bool hasInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3540
3541 <p>Returns true if this <tt>GlobalVariable</tt> has an intializer.</p></li>
3542
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3543 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3544
3545 <p>Returns the intial value for a <tt>GlobalVariable</tt>. It is not legal
3546 to call this method if there is no initializer.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3547</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3548
3549</div>
3550
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3551
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3552<!-- ======================================================================= -->
3553<div class="doc_subsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3554 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3555</div>
3556
3557<div class="doc_text">
3558
3559<p><tt>#include "<a
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3560href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt><br>
Stefanus Du Toit24e04112009-06-17 21:12:26 +0000[diff] [blame]3561doxygen info: <a href="/doxygen/classllvm_1_1BasicBlock.html">BasicBlock
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3562Class</a><br>
3563Superclass: <a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3564
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3565<p>This class represents a single entry multiple exit section of the code,
3566commonly known as a basic block by the compiler community. The
3567<tt>BasicBlock</tt> class maintains a list of <a
3568href="#Instruction"><tt>Instruction</tt></a>s, which form the body of the block.
3569Matching the language definition, the last element of this list of instructions
3570is always a terminator instruction (a subclass of the <a
3571href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).</p>
3572
3573<p>In addition to tracking the list of instructions that make up the block, the
3574<tt>BasicBlock</tt> class also keeps track of the <a
3575href="#Function"><tt>Function</tt></a> that it is embedded into.</p>
3576
3577<p>Note that <tt>BasicBlock</tt>s themselves are <a
3578href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
3579like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
3580<tt>label</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3581
3582</div>
3583
3584<!-- _______________________________________________________________________ -->
3585<div class="doc_subsubsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3586 <a name="m_BasicBlock">Important Public Members of the <tt>BasicBlock</tt>
3587 class</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3588</div>
3589
3590<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3591<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3592
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3593<li><tt>BasicBlock(const std::string &amp;Name = "", </tt><tt><a
3594 href="#Function">Function</a> *Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3595
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3596<p>The <tt>BasicBlock</tt> constructor is used to create new basic blocks for
3597insertion into a function. The constructor optionally takes a name for the new
3598block, and a <a href="#Function"><tt>Function</tt></a> to insert it into. If
3599the <tt>Parent</tt> parameter is specified, the new <tt>BasicBlock</tt> is
3600automatically inserted at the end of the specified <a
3601href="#Function"><tt>Function</tt></a>, if not specified, the BasicBlock must be
3602manually inserted into the <a href="#Function"><tt>Function</tt></a>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3603
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3604<li><tt>BasicBlock::iterator</tt> - Typedef for instruction list iterator<br>
3605<tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
3606<tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
3607<tt>size()</tt>, <tt>empty()</tt>
3608STL-style functions for accessing the instruction list.
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3609
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3610<p>These methods and typedefs are forwarding functions that have the same
3611semantics as the standard library methods of the same names. These methods
3612expose the underlying instruction list of a basic block in a way that is easy to
3613manipulate. To get the full complement of container operations (including
3614operations to update the list), you must use the <tt>getInstList()</tt>
3615method.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3616
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3617<li><tt>BasicBlock::InstListType &amp;getInstList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3618
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3619<p>This method is used to get access to the underlying container that actually
3620holds the Instructions. This method must be used when there isn't a forwarding
3621function in the <tt>BasicBlock</tt> class for the operation that you would like
3622to perform. Because there are no forwarding functions for "updating"
3623operations, you need to use this if you want to update the contents of a
3624<tt>BasicBlock</tt>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3625
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3626<li><tt><a href="#Function">Function</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3627
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3628<p> Returns a pointer to <a href="#Function"><tt>Function</tt></a> the block is
3629embedded into, or a null pointer if it is homeless.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3630
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3631<li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3632
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3633<p> Returns a pointer to the terminator instruction that appears at the end of
3634the <tt>BasicBlock</tt>. If there is no terminator instruction, or if the last
3635instruction in the block is not a terminator, then a null pointer is
3636returned.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3637
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3638</ul>
3639
3640</div>
3641
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3642
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3643<!-- ======================================================================= -->
3644<div class="doc_subsection">
3645 <a name="Argument">The <tt>Argument</tt> class</a>
3646</div>
3647
3648<div class="doc_text">
3649
3650<p>This subclass of Value defines the interface for incoming formal
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]3651arguments to a function. A Function maintains a list of its formal
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3652arguments. An argument has a pointer to the parent Function.</p>
3653
3654</div>
3655
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3656<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3657<hr>
3658<address>
3659 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]3660 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3661 <a href="http://validator.w3.org/check/referer"><img
Gabor Greifa9c0f2b2008-06-18 14:05:31 +0000[diff] [blame]3662 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01 Strict"></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3663
3664 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
3665 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]3666 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3667 Last modified: $Date$
3668</address>
3669
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3670</body>
3671</html>