Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / llvm / 81df089f860eebb4e1e44d089a441d62e4ff668b / . / docs / ProgrammersManual.html
blob: 49a76ee414767692a614a0b7e6bc69ba031eaa56 [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
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]11<h1>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]12 LLVM Programmer's Manual
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]13</h1>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]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>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]32 <li><a href="#string_apis">Passing strings (the <tt>StringRef</tt>
Benjamin Kramere15192b2009-08-05 15:42:44 +0000[diff] [blame]33and <tt>Twine</tt> classes)</a>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]34 <ul>
35 <li><a href="#StringRef">The <tt>StringRef</tt> class</a> </li>
36 <li><a href="#Twine">The <tt>Twine</tt> class</a> </li>
37 </ul>
Benjamin Kramere15192b2009-08-05 15:42:44 +0000[diff] [blame]38 </li>
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]39 <li><a href="#DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]40option</a>
41 <ul>
42 <li><a href="#DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt>
43and the <tt>-debug-only</tt> option</a> </li>
44 </ul>
45 </li>
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]46 <li><a href="#Statistic">The <tt>Statistic</tt> class &amp; <tt>-stats</tt>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]47option</a></li>
48<!--
49 <li>The <tt>InstVisitor</tt> template
50 <li>The general graph API
51-->
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]52 <li><a href="#ViewGraph">Viewing graphs while debugging code</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]53 </ul>
54 </li>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]55 <li><a href="#datastructure">Picking the Right Data Structure for a Task</a>
56 <ul>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]57 <li><a href="#ds_sequential">Sequential Containers (std::vector, std::list, etc)</a>
58 <ul>
Chris Lattner8ae42612011-04-05 23:18:20 +0000[diff] [blame]59 <li><a href="#dss_arrayref">llvm/ADT/ArrayRef.h</a></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]60 <li><a href="#dss_fixedarrays">Fixed Size Arrays</a></li>
61 <li><a href="#dss_heaparrays">Heap Allocated Arrays</a></li>
62 <li><a href="#dss_smallvector">"llvm/ADT/SmallVector.h"</a></li>
63 <li><a href="#dss_vector">&lt;vector&gt;</a></li>
64 <li><a href="#dss_deque">&lt;deque&gt;</a></li>
65 <li><a href="#dss_list">&lt;list&gt;</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]66 <li><a href="#dss_ilist">llvm/ADT/ilist.h</a></li>
Argyrios Kyrtzidis81df0892011-06-15 19:56:01 +0000[diff] [blame^]67 <li><a href="#dss_packedvector">llvm/ADT/PackedVector.h</a></li>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]68 <li><a href="#dss_other">Other Sequential Container Options</a></li>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]69 </ul></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]70 <li><a href="#ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
71 <ul>
72 <li><a href="#dss_sortedvectorset">A sorted 'vector'</a></li>
73 <li><a href="#dss_smallset">"llvm/ADT/SmallSet.h"</a></li>
74 <li><a href="#dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a></li>
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]75 <li><a href="#dss_denseset">"llvm/ADT/DenseSet.h"</a></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]76 <li><a href="#dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a></li>
77 <li><a href="#dss_set">&lt;set&gt;</a></li>
78 <li><a href="#dss_setvector">"llvm/ADT/SetVector.h"</a></li>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]79 <li><a href="#dss_uniquevector">"llvm/ADT/UniqueVector.h"</a></li>
80 <li><a href="#dss_otherset">Other Set-Like ContainerOptions</a></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]81 </ul></li>
Chris Lattnerf3692522007-02-03 19:51:56 +0000[diff] [blame]82 <li><a href="#ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
83 <ul>
84 <li><a href="#dss_sortedvectormap">A sorted 'vector'</a></li>
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]85 <li><a href="#dss_stringmap">"llvm/ADT/StringMap.h"</a></li>
Chris Lattnerf3692522007-02-03 19:51:56 +0000[diff] [blame]86 <li><a href="#dss_indexedmap">"llvm/ADT/IndexedMap.h"</a></li>
87 <li><a href="#dss_densemap">"llvm/ADT/DenseMap.h"</a></li>
Jeffrey Yasskin71a5c222009-10-22 22:11:22 +0000[diff] [blame]88 <li><a href="#dss_valuemap">"llvm/ADT/ValueMap.h"</a></li>
Jakob Stoklund Olesenaca0da62010-12-14 00:55:51 +0000[diff] [blame]89 <li><a href="#dss_intervalmap">"llvm/ADT/IntervalMap.h"</a></li>
Chris Lattnerf3692522007-02-03 19:51:56 +0000[diff] [blame]90 <li><a href="#dss_map">&lt;map&gt;</a></li>
Jakob Stoklund Olesen4359e5e2011-04-05 20:56:08 +0000[diff] [blame]91 <li><a href="#dss_inteqclasses">"llvm/ADT/IntEqClasses.h"</a></li>
Chris Lattnerf3692522007-02-03 19:51:56 +0000[diff] [blame]92 <li><a href="#dss_othermap">Other Map-Like Container Options</a></li>
93 </ul></li>
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]94 <li><a href="#ds_string">String-like containers</a>
Benjamin Kramere15192b2009-08-05 15:42:44 +0000[diff] [blame]95 <!--<ul>
96 todo
97 </ul>--></li>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]98 <li><a href="#ds_bit">BitVector-like containers</a>
99 <ul>
100 <li><a href="#dss_bitvector">A dense bitvector</a></li>
Dan Gohman5f7775c2010-01-05 18:24:00 +0000[diff] [blame]101 <li><a href="#dss_smallbitvector">A "small" dense bitvector</a></li>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]102 <li><a href="#dss_sparsebitvector">A sparse bitvector</a></li>
103 </ul></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]104 </ul>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]105 </li>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]106 <li><a href="#common">Helpful Hints for Common Operations</a>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]107 <ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]108 <li><a href="#inspection">Basic Inspection and Traversal Routines</a>
109 <ul>
110 <li><a href="#iterate_function">Iterating over the <tt>BasicBlock</tt>s
111in a <tt>Function</tt></a> </li>
112 <li><a href="#iterate_basicblock">Iterating over the <tt>Instruction</tt>s
113in a <tt>BasicBlock</tt></a> </li>
114 <li><a href="#iterate_institer">Iterating over the <tt>Instruction</tt>s
115in a <tt>Function</tt></a> </li>
116 <li><a href="#iterate_convert">Turning an iterator into a
117class pointer</a> </li>
118 <li><a href="#iterate_complex">Finding call sites: a more
119complex example</a> </li>
120 <li><a href="#calls_and_invokes">Treating calls and invokes
121the same way</a> </li>
122 <li><a href="#iterate_chains">Iterating over def-use &amp;
123use-def chains</a> </li>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]124 <li><a href="#iterate_preds">Iterating over predecessors &amp;
125successors of blocks</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]126 </ul>
127 </li>
128 <li><a href="#simplechanges">Making simple changes</a>
129 <ul>
130 <li><a href="#schanges_creating">Creating and inserting new
131 <tt>Instruction</tt>s</a> </li>
132 <li><a href="#schanges_deleting">Deleting <tt>Instruction</tt>s</a> </li>
133 <li><a href="#schanges_replacing">Replacing an <tt>Instruction</tt>
134with another <tt>Value</tt></a> </li>
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]135 <li><a href="#schanges_deletingGV">Deleting <tt>GlobalVariable</tt>s</a> </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]136 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]137 </li>
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]138 <li><a href="#create_types">How to Create Types</a></li>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]139<!--
140 <li>Working with the Control Flow Graph
141 <ul>
142 <li>Accessing predecessors and successors of a <tt>BasicBlock</tt>
143 <li>
144 <li>
145 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]146-->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]147 </ul>
148 </li>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]149
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]150 <li><a href="#threading">Threads and LLVM</a>
151 <ul>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]152 <li><a href="#startmultithreaded">Entering and Exiting Multithreaded Mode
153 </a></li>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]154 <li><a href="#shutdown">Ending execution with <tt>llvm_shutdown()</tt></a></li>
155 <li><a href="#managedstatic">Lazy initialization with <tt>ManagedStatic</tt></a></li>
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]156 <li><a href="#llvmcontext">Achieving Isolation with <tt>LLVMContext</tt></a></li>
Jeffrey Yasskin01eba392010-01-29 19:10:38 +0000[diff] [blame]157 <li><a href="#jitthreading">Threads and the JIT</a></li>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]158 </ul>
159 </li>
160
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]161 <li><a href="#advanced">Advanced Topics</a>
162 <ul>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]163 <li><a href="#TypeResolve">LLVM Type Resolution</a>
164 <ul>
165 <li><a href="#BuildRecType">Basic Recursive Type Construction</a></li>
166 <li><a href="#refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a></li>
167 <li><a href="#PATypeHolder">The PATypeHolder Class</a></li>
168 <li><a href="#AbstractTypeUser">The AbstractTypeUser Class</a></li>
169 </ul></li>
170
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]171 <li><a href="#SymbolTable">The <tt>ValueSymbolTable</tt> and <tt>TypeSymbolTable</tt> classes</a></li>
172 <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]173 </ul></li>
174
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]175 <li><a href="#coreclasses">The Core LLVM Class Hierarchy Reference</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]176 <ul>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]177 <li><a href="#Type">The <tt>Type</tt> class</a> </li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]178 <li><a href="#Module">The <tt>Module</tt> class</a></li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]179 <li><a href="#Value">The <tt>Value</tt> class</a>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]180 <ul>
181 <li><a href="#User">The <tt>User</tt> class</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]182 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]183 <li><a href="#Instruction">The <tt>Instruction</tt> class</a></li>
184 <li><a href="#Constant">The <tt>Constant</tt> class</a>
185 <ul>
186 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]187 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]188 <li><a href="#Function">The <tt>Function</tt> class</a></li>
189 <li><a href="#GlobalVariable">The <tt>GlobalVariable</tt> class</a></li>
190 </ul>
191 </li>
192 </ul>
193 </li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]194 </ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]195 </li>
196 <li><a href="#BasicBlock">The <tt>BasicBlock</tt> class</a></li>
197 <li><a href="#Argument">The <tt>Argument</tt> class</a></li>
198 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]199 </li>
200 </ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]201 </li>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]202</ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]203
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]204<div class="doc_author">
205 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
Chris Lattner94c43592004-05-26 16:52:55 +0000[diff] [blame]206 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>,
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]207 <a href="mailto:ggreif@gmail.com">Gabor Greif</a>,
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]208 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a>,
209 <a href="mailto:rspencer@x10sys.com">Reid Spencer</a> and
210 <a href="mailto:owen@apple.com">Owen Anderson</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]211</div>
212
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]213<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]214<h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]215 <a name="introduction">Introduction </a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]216</h2>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]217<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]218
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]219<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]220
221<p>This document is meant to highlight some of the important classes and
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]222interfaces available in the LLVM source-base. This manual is not
223intended to explain what LLVM is, how it works, and what LLVM code looks
224like. It assumes that you know the basics of LLVM and are interested
225in writing transformations or otherwise analyzing or manipulating the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]226code.</p>
227
228<p>This document should get you oriented so that you can find your
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]229way in the continuously growing source code that makes up the LLVM
230infrastructure. Note that this manual is not intended to serve as a
231replacement for reading the source code, so if you think there should be
232a method in one of these classes to do something, but it's not listed,
233check the source. Links to the <a href="/doxygen/">doxygen</a> sources
234are provided to make this as easy as possible.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]235
236<p>The first section of this document describes general information that is
237useful to know when working in the LLVM infrastructure, and the second describes
238the Core LLVM classes. In the future this manual will be extended with
239information describing how to use extension libraries, such as dominator
240information, CFG traversal routines, and useful utilities like the <tt><a
241href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.</p>
242
243</div>
244
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]245<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]246<h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]247 <a name="general">General Information</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]248</h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]249<!-- *********************************************************************** -->
250
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]251<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]252
253<p>This section contains general information that is useful if you are working
254in the LLVM source-base, but that isn't specific to any particular API.</p>
255
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]256<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]257<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]258 <a name="stl">The C++ Standard Template Library</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]259</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]260
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]261<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]262
263<p>LLVM makes heavy use of the C++ Standard Template Library (STL),
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]264perhaps much more than you are used to, or have seen before. Because of
265this, you might want to do a little background reading in the
266techniques used and capabilities of the library. There are many good
267pages that discuss the STL, and several books on the subject that you
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]268can get, so it will not be discussed in this document.</p>
269
270<p>Here are some useful links:</p>
271
272<ol>
273
Nick Lewyckyea1fe2c2010-10-09 21:12:29 +0000[diff] [blame]274<li><a href="http://www.dinkumware.com/manuals/#Standard C++ Library">Dinkumware
275C++ Library reference</a> - an excellent reference for the STL and other parts
276of the standard C++ library.</li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]277
278<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]279O'Reilly book in the making. It has a decent Standard Library
280Reference that rivals Dinkumware's, and is unfortunately no longer free since the
281book has been published.</li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]282
283<li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
284Questions</a></li>
285
286<li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
287Contains a useful <a
288href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
289STL</a>.</li>
290
291<li><a href="http://www.research.att.com/%7Ebs/C++.html">Bjarne Stroustrup's C++
292Page</a></li>
293
Tanya Lattner79445ba2004-12-08 18:34:56 +0000[diff] [blame]294<li><a href="http://64.78.49.204/">
Reid Spencer096603a2004-05-26 08:41:35 +0000[diff] [blame]295Bruce Eckel's Thinking in C++, 2nd ed. Volume 2 Revision 4.0 (even better, get
296the book).</a></li>
297
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]298</ol>
299
300<p>You are also encouraged to take a look at the <a
301href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
302to write maintainable code more than where to put your curly braces.</p>
303
304</div>
305
306<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]307<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]308 <a name="stl">Other useful references</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]309</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]310
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]311<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]312
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]313<ol>
Misha Brukmana0f71e42004-06-18 18:39:00 +0000[diff] [blame]314<li><a href="http://www.fortran-2000.com/ArnaudRecipes/sharedlib.html">Using
315static and shared libraries across platforms</a></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]316</ol>
317
318</div>
319
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]320</div>
321
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]322<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]323<h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]324 <a name="apis">Important and useful LLVM APIs</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]325</h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]326<!-- *********************************************************************** -->
327
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]328<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]329
330<p>Here we highlight some LLVM APIs that are generally useful and good to
331know about when writing transformations.</p>
332
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]333<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]334<h3>
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]335 <a name="isa">The <tt>isa&lt;&gt;</tt>, <tt>cast&lt;&gt;</tt> and
336 <tt>dyn_cast&lt;&gt;</tt> templates</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]337</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]338
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]339<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]340
341<p>The LLVM source-base makes extensive use of a custom form of RTTI.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]342These templates have many similarities to the C++ <tt>dynamic_cast&lt;&gt;</tt>
343operator, but they don't have some drawbacks (primarily stemming from
344the fact that <tt>dynamic_cast&lt;&gt;</tt> only works on classes that
345have a v-table). Because they are used so often, you must know what they
346do and how they work. All of these templates are defined in the <a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]347 href="/doxygen/Casting_8h-source.html"><tt>llvm/Support/Casting.h</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]348file (note that you very rarely have to include this file directly).</p>
349
350<dl>
351 <dt><tt>isa&lt;&gt;</tt>: </dt>
352
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]353 <dd><p>The <tt>isa&lt;&gt;</tt> operator works exactly like the Java
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]354 "<tt>instanceof</tt>" operator. It returns true or false depending on whether
355 a reference or pointer points to an instance of the specified class. This can
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]356 be very useful for constraint checking of various sorts (example below).</p>
357 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]358
359 <dt><tt>cast&lt;&gt;</tt>: </dt>
360
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]361 <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]362 converts a pointer or reference from a base class to a derived class, causing
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]363 an assertion failure if it is not really an instance of the right type. This
364 should be used in cases where you have some information that makes you believe
365 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]366 and <tt>cast&lt;&gt;</tt> template is:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]367
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]368<div class="doc_code">
369<pre>
370static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
371 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))
372 return true;
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]373
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]374 // <i>Otherwise, it must be an instruction...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]375 return !L-&gt;contains(cast&lt;<a href="#Instruction">Instruction</a>&gt;(V)-&gt;getParent());
376}
377</pre>
378</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]379
380 <p>Note that you should <b>not</b> use an <tt>isa&lt;&gt;</tt> test followed
381 by a <tt>cast&lt;&gt;</tt>, for that use the <tt>dyn_cast&lt;&gt;</tt>
382 operator.</p>
383
384 </dd>
385
386 <dt><tt>dyn_cast&lt;&gt;</tt>:</dt>
387
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]388 <dd><p>The <tt>dyn_cast&lt;&gt;</tt> operator is a "checking cast" operation.
389 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]390 pointer to it (this operator does not work with references). If the operand is
391 not of the correct type, a null pointer is returned. Thus, this works very
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]392 much like the <tt>dynamic_cast&lt;&gt;</tt> operator in C++, and should be
393 used in the same circumstances. Typically, the <tt>dyn_cast&lt;&gt;</tt>
394 operator is used in an <tt>if</tt> statement or some other flow control
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]395 statement like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]396
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]397<div class="doc_code">
398<pre>
399if (<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]400 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]401}
402</pre>
403</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]404
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]405 <p>This form of the <tt>if</tt> statement effectively combines together a call
406 to <tt>isa&lt;&gt;</tt> and a call to <tt>cast&lt;&gt;</tt> into one
407 statement, which is very convenient.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]408
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]409 <p>Note that the <tt>dyn_cast&lt;&gt;</tt> operator, like C++'s
410 <tt>dynamic_cast&lt;&gt;</tt> or Java's <tt>instanceof</tt> operator, can be
411 abused. In particular, you should not use big chained <tt>if/then/else</tt>
412 blocks to check for lots of different variants of classes. If you find
413 yourself wanting to do this, it is much cleaner and more efficient to use the
414 <tt>InstVisitor</tt> class to dispatch over the instruction type directly.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]415
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]416 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]417
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]418 <dt><tt>cast_or_null&lt;&gt;</tt>: </dt>
419
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]420 <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]421 <tt>cast&lt;&gt;</tt> operator, except that it allows for a null pointer as an
422 argument (which it then propagates). This can sometimes be useful, allowing
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]423 you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]424
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]425 <dt><tt>dyn_cast_or_null&lt;&gt;</tt>: </dt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]426
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]427 <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]428 <tt>dyn_cast&lt;&gt;</tt> operator, except that it allows for a null pointer
429 as an argument (which it then propagates). This can sometimes be useful,
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]430 allowing you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]431
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]432</dl>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]433
434<p>These five templates can be used with any classes, whether they have a
435v-table or not. To add support for these templates, you simply need to add
436<tt>classof</tt> static methods to the class you are interested casting
437to. Describing this is currently outside the scope of this document, but there
438are lots of examples in the LLVM source base.</p>
439
440</div>
441
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]442
443<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]444<h3>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]445 <a name="string_apis">Passing strings (the <tt>StringRef</tt>
446and <tt>Twine</tt> classes)</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]447</h3>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]448
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]449<div>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]450
451<p>Although LLVM generally does not do much string manipulation, we do have
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]452several important APIs which take strings. Two important examples are the
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]453Value class -- which has names for instructions, functions, etc. -- and the
454StringMap class which is used extensively in LLVM and Clang.</p>
455
456<p>These are generic classes, and they need to be able to accept strings which
457may have embedded null characters. Therefore, they cannot simply take
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]458a <tt>const char *</tt>, and taking a <tt>const std::string&amp;</tt> requires
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]459clients to perform a heap allocation which is usually unnecessary. Instead,
Benjamin Kramer38e59892010-07-14 22:38:02 +0000[diff] [blame]460many LLVM APIs use a <tt>StringRef</tt> or a <tt>const Twine&amp;</tt> for
461passing strings efficiently.</p>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]462
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]463<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]464<h4>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]465 <a name="StringRef">The <tt>StringRef</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]466</h4>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]467
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]468<div>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]469
470<p>The <tt>StringRef</tt> data type represents a reference to a constant string
471(a character array and a length) and supports the common operations available
472on <tt>std:string</tt>, but does not require heap allocation.</p>
473
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]474<p>It can be implicitly constructed using a C style null-terminated string,
475an <tt>std::string</tt>, or explicitly with a character pointer and length.
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]476For example, the <tt>StringRef</tt> find function is declared as:</p>
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]477
Benjamin Kramer38e59892010-07-14 22:38:02 +0000[diff] [blame]478<pre class="doc_code">
479 iterator find(StringRef Key);
480</pre>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]481
482<p>and clients can call it using any one of:</p>
483
Benjamin Kramer38e59892010-07-14 22:38:02 +0000[diff] [blame]484<pre class="doc_code">
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]485 Map.find("foo"); <i>// Lookup "foo"</i>
486 Map.find(std::string("bar")); <i>// Lookup "bar"</i>
487 Map.find(StringRef("\0baz", 4)); <i>// Lookup "\0baz"</i>
488</pre>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]489
490<p>Similarly, APIs which need to return a string may return a <tt>StringRef</tt>
491instance, which can be used directly or converted to an <tt>std::string</tt>
492using the <tt>str</tt> member function. See
493"<tt><a href="/doxygen/classllvm_1_1StringRef_8h-source.html">llvm/ADT/StringRef.h</a></tt>"
494for more information.</p>
495
496<p>You should rarely use the <tt>StringRef</tt> class directly, because it contains
497pointers to external memory it is not generally safe to store an instance of the
Benjamin Kramer38e59892010-07-14 22:38:02 +0000[diff] [blame]498class (unless you know that the external storage will not be freed). StringRef is
499small and pervasive enough in LLVM that it should always be passed by value.</p>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]500
501</div>
502
503<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]504<h4>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]505 <a name="Twine">The <tt>Twine</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]506</h4>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]507
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]508<div>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]509
510<p>The <tt>Twine</tt> class is an efficient way for APIs to accept concatenated
511strings. For example, a common LLVM paradigm is to name one instruction based on
512the name of another instruction with a suffix, for example:</p>
513
514<div class="doc_code">
515<pre>
516 New = CmpInst::Create(<i>...</i>, SO->getName() + ".cmp");
517</pre>
518</div>
519
520<p>The <tt>Twine</tt> class is effectively a
521lightweight <a href="http://en.wikipedia.org/wiki/Rope_(computer_science)">rope</a>
522which points to temporary (stack allocated) objects. Twines can be implicitly
523constructed as the result of the plus operator applied to strings (i.e., a C
524strings, an <tt>std::string</tt>, or a <tt>StringRef</tt>). The twine delays the
Dan Gohmancf0c9bc2010-02-25 23:51:27 +0000[diff] [blame]525actual concatenation of strings until it is actually required, at which point
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]526it can be efficiently rendered directly into a character array. This avoids
527unnecessary heap allocation involved in constructing the temporary results of
528string concatenation. See
529"<tt><a href="/doxygen/classllvm_1_1Twine_8h-source.html">llvm/ADT/Twine.h</a></tt>"
Benjamin Kramere15192b2009-08-05 15:42:44 +0000[diff] [blame]530for more information.</p>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]531
532<p>As with a <tt>StringRef</tt>, <tt>Twine</tt> objects point to external memory
533and should almost never be stored or mentioned directly. They are intended
534solely for use when defining a function which should be able to efficiently
535accept concatenated strings.</p>
536
537</div>
538
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]539</div>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]540
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]541<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]542<h3>
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]543 <a name="DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt> option</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]544</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]545
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]546<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]547
548<p>Often when working on your pass you will put a bunch of debugging printouts
549and other code into your pass. After you get it working, you want to remove
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]550it, 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]551across).</p>
552
553<p> Naturally, because of this, you don't want to delete the debug printouts,
554but you don't want them to always be noisy. A standard compromise is to comment
555them out, allowing you to enable them if you need them in the future.</p>
556
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]557<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]558file provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to
559this problem. Basically, you can put arbitrary code into the argument of the
560<tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
561tool) is run with the '<tt>-debug</tt>' command line argument:</p>
562
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]563<div class="doc_code">
564<pre>
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]565DEBUG(errs() &lt;&lt; "I am here!\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]566</pre>
567</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]568
569<p>Then you can run your pass like this:</p>
570
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]571<div class="doc_code">
572<pre>
573$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]574<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]575$ opt &lt; a.bc &gt; /dev/null -mypass -debug
576I am here!
577</pre>
578</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]579
580<p>Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you
581to not have to create "yet another" command line option for the debug output for
582your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
583so they do not cause a performance impact at all (for the same reason, they
584should also not contain side-effects!).</p>
585
586<p>One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
587enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
588"<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
589program hasn't been started yet, you can always just run it with
590<tt>-debug</tt>.</p>
591
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]592<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]593<h4>
Chris Lattnerc9151082005-04-26 22:57:07 +0000[diff] [blame]594 <a name="DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt> and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]595 the <tt>-debug-only</tt> option</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]596</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]597
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]598<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]599
600<p>Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
601just turns on <b>too much</b> information (such as when working on the code
602generator). If you want to enable debug information with more fine-grained
603control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
604option as follows:</p>
605
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]606<div class="doc_code">
607<pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]608#undef DEBUG_TYPE
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]609DEBUG(errs() &lt;&lt; "No debug type\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]610#define DEBUG_TYPE "foo"
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]611DEBUG(errs() &lt;&lt; "'foo' debug type\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]612#undef DEBUG_TYPE
613#define DEBUG_TYPE "bar"
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]614DEBUG(errs() &lt;&lt; "'bar' debug type\n"));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]615#undef DEBUG_TYPE
616#define DEBUG_TYPE ""
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]617DEBUG(errs() &lt;&lt; "No debug type (2)\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]618</pre>
619</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]620
621<p>Then you can run your pass like this:</p>
622
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]623<div class="doc_code">
624<pre>
625$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]626<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]627$ opt &lt; a.bc &gt; /dev/null -mypass -debug
628No debug type
629'foo' debug type
630'bar' debug type
631No debug type (2)
632$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=foo
633'foo' debug type
634$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=bar
635'bar' debug type
636</pre>
637</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]638
639<p>Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of
640a 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]641you <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]642<tt>#undef</tt>'s). Also, you should use names more meaningful than "foo" and
643"bar", because there is no system in place to ensure that names do not
644conflict. If two different modules use the same string, they will all be turned
645on when the name is specified. This allows, for example, all debug information
646for instruction scheduling to be enabled with <tt>-debug-type=InstrSched</tt>,
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]647even if the source lives in multiple files.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]648
Daniel Dunbarc3c92392009-08-07 23:48:59 +0000[diff] [blame]649<p>The <tt>DEBUG_WITH_TYPE</tt> macro is also available for situations where you
650would like to set <tt>DEBUG_TYPE</tt>, but only for one specific <tt>DEBUG</tt>
651statement. It takes an additional first parameter, which is the type to use. For
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]652example, the preceding example could be written as:</p>
Daniel Dunbarc3c92392009-08-07 23:48:59 +0000[diff] [blame]653
654
655<div class="doc_code">
656<pre>
657DEBUG_WITH_TYPE("", errs() &lt;&lt; "No debug type\n");
658DEBUG_WITH_TYPE("foo", errs() &lt;&lt; "'foo' debug type\n");
659DEBUG_WITH_TYPE("bar", errs() &lt;&lt; "'bar' debug type\n"));
660DEBUG_WITH_TYPE("", errs() &lt;&lt; "No debug type (2)\n");
661</pre>
662</div>
663
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]664</div>
665
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]666</div>
667
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]668<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]669<h3>
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]670 <a name="Statistic">The <tt>Statistic</tt> class &amp; <tt>-stats</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]671 option</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]672</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]673
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]674<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]675
676<p>The "<tt><a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]677href="/doxygen/Statistic_8h-source.html">llvm/ADT/Statistic.h</a></tt>" file
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]678provides a class named <tt>Statistic</tt> that is used as a unified way to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]679keep track of what the LLVM compiler is doing and how effective various
680optimizations are. It is useful to see what optimizations are contributing to
681making a particular program run faster.</p>
682
683<p>Often you may run your pass on some big program, and you're interested to see
684how many times it makes a certain transformation. Although you can do this with
685hand 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]686for big programs. Using the <tt>Statistic</tt> class makes it very easy to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]687keep track of this information, and the calculated information is presented in a
688uniform manner with the rest of the passes being executed.</p>
689
690<p>There are many examples of <tt>Statistic</tt> uses, but the basics of using
691it are as follows:</p>
692
693<ol>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]694 <li><p>Define your statistic like this:</p>
695
696<div class="doc_code">
697<pre>
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]698#define <a href="#DEBUG_TYPE">DEBUG_TYPE</a> "mypassname" <i>// This goes before any #includes.</i>
699STATISTIC(NumXForms, "The # of times I did stuff");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]700</pre>
701</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]702
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]703 <p>The <tt>STATISTIC</tt> macro defines a static variable, whose name is
704 specified by the first argument. The pass name is taken from the DEBUG_TYPE
705 macro, and the description is taken from the second argument. The variable
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]706 defined ("NumXForms" in this case) acts like an unsigned integer.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]707
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]708 <li><p>Whenever you make a transformation, bump the counter:</p>
709
710<div class="doc_code">
711<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]712++NumXForms; // <i>I did stuff!</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]713</pre>
714</div>
715
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]716 </li>
717 </ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]718
719 <p>That's all you have to do. To get '<tt>opt</tt>' to print out the
720 statistics gathered, use the '<tt>-stats</tt>' option:</p>
721
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]722<div class="doc_code">
723<pre>
724$ opt -stats -mypassname &lt; program.bc &gt; /dev/null
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]725<i>... statistics output ...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]726</pre>
727</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]728
Reid Spencer6b6c73e2007-02-09 16:00:28 +0000[diff] [blame]729 <p> When running <tt>opt</tt> on a C file from the SPEC benchmark
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]730suite, it gives a report that looks like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]731
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]732<div class="doc_code">
733<pre>
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]734 7646 bitcodewriter - Number of normal instructions
735 725 bitcodewriter - Number of oversized instructions
736 129996 bitcodewriter - Number of bitcode bytes written
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]737 2817 raise - Number of insts DCEd or constprop'd
738 3213 raise - Number of cast-of-self removed
739 5046 raise - Number of expression trees converted
740 75 raise - Number of other getelementptr's formed
741 138 raise - Number of load/store peepholes
742 42 deadtypeelim - Number of unused typenames removed from symtab
743 392 funcresolve - Number of varargs functions resolved
744 27 globaldce - Number of global variables removed
745 2 adce - Number of basic blocks removed
746 134 cee - Number of branches revectored
747 49 cee - Number of setcc instruction eliminated
748 532 gcse - Number of loads removed
749 2919 gcse - Number of instructions removed
750 86 indvars - Number of canonical indvars added
751 87 indvars - Number of aux indvars removed
752 25 instcombine - Number of dead inst eliminate
753 434 instcombine - Number of insts combined
754 248 licm - Number of load insts hoisted
755 1298 licm - Number of insts hoisted to a loop pre-header
756 3 licm - Number of insts hoisted to multiple loop preds (bad, no loop pre-header)
757 75 mem2reg - Number of alloca's promoted
758 1444 cfgsimplify - Number of blocks simplified
759</pre>
760</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]761
762<p>Obviously, with so many optimizations, having a unified framework for this
763stuff is very nice. Making your pass fit well into the framework makes it more
764maintainable and useful.</p>
765
766</div>
767
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]768<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]769<h3>
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]770 <a name="ViewGraph">Viewing graphs while debugging code</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]771</h3>
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]772
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]773<div>
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]774
775<p>Several of the important data structures in LLVM are graphs: for example
776CFGs made out of LLVM <a href="#BasicBlock">BasicBlock</a>s, CFGs made out of
777LLVM <a href="CodeGenerator.html#machinebasicblock">MachineBasicBlock</a>s, and
778<a href="CodeGenerator.html#selectiondag_intro">Instruction Selection
779DAGs</a>. In many cases, while debugging various parts of the compiler, it is
780nice to instantly visualize these graphs.</p>
781
782<p>LLVM provides several callbacks that are available in a debug build to do
783exactly that. If you call the <tt>Function::viewCFG()</tt> method, for example,
784the current LLVM tool will pop up a window containing the CFG for the function
785where each basic block is a node in the graph, and each node contains the
786instructions in the block. Similarly, there also exists
787<tt>Function::viewCFGOnly()</tt> (does not include the instructions), the
788<tt>MachineFunction::viewCFG()</tt> and <tt>MachineFunction::viewCFGOnly()</tt>,
789and the <tt>SelectionDAG::viewGraph()</tt> methods. Within GDB, for example,
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]790you can usually use something like <tt>call DAG.viewGraph()</tt> to pop
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]791up a window. Alternatively, you can sprinkle calls to these functions in your
792code in places you want to debug.</p>
793
794<p>Getting this to work requires a small amount of configuration. On Unix
795systems with X11, install the <a href="http://www.graphviz.org">graphviz</a>
796toolkit, and make sure 'dot' and 'gv' are in your path. If you are running on
797Mac OS/X, download and install the Mac OS/X <a
798href="http://www.pixelglow.com/graphviz/">Graphviz program</a>, and add
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]799<tt>/Applications/Graphviz.app/Contents/MacOS/</tt> (or wherever you install
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]800it) to your path. Once in your system and path are set up, rerun the LLVM
801configure script and rebuild LLVM to enable this functionality.</p>
802
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]803<p><tt>SelectionDAG</tt> has been extended to make it easier to locate
804<i>interesting</i> nodes in large complex graphs. From gdb, if you
805<tt>call DAG.setGraphColor(<i>node</i>, "<i>color</i>")</tt>, then the
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]806next <tt>call DAG.viewGraph()</tt> would highlight the node in the
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]807specified color (choices of colors can be found at <a
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]808href="http://www.graphviz.org/doc/info/colors.html">colors</a>.) More
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]809complex node attributes can be provided with <tt>call
810DAG.setGraphAttrs(<i>node</i>, "<i>attributes</i>")</tt> (choices can be
811found at <a href="http://www.graphviz.org/doc/info/attrs.html">Graph
812Attributes</a>.) If you want to restart and clear all the current graph
813attributes, then you can <tt>call DAG.clearGraphAttrs()</tt>. </p>
814
Chris Lattner83f94672011-06-13 15:59:35 +0000[diff] [blame]815<p>Note that graph visualization features are compiled out of Release builds
816to reduce file size. This means that you need a Debug+Asserts or
817Release+Asserts build to use these features.</p>
818
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]819</div>
820
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]821</div>
822
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]823<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]824<h2>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]825 <a name="datastructure">Picking the Right Data Structure for a Task</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]826</h2>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]827<!-- *********************************************************************** -->
828
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]829<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]830
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]831<p>LLVM has a plethora of data structures in the <tt>llvm/ADT/</tt> directory,
832 and we commonly use STL data structures. This section describes the trade-offs
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]833 you should consider when you pick one.</p>
834
835<p>
836The first step is a choose your own adventure: do you want a sequential
837container, a set-like container, or a map-like container? The most important
838thing when choosing a container is the algorithmic properties of how you plan to
839access the container. Based on that, you should use:</p>
840
841<ul>
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]842<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]843 of an value based on another value. Map-like containers also support
844 efficient queries for containment (whether a key is in the map). Map-like
845 containers generally do not support efficient reverse mapping (values to
846 keys). If you need that, use two maps. Some map-like containers also
847 support efficient iteration through the keys in sorted order. Map-like
848 containers are the most expensive sort, only use them if you need one of
849 these capabilities.</li>
850
851<li>a <a href="#ds_set">set-like</a> container if you need to put a bunch of
852 stuff into a container that automatically eliminates duplicates. Some
853 set-like containers support efficient iteration through the elements in
854 sorted order. Set-like containers are more expensive than sequential
855 containers.
856</li>
857
858<li>a <a href="#ds_sequential">sequential</a> container provides
859 the most efficient way to add elements and keeps track of the order they are
860 added to the collection. They permit duplicates and support efficient
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]861 iteration, but do not support efficient look-up based on a key.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]862</li>
863
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]864<li>a <a href="#ds_string">string</a> container is a specialized sequential
865 container or reference structure that is used for character or byte
866 arrays.</li>
867
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]868<li>a <a href="#ds_bit">bit</a> container provides an efficient way to store and
869 perform set operations on sets of numeric id's, while automatically
870 eliminating duplicates. Bit containers require a maximum of 1 bit for each
871 identifier you want to store.
872</li>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]873</ul>
874
875<p>
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]876Once the proper category of container is determined, you can fine tune the
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]877memory use, constant factors, and cache behaviors of access by intelligently
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]878picking a member of the category. Note that constant factors and cache behavior
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]879can be a big deal. If you have a vector that usually only contains a few
880elements (but could contain many), for example, it's much better to use
881<a href="#dss_smallvector">SmallVector</a> than <a href="#dss_vector">vector</a>
882. Doing so avoids (relatively) expensive malloc/free calls, which dwarf the
883cost of adding the elements to the container. </p>
884
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]885<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]886<h3>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]887 <a name="ds_sequential">Sequential Containers (std::vector, std::list, etc)</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]888</h3>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]889
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]890<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]891There are a variety of sequential containers available for you, based on your
892needs. Pick the first in this section that will do what you want.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]893
894<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]895<h4>
Chris Lattner8ae42612011-04-05 23:18:20 +0000[diff] [blame]896 <a name="dss_arrayref">llvm/ADT/ArrayRef.h</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]897</h4>
Chris Lattner8ae42612011-04-05 23:18:20 +0000[diff] [blame]898
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]899<div>
Chris Lattner8ae42612011-04-05 23:18:20 +0000[diff] [blame]900<p>The llvm::ArrayRef class is the preferred class to use in an interface that
901 accepts a sequential list of elements in memory and just reads from them. By
902 taking an ArrayRef, the API can be passed a fixed size array, an std::vector,
903 an llvm::SmallVector and anything else that is contiguous in memory.
904</p>
905</div>
906
907
908
909<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]910<h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]911 <a name="dss_fixedarrays">Fixed Size Arrays</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]912</h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]913
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]914<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]915<p>Fixed size arrays are very simple and very fast. They are good if you know
916exactly how many elements you have, or you have a (low) upper bound on how many
917you have.</p>
918</div>
919
920<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]921<h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]922 <a name="dss_heaparrays">Heap Allocated Arrays</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]923</h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]924
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]925<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]926<p>Heap allocated arrays (new[] + delete[]) are also simple. They are good if
927the number of elements is variable, if you know how many elements you will need
928before the array is allocated, and if the array is usually large (if not,
929consider a <a href="#dss_smallvector">SmallVector</a>). The cost of a heap
930allocated array is the cost of the new/delete (aka malloc/free). Also note that
931if you are allocating an array of a type with a constructor, the constructor and
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]932destructors will be run for every element in the array (re-sizable vectors only
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]933construct those elements actually used).</p>
934</div>
935
936<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]937<h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]938 <a name="dss_smallvector">"llvm/ADT/SmallVector.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]939</h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]940
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]941<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]942<p><tt>SmallVector&lt;Type, N&gt;</tt> is a simple class that looks and smells
943just like <tt>vector&lt;Type&gt;</tt>:
944it supports efficient iteration, lays out elements in memory order (so you can
945do pointer arithmetic between elements), supports efficient push_back/pop_back
946operations, supports efficient random access to its elements, etc.</p>
947
948<p>The advantage of SmallVector is that it allocates space for
949some number of elements (N) <b>in the object itself</b>. Because of this, if
950the SmallVector is dynamically smaller than N, no malloc is performed. This can
951be a big win in cases where the malloc/free call is far more expensive than the
952code that fiddles around with the elements.</p>
953
954<p>This is good for vectors that are "usually small" (e.g. the number of
955predecessors/successors of a block is usually less than 8). On the other hand,
956this makes the size of the SmallVector itself large, so you don't want to
957allocate lots of them (doing so will waste a lot of space). As such,
958SmallVectors are most useful when on the stack.</p>
959
960<p>SmallVector also provides a nice portable and efficient replacement for
961<tt>alloca</tt>.</p>
962
963</div>
964
965<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]966<h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]967 <a name="dss_vector">&lt;vector&gt;</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]968</h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]969
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]970<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]971<p>
972std::vector is well loved and respected. It is useful when SmallVector isn't:
973when the size of the vector is often large (thus the small optimization will
974rarely be a benefit) or if you will be allocating many instances of the vector
975itself (which would waste space for elements that aren't in the container).
976vector is also useful when interfacing with code that expects vectors :).
977</p>
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]978
979<p>One worthwhile note about std::vector: avoid code like this:</p>
980
981<div class="doc_code">
982<pre>
983for ( ... ) {
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]984 std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]985 use V;
986}
987</pre>
988</div>
989
990<p>Instead, write this as:</p>
991
992<div class="doc_code">
993<pre>
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]994std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]995for ( ... ) {
996 use V;
997 V.clear();
998}
999</pre>
1000</div>
1001
1002<p>Doing so will save (at least) one heap allocation and free per iteration of
1003the loop.</p>
1004
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1005</div>
1006
1007<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1008<h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1009 <a name="dss_deque">&lt;deque&gt;</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1010</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1011
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1012<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1013<p>std::deque is, in some senses, a generalized version of std::vector. Like
1014std::vector, it provides constant time random access and other similar
1015properties, but it also provides efficient access to the front of the list. It
1016does not guarantee continuity of elements within memory.</p>
1017
1018<p>In exchange for this extra flexibility, std::deque has significantly higher
1019constant factor costs than std::vector. If possible, use std::vector or
1020something cheaper.</p>
1021</div>
1022
1023<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1024<h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1025 <a name="dss_list">&lt;list&gt;</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1026</h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1027
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1028<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1029<p>std::list is an extremely inefficient class that is rarely useful.
1030It performs a heap allocation for every element inserted into it, thus having an
1031extremely high constant factor, particularly for small data types. std::list
1032also only supports bidirectional iteration, not random access iteration.</p>
1033
1034<p>In exchange for this high cost, std::list supports efficient access to both
1035ends of the list (like std::deque, but unlike std::vector or SmallVector). In
1036addition, the iterator invalidation characteristics of std::list are stronger
1037than that of a vector class: inserting or removing an element into the list does
1038not invalidate iterator or pointers to other elements in the list.</p>
1039</div>
1040
1041<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1042<h4>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1043 <a name="dss_ilist">llvm/ADT/ilist.h</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1044</h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1045
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1046<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1047<p><tt>ilist&lt;T&gt;</tt> implements an 'intrusive' doubly-linked list. It is
1048intrusive, because it requires the element to store and provide access to the
1049prev/next pointers for the list.</p>
1050
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1051<p><tt>ilist</tt> has the same drawbacks as <tt>std::list</tt>, and additionally
1052requires an <tt>ilist_traits</tt> implementation for the element type, but it
1053provides some novel characteristics. In particular, it can efficiently store
1054polymorphic objects, the traits class is informed when an element is inserted or
Gabor Greif0cbcabe2009-03-12 09:47:03 +0000[diff] [blame]1055removed from the list, and <tt>ilist</tt>s are guaranteed to support a
1056constant-time splice operation.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1057
Gabor Greif0cbcabe2009-03-12 09:47:03 +0000[diff] [blame]1058<p>These properties are exactly what we want for things like
1059<tt>Instruction</tt>s and basic blocks, which is why these are implemented with
1060<tt>ilist</tt>s.</p>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1061
1062Related classes of interest are explained in the following subsections:
1063 <ul>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1064 <li><a href="#dss_ilist_traits">ilist_traits</a></li>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1065 <li><a href="#dss_iplist">iplist</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1066 <li><a href="#dss_ilist_node">llvm/ADT/ilist_node.h</a></li>
Gabor Greif6a65f422009-03-12 10:30:31 +0000[diff] [blame]1067 <li><a href="#dss_ilist_sentinel">Sentinels</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1068 </ul>
1069</div>
1070
1071<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1072<h4>
Argyrios Kyrtzidis81df0892011-06-15 19:56:01 +0000[diff] [blame^]1073 <a name="dss_packedvector">llvm/ADT/PackedVector.h</a>
1074</h4>
1075
1076<div>
1077<p>
1078Useful for storing a vector of values using only a few number of bits for each
1079value. Apart from the standard operations of a vector-like container, it can
1080also perform an 'or' set operation.
1081</p>
1082
1083<p>For example:</p>
1084
1085<div class="doc_code">
1086<pre>
1087enum State {
1088 None = 0x0,
1089 FirstCondition = 0x1,
1090 SecondCondition = 0x2,
1091 Both = 0x3
1092};
1093
1094State get() {
1095 PackedVector&lt;State, 2&gt; Vec1;
1096 Vec1.push_back(FirstCondition);
1097
1098 PackedVector&lt;State, 2&gt; Vec2;
1099 Vec2.push_back(SecondCondition);
1100
1101 Vec1 |= Vec2;
1102 return Vec1[0]; // returns 'Both'.
1103}
1104</pre>
1105</div>
1106
1107</div>
1108
1109<!-- _______________________________________________________________________ -->
1110<h4>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1111 <a name="dss_ilist_traits">ilist_traits</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1112</h4>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1113
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1114<div>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1115<p><tt>ilist_traits&lt;T&gt;</tt> is <tt>ilist&lt;T&gt;</tt>'s customization
1116mechanism. <tt>iplist&lt;T&gt;</tt> (and consequently <tt>ilist&lt;T&gt;</tt>)
1117publicly derive from this traits class.</p>
1118</div>
1119
1120<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1121<h4>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1122 <a name="dss_iplist">iplist</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1123</h4>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1124
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1125<div>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1126<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]1127supports a slightly narrower interface. Notably, inserters from
1128<tt>T&amp;</tt> are absent.</p>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1129
1130<p><tt>ilist_traits&lt;T&gt;</tt> is a public base of this class and can be
1131used for a wide variety of customizations.</p>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1132</div>
1133
1134<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1135<h4>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1136 <a name="dss_ilist_node">llvm/ADT/ilist_node.h</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1137</h4>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1138
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1139<div>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1140<p><tt>ilist_node&lt;T&gt;</tt> implements a the forward and backward links
1141that are expected by the <tt>ilist&lt;T&gt;</tt> (and analogous containers)
1142in the default manner.</p>
1143
1144<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]1145<tt>T</tt>, usually <tt>T</tt> publicly derives from
1146<tt>ilist_node&lt;T&gt;</tt>.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1147</div>
1148
1149<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1150<h4>
Gabor Greif6a65f422009-03-12 10:30:31 +0000[diff] [blame]1151 <a name="dss_ilist_sentinel">Sentinels</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1152</h4>
Gabor Greif6a65f422009-03-12 10:30:31 +0000[diff] [blame]1153
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1154<div>
Dan Gohmancf0c9bc2010-02-25 23:51:27 +0000[diff] [blame]1155<p><tt>ilist</tt>s have another specialty that must be considered. To be a good
Gabor Greif6a65f422009-03-12 10:30:31 +0000[diff] [blame]1156citizen in the C++ ecosystem, it needs to support the standard container
1157operations, such as <tt>begin</tt> and <tt>end</tt> iterators, etc. Also, the
1158<tt>operator--</tt> must work correctly on the <tt>end</tt> iterator in the
1159case of non-empty <tt>ilist</tt>s.</p>
1160
1161<p>The only sensible solution to this problem is to allocate a so-called
1162<i>sentinel</i> along with the intrusive list, which serves as the <tt>end</tt>
1163iterator, providing the back-link to the last element. However conforming to the
1164C++ convention it is illegal to <tt>operator++</tt> beyond the sentinel and it
1165also must not be dereferenced.</p>
1166
1167<p>These constraints allow for some implementation freedom to the <tt>ilist</tt>
1168how to allocate and store the sentinel. The corresponding policy is dictated
1169by <tt>ilist_traits&lt;T&gt;</tt>. By default a <tt>T</tt> gets heap-allocated
1170whenever the need for a sentinel arises.</p>
1171
1172<p>While the default policy is sufficient in most cases, it may break down when
1173<tt>T</tt> does not provide a default constructor. Also, in the case of many
1174instances of <tt>ilist</tt>s, the memory overhead of the associated sentinels
1175is wasted. To alleviate the situation with numerous and voluminous
1176<tt>T</tt>-sentinels, sometimes a trick is employed, leading to <i>ghostly
1177sentinels</i>.</p>
1178
1179<p>Ghostly sentinels are obtained by specially-crafted <tt>ilist_traits&lt;T&gt;</tt>
1180which superpose the sentinel with the <tt>ilist</tt> instance in memory. Pointer
1181arithmetic is used to obtain the sentinel, which is relative to the
1182<tt>ilist</tt>'s <tt>this</tt> pointer. The <tt>ilist</tt> is augmented by an
1183extra pointer, which serves as the back-link of the sentinel. This is the only
1184field in the ghostly sentinel which can be legally accessed.</p>
1185</div>
1186
1187<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1188<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1189 <a name="dss_other">Other Sequential Container options</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1190</h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1191
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1192<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1193<p>Other STL containers are available, such as std::string.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1194
1195<p>There are also various STL adapter classes such as std::queue,
1196std::priority_queue, std::stack, etc. These provide simplified access to an
1197underlying container but don't affect the cost of the container itself.</p>
1198
1199</div>
1200
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1201</div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1202
1203<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1204<h3>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1205 <a name="ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1206</h3>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1207
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1208<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1209
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1210<p>Set-like containers are useful when you need to canonicalize multiple values
1211into a single representation. There are several different choices for how to do
1212this, providing various trade-offs.</p>
1213
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1214<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1215<h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1216 <a name="dss_sortedvectorset">A sorted 'vector'</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1217</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1218
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1219<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1220
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]1221<p>If you intend to insert a lot of elements, then do a lot of queries, a
1222great approach is to use a vector (or other sequential container) with
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1223std::sort+std::unique to remove duplicates. This approach works really well if
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]1224your usage pattern has these two distinct phases (insert then query), and can be
1225coupled with a good choice of <a href="#ds_sequential">sequential container</a>.
1226</p>
1227
1228<p>
1229This combination provides the several nice properties: the result data is
1230contiguous in memory (good for cache locality), has few allocations, is easy to
1231address (iterators in the final vector are just indices or pointers), and can be
1232efficiently queried with a standard binary or radix search.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1233
1234</div>
1235
1236<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1237<h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1238 <a name="dss_smallset">"llvm/ADT/SmallSet.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1239</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1240
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1241<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1242
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1243<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]1244are 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]1245has space for N elements in place (thus, if the set is dynamically smaller than
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1246N, no malloc traffic is required) and accesses them with a simple linear search.
1247When the set grows beyond 'N' elements, it allocates a more expensive representation that
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1248guarantees efficient access (for most types, it falls back to std::set, but for
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1249pointers it uses something far better, <a
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1250href="#dss_smallptrset">SmallPtrSet</a>).</p>
1251
1252<p>The magic of this class is that it handles small sets extremely efficiently,
1253but gracefully handles extremely large sets without loss of efficiency. The
1254drawback is that the interface is quite small: it supports insertion, queries
1255and erasing, but does not support iteration.</p>
1256
1257</div>
1258
1259<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1260<h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1261 <a name="dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1262</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1263
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1264<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1265
Gabor Greif4de73682010-03-26 19:30:47 +0000[diff] [blame]1266<p>SmallPtrSet has all the advantages of <tt>SmallSet</tt> (and a <tt>SmallSet</tt> of pointers is
1267transparently implemented with a <tt>SmallPtrSet</tt>), but also supports iterators. If
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1268more than 'N' insertions are performed, a single quadratically
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1269probed hash table is allocated and grows as needed, providing extremely
1270efficient access (constant time insertion/deleting/queries with low constant
1271factors) and is very stingy with malloc traffic.</p>
1272
Gabor Greif4de73682010-03-26 19:30:47 +0000[diff] [blame]1273<p>Note that, unlike <tt>std::set</tt>, the iterators of <tt>SmallPtrSet</tt> are invalidated
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1274whenever an insertion occurs. Also, the values visited by the iterators are not
1275visited in sorted order.</p>
1276
1277</div>
1278
1279<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1280<h4>
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]1281 <a name="dss_denseset">"llvm/ADT/DenseSet.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1282</h4>
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]1283
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1284<div>
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]1285
1286<p>
1287DenseSet is a simple quadratically probed hash table. It excels at supporting
1288small values: it uses a single allocation to hold all of the pairs that
1289are currently inserted in the set. DenseSet is a great way to unique small
1290values that are not simple pointers (use <a
1291href="#dss_smallptrset">SmallPtrSet</a> for pointers). Note that DenseSet has
1292the same requirements for the value type that <a
1293href="#dss_densemap">DenseMap</a> has.
1294</p>
1295
1296</div>
1297
1298<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1299<h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1300 <a name="dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1301</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1302
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1303<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1304
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1305<p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1306FoldingSet is an aggregate class that is really good at uniquing
1307expensive-to-create or polymorphic objects. It is a combination of a chained
1308hash table with intrusive links (uniqued objects are required to inherit from
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1309FoldingSetNode) that uses <a href="#dss_smallvector">SmallVector</a> as part of
1310its ID process.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1311
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1312<p>Consider a case where you want to implement a "getOrCreateFoo" method for
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1313a complex object (for example, a node in the code generator). The client has a
1314description of *what* it wants to generate (it knows the opcode and all the
1315operands), 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]1316only to find out it already exists, at which point we would have to delete it
1317and return the node that already exists.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1318</p>
1319
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1320<p>To support this style of client, FoldingSet perform a query with a
1321FoldingSetNodeID (which wraps SmallVector) that can be used to describe the
1322element that we want to query for. The query either returns the element
1323matching the ID or it returns an opaque ID that indicates where insertion should
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1324take place. Construction of the ID usually does not require heap traffic.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1325
1326<p>Because FoldingSet uses intrusive links, it can support polymorphic objects
1327in the set (for example, you can have SDNode instances mixed with LoadSDNodes).
1328Because the elements are individually allocated, pointers to the elements are
1329stable: inserting or removing elements does not invalidate any pointers to other
1330elements.
1331</p>
1332
1333</div>
1334
1335<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1336<h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1337 <a name="dss_set">&lt;set&gt;</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1338</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1339
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1340<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1341
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1342<p><tt>std::set</tt> is a reasonable all-around set class, which is decent at
1343many things but great at nothing. std::set allocates memory for each element
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1344inserted (thus it is very malloc intensive) and typically stores three pointers
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1345per element in the set (thus adding a large amount of per-element space
1346overhead). It offers guaranteed log(n) performance, which is not particularly
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1347fast from a complexity standpoint (particularly if the elements of the set are
1348expensive to compare, like strings), and has extremely high constant factors for
1349lookup, insertion and removal.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1350
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1351<p>The advantages of std::set are that its iterators are stable (deleting or
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1352inserting an element from the set does not affect iterators or pointers to other
1353elements) and that iteration over the set is guaranteed to be in sorted order.
1354If the elements in the set are large, then the relative overhead of the pointers
1355and malloc traffic is not a big deal, but if the elements of the set are small,
1356std::set is almost never a good choice.</p>
1357
1358</div>
1359
1360<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1361<h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1362 <a name="dss_setvector">"llvm/ADT/SetVector.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1363</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1364
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1365<div>
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1366<p>LLVM's SetVector&lt;Type&gt; is an adapter class that combines your choice of
1367a set-like container along with a <a href="#ds_sequential">Sequential
1368Container</a>. The important property
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1369that this provides is efficient insertion with uniquing (duplicate elements are
1370ignored) with iteration support. It implements this by inserting elements into
1371both a set-like container and the sequential container, using the set-like
1372container for uniquing and the sequential container for iteration.
1373</p>
1374
1375<p>The difference between SetVector and other sets is that the order of
1376iteration is guaranteed to match the order of insertion into the SetVector.
1377This property is really important for things like sets of pointers. Because
1378pointer values are non-deterministic (e.g. vary across runs of the program on
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1379different machines), iterating over the pointers in the set will
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1380not be in a well-defined order.</p>
1381
1382<p>
1383The drawback of SetVector is that it requires twice as much space as a normal
1384set and has the sum of constant factors from the set-like container and the
1385sequential container that it uses. Use it *only* if you need to iterate over
1386the elements in a deterministic order. SetVector is also expensive to delete
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1387elements out of (linear time), unless you use it's "pop_back" method, which is
1388faster.
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1389</p>
1390
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1391<p>SetVector is an adapter class that defaults to using std::vector and std::set
1392for the underlying containers, so it is quite expensive. However,
1393<tt>"llvm/ADT/SetVector.h"</tt> also provides a SmallSetVector class, which
1394defaults to using a SmallVector and SmallSet of a specified size. If you use
1395this, and if your sets are dynamically smaller than N, you will save a lot of
1396heap traffic.</p>
1397
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1398</div>
1399
1400<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1401<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1402 <a name="dss_uniquevector">"llvm/ADT/UniqueVector.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1403</h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1404
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1405<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1406
1407<p>
1408UniqueVector is similar to <a href="#dss_setvector">SetVector</a>, but it
1409retains a unique ID for each element inserted into the set. It internally
1410contains a map and a vector, and it assigns a unique ID for each value inserted
1411into the set.</p>
1412
1413<p>UniqueVector is very expensive: its cost is the sum of the cost of
1414maintaining both the map and vector, it has high complexity, high constant
1415factors, and produces a lot of malloc traffic. It should be avoided.</p>
1416
1417</div>
1418
1419
1420<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1421<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1422 <a name="dss_otherset">Other Set-Like Container Options</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1423</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1424
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1425<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1426
1427<p>
1428The STL provides several other options, such as std::multiset and the various
Chris Lattnerf1a30822009-03-09 05:20:45 +0000[diff] [blame]1429"hash_set" like containers (whether from C++ TR1 or from the SGI library). We
1430never use hash_set and unordered_set because they are generally very expensive
1431(each insertion requires a malloc) and very non-portable.
1432</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1433
1434<p>std::multiset is useful if you're not interested in elimination of
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1435duplicates, but has all the drawbacks of std::set. A sorted vector (where you
1436don't delete duplicate entries) or some other approach is almost always
1437better.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1438
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1439</div>
1440
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1441</div>
1442
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1443<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1444<h3>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1445 <a name="ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1446</h3>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1447
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1448<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1449Map-like containers are useful when you want to associate data to a key. As
1450usual, there are a lot of different ways to do this. :)
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1451
1452<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1453<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1454 <a name="dss_sortedvectormap">A sorted 'vector'</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1455</h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1456
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1457<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1458
1459<p>
1460If your usage pattern follows a strict insert-then-query approach, you can
1461trivially use the same approach as <a href="#dss_sortedvectorset">sorted vectors
1462for set-like containers</a>. The only difference is that your query function
1463(which uses std::lower_bound to get efficient log(n) lookup) should only compare
1464the key, not both the key and value. This yields the same advantages as sorted
1465vectors for sets.
1466</p>
1467</div>
1468
1469<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1470<h4>
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1471 <a name="dss_stringmap">"llvm/ADT/StringMap.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1472</h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1473
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1474<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1475
1476<p>
1477Strings are commonly used as keys in maps, and they are difficult to support
1478efficiently: they are variable length, inefficient to hash and compare when
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1479long, expensive to copy, etc. StringMap is a specialized container designed to
1480cope with these issues. It supports mapping an arbitrary range of bytes to an
1481arbitrary other object.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1482
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1483<p>The StringMap implementation uses a quadratically-probed hash table, where
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1484the buckets store a pointer to the heap allocated entries (and some other
1485stuff). The entries in the map must be heap allocated because the strings are
1486variable length. The string data (key) and the element object (value) are
1487stored in the same allocation with the string data immediately after the element
1488object. This container guarantees the "<tt>(char*)(&amp;Value+1)</tt>" points
1489to the key string for a value.</p>
1490
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1491<p>The StringMap is very fast for several reasons: quadratic probing is very
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1492cache efficient for lookups, the hash value of strings in buckets is not
Nick Lewycky2a80aca2010-08-01 23:18:45 +0000[diff] [blame]1493recomputed when looking up an element, StringMap rarely has to touch the
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1494memory for unrelated objects when looking up a value (even when hash collisions
1495happen), hash table growth does not recompute the hash values for strings
1496already in the table, and each pair in the map is store in a single allocation
1497(the string data is stored in the same allocation as the Value of a pair).</p>
1498
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1499<p>StringMap also provides query methods that take byte ranges, so it only ever
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1500copies a string if a value is inserted into the table.</p>
1501</div>
1502
1503<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1504<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1505 <a name="dss_indexedmap">"llvm/ADT/IndexedMap.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1506</h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1507
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1508<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1509<p>
1510IndexedMap is a specialized container for mapping small dense integers (or
1511values that can be mapped to small dense integers) to some other type. It is
1512internally implemented as a vector with a mapping function that maps the keys to
1513the dense integer range.
1514</p>
1515
1516<p>
1517This is useful for cases like virtual registers in the LLVM code generator: they
1518have a dense mapping that is offset by a compile-time constant (the first
1519virtual register ID).</p>
1520
1521</div>
1522
1523<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1524<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1525 <a name="dss_densemap">"llvm/ADT/DenseMap.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1526</h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1527
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1528<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1529
1530<p>
1531DenseMap is a simple quadratically probed hash table. It excels at supporting
1532small keys and values: it uses a single allocation to hold all of the pairs that
1533are currently inserted in the map. DenseMap is a great way to map pointers to
1534pointers, or map other small types to each other.
1535</p>
1536
1537<p>
1538There are several aspects of DenseMap that you should be aware of, however. The
1539iterators in a densemap are invalidated whenever an insertion occurs, unlike
1540map. Also, because DenseMap allocates space for a large number of key/value
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1541pairs (it starts with 64 by default), it will waste a lot of space if your keys
1542or values are large. Finally, you must implement a partial specialization of
Chris Lattner76c1b972007-09-17 18:34:04 +0000[diff] [blame]1543DenseMapInfo for the key that you want, if it isn't already supported. This
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1544is required to tell DenseMap about two special marker values (which can never be
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1545inserted into the map) that it needs internally.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1546
1547</div>
1548
1549<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1550<h4>
Jeffrey Yasskin71a5c222009-10-22 22:11:22 +0000[diff] [blame]1551 <a name="dss_valuemap">"llvm/ADT/ValueMap.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1552</h4>
Jeffrey Yasskin71a5c222009-10-22 22:11:22 +0000[diff] [blame]1553
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1554<div>
Jeffrey Yasskin71a5c222009-10-22 22:11:22 +0000[diff] [blame]1555
1556<p>
1557ValueMap is a wrapper around a <a href="#dss_densemap">DenseMap</a> mapping
1558Value*s (or subclasses) to another type. When a Value is deleted or RAUW'ed,
1559ValueMap will update itself so the new version of the key is mapped to the same
1560value, just as if the key were a WeakVH. You can configure exactly how this
1561happens, and what else happens on these two events, by passing
1562a <code>Config</code> parameter to the ValueMap template.</p>
1563
1564</div>
1565
1566<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1567<h4>
Jakob Stoklund Olesenaca0da62010-12-14 00:55:51 +0000[diff] [blame]1568 <a name="dss_intervalmap">"llvm/ADT/IntervalMap.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1569</h4>
Jakob Stoklund Olesenaca0da62010-12-14 00:55:51 +0000[diff] [blame]1570
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1571<div>
Jakob Stoklund Olesenaca0da62010-12-14 00:55:51 +0000[diff] [blame]1572
1573<p> IntervalMap is a compact map for small keys and values. It maps key
1574intervals instead of single keys, and it will automatically coalesce adjacent
1575intervals. When then map only contains a few intervals, they are stored in the
1576map object itself to avoid allocations.</p>
1577
1578<p> The IntervalMap iterators are quite big, so they should not be passed around
1579as STL iterators. The heavyweight iterators allow a smaller data structure.</p>
1580
1581</div>
1582
1583<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1584<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1585 <a name="dss_map">&lt;map&gt;</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1586</h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1587
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1588<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1589
1590<p>
1591std::map has similar characteristics to <a href="#dss_set">std::set</a>: it uses
1592a single allocation per pair inserted into the map, it offers log(n) lookup with
1593an extremely large constant factor, imposes a space penalty of 3 pointers per
1594pair in the map, etc.</p>
1595
1596<p>std::map is most useful when your keys or values are very large, if you need
1597to iterate over the collection in sorted order, or if you need stable iterators
1598into the map (i.e. they don't get invalidated if an insertion or deletion of
1599another element takes place).</p>
1600
1601</div>
1602
1603<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1604<h4>
Jakob Stoklund Olesen4359e5e2011-04-05 20:56:08 +0000[diff] [blame]1605 <a name="dss_inteqclasses">"llvm/ADT/IntEqClasses.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1606</h4>
Jakob Stoklund Olesen4359e5e2011-04-05 20:56:08 +0000[diff] [blame]1607
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1608<div>
Jakob Stoklund Olesen4359e5e2011-04-05 20:56:08 +0000[diff] [blame]1609
1610<p>IntEqClasses provides a compact representation of equivalence classes of
1611small integers. Initially, each integer in the range 0..n-1 has its own
1612equivalence class. Classes can be joined by passing two class representatives to
1613the join(a, b) method. Two integers are in the same class when findLeader()
1614returns the same representative.</p>
1615
1616<p>Once all equivalence classes are formed, the map can be compressed so each
1617integer 0..n-1 maps to an equivalence class number in the range 0..m-1, where m
1618is the total number of equivalence classes. The map must be uncompressed before
1619it can be edited again.</p>
1620
1621</div>
1622
1623<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1624<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1625 <a name="dss_othermap">Other Map-Like Container Options</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1626</h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1627
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1628<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1629
1630<p>
1631The STL provides several other options, such as std::multimap and the various
Chris Lattnerf1a30822009-03-09 05:20:45 +0000[diff] [blame]1632"hash_map" like containers (whether from C++ TR1 or from the SGI library). We
1633never use hash_set and unordered_set because they are generally very expensive
1634(each insertion requires a malloc) and very non-portable.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1635
1636<p>std::multimap is useful if you want to map a key to multiple values, but has
1637all the drawbacks of std::map. A sorted vector or some other approach is almost
1638always better.</p>
1639
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1640</div>
1641
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1642</div>
1643
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1644<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1645<h3>
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]1646 <a name="ds_string">String-like containers</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1647</h3>
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]1648
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1649<div>
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]1650
1651<p>
1652TODO: const char* vs stringref vs smallstring vs std::string. Describe twine,
1653xref to #string_apis.
1654</p>
1655
1656</div>
1657
1658<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1659<h3>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1660 <a name="ds_bit">Bit storage containers (BitVector, SparseBitVector)</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1661</h3>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1662
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1663<div>
Chris Lattner7086ce72007-09-25 22:37:50 +0000[diff] [blame]1664<p>Unlike the other containers, there are only two bit storage containers, and
1665choosing when to use each is relatively straightforward.</p>
1666
1667<p>One additional option is
1668<tt>std::vector&lt;bool&gt;</tt>: we discourage its use for two reasons 1) the
1669implementation in many common compilers (e.g. commonly available versions of
1670GCC) is extremely inefficient and 2) the C++ standards committee is likely to
1671deprecate this container and/or change it significantly somehow. In any case,
1672please don't use it.</p>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1673
1674<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1675<h4>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1676 <a name="dss_bitvector">BitVector</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1677</h4>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1678
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1679<div>
Dan Gohman5f7775c2010-01-05 18:24:00 +0000[diff] [blame]1680<p> The BitVector container provides a dynamic size set of bits for manipulation.
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1681It supports individual bit setting/testing, as well as set operations. The set
1682operations take time O(size of bitvector), but operations are performed one word
1683at a time, instead of one bit at a time. This makes the BitVector very fast for
1684set operations compared to other containers. Use the BitVector when you expect
1685the number of set bits to be high (IE a dense set).
1686</p>
1687</div>
1688
1689<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1690<h4>
Dan Gohman5f7775c2010-01-05 18:24:00 +0000[diff] [blame]1691 <a name="dss_smallbitvector">SmallBitVector</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1692</h4>
Dan Gohman5f7775c2010-01-05 18:24:00 +0000[diff] [blame]1693
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1694<div>
Dan Gohman5f7775c2010-01-05 18:24:00 +0000[diff] [blame]1695<p> The SmallBitVector container provides the same interface as BitVector, but
1696it is optimized for the case where only a small number of bits, less than
169725 or so, are needed. It also transparently supports larger bit counts, but
1698slightly less efficiently than a plain BitVector, so SmallBitVector should
1699only be used when larger counts are rare.
1700</p>
1701
1702<p>
1703At this time, SmallBitVector does not support set operations (and, or, xor),
1704and its operator[] does not provide an assignable lvalue.
1705</p>
1706</div>
1707
1708<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1709<h4>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1710 <a name="dss_sparsebitvector">SparseBitVector</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1711</h4>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1712
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1713<div>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1714<p> The SparseBitVector container is much like BitVector, with one major
1715difference: Only the bits that are set, are stored. This makes the
1716SparseBitVector much more space efficient than BitVector when the set is sparse,
1717as well as making set operations O(number of set bits) instead of O(size of
1718universe). 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
1719(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).
1720</p>
1721</div>
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]1722
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1723</div>
1724
1725</div>
1726
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1727<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1728<h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1729 <a name="common">Helpful Hints for Common Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1730</h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1731<!-- *********************************************************************** -->
1732
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1733<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1734
1735<p>This section describes how to perform some very simple transformations of
1736LLVM code. This is meant to give examples of common idioms used, showing the
1737practical side of LLVM transformations. <p> Because this is a "how-to" section,
1738you should also read about the main classes that you will be working with. The
1739<a href="#coreclasses">Core LLVM Class Hierarchy Reference</a> contains details
1740and descriptions of the main classes that you should know about.</p>
1741
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1742<!-- NOTE: this section should be heavy on example code -->
1743<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1744<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1745 <a name="inspection">Basic Inspection and Traversal Routines</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1746</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1747
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1748<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1749
1750<p>The LLVM compiler infrastructure have many different data structures that may
1751be traversed. Following the example of the C++ standard template library, the
1752techniques used to traverse these various data structures are all basically the
1753same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
1754method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
1755function returns an iterator pointing to one past the last valid element of the
1756sequence, and there is some <tt>XXXiterator</tt> data type that is common
1757between the two operations.</p>
1758
1759<p>Because the pattern for iteration is common across many different aspects of
1760the program representation, the standard template library algorithms may be used
1761on them, and it is easier to remember how to iterate. First we show a few common
1762examples of the data structures that need to be traversed. Other data
1763structures are traversed in very similar ways.</p>
1764
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1765<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1766<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1767 <a name="iterate_function">Iterating over the </a><a
1768 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
1769 href="#Function"><tt>Function</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1770</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1771
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1772<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1773
1774<p>It's quite common to have a <tt>Function</tt> instance that you'd like to
1775transform in some way; in particular, you'd like to manipulate its
1776<tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over all of
1777the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>. The following is
1778an example that prints the name of a <tt>BasicBlock</tt> and the number of
1779<tt>Instruction</tt>s it contains:</p>
1780
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1781<div class="doc_code">
1782<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1783// <i>func is a pointer to a Function instance</i>
1784for (Function::iterator i = func-&gt;begin(), e = func-&gt;end(); i != e; ++i)
1785 // <i>Print out the name of the basic block if it has one, and then the</i>
1786 // <i>number of instructions that it contains</i>
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1787 errs() &lt;&lt; "Basic block (name=" &lt;&lt; i-&gt;getName() &lt;&lt; ") has "
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1788 &lt;&lt; i-&gt;size() &lt;&lt; " instructions.\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1789</pre>
1790</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1791
1792<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]1793invoking member functions of the <tt>Instruction</tt> class. This is
1794because the indirection operator is overloaded for the iterator
Chris Lattner7496ec52003-08-05 22:54:23 +0000[diff] [blame]1795classes. In the above code, the expression <tt>i-&gt;size()</tt> is
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1796exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.</p>
1797
1798</div>
1799
1800<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1801<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1802 <a name="iterate_basicblock">Iterating over the </a><a
1803 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1804 href="#BasicBlock"><tt>BasicBlock</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1805</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1806
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1807<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1808
1809<p>Just like when dealing with <tt>BasicBlock</tt>s in <tt>Function</tt>s, it's
1810easy to iterate over the individual instructions that make up
1811<tt>BasicBlock</tt>s. Here's a code snippet that prints out each instruction in
1812a <tt>BasicBlock</tt>:</p>
1813
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1814<div class="doc_code">
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1815<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1816// <i>blk is a pointer to a BasicBlock instance</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1817for (BasicBlock::iterator i = blk-&gt;begin(), e = blk-&gt;end(); i != e; ++i)
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1818 // <i>The next statement works since operator&lt;&lt;(ostream&amp;,...)</i>
1819 // <i>is overloaded for Instruction&amp;</i>
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1820 errs() &lt;&lt; *i &lt;&lt; "\n";
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1821</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1822</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1823
1824<p>However, this isn't really the best way to print out the contents of a
1825<tt>BasicBlock</tt>! Since the ostream operators are overloaded for virtually
1826anything you'll care about, you could have just invoked the print routine on the
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1827basic block itself: <tt>errs() &lt;&lt; *blk &lt;&lt; "\n";</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1828
1829</div>
1830
1831<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1832<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1833 <a name="iterate_institer">Iterating over the </a><a
1834 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1835 href="#Function"><tt>Function</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1836</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1837
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1838<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1839
1840<p>If you're finding that you commonly iterate over a <tt>Function</tt>'s
1841<tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s <tt>Instruction</tt>s,
1842<tt>InstIterator</tt> should be used instead. You'll need to include <a
1843href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>,
1844and then instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1845small 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]1846
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1847<div class="doc_code">
1848<pre>
1849#include "<a href="/doxygen/InstIterator_8h-source.html">llvm/Support/InstIterator.h</a>"
1850
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1851// <i>F is a pointer to a Function instance</i>
Chris Lattnerda021aa2008-06-04 18:20:42 +0000[diff] [blame]1852for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1853 errs() &lt;&lt; *I &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1854</pre>
1855</div>
1856
1857<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]1858work list with its initial contents. For example, if you wanted to
1859initialize a work list to contain all instructions in a <tt>Function</tt>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1860F, all you would need to do is something like:</p>
1861
1862<div class="doc_code">
1863<pre>
1864std::set&lt;Instruction*&gt; worklist;
Chris Lattnerda021aa2008-06-04 18:20:42 +0000[diff] [blame]1865// or better yet, SmallPtrSet&lt;Instruction*, 64&gt; worklist;
1866
1867for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
1868 worklist.insert(&amp;*I);
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>The STL set <tt>worklist</tt> would now contain all instructions in the
1873<tt>Function</tt> pointed to by F.</p>
1874
1875</div>
1876
1877<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1878<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1879 <a name="iterate_convert">Turning an iterator into a class pointer (and
1880 vice-versa)</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1881</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1882
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1883<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1884
1885<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]1886instance when all you've got at hand is an iterator. Well, extracting
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1887a reference or a pointer from an iterator is very straight-forward.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1888Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and <tt>j</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1889is a <tt>BasicBlock::const_iterator</tt>:</p>
1890
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1891<div class="doc_code">
1892<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1893Instruction&amp; inst = *i; // <i>Grab reference to instruction reference</i>
1894Instruction* pinst = &amp;*i; // <i>Grab pointer to instruction reference</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1895const Instruction&amp; inst = *j;
1896</pre>
1897</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1898
1899<p>However, the iterators you'll be working with in the LLVM framework are
1900special: they will automatically convert to a ptr-to-instance type whenever they
1901need to. Instead of dereferencing the iterator and then taking the address of
1902the result, you can simply assign the iterator to the proper pointer type and
1903you get the dereference and address-of operation as a result of the assignment
1904(behind the scenes, this is a result of overloading casting mechanisms). Thus
1905the last line of the last example,</p>
1906
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1907<div class="doc_code">
1908<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1909Instruction *pinst = &amp;*i;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1910</pre>
1911</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1912
1913<p>is semantically equivalent to</p>
1914
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1915<div class="doc_code">
1916<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1917Instruction *pinst = i;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1918</pre>
1919</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1920
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1921<p>It's also possible to turn a class pointer into the corresponding iterator,
1922and this is a constant time operation (very efficient). The following code
1923snippet illustrates use of the conversion constructors provided by LLVM
1924iterators. By using these, you can explicitly grab the iterator of something
1925without actually obtaining it via iteration over some structure:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1926
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1927<div class="doc_code">
1928<pre>
1929void printNextInstruction(Instruction* inst) {
1930 BasicBlock::iterator it(inst);
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1931 ++it; // <i>After this line, it refers to the instruction after *inst</i>
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1932 if (it != inst-&gt;getParent()-&gt;end()) errs() &lt;&lt; *it &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1933}
1934</pre>
1935</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1936
Dan Gohman525bf8e2010-03-26 19:39:05 +0000[diff] [blame]1937<p>Unfortunately, these implicit conversions come at a cost; they prevent
1938these iterators from conforming to standard iterator conventions, and thus
Dan Gohman0d91c112010-03-26 19:51:14 +0000[diff] [blame]1939from being usable with standard algorithms and containers. For example, they
1940prevent the following code, where <tt>B</tt> is a <tt>BasicBlock</tt>,
Dan Gohman525bf8e2010-03-26 19:39:05 +0000[diff] [blame]1941from compiling:</p>
1942
1943<div class="doc_code">
1944<pre>
1945 llvm::SmallVector&lt;llvm::Instruction *, 16&gt;(B-&gt;begin(), B-&gt;end());
1946</pre>
1947</div>
1948
1949<p>Because of this, these implicit conversions may be removed some day,
Dan Gohman0d91c112010-03-26 19:51:14 +0000[diff] [blame]1950and <tt>operator*</tt> changed to return a pointer instead of a reference.</p>
Dan Gohman525bf8e2010-03-26 19:39:05 +0000[diff] [blame]1951
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1952</div>
1953
1954<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1955<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1956 <a name="iterate_complex">Finding call sites: a slightly more complex
1957 example</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1958</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1959
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1960<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1961
1962<p>Say that you're writing a FunctionPass and would like to count all the
1963locations in the entire module (that is, across every <tt>Function</tt>) where a
1964certain function (i.e., some <tt>Function</tt>*) is already in scope. As you'll
1965learn 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]1966much more straight-forward manner, but this example will allow us to explore how
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1967you'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]1968is what we want to do:</p>
1969
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1970<div class="doc_code">
1971<pre>
1972initialize callCounter to zero
1973for each Function f in the Module
1974 for each BasicBlock b in f
1975 for each Instruction i in b
1976 if (i is a CallInst and calls the given function)
1977 increment callCounter
1978</pre>
1979</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1980
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1981<p>And the actual code is (remember, because we're writing a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1982<tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply has to
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1983override the <tt>runOnFunction</tt> method):</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1984
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1985<div class="doc_code">
1986<pre>
1987Function* targetFunc = ...;
1988
1989class OurFunctionPass : public FunctionPass {
1990 public:
1991 OurFunctionPass(): callCounter(0) { }
1992
1993 virtual runOnFunction(Function&amp; F) {
1994 for (Function::iterator b = F.begin(), be = F.end(); b != be; ++b) {
Eric Christopher203e71d2008-11-08 08:20:49 +0000[diff] [blame]1995 for (BasicBlock::iterator i = b-&gt;begin(), ie = b-&gt;end(); i != ie; ++i) {
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1996 if (<a href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a>&lt;<a
1997 href="#CallInst">CallInst</a>&gt;(&amp;*i)) {
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1998 // <i>We know we've encountered a call instruction, so we</i>
1999 // <i>need to determine if it's a call to the</i>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2000 // <i>function pointed to by m_func or not.</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2001 if (callInst-&gt;getCalledFunction() == targetFunc)
2002 ++callCounter;
2003 }
2004 }
2005 }
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2006 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2007
2008 private:
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2009 unsigned callCounter;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2010};
2011</pre>
2012</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2013
2014</div>
2015
Brian Gaekef1972c62003-11-07 19:25:45 +0000[diff] [blame]2016<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2017<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2018 <a name="calls_and_invokes">Treating calls and invokes the same way</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2019</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2020
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2021<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2022
2023<p>You may have noticed that the previous example was a bit oversimplified in
2024that it did not deal with call sites generated by 'invoke' instructions. In
2025this, and in other situations, you may find that you want to treat
2026<tt>CallInst</tt>s and <tt>InvokeInst</tt>s the same way, even though their
2027most-specific common base class is <tt>Instruction</tt>, which includes lots of
2028less closely-related things. For these cases, LLVM provides a handy wrapper
2029class called <a
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]2030href="http://llvm.org/doxygen/classllvm_1_1CallSite.html"><tt>CallSite</tt></a>.
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]2031It is essentially a wrapper around an <tt>Instruction</tt> pointer, with some
2032methods that provide functionality common to <tt>CallInst</tt>s and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2033<tt>InvokeInst</tt>s.</p>
2034
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]2035<p>This class has "value semantics": it should be passed by value, not by
2036reference and it should not be dynamically allocated or deallocated using
2037<tt>operator new</tt> or <tt>operator delete</tt>. It is efficiently copyable,
2038assignable and constructable, with costs equivalents to that of a bare pointer.
2039If you look at its definition, it has only a single pointer member.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2040
2041</div>
2042
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]2043<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2044<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2045 <a name="iterate_chains">Iterating over def-use &amp; use-def chains</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2046</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2047
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2048<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2049
2050<p>Frequently, we might have an instance of the <a
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]2051href="/doxygen/classllvm_1_1Value.html">Value Class</a> and we want to
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]2052determine which <tt>User</tt>s use the <tt>Value</tt>. The list of all
2053<tt>User</tt>s of a particular <tt>Value</tt> is called a <i>def-use</i> chain.
2054For example, let's say we have a <tt>Function*</tt> named <tt>F</tt> to a
2055particular function <tt>foo</tt>. Finding all of the instructions that
2056<i>use</i> <tt>foo</tt> is as simple as iterating over the <i>def-use</i> chain
2057of <tt>F</tt>:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2058
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2059<div class="doc_code">
2060<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2061Function *F = ...;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2062
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2063for (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]2064 if (Instruction *Inst = dyn_cast&lt;Instruction&gt;(*i)) {
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]2065 errs() &lt;&lt; "F is used in instruction:\n";
2066 errs() &lt;&lt; *Inst &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2067 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2068</pre>
Gabor Greif394fdfb2010-03-26 19:35:48 +0000[diff] [blame]2069</div>
2070
Gabor Greifce94319532010-03-26 19:40:38 +0000[diff] [blame]2071<p>Note that dereferencing a <tt>Value::use_iterator</tt> is not a very cheap
Gabor Greif4de73682010-03-26 19:30:47 +0000[diff] [blame]2072operation. Instead of performing <tt>*i</tt> above several times, consider
Gabor Greifce94319532010-03-26 19:40:38 +0000[diff] [blame]2073doing it only once in the loop body and reusing its result.</p>
Gabor Greif4de73682010-03-26 19:30:47 +0000[diff] [blame]2074
Gabor Greif6091ff32010-03-26 19:04:42 +0000[diff] [blame]2075<p>Alternatively, it's common to have an instance of the <a
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]2076href="/doxygen/classllvm_1_1User.html">User Class</a> and need to know what
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2077<tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used by a
2078<tt>User</tt> is known as a <i>use-def</i> chain. Instances of class
2079<tt>Instruction</tt> are common <tt>User</tt>s, so we might want to iterate over
2080all of the values that a particular instruction uses (that is, the operands of
2081the particular <tt>Instruction</tt>):</p>
2082
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2083<div class="doc_code">
2084<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2085Instruction *pi = ...;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2086
2087for (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]2088 Value *v = *i;
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2089 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2090}
2091</pre>
2092</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2093
Gabor Greif4de73682010-03-26 19:30:47 +0000[diff] [blame]2094<p>Declaring objects as <tt>const</tt> is an important tool of enforcing
Gabor Greifce94319532010-03-26 19:40:38 +0000[diff] [blame]2095mutation free algorithms (such as analyses, etc.). For this purpose above
Gabor Greif4de73682010-03-26 19:30:47 +0000[diff] [blame]2096iterators come in constant flavors as <tt>Value::const_use_iterator</tt>
2097and <tt>Value::const_op_iterator</tt>. They automatically arise when
2098calling <tt>use/op_begin()</tt> on <tt>const Value*</tt>s or
2099<tt>const User*</tt>s respectively. Upon dereferencing, they return
Gabor Greifce94319532010-03-26 19:40:38 +0000[diff] [blame]2100<tt>const Use*</tt>s. Otherwise the above patterns remain unchanged.</p>
2101
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2102</div>
2103
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2104<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2105<h4>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2106 <a name="iterate_preds">Iterating over predecessors &amp;
2107successors of blocks</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2108</h4>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2109
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2110<div>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2111
2112<p>Iterating over the predecessors and successors of a block is quite easy
2113with the routines defined in <tt>"llvm/Support/CFG.h"</tt>. Just use code like
2114this to iterate over all predecessors of BB:</p>
2115
2116<div class="doc_code">
2117<pre>
2118#include "llvm/Support/CFG.h"
2119BasicBlock *BB = ...;
2120
2121for (pred_iterator PI = pred_begin(BB), E = pred_end(BB); PI != E; ++PI) {
2122 BasicBlock *Pred = *PI;
2123 // <i>...</i>
2124}
2125</pre>
2126</div>
2127
2128<p>Similarly, to iterate over successors use
2129succ_iterator/succ_begin/succ_end.</p>
2130
2131</div>
2132
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2133</div>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2134
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2135<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2136<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2137 <a name="simplechanges">Making simple changes</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2138</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2139
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2140<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2141
2142<p>There are some primitive transformation operations present in the LLVM
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]2143infrastructure that are worth knowing about. When performing
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2144transformations, it's fairly common to manipulate the contents of basic
2145blocks. This section describes some of the common methods for doing so
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2146and gives example code.</p>
2147
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2148<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2149<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2150 <a name="schanges_creating">Creating and inserting new
2151 <tt>Instruction</tt>s</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2152</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2153
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2154<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2155
2156<p><i>Instantiating Instructions</i></p>
2157
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]2158<p>Creation of <tt>Instruction</tt>s is straight-forward: simply call the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2159constructor for the kind of instruction to instantiate and provide the necessary
2160parameters. For example, an <tt>AllocaInst</tt> only <i>requires</i> a
2161(const-ptr-to) <tt>Type</tt>. Thus:</p>
2162
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2163<div class="doc_code">
2164<pre>
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2165AllocaInst* ai = new AllocaInst(Type::Int32Ty);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2166</pre>
2167</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2168
2169<p>will create an <tt>AllocaInst</tt> instance that represents the allocation of
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]2170one integer in the current stack frame, at run time. Each <tt>Instruction</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2171subclass is likely to have varying default parameters which change the semantics
2172of the instruction, so refer to the <a
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]2173href="/doxygen/classllvm_1_1Instruction.html">doxygen documentation for the subclass of
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2174Instruction</a> that you're interested in instantiating.</p>
2175
2176<p><i>Naming values</i></p>
2177
2178<p>It is very useful to name the values of instructions when you're able to, as
2179this facilitates the debugging of your transformations. If you end up looking
2180at generated LLVM machine code, you definitely want to have logical names
2181associated with the results of instructions! By supplying a value for the
2182<tt>Name</tt> (default) parameter of the <tt>Instruction</tt> constructor, you
2183associate a logical name with the result of the instruction's execution at
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]2184run time. For example, say that I'm writing a transformation that dynamically
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2185allocates space for an integer on the stack, and that integer is going to be
2186used as some kind of index by some other code. To accomplish this, I place an
2187<tt>AllocaInst</tt> at the first point in the first <tt>BasicBlock</tt> of some
2188<tt>Function</tt>, and I'm intending to use it within the same
2189<tt>Function</tt>. I might do:</p>
2190
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2191<div class="doc_code">
2192<pre>
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2193AllocaInst* pa = new AllocaInst(Type::Int32Ty, 0, "indexLoc");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2194</pre>
2195</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2196
2197<p>where <tt>indexLoc</tt> is now the logical name of the instruction's
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]2198execution value, which is a pointer to an integer on the run time stack.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2199
2200<p><i>Inserting instructions</i></p>
2201
2202<p>There are essentially two ways to insert an <tt>Instruction</tt>
2203into an existing sequence of instructions that form a <tt>BasicBlock</tt>:</p>
2204
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]2205<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2206 <li>Insertion into an explicit instruction list
2207
2208 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within that
2209 <tt>BasicBlock</tt>, and a newly-created instruction we wish to insert
2210 before <tt>*pi</tt>, we do the following: </p>
2211
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2212<div class="doc_code">
2213<pre>
2214BasicBlock *pb = ...;
2215Instruction *pi = ...;
2216Instruction *newInst = new Instruction(...);
2217
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2218pb-&gt;getInstList().insert(pi, newInst); // <i>Inserts newInst before pi in pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2219</pre>
2220</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]2221
2222 <p>Appending to the end of a <tt>BasicBlock</tt> is so common that
2223 the <tt>Instruction</tt> class and <tt>Instruction</tt>-derived
2224 classes provide constructors which take a pointer to a
2225 <tt>BasicBlock</tt> to be appended to. For example code that
2226 looked like: </p>
2227
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2228<div class="doc_code">
2229<pre>
2230BasicBlock *pb = ...;
2231Instruction *newInst = new Instruction(...);
2232
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2233pb-&gt;getInstList().push_back(newInst); // <i>Appends newInst to pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2234</pre>
2235</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]2236
2237 <p>becomes: </p>
2238
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2239<div class="doc_code">
2240<pre>
2241BasicBlock *pb = ...;
2242Instruction *newInst = new Instruction(..., pb);
2243</pre>
2244</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]2245
2246 <p>which is much cleaner, especially if you are creating
2247 long instruction streams.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2248
2249 <li>Insertion into an implicit instruction list
2250
2251 <p><tt>Instruction</tt> instances that are already in <tt>BasicBlock</tt>s
2252 are implicitly associated with an existing instruction list: the instruction
2253 list of the enclosing basic block. Thus, we could have accomplished the same
2254 thing as the above code without being given a <tt>BasicBlock</tt> by doing:
2255 </p>
2256
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2257<div class="doc_code">
2258<pre>
2259Instruction *pi = ...;
2260Instruction *newInst = new Instruction(...);
2261
2262pi-&gt;getParent()-&gt;getInstList().insert(pi, newInst);
2263</pre>
2264</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2265
2266 <p>In fact, this sequence of steps occurs so frequently that the
2267 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes provide
2268 constructors which take (as a default parameter) a pointer to an
2269 <tt>Instruction</tt> which the newly-created <tt>Instruction</tt> should
2270 precede. That is, <tt>Instruction</tt> constructors are capable of
2271 inserting the newly-created instance into the <tt>BasicBlock</tt> of a
2272 provided instruction, immediately before that instruction. Using an
2273 <tt>Instruction</tt> constructor with a <tt>insertBefore</tt> (default)
2274 parameter, the above code becomes:</p>
2275
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2276<div class="doc_code">
2277<pre>
2278Instruction* pi = ...;
2279Instruction* newInst = new Instruction(..., pi);
2280</pre>
2281</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2282
2283 <p>which is much cleaner, especially if you're creating a lot of
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2284 instructions and adding them to <tt>BasicBlock</tt>s.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2285</ul>
2286
2287</div>
2288
2289<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2290<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2291 <a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2292</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2293
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2294<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2295
2296<p>Deleting an instruction from an existing sequence of instructions that form a
Chris Lattner49e0ccf2011-03-24 16:13:31 +0000[diff] [blame]2297<a href="#BasicBlock"><tt>BasicBlock</tt></a> is very straight-forward: just
2298call the instruction's eraseFromParent() method. For example:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2299
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2300<div class="doc_code">
2301<pre>
2302<a href="#Instruction">Instruction</a> *I = .. ;
Chris Lattner9f8ec252008-02-15 22:57:17 +0000[diff] [blame]2303I-&gt;eraseFromParent();
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2304</pre>
2305</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2306
Chris Lattner49e0ccf2011-03-24 16:13:31 +0000[diff] [blame]2307<p>This unlinks the instruction from its containing basic block and deletes
2308it. If you'd just like to unlink the instruction from its containing basic
2309block but not delete it, you can use the <tt>removeFromParent()</tt> method.</p>
2310
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2311</div>
2312
2313<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2314<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2315 <a name="schanges_replacing">Replacing an <tt>Instruction</tt> with another
2316 <tt>Value</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2317</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2318
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2319<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2320
2321<p><i>Replacing individual instructions</i></p>
2322
2323<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]2324permits use of two very useful replace functions: <tt>ReplaceInstWithValue</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2325and <tt>ReplaceInstWithInst</tt>.</p>
2326
NAKAMURA Takumi06c6d9a2011-04-18 01:17:51 +0000[diff] [blame]2327<h5><a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a></h5>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2328
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2329<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2330 <li><tt>ReplaceInstWithValue</tt>
2331
Nick Lewyckyb6d1f392008-09-15 06:31:52 +0000[diff] [blame]2332 <p>This function replaces all uses of a given instruction with a value,
2333 and then removes the original instruction. The following example
2334 illustrates the replacement of the result of a particular
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]2335 <tt>AllocaInst</tt> that allocates memory for a single integer with a null
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2336 pointer to an integer.</p>
2337
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2338<div class="doc_code">
2339<pre>
2340AllocaInst* instToReplace = ...;
2341BasicBlock::iterator ii(instToReplace);
2342
2343ReplaceInstWithValue(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
Daniel Dunbar58c2ac02008-10-03 22:17:25 +0000[diff] [blame]2344 Constant::getNullValue(PointerType::getUnqual(Type::Int32Ty)));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2345</pre></div></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2346
2347 <li><tt>ReplaceInstWithInst</tt>
2348
2349 <p>This function replaces a particular instruction with another
Nick Lewyckyb6d1f392008-09-15 06:31:52 +0000[diff] [blame]2350 instruction, inserting the new instruction into the basic block at the
2351 location where the old instruction was, and replacing any uses of the old
2352 instruction with the new instruction. The following example illustrates
2353 the replacement of one <tt>AllocaInst</tt> with another.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2354
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2355<div class="doc_code">
2356<pre>
2357AllocaInst* instToReplace = ...;
2358BasicBlock::iterator ii(instToReplace);
2359
2360ReplaceInstWithInst(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2361 new AllocaInst(Type::Int32Ty, 0, "ptrToReplacedInt"));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2362</pre></div></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2363</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2364
2365<p><i>Replacing multiple uses of <tt>User</tt>s and <tt>Value</tt>s</i></p>
2366
2367<p>You can use <tt>Value::replaceAllUsesWith</tt> and
2368<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]2369doxygen documentation for the <a href="/doxygen/classllvm_1_1Value.html">Value Class</a>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]2370and <a href="/doxygen/classllvm_1_1User.html">User Class</a>, respectively, for more
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2371information.</p>
2372
2373<!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
2374include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
2375ReplaceInstWithValue, ReplaceInstWithInst -->
2376
2377</div>
2378
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2379<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2380<h4>
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2381 <a name="schanges_deletingGV">Deleting <tt>GlobalVariable</tt>s</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2382</h4>
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2383
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2384<div>
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2385
Tanya Lattnerc5dfcdb2007-06-20 20:46:37 +0000[diff] [blame]2386<p>Deleting a global variable from a module is just as easy as deleting an
2387Instruction. First, you must have a pointer to the global variable that you wish
2388 to delete. You use this pointer to erase it from its parent, the module.
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2389 For example:</p>
2390
2391<div class="doc_code">
2392<pre>
2393<a href="#GlobalVariable">GlobalVariable</a> *GV = .. ;
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2394
Tanya Lattnerc5dfcdb2007-06-20 20:46:37 +0000[diff] [blame]2395GV-&gt;eraseFromParent();
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2396</pre>
2397</div>
2398
2399</div>
2400
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2401</div>
2402
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2403<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2404<h3>
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2405 <a name="create_types">How to Create Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2406</h3>
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2407
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2408<div>
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2409
2410<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]2411statically, you can use <tt>TypeBuilder&lt;...&gt;::get()</tt>, defined
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2412in <tt>llvm/Support/TypeBuilder.h</tt>, to retrieve them. <tt>TypeBuilder</tt>
2413has two forms depending on whether you're building types for cross-compilation
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2414or native library use. <tt>TypeBuilder&lt;T, true&gt;</tt> requires
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2415that <tt>T</tt> be independent of the host environment, meaning that it's built
2416out of types from
2417the <a href="/doxygen/namespacellvm_1_1types.html"><tt>llvm::types</tt></a>
2418namespace and pointers, functions, arrays, etc. built of
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2419those. <tt>TypeBuilder&lt;T, false&gt;</tt> additionally allows native C types
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2420whose size may depend on the host compiler. For example,</p>
2421
2422<div class="doc_code">
2423<pre>
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2424FunctionType *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]2425</pre>
2426</div>
2427
2428<p>is easier to read and write than the equivalent</p>
2429
2430<div class="doc_code">
2431<pre>
Owen Anderson5e8c50e2009-06-16 17:40:28 +0000[diff] [blame]2432std::vector&lt;const Type*&gt; params;
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2433params.push_back(PointerType::getUnqual(Type::Int32Ty));
2434FunctionType *ft = FunctionType::get(Type::Int8Ty, params, false);
2435</pre>
2436</div>
2437
2438<p>See the <a href="/doxygen/TypeBuilder_8h-source.html#l00001">class
2439comment</a> for more details.</p>
2440
2441</div>
2442
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2443</div>
2444
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]2445<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2446<h2>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2447 <a name="threading">Threads and LLVM</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2448</h2>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2449<!-- *********************************************************************** -->
2450
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2451<div>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2452<p>
2453This section describes the interaction of the LLVM APIs with multithreading,
2454both on the part of client applications, and in the JIT, in the hosted
2455application.
2456</p>
2457
2458<p>
2459Note that LLVM's support for multithreading is still relatively young. Up
2460through version 2.5, the execution of threaded hosted applications was
2461supported, but not threaded client access to the APIs. While this use case is
2462now supported, clients <em>must</em> adhere to the guidelines specified below to
2463ensure proper operation in multithreaded mode.
2464</p>
2465
2466<p>
2467Note that, on Unix-like platforms, LLVM requires the presence of GCC's atomic
2468intrinsics in order to support threaded operation. If you need a
2469multhreading-capable LLVM on a platform without a suitably modern system
2470compiler, consider compiling LLVM and LLVM-GCC in single-threaded mode, and
2471using the resultant compiler to build a copy of LLVM with multithreading
2472support.
2473</p>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2474
2475<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2476<h3>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2477 <a name="startmultithreaded">Entering and Exiting Multithreaded Mode</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2478</h3>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2479
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2480<div>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2481
2482<p>
2483In order to properly protect its internal data structures while avoiding
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2484excessive locking overhead in the single-threaded case, the LLVM must intialize
2485certain data structures necessary to provide guards around its internals. To do
2486so, the client program must invoke <tt>llvm_start_multithreaded()</tt> before
2487making any concurrent LLVM API calls. To subsequently tear down these
2488structures, use the <tt>llvm_stop_multithreaded()</tt> call. You can also use
2489the <tt>llvm_is_multithreaded()</tt> call to check the status of multithreaded
2490mode.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2491</p>
2492
2493<p>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2494Note that both of these calls must be made <em>in isolation</em>. That is to
2495say that no other LLVM API calls may be executing at any time during the
2496execution of <tt>llvm_start_multithreaded()</tt> or <tt>llvm_stop_multithreaded
2497</tt>. It's is the client's responsibility to enforce this isolation.
2498</p>
2499
2500<p>
2501The return value of <tt>llvm_start_multithreaded()</tt> indicates the success or
2502failure of the initialization. Failure typically indicates that your copy of
2503LLVM was built without multithreading support, typically because GCC atomic
2504intrinsics were not found in your system compiler. In this case, the LLVM API
2505will not be safe for concurrent calls. However, it <em>will</em> be safe for
Jeffrey Yasskin01eba392010-01-29 19:10:38 +0000[diff] [blame]2506hosting threaded applications in the JIT, though <a href="#jitthreading">care
2507must be taken</a> to ensure that side exits and the like do not accidentally
2508result in concurrent LLVM API calls.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2509</p>
2510</div>
2511
2512<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2513<h3>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2514 <a name="shutdown">Ending Execution with <tt>llvm_shutdown()</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2515</h3>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2516
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2517<div>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2518<p>
2519When you are done using the LLVM APIs, you should call <tt>llvm_shutdown()</tt>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2520to deallocate memory used for internal structures. This will also invoke
2521<tt>llvm_stop_multithreaded()</tt> if LLVM is operating in multithreaded mode.
2522As such, <tt>llvm_shutdown()</tt> requires the same isolation guarantees as
2523<tt>llvm_stop_multithreaded()</tt>.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2524</p>
2525
2526<p>
2527Note that, if you use scope-based shutdown, you can use the
2528<tt>llvm_shutdown_obj</tt> class, which calls <tt>llvm_shutdown()</tt> in its
2529destructor.
2530</div>
2531
2532<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2533<h3>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2534 <a name="managedstatic">Lazy Initialization with <tt>ManagedStatic</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2535</h3>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2536
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2537<div>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2538<p>
2539<tt>ManagedStatic</tt> is a utility class in LLVM used to implement static
2540initialization of static resources, such as the global type tables. Before the
2541invocation of <tt>llvm_shutdown()</tt>, it implements a simple lazy
2542initialization scheme. Once <tt>llvm_start_multithreaded()</tt> returns,
2543however, it uses double-checked locking to implement thread-safe lazy
2544initialization.
2545</p>
2546
2547<p>
2548Note that, because no other threads are allowed to issue LLVM API calls before
2549<tt>llvm_start_multithreaded()</tt> returns, it is possible to have
2550<tt>ManagedStatic</tt>s of <tt>llvm::sys::Mutex</tt>s.
2551</p>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2552
2553<p>
2554The <tt>llvm_acquire_global_lock()</tt> and <tt>llvm_release_global_lock</tt>
2555APIs provide access to the global lock used to implement the double-checked
2556locking for lazy initialization. These should only be used internally to LLVM,
2557and only if you know what you're doing!
2558</p>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2559</div>
2560
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]2561<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2562<h3>
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]2563 <a name="llvmcontext">Achieving Isolation with <tt>LLVMContext</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2564</h3>
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]2565
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2566<div>
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]2567<p>
2568<tt>LLVMContext</tt> is an opaque class in the LLVM API which clients can use
2569to operate multiple, isolated instances of LLVM concurrently within the same
2570address space. For instance, in a hypothetical compile-server, the compilation
2571of an individual translation unit is conceptually independent from all the
2572others, and it would be desirable to be able to compile incoming translation
2573units concurrently on independent server threads. Fortunately,
2574<tt>LLVMContext</tt> exists to enable just this kind of scenario!
2575</p>
2576
2577<p>
2578Conceptually, <tt>LLVMContext</tt> provides isolation. Every LLVM entity
2579(<tt>Module</tt>s, <tt>Value</tt>s, <tt>Type</tt>s, <tt>Constant</tt>s, etc.)
Chris Lattner38eee3c2009-08-20 03:10:14 +0000[diff] [blame]2580in LLVM's in-memory IR belongs to an <tt>LLVMContext</tt>. Entities in
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]2581different contexts <em>cannot</em> interact with each other: <tt>Module</tt>s in
2582different contexts cannot be linked together, <tt>Function</tt>s cannot be added
2583to <tt>Module</tt>s in different contexts, etc. What this means is that is is
2584safe to compile on multiple threads simultaneously, as long as no two threads
2585operate on entities within the same context.
2586</p>
2587
2588<p>
2589In practice, very few places in the API require the explicit specification of a
2590<tt>LLVMContext</tt>, other than the <tt>Type</tt> creation/lookup APIs.
2591Because every <tt>Type</tt> carries a reference to its owning context, most
2592other entities can determine what context they belong to by looking at their
2593own <tt>Type</tt>. If you are adding new entities to LLVM IR, please try to
2594maintain this interface design.
2595</p>
2596
2597<p>
2598For clients that do <em>not</em> require the benefits of isolation, LLVM
2599provides a convenience API <tt>getGlobalContext()</tt>. This returns a global,
2600lazily initialized <tt>LLVMContext</tt> that may be used in situations where
2601isolation is not a concern.
2602</p>
2603</div>
2604
Jeffrey Yasskin01eba392010-01-29 19:10:38 +0000[diff] [blame]2605<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2606<h3>
Jeffrey Yasskin01eba392010-01-29 19:10:38 +0000[diff] [blame]2607 <a name="jitthreading">Threads and the JIT</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2608</h3>
Jeffrey Yasskin01eba392010-01-29 19:10:38 +0000[diff] [blame]2609
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2610<div>
Jeffrey Yasskin01eba392010-01-29 19:10:38 +0000[diff] [blame]2611<p>
2612LLVM's "eager" JIT compiler is safe to use in threaded programs. Multiple
2613threads can call <tt>ExecutionEngine::getPointerToFunction()</tt> or
2614<tt>ExecutionEngine::runFunction()</tt> concurrently, and multiple threads can
2615run code output by the JIT concurrently. The user must still ensure that only
2616one thread accesses IR in a given <tt>LLVMContext</tt> while another thread
2617might be modifying it. One way to do that is to always hold the JIT lock while
2618accessing IR outside the JIT (the JIT <em>modifies</em> the IR by adding
2619<tt>CallbackVH</tt>s). Another way is to only
2620call <tt>getPointerToFunction()</tt> from the <tt>LLVMContext</tt>'s thread.
2621</p>
2622
2623<p>When the JIT is configured to compile lazily (using
2624<tt>ExecutionEngine::DisableLazyCompilation(false)</tt>), there is currently a
2625<a href="http://llvm.org/bugs/show_bug.cgi?id=5184">race condition</a> in
2626updating call sites after a function is lazily-jitted. It's still possible to
2627use the lazy JIT in a threaded program if you ensure that only one thread at a
2628time can call any particular lazy stub and that the JIT lock guards any IR
2629access, but we suggest using only the eager JIT in threaded programs.
2630</p>
2631</div>
2632
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2633</div>
2634
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2635<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2636<h2>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2637 <a name="advanced">Advanced Topics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2638</h2>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2639<!-- *********************************************************************** -->
2640
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2641<div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2642<p>
2643This section describes some of the advanced or obscure API's that most clients
2644do not need to be aware of. These API's tend manage the inner workings of the
2645LLVM system, and only need to be accessed in unusual circumstances.
2646</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2647
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2648<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2649<h3>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2650 <a name="TypeResolve">LLVM Type Resolution</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2651</h3>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2652
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2653<div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2654
2655<p>
2656The LLVM type system has a very simple goal: allow clients to compare types for
2657structural equality with a simple pointer comparison (aka a shallow compare).
2658This goal makes clients much simpler and faster, and is used throughout the LLVM
2659system.
2660</p>
2661
2662<p>
2663Unfortunately achieving this goal is not a simple matter. In particular,
2664recursive types and late resolution of opaque types makes the situation very
2665difficult to handle. Fortunately, for the most part, our implementation makes
2666most clients able to be completely unaware of the nasty internal details. The
2667primary case where clients are exposed to the inner workings of it are when
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]2668building a recursive type. In addition to this case, the LLVM bitcode reader,
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2669assembly parser, and linker also have to be aware of the inner workings of this
2670system.
2671</p>
2672
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]2673<p>
2674For our purposes below, we need three concepts. First, an "Opaque Type" is
2675exactly as defined in the <a href="LangRef.html#t_opaque">language
2676reference</a>. Second an "Abstract Type" is any type which includes an
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2677opaque type as part of its type graph (for example "<tt>{ opaque, i32 }</tt>").
2678Third, 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]2679float }</tt>").
2680</p>
2681
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2682<!-- ______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2683<h4>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2684 <a name="BuildRecType">Basic Recursive Type Construction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2685</h4>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2686
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2687<div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2688
2689<p>
2690Because the most common question is "how do I build a recursive type with LLVM",
2691we answer it now and explain it as we go. Here we include enough to cause this
2692to be emitted to an output .ll file:
2693</p>
2694
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2695<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2696<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2697%mylist = type { %mylist*, i32 }
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2698</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2699</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2700
2701<p>
2702To build this, use the following LLVM APIs:
2703</p>
2704
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2705<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2706<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2707// <i>Create the initial outer struct</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2708<a href="#PATypeHolder">PATypeHolder</a> StructTy = OpaqueType::get();
2709std::vector&lt;const Type*&gt; Elts;
Daniel Dunbar58c2ac02008-10-03 22:17:25 +0000[diff] [blame]2710Elts.push_back(PointerType::getUnqual(StructTy));
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2711Elts.push_back(Type::Int32Ty);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2712StructType *NewSTy = StructType::get(Elts);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2713
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2714// <i>At this point, NewSTy = "{ opaque*, i32 }". Tell VMCore that</i>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2715// <i>the struct and the opaque type are actually the same.</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2716cast&lt;OpaqueType&gt;(StructTy.get())-&gt;<a href="#refineAbstractTypeTo">refineAbstractTypeTo</a>(NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2717
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2718// <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]2719// <i>kept up-to-date</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2720NewSTy = cast&lt;StructType&gt;(StructTy.get());
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2721
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2722// <i>Add a name for the type to the module symbol table (optional)</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2723MyModule-&gt;addTypeName("mylist", NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2724</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2725</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2726
2727<p>
2728This code shows the basic approach used to build recursive types: build a
2729non-recursive type using 'opaque', then use type unification to close the cycle.
2730The type unification step is performed by the <tt><a
Chris Lattneraff26d12007-02-03 03:06:52 +0000[diff] [blame]2731href="#refineAbstractTypeTo">refineAbstractTypeTo</a></tt> method, which is
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2732described next. After that, we describe the <a
2733href="#PATypeHolder">PATypeHolder class</a>.
2734</p>
2735
2736</div>
2737
2738<!-- ______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2739<h4>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2740 <a name="refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2741</h4>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2742
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2743<div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2744<p>
2745The <tt>refineAbstractTypeTo</tt> method starts the type unification process.
2746While this method is actually a member of the DerivedType class, it is most
2747often used on OpaqueType instances. Type unification is actually a recursive
2748process. After unification, types can become structurally isomorphic to
2749existing types, and all duplicates are deleted (to preserve pointer equality).
2750</p>
2751
2752<p>
2753In the example above, the OpaqueType object is definitely deleted.
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2754Additionally, if there is an "{ \2*, i32}" type already created in the system,
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2755the pointer and struct type created are <b>also</b> deleted. Obviously whenever
2756a type is deleted, any "Type*" pointers in the program are invalidated. As
2757such, it is safest to avoid having <i>any</i> "Type*" pointers to abstract types
2758live across a call to <tt>refineAbstractTypeTo</tt> (note that non-abstract
2759types can never move or be deleted). To deal with this, the <a
2760href="#PATypeHolder">PATypeHolder</a> class is used to maintain a stable
2761reference to a possibly refined type, and the <a
2762href="#AbstractTypeUser">AbstractTypeUser</a> class is used to update more
2763complex datastructures.
2764</p>
2765
2766</div>
2767
2768<!-- ______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2769<h4>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2770 <a name="PATypeHolder">The PATypeHolder Class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2771</h4>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2772
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2773<div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2774<p>
2775PATypeHolder is a form of a "smart pointer" for Type objects. When VMCore
2776happily goes about nuking types that become isomorphic to existing types, it
2777automatically updates all PATypeHolder objects to point to the new type. In the
2778example above, this allows the code to maintain a pointer to the resultant
2779resolved recursive type, even though the Type*'s are potentially invalidated.
2780</p>
2781
2782<p>
2783PATypeHolder is an extremely light-weight object that uses a lazy union-find
2784implementation to update pointers. For example the pointer from a Value to its
2785Type is maintained by PATypeHolder objects.
2786</p>
2787
2788</div>
2789
2790<!-- ______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2791<h4>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2792 <a name="AbstractTypeUser">The AbstractTypeUser Class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2793</h4>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2794
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2795<div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2796
2797<p>
2798Some data structures need more to perform more complex updates when types get
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2799resolved. To support this, a class can derive from the AbstractTypeUser class.
2800This class
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2801allows it to get callbacks when certain types are resolved. To register to get
2802callbacks for a particular type, the DerivedType::{add/remove}AbstractTypeUser
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]2803methods can be called on a type. Note that these methods only work for <i>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2804 abstract</i> types. Concrete types (those that do not include any opaque
2805objects) can never be refined.
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2806</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2807</div>
2808
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2809</div>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2810
2811<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2812<h3>
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2813 <a name="SymbolTable">The <tt>ValueSymbolTable</tt> and
2814 <tt>TypeSymbolTable</tt> classes</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2815</h3>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2816
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2817<div>
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2818<p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1ValueSymbolTable.html">
2819ValueSymbolTable</a></tt> class provides a symbol table that the <a
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2820href="#Function"><tt>Function</tt></a> and <a href="#Module">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2821<tt>Module</tt></a> classes use for naming value definitions. The symbol table
2822can provide a name for any <a href="#Value"><tt>Value</tt></a>.
2823The <tt><a href="http://llvm.org/doxygen/classllvm_1_1TypeSymbolTable.html">
2824TypeSymbolTable</a></tt> class is used by the <tt>Module</tt> class to store
2825names for types.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2826
Reid Spencera6362242007-01-07 00:41:39 +0000[diff] [blame]2827<p>Note that the <tt>SymbolTable</tt> class should not be directly accessed
2828by most clients. It should only be used when iteration over the symbol table
2829names themselves are required, which is very special purpose. Note that not
2830all LLVM
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2831<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]2832an empty name) do not exist in the symbol table.
2833</p>
2834
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2835<p>These symbol tables support iteration over the values/types in the symbol
2836table with <tt>begin/end/iterator</tt> and supports querying to see if a
2837specific name is in the symbol table (with <tt>lookup</tt>). The
2838<tt>ValueSymbolTable</tt> class exposes no public mutator methods, instead,
2839simply call <tt>setName</tt> on a value, which will autoinsert it into the
2840appropriate symbol table. For types, use the Module::addTypeName method to
2841insert entries into the symbol table.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2842
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2843</div>
2844
2845
2846
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2847<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2848<h3>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2849 <a name="UserLayout">The <tt>User</tt> and owned <tt>Use</tt> classes' memory layout</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2850</h3>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2851
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2852<div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2853<p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1User.html">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2854User</a></tt> class provides a basis for expressing the ownership of <tt>User</tt>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2855towards other <tt><a href="http://llvm.org/doxygen/classllvm_1_1Value.html">
2856Value</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]2857Use</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]2858addition and removal.</p>
2859
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2860<!-- ______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2861<h4>
2862 <a name="Use2User">
2863 Interaction and relationship between <tt>User</tt> and <tt>Use</tt> objects
2864 </a>
2865</h4>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2866
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2867<div>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2868<p>
2869A subclass of <tt>User</tt> can choose between incorporating its <tt>Use</tt> objects
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2870or refer to them out-of-line by means of a pointer. A mixed variant
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2871(some <tt>Use</tt>s inline others hung off) is impractical and breaks the invariant
2872that the <tt>Use</tt> objects belonging to the same <tt>User</tt> form a contiguous array.
2873</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2874
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2875<p>
2876We have 2 different layouts in the <tt>User</tt> (sub)classes:
2877<ul>
2878<li><p>Layout a)
2879The <tt>Use</tt> object(s) are inside (resp. at fixed offset) of the <tt>User</tt>
2880object and there are a fixed number of them.</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2881
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2882<li><p>Layout b)
2883The <tt>Use</tt> object(s) are referenced by a pointer to an
2884array from the <tt>User</tt> object and there may be a variable
2885number of them.</p>
2886</ul>
2887<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2888As of v2.4 each layout still possesses a direct pointer to the
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2889start of the array of <tt>Use</tt>s. Though not mandatory for layout a),
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2890we stick to this redundancy for the sake of simplicity.
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2891The <tt>User</tt> object also stores the number of <tt>Use</tt> objects it
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2892has. (Theoretically this information can also be calculated
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2893given the scheme presented below.)</p>
2894<p>
2895Special forms of allocation operators (<tt>operator new</tt>)
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2896enforce the following memory layouts:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2897
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2898<ul>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2899<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]2900
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2901<pre>
2902...---.---.---.---.-------...
2903 | P | P | P | P | User
2904'''---'---'---'---'-------'''
2905</pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2906
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2907<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]2908<pre>
2909.-------...
2910| User
2911'-------'''
2912 |
2913 v
2914 .---.---.---.---...
2915 | P | P | P | P |
2916 '---'---'---'---'''
2917</pre>
2918</ul>
2919<i>(In the above figures '<tt>P</tt>' stands for the <tt>Use**</tt> that
2920 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]2921
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2922</div>
2923
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2924<!-- ______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2925<h4>
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2926 <a name="Waymarking">The waymarking algorithm</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2927</h4>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2928
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2929<div>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2930<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2931Since the <tt>Use</tt> objects are deprived of the direct (back)pointer to
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2932their <tt>User</tt> objects, there must be a fast and exact method to
2933recover it. This is accomplished by the following scheme:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2934
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2935A 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]2936start of the <tt>User</tt> object:
2937<ul>
2938<li><tt>00</tt> &mdash;&gt; binary digit 0</li>
2939<li><tt>01</tt> &mdash;&gt; binary digit 1</li>
2940<li><tt>10</tt> &mdash;&gt; stop and calculate (<tt>s</tt>)</li>
2941<li><tt>11</tt> &mdash;&gt; full stop (<tt>S</tt>)</li>
2942</ul>
2943<p>
2944Given a <tt>Use*</tt>, all we have to do is to walk till we get
2945a stop and we either have a <tt>User</tt> immediately behind or
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2946we have to walk to the next stop picking up digits
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2947and calculating the offset:</p>
2948<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2949.---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.----------------
2950| 1 | s | 1 | 0 | 1 | 0 | s | 1 | 1 | 0 | s | 1 | 1 | s | 1 | S | User (or User*)
2951'---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'----------------
2952 |+15 |+10 |+6 |+3 |+1
2953 | | | | |__>
2954 | | | |__________>
2955 | | |______________________>
2956 | |______________________________________>
2957 |__________________________________________________________>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2958</pre>
2959<p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2960Only the significant number of bits need to be stored between the
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2961stops, so that the <i>worst case is 20 memory accesses</i> when there are
29621000 <tt>Use</tt> objects associated with a <tt>User</tt>.</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2963
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2964</div>
2965
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2966<!-- ______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2967<h4>
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2968 <a name="ReferenceImpl">Reference implementation</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2969</h4>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2970
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2971<div>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2972<p>
2973The following literate Haskell fragment demonstrates the concept:</p>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2974
2975<div class="doc_code">
2976<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2977> import Test.QuickCheck
2978>
2979> digits :: Int -> [Char] -> [Char]
2980> digits 0 acc = '0' : acc
2981> digits 1 acc = '1' : acc
2982> digits n acc = digits (n `div` 2) $ digits (n `mod` 2) acc
2983>
2984> dist :: Int -> [Char] -> [Char]
2985> dist 0 [] = ['S']
2986> dist 0 acc = acc
2987> dist 1 acc = let r = dist 0 acc in 's' : digits (length r) r
2988> dist n acc = dist (n - 1) $ dist 1 acc
2989>
2990> takeLast n ss = reverse $ take n $ reverse ss
2991>
2992> test = takeLast 40 $ dist 20 []
2993>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2994</pre>
2995</div>
2996<p>
2997Printing &lt;test&gt; gives: <tt>"1s100000s11010s10100s1111s1010s110s11s1S"</tt></p>
2998<p>
2999The reverse algorithm computes the length of the string just by examining
3000a certain prefix:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]3001
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]3002<div class="doc_code">
3003<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]3004> pref :: [Char] -> Int
3005> pref "S" = 1
3006> pref ('s':'1':rest) = decode 2 1 rest
3007> pref (_:rest) = 1 + pref rest
3008>
3009> decode walk acc ('0':rest) = decode (walk + 1) (acc * 2) rest
3010> decode walk acc ('1':rest) = decode (walk + 1) (acc * 2 + 1) rest
3011> decode walk acc _ = walk + acc
3012>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]3013</pre>
3014</div>
3015<p>
3016Now, as expected, printing &lt;pref test&gt; gives <tt>40</tt>.</p>
3017<p>
3018We can <i>quickCheck</i> this with following property:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]3019
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]3020<div class="doc_code">
3021<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]3022> testcase = dist 2000 []
3023> testcaseLength = length testcase
3024>
3025> identityProp n = n > 0 && n <= testcaseLength ==> length arr == pref arr
3026> where arr = takeLast n testcase
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]3027>
3028</pre>
3029</div>
3030<p>
3031As expected &lt;quickCheck identityProp&gt; gives:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]3032
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]3033<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]3034*Main> quickCheck identityProp
3035OK, passed 100 tests.
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]3036</pre>
3037<p>
3038Let's be a bit more exhaustive:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]3039
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]3040<div class="doc_code">
3041<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]3042>
3043> deepCheck p = check (defaultConfig { configMaxTest = 500 }) p
3044>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]3045</pre>
3046</div>
3047<p>
3048And here is the result of &lt;deepCheck identityProp&gt;:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]3049
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]3050<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]3051*Main> deepCheck identityProp
3052OK, passed 500 tests.
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]3053</pre>
3054
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3055</div>
3056
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]3057<!-- ______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3058<h4>
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]3059 <a name="Tagging">Tagging considerations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3060</h4>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]3061
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3062<div>
3063
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]3064<p>
3065To maintain the invariant that the 2 LSBits of each <tt>Use**</tt> in <tt>Use</tt>
3066never change after being set up, setters of <tt>Use::Prev</tt> must re-tag the
3067new <tt>Use**</tt> on every modification. Accordingly getters must strip the
3068tag bits.</p>
3069<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]3070For layout b) instead of the <tt>User</tt> we find a pointer (<tt>User*</tt> with LSBit set).
3071Following this pointer brings us to the <tt>User</tt>. A portable trick ensures
3072that the first bytes of <tt>User</tt> (if interpreted as a pointer) never has
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]3073the LSBit set. (Portability is relying on the fact that all known compilers place the
3074<tt>vptr</tt> in the first word of the instances.)</p>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]3075
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]3076</div>
3077
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3078</div>
3079
3080</div>
3081
3082<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3083<h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3084 <a name="coreclasses">The Core LLVM Class Hierarchy Reference </a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3085</h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3086<!-- *********************************************************************** -->
3087
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3088<div>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3089<p><tt>#include "<a href="/doxygen/Type_8h-source.html">llvm/Type.h</a>"</tt>
3090<br>doxygen info: <a href="/doxygen/classllvm_1_1Type.html">Type Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3091
3092<p>The Core LLVM classes are the primary means of representing the program
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3093being inspected or transformed. The core LLVM classes are defined in
3094header files in the <tt>include/llvm/</tt> directory, and implemented in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3095the <tt>lib/VMCore</tt> directory.</p>
3096
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3097<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3098<h3>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3099 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3100</h3>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3101
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3102<div>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3103
3104 <p><tt>Type</tt> is a superclass of all type classes. Every <tt>Value</tt> has
3105 a <tt>Type</tt>. <tt>Type</tt> cannot be instantiated directly but only
3106 through its subclasses. Certain primitive types (<tt>VoidType</tt>,
3107 <tt>LabelType</tt>, <tt>FloatType</tt> and <tt>DoubleType</tt>) have hidden
3108 subclasses. They are hidden because they offer no useful functionality beyond
3109 what the <tt>Type</tt> class offers except to distinguish themselves from
3110 other subclasses of <tt>Type</tt>.</p>
3111 <p>All other types are subclasses of <tt>DerivedType</tt>. Types can be
3112 named, but this is not a requirement. There exists exactly
3113 one instance of a given shape at any one time. This allows type equality to
3114 be performed with address equality of the Type Instance. That is, given two
3115 <tt>Type*</tt> values, the types are identical if the pointers are identical.
3116 </p>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3117
3118<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3119<h4>
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]3120 <a name="m_Type">Important Public Methods</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3121</h4>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3122
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3123<div>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3124
3125<ul>
Duncan Sandsb0bc6c32010-02-15 16:12:20 +0000[diff] [blame]3126 <li><tt>bool isIntegerTy() const</tt>: Returns true for any integer type.</li>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3127
Duncan Sandsb0bc6c32010-02-15 16:12:20 +0000[diff] [blame]3128 <li><tt>bool isFloatingPointTy()</tt>: Return true if this is one of the five
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3129 floating point types.</li>
3130
3131 <li><tt>bool isAbstract()</tt>: Return true if the type is abstract (contains
3132 an OpaqueType anywhere in its definition).</li>
3133
3134 <li><tt>bool isSized()</tt>: Return true if the type has known size. Things
3135 that don't have a size are abstract types, labels and void.</li>
3136
3137</ul>
3138</div>
3139
3140<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3141<h4>
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]3142 <a name="derivedtypes">Important Derived Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3143</h4>
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3144<div>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3145<dl>
3146 <dt><tt>IntegerType</tt></dt>
3147 <dd>Subclass of DerivedType that represents integer types of any bit width.
3148 Any bit width between <tt>IntegerType::MIN_INT_BITS</tt> (1) and
3149 <tt>IntegerType::MAX_INT_BITS</tt> (~8 million) can be represented.
3150 <ul>
3151 <li><tt>static const IntegerType* get(unsigned NumBits)</tt>: get an integer
3152 type of a specific bit width.</li>
3153 <li><tt>unsigned getBitWidth() const</tt>: Get the bit width of an integer
3154 type.</li>
3155 </ul>
3156 </dd>
3157 <dt><tt>SequentialType</tt></dt>
3158 <dd>This is subclassed by ArrayType and PointerType
3159 <ul>
3160 <li><tt>const Type * getElementType() const</tt>: Returns the type of each
3161 of the elements in the sequential type. </li>
3162 </ul>
3163 </dd>
3164 <dt><tt>ArrayType</tt></dt>
3165 <dd>This is a subclass of SequentialType and defines the interface for array
3166 types.
3167 <ul>
3168 <li><tt>unsigned getNumElements() const</tt>: Returns the number of
3169 elements in the array. </li>
3170 </ul>
3171 </dd>
3172 <dt><tt>PointerType</tt></dt>
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]3173 <dd>Subclass of SequentialType for pointer types.</dd>
Reid Spencer9d6565a2007-02-15 02:26:10 +0000[diff] [blame]3174 <dt><tt>VectorType</tt></dt>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]3175 <dd>Subclass of SequentialType for vector types. A
3176 vector type is similar to an ArrayType but is distinguished because it is
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3177 a first class type whereas ArrayType is not. Vector types are used for
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3178 vector operations and are usually small vectors of of an integer or floating
3179 point type.</dd>
3180 <dt><tt>StructType</tt></dt>
3181 <dd>Subclass of DerivedTypes for struct types.</dd>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3182 <dt><tt><a name="FunctionType">FunctionType</a></tt></dt>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3183 <dd>Subclass of DerivedTypes for function types.
3184 <ul>
Dan Gohman4bb31bf2010-03-30 20:04:57 +0000[diff] [blame]3185 <li><tt>bool isVarArg() const</tt>: Returns true if it's a vararg
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3186 function</li>
3187 <li><tt> const Type * getReturnType() const</tt>: Returns the
3188 return type of the function.</li>
3189 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns
3190 the type of the ith parameter.</li>
3191 <li><tt> const unsigned getNumParams() const</tt>: Returns the
3192 number of formal parameters.</li>
3193 </ul>
3194 </dd>
3195 <dt><tt>OpaqueType</tt></dt>
3196 <dd>Sublcass of DerivedType for abstract types. This class
3197 defines no content and is used as a placeholder for some other type. Note
3198 that OpaqueType is used (temporarily) during type resolution for forward
3199 references of types. Once the referenced type is resolved, the OpaqueType
3200 is replaced with the actual type. OpaqueType can also be used for data
3201 abstraction. At link time opaque types can be resolved to actual types
3202 of the same name.</dd>
3203</dl>
3204</div>
3205
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3206</div>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3207
3208<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3209<h3>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3210 <a name="Module">The <tt>Module</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3211</h3>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3212
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3213<div>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3214
3215<p><tt>#include "<a
3216href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt><br> doxygen info:
3217<a href="/doxygen/classllvm_1_1Module.html">Module Class</a></p>
3218
3219<p>The <tt>Module</tt> class represents the top level structure present in LLVM
3220programs. An LLVM module is effectively either a translation unit of the
3221original program or a combination of several translation units merged by the
3222linker. The <tt>Module</tt> class keeps track of a list of <a
3223href="#Function"><tt>Function</tt></a>s, a list of <a
3224href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
3225href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
3226helpful member functions that try to make common operations easy.</p>
3227
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3228<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3229<h4>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3230 <a name="m_Module">Important Public Members of the <tt>Module</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3231</h4>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3232
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3233<div>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3234
3235<ul>
3236 <li><tt>Module::Module(std::string name = "")</tt></li>
3237</ul>
3238
3239<p>Constructing a <a href="#Module">Module</a> is easy. You can optionally
3240provide a name for it (probably based on the name of the translation unit).</p>
3241
3242<ul>
3243 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
3244 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
3245
3246 <tt>begin()</tt>, <tt>end()</tt>
3247 <tt>size()</tt>, <tt>empty()</tt>
3248
3249 <p>These are forwarding methods that make it easy to access the contents of
3250 a <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
3251 list.</p></li>
3252
3253 <li><tt>Module::FunctionListType &amp;getFunctionList()</tt>
3254
3255 <p> Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
3256 necessary to use when you need to update the list or perform a complex
3257 action that doesn't have a forwarding method.</p>
3258
3259 <p><!-- Global Variable --></p></li>
3260</ul>
3261
3262<hr>
3263
3264<ul>
3265 <li><tt>Module::global_iterator</tt> - Typedef for global variable list iterator<br>
3266
3267 <tt>Module::const_global_iterator</tt> - Typedef for const_iterator.<br>
3268
3269 <tt>global_begin()</tt>, <tt>global_end()</tt>
3270 <tt>global_size()</tt>, <tt>global_empty()</tt>
3271
3272 <p> These are forwarding methods that make it easy to access the contents of
3273 a <tt>Module</tt> object's <a
3274 href="#GlobalVariable"><tt>GlobalVariable</tt></a> list.</p></li>
3275
3276 <li><tt>Module::GlobalListType &amp;getGlobalList()</tt>
3277
3278 <p>Returns the list of <a
3279 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s. This is necessary to
3280 use when you need to update the list or perform a complex action that
3281 doesn't have a forwarding method.</p>
3282
3283 <p><!-- Symbol table stuff --> </p></li>
3284</ul>
3285
3286<hr>
3287
3288<ul>
3289 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
3290
3291 <p>Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
3292 for this <tt>Module</tt>.</p>
3293
3294 <p><!-- Convenience methods --></p></li>
3295</ul>
3296
3297<hr>
3298
3299<ul>
3300 <li><tt><a href="#Function">Function</a> *getFunction(const std::string
3301 &amp;Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt>
3302
3303 <p>Look up the specified function in the <tt>Module</tt> <a
3304 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
3305 <tt>null</tt>.</p></li>
3306
3307 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const
3308 std::string &amp;Name, const <a href="#FunctionType">FunctionType</a> *T)</tt>
3309
3310 <p>Look up the specified function in the <tt>Module</tt> <a
3311 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
3312 external declaration for the function and return it.</p></li>
3313
3314 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt>
3315
3316 <p>If there is at least one entry in the <a
3317 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
3318 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
3319 string.</p></li>
3320
3321 <li><tt>bool addTypeName(const std::string &amp;Name, const <a
3322 href="#Type">Type</a> *Ty)</tt>
3323
3324 <p>Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
3325 mapping <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this
3326 name, true is returned and the <a
3327 href="#SymbolTable"><tt>SymbolTable</tt></a> is not modified.</p></li>
3328</ul>
3329
3330</div>
3331
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3332</div>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3333
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3334<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3335<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3336 <a name="Value">The <tt>Value</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3337</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3338
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3339<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3340
3341<p><tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt>
3342<br>
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]3343doxygen info: <a href="/doxygen/classllvm_1_1Value.html">Value Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3344
3345<p>The <tt>Value</tt> class is the most important class in the LLVM Source
3346base. It represents a typed value that may be used (among other things) as an
3347operand to an instruction. There are many different types of <tt>Value</tt>s,
3348such as <a href="#Constant"><tt>Constant</tt></a>s,<a
3349href="#Argument"><tt>Argument</tt></a>s. Even <a
3350href="#Instruction"><tt>Instruction</tt></a>s and <a
3351href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.</p>
3352
3353<p>A particular <tt>Value</tt> may be used many times in the LLVM representation
3354for a program. For example, an incoming argument to a function (represented
3355with an instance of the <a href="#Argument">Argument</a> class) is "used" by
3356every instruction in the function that references the argument. To keep track
3357of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
3358href="#User"><tt>User</tt></a>s that is using it (the <a
3359href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
3360graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
3361def-use information in the program, and is accessible through the <tt>use_</tt>*
3362methods, shown below.</p>
3363
3364<p>Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed,
3365and this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
3366method. In addition, all LLVM values can be named. The "name" of the
3367<tt>Value</tt> is a symbolic string printed in the LLVM code:</p>
3368
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3369<div class="doc_code">
3370<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]3371%<b>foo</b> = add i32 1, 2
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3372</pre>
3373</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3374
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3375<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]3376that the name of any value may be missing (an empty string), so names should
3377<b>ONLY</b> be used for debugging (making the source code easier to read,
3378debugging printouts), they should not be used to keep track of values or map
3379between them. For this purpose, use a <tt>std::map</tt> of pointers to the
3380<tt>Value</tt> itself instead.</p>
3381
3382<p>One important aspect of LLVM is that there is no distinction between an SSA
3383variable and the operation that produces it. Because of this, any reference to
3384the value produced by an instruction (or the value available as an incoming
Chris Lattnerd5fc4fc2004-03-18 14:58:55 +0000[diff] [blame]3385argument, for example) is represented as a direct pointer to the instance of
3386the class that
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3387represents this value. Although this may take some getting used to, it
3388simplifies the representation and makes it easier to manipulate.</p>
3389
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3390<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3391<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3392 <a name="m_Value">Important Public Members of the <tt>Value</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3393</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3394
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3395<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3396
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3397<ul>
3398 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the
3399use-list<br>
Gabor Greifbbbf9a22010-03-26 19:59:25 +0000[diff] [blame]3400 <tt>Value::const_use_iterator</tt> - Typedef for const_iterator over
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3401the use-list<br>
3402 <tt>unsigned use_size()</tt> - Returns the number of users of the
3403value.<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3404 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3405 <tt>use_iterator use_begin()</tt> - Get an iterator to the start of
3406the use-list.<br>
3407 <tt>use_iterator use_end()</tt> - Get an iterator to the end of the
3408use-list.<br>
3409 <tt><a href="#User">User</a> *use_back()</tt> - Returns the last
3410element in the list.
3411 <p> These methods are the interface to access the def-use
3412information in LLVM. As with all other iterators in LLVM, the naming
3413conventions follow the conventions defined by the <a href="#stl">STL</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3414 </li>
3415 <li><tt><a href="#Type">Type</a> *getType() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3416 <p>This method returns the Type of the Value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3417 </li>
3418 <li><tt>bool hasName() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3419 <tt>std::string getName() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3420 <tt>void setName(const std::string &amp;Name)</tt>
3421 <p> This family of methods is used to access and assign a name to a <tt>Value</tt>,
3422be aware of the <a href="#nameWarning">precaution above</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3423 </li>
3424 <li><tt>void replaceAllUsesWith(Value *V)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3425
3426 <p>This method traverses the use list of a <tt>Value</tt> changing all <a
3427 href="#User"><tt>User</tt>s</a> of the current value to refer to
3428 "<tt>V</tt>" instead. For example, if you detect that an instruction always
3429 produces a constant value (for example through constant folding), you can
3430 replace all uses of the instruction with the constant like this:</p>
3431
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3432<div class="doc_code">
3433<pre>
3434Inst-&gt;replaceAllUsesWith(ConstVal);
3435</pre>
3436</div>
3437
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3438</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3439
3440</div>
3441
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3442</div>
3443
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3444<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3445<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3446 <a name="User">The <tt>User</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3447</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3448
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3449<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3450
3451<p>
3452<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]3453doxygen info: <a href="/doxygen/classllvm_1_1User.html">User Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3454Superclass: <a href="#Value"><tt>Value</tt></a></p>
3455
3456<p>The <tt>User</tt> class is the common base class of all LLVM nodes that may
3457refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
3458that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
3459referring to. The <tt>User</tt> class itself is a subclass of
3460<tt>Value</tt>.</p>
3461
3462<p>The operands of a <tt>User</tt> point directly to the LLVM <a
3463href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
3464Single Assignment (SSA) form, there can only be one definition referred to,
3465allowing this direct connection. This connection provides the use-def
3466information in LLVM.</p>
3467
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3468<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3469<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3470 <a name="m_User">Important Public Members of the <tt>User</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3471</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3472
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3473<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3474
3475<p>The <tt>User</tt> class exposes the operand list in two ways: through
3476an index access interface and through an iterator based interface.</p>
3477
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3478<ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3479 <li><tt>Value *getOperand(unsigned i)</tt><br>
3480 <tt>unsigned getNumOperands()</tt>
3481 <p> These two methods expose the operands of the <tt>User</tt> in a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3482convenient form for direct access.</p></li>
3483
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3484 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand
3485list<br>
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]3486 <tt>op_iterator op_begin()</tt> - Get an iterator to the start of
3487the operand list.<br>
3488 <tt>op_iterator op_end()</tt> - Get an iterator to the end of the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3489operand list.
3490 <p> Together, these methods make up the iterator based interface to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3491the operands of a <tt>User</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3492</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3493
3494</div>
3495
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3496</div>
3497
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3498<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3499<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3500 <a name="Instruction">The <tt>Instruction</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3501</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3502
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3503<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3504
3505<p><tt>#include "</tt><tt><a
3506href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt><br>
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3507doxygen info: <a href="/doxygen/classllvm_1_1Instruction.html">Instruction Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3508Superclasses: <a href="#User"><tt>User</tt></a>, <a
3509href="#Value"><tt>Value</tt></a></p>
3510
3511<p>The <tt>Instruction</tt> class is the common base class for all LLVM
3512instructions. It provides only a few methods, but is a very commonly used
3513class. The primary data tracked by the <tt>Instruction</tt> class itself is the
3514opcode (instruction type) and the parent <a
3515href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
3516into. To represent a specific type of instruction, one of many subclasses of
3517<tt>Instruction</tt> are used.</p>
3518
3519<p> Because the <tt>Instruction</tt> class subclasses the <a
3520href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
3521way as for other <a href="#User"><tt>User</tt></a>s (with the
3522<tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
3523<tt>op_begin()</tt>/<tt>op_end()</tt> methods).</p> <p> An important file for
3524the <tt>Instruction</tt> class is the <tt>llvm/Instruction.def</tt> file. This
3525file contains some meta-data about the various different types of instructions
3526in LLVM. It describes the enum values that are used as opcodes (for example
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3527<tt>Instruction::Add</tt> and <tt>Instruction::ICmp</tt>), as well as the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3528concrete sub-classes of <tt>Instruction</tt> that implement the instruction (for
3529example <tt><a href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3530href="#CmpInst">CmpInst</a></tt>). Unfortunately, the use of macros in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3531this file confuses doxygen, so these enum values don't show up correctly in the
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3532<a href="/doxygen/classllvm_1_1Instruction.html">doxygen output</a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3533
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3534<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3535<h4>
3536 <a name="s_Instruction">
3537 Important Subclasses of the <tt>Instruction</tt> class
3538 </a>
3539</h4>
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3540<div>
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3541 <ul>
3542 <li><tt><a name="BinaryOperator">BinaryOperator</a></tt>
3543 <p>This subclasses represents all two operand instructions whose operands
3544 must be the same type, except for the comparison instructions.</p></li>
3545 <li><tt><a name="CastInst">CastInst</a></tt>
3546 <p>This subclass is the parent of the 12 casting instructions. It provides
3547 common operations on cast instructions.</p>
3548 <li><tt><a name="CmpInst">CmpInst</a></tt>
3549 <p>This subclass respresents the two comparison instructions,
3550 <a href="LangRef.html#i_icmp">ICmpInst</a> (integer opreands), and
3551 <a href="LangRef.html#i_fcmp">FCmpInst</a> (floating point operands).</p>
3552 <li><tt><a name="TerminatorInst">TerminatorInst</a></tt>
3553 <p>This subclass is the parent of all terminator instructions (those which
3554 can terminate a block).</p>
3555 </ul>
3556 </div>
3557
3558<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3559<h4>
3560 <a name="m_Instruction">
3561 Important Public Members of the <tt>Instruction</tt> class
3562 </a>
3563</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3564
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3565<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3566
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3567<ul>
3568 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3569 <p>Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that
3570this <tt>Instruction</tt> is embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3571 <li><tt>bool mayWriteToMemory()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3572 <p>Returns true if the instruction writes to memory, i.e. it is a
3573 <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]3574 <li><tt>unsigned getOpcode()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3575 <p>Returns the opcode for the <tt>Instruction</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3576 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3577 <p>Returns another instance of the specified instruction, identical
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3578in all ways to the original except that the instruction has no parent
3579(ie it's not embedded into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>),
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3580and it has no name</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3581</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3582
3583</div>
3584
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3585</div>
3586
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3587<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3588<h3>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3589 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3590</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3591
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3592<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3593
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3594<p>Constant represents a base class for different types of constants. It
3595is subclassed by ConstantInt, ConstantArray, etc. for representing
3596the various types of Constants. <a href="#GlobalValue">GlobalValue</a> is also
3597a subclass, which represents the address of a global variable or function.
3598</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3599
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3600<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3601<h4>Important Subclasses of Constant</h4>
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3602<div>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3603<ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3604 <li>ConstantInt : This subclass of Constant represents an integer constant of
3605 any width.
3606 <ul>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]3607 <li><tt>const APInt&amp; getValue() const</tt>: Returns the underlying
3608 value of this constant, an APInt value.</li>
3609 <li><tt>int64_t getSExtValue() const</tt>: Converts the underlying APInt
3610 value to an int64_t via sign extension. If the value (not the bit width)
3611 of the APInt is too large to fit in an int64_t, an assertion will result.
3612 For this reason, use of this method is discouraged.</li>
3613 <li><tt>uint64_t getZExtValue() const</tt>: Converts the underlying APInt
3614 value to a uint64_t via zero extension. IF the value (not the bit width)
3615 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]3616 For this reason, use of this method is discouraged.</li>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]3617 <li><tt>static ConstantInt* get(const APInt&amp; Val)</tt>: Returns the
3618 ConstantInt object that represents the value provided by <tt>Val</tt>.
3619 The type is implied as the IntegerType that corresponds to the bit width
3620 of <tt>Val</tt>.</li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3621 <li><tt>static ConstantInt* get(const Type *Ty, uint64_t Val)</tt>:
3622 Returns the ConstantInt object that represents the value provided by
3623 <tt>Val</tt> for integer type <tt>Ty</tt>.</li>
3624 </ul>
3625 </li>
3626 <li>ConstantFP : This class represents a floating point constant.
3627 <ul>
3628 <li><tt>double getValue() const</tt>: Returns the underlying value of
3629 this constant. </li>
3630 </ul>
3631 </li>
3632 <li>ConstantArray : This represents a constant array.
3633 <ul>
3634 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
3635 a vector of component constants that makeup this array. </li>
3636 </ul>
3637 </li>
3638 <li>ConstantStruct : This represents a constant struct.
3639 <ul>
3640 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
3641 a vector of component constants that makeup this array. </li>
3642 </ul>
3643 </li>
3644 <li>GlobalValue : This represents either a global variable or a function. In
3645 either case, the value is a constant fixed address (after linking).
3646 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3647</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3648</div>
3649
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3650</div>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3651
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3652<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3653<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3654 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3655</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3656
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3657<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3658
3659<p><tt>#include "<a
3660href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt><br>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]3661doxygen info: <a href="/doxygen/classllvm_1_1GlobalValue.html">GlobalValue
3662Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3663Superclasses: <a href="#Constant"><tt>Constant</tt></a>,
3664<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]3665
3666<p>Global values (<a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
3667href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
3668visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
3669Because they are visible at global scope, they are also subject to linking with
3670other globals defined in different translation units. To control the linking
3671process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
3672<tt>GlobalValue</tt>s know whether they have internal or external linkage, as
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]3673defined by the <tt>LinkageTypes</tt> enumeration.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3674
3675<p>If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
3676<tt>static</tt> in C), it is not visible to code outside the current translation
3677unit, and does not participate in linking. If it has external linkage, it is
3678visible to external code, and does participate in linking. In addition to
3679linkage information, <tt>GlobalValue</tt>s keep track of which <a
3680href="#Module"><tt>Module</tt></a> they are currently part of.</p>
3681
3682<p>Because <tt>GlobalValue</tt>s are memory objects, they are always referred to
3683by their <b>address</b>. As such, the <a href="#Type"><tt>Type</tt></a> of a
3684global is always a pointer to its contents. It is important to remember this
3685when using the <tt>GetElementPtrInst</tt> instruction because this pointer must
3686be dereferenced first. For example, if you have a <tt>GlobalVariable</tt> (a
3687subclass 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]3688i32]</tt>, then the <tt>GlobalVariable</tt> is a pointer to that array. Although
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3689the address of the first element of this array and the value of the
3690<tt>GlobalVariable</tt> are the same, they have different types. The
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]3691<tt>GlobalVariable</tt>'s type is <tt>[24 x i32]</tt>. The first element's type
3692is <tt>i32.</tt> Because of this, accessing a global value requires you to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3693dereference the pointer with <tt>GetElementPtrInst</tt> first, then its elements
3694can be accessed. This is explained in the <a href="LangRef.html#globalvars">LLVM
3695Language Reference Manual</a>.</p>
3696
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3697<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3698<h4>
3699 <a name="m_GlobalValue">
3700 Important Public Members of the <tt>GlobalValue</tt> class
3701 </a>
3702</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3703
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3704<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3705
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3706<ul>
3707 <li><tt>bool hasInternalLinkage() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3708 <tt>bool hasExternalLinkage() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3709 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt>
3710 <p> These methods manipulate the linkage characteristics of the <tt>GlobalValue</tt>.</p>
3711 <p> </p>
3712 </li>
3713 <li><tt><a href="#Module">Module</a> *getParent()</tt>
3714 <p> This returns the <a href="#Module"><tt>Module</tt></a> that the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3715GlobalValue is currently embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3716</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3717
3718</div>
3719
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3720</div>
3721
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3722<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3723<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3724 <a name="Function">The <tt>Function</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3725</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3726
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3727<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3728
3729<p><tt>#include "<a
3730href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt><br> doxygen
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3731info: <a href="/doxygen/classllvm_1_1Function.html">Function Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3732Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
3733<a href="#Constant"><tt>Constant</tt></a>,
3734<a href="#User"><tt>User</tt></a>,
3735<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3736
3737<p>The <tt>Function</tt> class represents a single procedure in LLVM. It is
Torok Edwin87469292009-10-12 13:37:29 +0000[diff] [blame]3738actually one of the more complex classes in the LLVM hierarchy because it must
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3739keep track of a large amount of data. The <tt>Function</tt> class keeps track
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3740of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal
3741<a href="#Argument"><tt>Argument</tt></a>s, and a
3742<a href="#SymbolTable"><tt>SymbolTable</tt></a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3743
3744<p>The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most
3745commonly used part of <tt>Function</tt> objects. The list imposes an implicit
3746ordering of the blocks in the function, which indicate how the code will be
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3747laid out by the backend. Additionally, the first <a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3748href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
3749<tt>Function</tt>. It is not legal in LLVM to explicitly branch to this initial
3750block. There are no implicit exit nodes, and in fact there may be multiple exit
3751nodes from a single <tt>Function</tt>. If the <a
3752href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
3753the <tt>Function</tt> is actually a function declaration: the actual body of the
3754function hasn't been linked in yet.</p>
3755
3756<p>In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
3757<tt>Function</tt> class also keeps track of the list of formal <a
3758href="#Argument"><tt>Argument</tt></a>s that the function receives. This
3759container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
3760nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
3761the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.</p>
3762
3763<p>The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used
3764LLVM feature that is only used when you have to look up a value by name. Aside
3765from that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used
3766internally to make sure that there are not conflicts between the names of <a
3767href="#Instruction"><tt>Instruction</tt></a>s, <a
3768href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
3769href="#Argument"><tt>Argument</tt></a>s in the function body.</p>
3770
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]3771<p>Note that <tt>Function</tt> is a <a href="#GlobalValue">GlobalValue</a>
3772and therefore also a <a href="#Constant">Constant</a>. The value of the function
3773is its address (after linking) which is guaranteed to be constant.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3774
3775<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3776<h4>
3777 <a name="m_Function">
3778 Important Public Members of the <tt>Function</tt> class
3779 </a>
3780</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3781
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3782<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3783
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3784<ul>
3785 <li><tt>Function(const </tt><tt><a href="#FunctionType">FunctionType</a>
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]3786 *Ty, LinkageTypes Linkage, const std::string &amp;N = "", Module* Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3787
3788 <p>Constructor used when you need to create new <tt>Function</tt>s to add
3789 the the program. The constructor must specify the type of the function to
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]3790 create and what type of linkage the function should have. The <a
3791 href="#FunctionType"><tt>FunctionType</tt></a> argument
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3792 specifies the formal arguments and return value for the function. The same
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3793 <a href="#FunctionType"><tt>FunctionType</tt></a> value can be used to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3794 create multiple functions. The <tt>Parent</tt> argument specifies the Module
3795 in which the function is defined. If this argument is provided, the function
3796 will automatically be inserted into that module's list of
3797 functions.</p></li>
3798
Chris Lattner62810e32008-11-25 18:34:50 +0000[diff] [blame]3799 <li><tt>bool isDeclaration()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3800
3801 <p>Return whether or not the <tt>Function</tt> has a body defined. If the
3802 function is "external", it does not have a body, and thus must be resolved
3803 by linking with a function defined in a different translation unit.</p></li>
3804
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3805 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3806 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3807
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]3808 <tt>begin()</tt>, <tt>end()</tt>
3809 <tt>size()</tt>, <tt>empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3810
3811 <p>These are forwarding methods that make it easy to access the contents of
3812 a <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
3813 list.</p></li>
3814
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3815 <li><tt>Function::BasicBlockListType &amp;getBasicBlockList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3816
3817 <p>Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This
3818 is necessary to use when you need to update the list or perform a complex
3819 action that doesn't have a forwarding method.</p></li>
3820
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3821 <li><tt>Function::arg_iterator</tt> - Typedef for the argument list
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3822iterator<br>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3823 <tt>Function::const_arg_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3824
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]3825 <tt>arg_begin()</tt>, <tt>arg_end()</tt>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3826 <tt>arg_size()</tt>, <tt>arg_empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3827
3828 <p>These are forwarding methods that make it easy to access the contents of
3829 a <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a>
3830 list.</p></li>
3831
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3832 <li><tt>Function::ArgumentListType &amp;getArgumentList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3833
3834 <p>Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
3835 necessary to use when you need to update the list or perform a complex
3836 action that doesn't have a forwarding method.</p></li>
3837
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3838 <li><tt><a href="#BasicBlock">BasicBlock</a> &amp;getEntryBlock()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3839
3840 <p>Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
3841 function. Because the entry block for the function is always the first
3842 block, this returns the first block of the <tt>Function</tt>.</p></li>
3843
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3844 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
3845 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3846
3847 <p>This traverses the <a href="#Type"><tt>Type</tt></a> of the
3848 <tt>Function</tt> and returns the return type of the function, or the <a
3849 href="#FunctionType"><tt>FunctionType</tt></a> of the actual
3850 function.</p></li>
3851
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3852 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3853
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3854 <p> Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3855 for this <tt>Function</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3856</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3857
3858</div>
3859
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3860</div>
3861
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3862<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3863<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3864 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3865</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3866
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3867<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3868
3869<p><tt>#include "<a
3870href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt>
3871<br>
Tanya Lattnera3da7772004-06-22 08:02:25 +0000[diff] [blame]3872doxygen info: <a href="/doxygen/classllvm_1_1GlobalVariable.html">GlobalVariable
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3873 Class</a><br>
3874Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
3875<a href="#Constant"><tt>Constant</tt></a>,
3876<a href="#User"><tt>User</tt></a>,
3877<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3878
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3879<p>Global variables are represented with the (surprise surprise)
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3880<tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are also
3881subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such are
3882always referenced by their address (global values must live in memory, so their
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3883"name" refers to their constant address). See
3884<a href="#GlobalValue"><tt>GlobalValue</tt></a> for more on this. Global
3885variables may have an initial value (which must be a
3886<a href="#Constant"><tt>Constant</tt></a>), and if they have an initializer,
3887they may be marked as "constant" themselves (indicating that their contents
3888never change at runtime).</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3889
3890<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3891<h4>
3892 <a name="m_GlobalVariable">
3893 Important Public Members of the <tt>GlobalVariable</tt> class
3894 </a>
3895</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3896
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3897<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3898
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3899<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3900 <li><tt>GlobalVariable(const </tt><tt><a href="#Type">Type</a> *Ty, bool
3901 isConstant, LinkageTypes&amp; Linkage, <a href="#Constant">Constant</a>
3902 *Initializer = 0, const std::string &amp;Name = "", Module* Parent = 0)</tt>
3903
3904 <p>Create a new global variable of the specified type. If
3905 <tt>isConstant</tt> is true then the global variable will be marked as
3906 unchanging for the program. The Linkage parameter specifies the type of
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]3907 linkage (internal, external, weak, linkonce, appending) for the variable.
3908 If the linkage is InternalLinkage, WeakAnyLinkage, WeakODRLinkage,
3909 LinkOnceAnyLinkage or LinkOnceODRLinkage,&nbsp; then the resultant
3910 global variable will have internal linkage. AppendingLinkage concatenates
3911 together all instances (in different translation units) of the variable
3912 into a single variable but is only applicable to arrays. &nbsp;See
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3913 the <a href="LangRef.html#modulestructure">LLVM Language Reference</a> for
3914 further details on linkage types. Optionally an initializer, a name, and the
3915 module to put the variable into may be specified for the global variable as
3916 well.</p></li>
3917
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3918 <li><tt>bool isConstant() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3919
3920 <p>Returns true if this is a global variable that is known not to
3921 be modified at runtime.</p></li>
3922
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3923 <li><tt>bool hasInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3924
3925 <p>Returns true if this <tt>GlobalVariable</tt> has an intializer.</p></li>
3926
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3927 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3928
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3929 <p>Returns the initial value for a <tt>GlobalVariable</tt>. It is not legal
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3930 to call this method if there is no initializer.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3931</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3932
3933</div>
3934
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3935</div>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3936
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3937<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3938<h3>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3939 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3940</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3941
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3942<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3943
3944<p><tt>#include "<a
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3945href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt><br>
Stefanus Du Toit24e04112009-06-17 21:12:26 +0000[diff] [blame]3946doxygen info: <a href="/doxygen/classllvm_1_1BasicBlock.html">BasicBlock
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3947Class</a><br>
3948Superclass: <a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3949
Nick Lewyckyccd279d2011-02-17 02:19:22 +0000[diff] [blame]3950<p>This class represents a single entry single exit section of the code,
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3951commonly known as a basic block by the compiler community. The
3952<tt>BasicBlock</tt> class maintains a list of <a
3953href="#Instruction"><tt>Instruction</tt></a>s, which form the body of the block.
3954Matching the language definition, the last element of this list of instructions
3955is always a terminator instruction (a subclass of the <a
3956href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).</p>
3957
3958<p>In addition to tracking the list of instructions that make up the block, the
3959<tt>BasicBlock</tt> class also keeps track of the <a
3960href="#Function"><tt>Function</tt></a> that it is embedded into.</p>
3961
3962<p>Note that <tt>BasicBlock</tt>s themselves are <a
3963href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
3964like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
3965<tt>label</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3966
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3967<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3968<h4>
3969 <a name="m_BasicBlock">
3970 Important Public Members of the <tt>BasicBlock</tt> class
3971 </a>
3972</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3973
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3974<div>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3975<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3976
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3977<li><tt>BasicBlock(const std::string &amp;Name = "", </tt><tt><a
3978 href="#Function">Function</a> *Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3979
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3980<p>The <tt>BasicBlock</tt> constructor is used to create new basic blocks for
3981insertion into a function. The constructor optionally takes a name for the new
3982block, and a <a href="#Function"><tt>Function</tt></a> to insert it into. If
3983the <tt>Parent</tt> parameter is specified, the new <tt>BasicBlock</tt> is
3984automatically inserted at the end of the specified <a
3985href="#Function"><tt>Function</tt></a>, if not specified, the BasicBlock must be
3986manually inserted into the <a href="#Function"><tt>Function</tt></a>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3987
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3988<li><tt>BasicBlock::iterator</tt> - Typedef for instruction list iterator<br>
3989<tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
3990<tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
3991<tt>size()</tt>, <tt>empty()</tt>
3992STL-style functions for accessing the instruction list.
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3993
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3994<p>These methods and typedefs are forwarding functions that have the same
3995semantics as the standard library methods of the same names. These methods
3996expose the underlying instruction list of a basic block in a way that is easy to
3997manipulate. To get the full complement of container operations (including
3998operations to update the list), you must use the <tt>getInstList()</tt>
3999method.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4000
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]4001<li><tt>BasicBlock::InstListType &amp;getInstList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4002
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]4003<p>This method is used to get access to the underlying container that actually
4004holds the Instructions. This method must be used when there isn't a forwarding
4005function in the <tt>BasicBlock</tt> class for the operation that you would like
4006to perform. Because there are no forwarding functions for "updating"
4007operations, you need to use this if you want to update the contents of a
4008<tt>BasicBlock</tt>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4009
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]4010<li><tt><a href="#Function">Function</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4011
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]4012<p> Returns a pointer to <a href="#Function"><tt>Function</tt></a> the block is
4013embedded into, or a null pointer if it is homeless.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4014
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]4015<li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4016
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]4017<p> Returns a pointer to the terminator instruction that appears at the end of
4018the <tt>BasicBlock</tt>. If there is no terminator instruction, or if the last
4019instruction in the block is not a terminator, then a null pointer is
4020returned.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4021
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4022</ul>
4023
4024</div>
4025
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4026</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4027
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4028<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4029<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4030 <a name="Argument">The <tt>Argument</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4031</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4032
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4033<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4034
4035<p>This subclass of Value defines the interface for incoming formal
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]4036arguments to a function. A Function maintains a list of its formal
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4037arguments. An argument has a pointer to the parent Function.</p>
4038
4039</div>
4040
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4041</div>
4042
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]4043<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4044<hr>
4045<address>
4046 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]4047 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4048 <a href="http://validator.w3.org/check/referer"><img
Gabor Greifa9c0f2b2008-06-18 14:05:31 +0000[diff] [blame]4049 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]4050
4051 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
4052 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
NAKAMURA Takumib9a33632011-04-09 02:13:37 +0000[diff] [blame]4053 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]4054 Last modified: $Date$
4055</address>
4056
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4057</body>
4058</html>