Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / llvm / e3f5ae7b9e68953635f604a8b437d440678bb18e / . / docs / ProgrammersManual.html
blob: c1edd2a60a6fd299cb2a2dc5f15096cafbef7320 [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
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]164 <li><a href="#SymbolTable">The <tt>ValueSymbolTable</tt> class</a></li>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]165 <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]166 </ul></li>
167
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]168 <li><a href="#coreclasses">The Core LLVM Class Hierarchy Reference</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]169 <ul>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]170 <li><a href="#Type">The <tt>Type</tt> class</a> </li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]171 <li><a href="#Module">The <tt>Module</tt> class</a></li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]172 <li><a href="#Value">The <tt>Value</tt> class</a>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]173 <ul>
174 <li><a href="#User">The <tt>User</tt> class</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]175 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]176 <li><a href="#Instruction">The <tt>Instruction</tt> class</a></li>
177 <li><a href="#Constant">The <tt>Constant</tt> class</a>
178 <ul>
179 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]180 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]181 <li><a href="#Function">The <tt>Function</tt> class</a></li>
182 <li><a href="#GlobalVariable">The <tt>GlobalVariable</tt> class</a></li>
183 </ul>
184 </li>
185 </ul>
186 </li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]187 </ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]188 </li>
189 <li><a href="#BasicBlock">The <tt>BasicBlock</tt> class</a></li>
190 <li><a href="#Argument">The <tt>Argument</tt> class</a></li>
191 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]192 </li>
193 </ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]194 </li>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]195</ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]196
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]197<div class="doc_author">
198 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
Chris Lattner94c43592004-05-26 16:52:55 +0000[diff] [blame]199 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>,
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]200 <a href="mailto:ggreif@gmail.com">Gabor Greif</a>,
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]201 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a>,
202 <a href="mailto:rspencer@x10sys.com">Reid Spencer</a> and
203 <a href="mailto:owen@apple.com">Owen Anderson</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]204</div>
205
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]206<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]207<h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]208 <a name="introduction">Introduction </a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]209</h2>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]210<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]211
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]212<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]213
214<p>This document is meant to highlight some of the important classes and
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]215interfaces available in the LLVM source-base. This manual is not
216intended to explain what LLVM is, how it works, and what LLVM code looks
217like. It assumes that you know the basics of LLVM and are interested
218in writing transformations or otherwise analyzing or manipulating the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]219code.</p>
220
221<p>This document should get you oriented so that you can find your
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]222way in the continuously growing source code that makes up the LLVM
223infrastructure. Note that this manual is not intended to serve as a
224replacement for reading the source code, so if you think there should be
225a method in one of these classes to do something, but it's not listed,
226check the source. Links to the <a href="/doxygen/">doxygen</a> sources
227are provided to make this as easy as possible.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]228
229<p>The first section of this document describes general information that is
230useful to know when working in the LLVM infrastructure, and the second describes
231the Core LLVM classes. In the future this manual will be extended with
232information describing how to use extension libraries, such as dominator
233information, CFG traversal routines, and useful utilities like the <tt><a
234href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.</p>
235
236</div>
237
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]238<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]239<h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]240 <a name="general">General Information</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]241</h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]242<!-- *********************************************************************** -->
243
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]244<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]245
246<p>This section contains general information that is useful if you are working
247in the LLVM source-base, but that isn't specific to any particular API.</p>
248
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]249<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]250<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]251 <a name="stl">The C++ Standard Template Library</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]252</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]253
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]254<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]255
256<p>LLVM makes heavy use of the C++ Standard Template Library (STL),
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]257perhaps much more than you are used to, or have seen before. Because of
258this, you might want to do a little background reading in the
259techniques used and capabilities of the library. There are many good
260pages that discuss the STL, and several books on the subject that you
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]261can get, so it will not be discussed in this document.</p>
262
263<p>Here are some useful links:</p>
264
265<ol>
266
Nick Lewyckyea1fe2c2010-10-09 21:12:29 +0000[diff] [blame]267<li><a href="http://www.dinkumware.com/manuals/#Standard C++ Library">Dinkumware
268C++ Library reference</a> - an excellent reference for the STL and other parts
269of the standard C++ library.</li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]270
271<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]272O'Reilly book in the making. It has a decent Standard Library
273Reference that rivals Dinkumware's, and is unfortunately no longer free since the
274book has been published.</li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]275
276<li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
277Questions</a></li>
278
279<li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
280Contains a useful <a
281href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
282STL</a>.</li>
283
284<li><a href="http://www.research.att.com/%7Ebs/C++.html">Bjarne Stroustrup's C++
285Page</a></li>
286
Tanya Lattner79445ba2004-12-08 18:34:56 +0000[diff] [blame]287<li><a href="http://64.78.49.204/">
Reid Spencer096603a2004-05-26 08:41:35 +0000[diff] [blame]288Bruce Eckel's Thinking in C++, 2nd ed. Volume 2 Revision 4.0 (even better, get
289the book).</a></li>
290
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]291</ol>
292
293<p>You are also encouraged to take a look at the <a
294href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
295to write maintainable code more than where to put your curly braces.</p>
296
297</div>
298
299<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]300<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]301 <a name="stl">Other useful references</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]302</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]303
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]304<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]305
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]306<ol>
Misha Brukmana0f71e42004-06-18 18:39:00 +0000[diff] [blame]307<li><a href="http://www.fortran-2000.com/ArnaudRecipes/sharedlib.html">Using
308static and shared libraries across platforms</a></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]309</ol>
310
311</div>
312
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]313</div>
314
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]315<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]316<h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]317 <a name="apis">Important and useful LLVM APIs</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]318</h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]319<!-- *********************************************************************** -->
320
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]321<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]322
323<p>Here we highlight some LLVM APIs that are generally useful and good to
324know about when writing transformations.</p>
325
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]326<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]327<h3>
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]328 <a name="isa">The <tt>isa&lt;&gt;</tt>, <tt>cast&lt;&gt;</tt> and
329 <tt>dyn_cast&lt;&gt;</tt> templates</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]330</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]331
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]332<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]333
334<p>The LLVM source-base makes extensive use of a custom form of RTTI.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]335These templates have many similarities to the C++ <tt>dynamic_cast&lt;&gt;</tt>
336operator, but they don't have some drawbacks (primarily stemming from
337the fact that <tt>dynamic_cast&lt;&gt;</tt> only works on classes that
338have a v-table). Because they are used so often, you must know what they
339do and how they work. All of these templates are defined in the <a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]340 href="/doxygen/Casting_8h-source.html"><tt>llvm/Support/Casting.h</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]341file (note that you very rarely have to include this file directly).</p>
342
343<dl>
344 <dt><tt>isa&lt;&gt;</tt>: </dt>
345
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]346 <dd><p>The <tt>isa&lt;&gt;</tt> operator works exactly like the Java
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]347 "<tt>instanceof</tt>" operator. It returns true or false depending on whether
348 a reference or pointer points to an instance of the specified class. This can
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]349 be very useful for constraint checking of various sorts (example below).</p>
350 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]351
352 <dt><tt>cast&lt;&gt;</tt>: </dt>
353
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]354 <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]355 converts a pointer or reference from a base class to a derived class, causing
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]356 an assertion failure if it is not really an instance of the right type. This
357 should be used in cases where you have some information that makes you believe
358 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]359 and <tt>cast&lt;&gt;</tt> template is:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]360
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]361<div class="doc_code">
362<pre>
363static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
364 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))
365 return true;
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]366
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]367 // <i>Otherwise, it must be an instruction...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]368 return !L-&gt;contains(cast&lt;<a href="#Instruction">Instruction</a>&gt;(V)-&gt;getParent());
369}
370</pre>
371</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]372
373 <p>Note that you should <b>not</b> use an <tt>isa&lt;&gt;</tt> test followed
374 by a <tt>cast&lt;&gt;</tt>, for that use the <tt>dyn_cast&lt;&gt;</tt>
375 operator.</p>
376
377 </dd>
378
379 <dt><tt>dyn_cast&lt;&gt;</tt>:</dt>
380
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]381 <dd><p>The <tt>dyn_cast&lt;&gt;</tt> operator is a "checking cast" operation.
382 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]383 pointer to it (this operator does not work with references). If the operand is
384 not of the correct type, a null pointer is returned. Thus, this works very
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]385 much like the <tt>dynamic_cast&lt;&gt;</tt> operator in C++, and should be
386 used in the same circumstances. Typically, the <tt>dyn_cast&lt;&gt;</tt>
387 operator is used in an <tt>if</tt> statement or some other flow control
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]388 statement like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]389
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]390<div class="doc_code">
391<pre>
392if (<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]393 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]394}
395</pre>
396</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]397
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]398 <p>This form of the <tt>if</tt> statement effectively combines together a call
399 to <tt>isa&lt;&gt;</tt> and a call to <tt>cast&lt;&gt;</tt> into one
400 statement, which is very convenient.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]401
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]402 <p>Note that the <tt>dyn_cast&lt;&gt;</tt> operator, like C++'s
403 <tt>dynamic_cast&lt;&gt;</tt> or Java's <tt>instanceof</tt> operator, can be
404 abused. In particular, you should not use big chained <tt>if/then/else</tt>
405 blocks to check for lots of different variants of classes. If you find
406 yourself wanting to do this, it is much cleaner and more efficient to use the
407 <tt>InstVisitor</tt> class to dispatch over the instruction type directly.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]408
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]409 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]410
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]411 <dt><tt>cast_or_null&lt;&gt;</tt>: </dt>
412
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]413 <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]414 <tt>cast&lt;&gt;</tt> operator, except that it allows for a null pointer as an
415 argument (which it then propagates). This can sometimes be useful, allowing
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]416 you to combine several null checks into one.</p></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>dyn_cast_or_null&lt;&gt;</tt>: </dt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]419
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]420 <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]421 <tt>dyn_cast&lt;&gt;</tt> operator, except that it allows for a null pointer
422 as an argument (which it then propagates). This can sometimes be useful,
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]423 allowing 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</dl>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]426
427<p>These five templates can be used with any classes, whether they have a
428v-table or not. To add support for these templates, you simply need to add
429<tt>classof</tt> static methods to the class you are interested casting
430to. Describing this is currently outside the scope of this document, but there
431are lots of examples in the LLVM source base.</p>
432
433</div>
434
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]435
436<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]437<h3>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]438 <a name="string_apis">Passing strings (the <tt>StringRef</tt>
439and <tt>Twine</tt> classes)</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]440</h3>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]441
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]442<div>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]443
444<p>Although LLVM generally does not do much string manipulation, we do have
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]445several important APIs which take strings. Two important examples are the
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]446Value class -- which has names for instructions, functions, etc. -- and the
447StringMap class which is used extensively in LLVM and Clang.</p>
448
449<p>These are generic classes, and they need to be able to accept strings which
450may have embedded null characters. Therefore, they cannot simply take
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]451a <tt>const char *</tt>, and taking a <tt>const std::string&amp;</tt> requires
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]452clients to perform a heap allocation which is usually unnecessary. Instead,
Benjamin Kramer38e59892010-07-14 22:38:02 +0000[diff] [blame]453many LLVM APIs use a <tt>StringRef</tt> or a <tt>const Twine&amp;</tt> for
454passing strings efficiently.</p>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]455
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]456<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]457<h4>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]458 <a name="StringRef">The <tt>StringRef</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]459</h4>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]460
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]461<div>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]462
463<p>The <tt>StringRef</tt> data type represents a reference to a constant string
464(a character array and a length) and supports the common operations available
465on <tt>std:string</tt>, but does not require heap allocation.</p>
466
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]467<p>It can be implicitly constructed using a C style null-terminated string,
468an <tt>std::string</tt>, or explicitly with a character pointer and length.
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]469For example, the <tt>StringRef</tt> find function is declared as:</p>
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]470
Benjamin Kramer38e59892010-07-14 22:38:02 +0000[diff] [blame]471<pre class="doc_code">
472 iterator find(StringRef Key);
473</pre>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]474
475<p>and clients can call it using any one of:</p>
476
Benjamin Kramer38e59892010-07-14 22:38:02 +0000[diff] [blame]477<pre class="doc_code">
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]478 Map.find("foo"); <i>// Lookup "foo"</i>
479 Map.find(std::string("bar")); <i>// Lookup "bar"</i>
480 Map.find(StringRef("\0baz", 4)); <i>// Lookup "\0baz"</i>
481</pre>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]482
483<p>Similarly, APIs which need to return a string may return a <tt>StringRef</tt>
484instance, which can be used directly or converted to an <tt>std::string</tt>
485using the <tt>str</tt> member function. See
486"<tt><a href="/doxygen/classllvm_1_1StringRef_8h-source.html">llvm/ADT/StringRef.h</a></tt>"
487for more information.</p>
488
489<p>You should rarely use the <tt>StringRef</tt> class directly, because it contains
490pointers to external memory it is not generally safe to store an instance of the
Benjamin Kramer38e59892010-07-14 22:38:02 +0000[diff] [blame]491class (unless you know that the external storage will not be freed). StringRef is
492small and pervasive enough in LLVM that it should always be passed by value.</p>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]493
494</div>
495
496<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]497<h4>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]498 <a name="Twine">The <tt>Twine</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]499</h4>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]500
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]501<div>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]502
503<p>The <tt>Twine</tt> class is an efficient way for APIs to accept concatenated
504strings. For example, a common LLVM paradigm is to name one instruction based on
505the name of another instruction with a suffix, for example:</p>
506
507<div class="doc_code">
508<pre>
509 New = CmpInst::Create(<i>...</i>, SO->getName() + ".cmp");
510</pre>
511</div>
512
513<p>The <tt>Twine</tt> class is effectively a
514lightweight <a href="http://en.wikipedia.org/wiki/Rope_(computer_science)">rope</a>
515which points to temporary (stack allocated) objects. Twines can be implicitly
516constructed as the result of the plus operator applied to strings (i.e., a C
517strings, an <tt>std::string</tt>, or a <tt>StringRef</tt>). The twine delays the
Dan Gohmancf0c9bc2010-02-25 23:51:27 +0000[diff] [blame]518actual concatenation of strings until it is actually required, at which point
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]519it can be efficiently rendered directly into a character array. This avoids
520unnecessary heap allocation involved in constructing the temporary results of
521string concatenation. See
522"<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]523for more information.</p>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]524
525<p>As with a <tt>StringRef</tt>, <tt>Twine</tt> objects point to external memory
526and should almost never be stored or mentioned directly. They are intended
527solely for use when defining a function which should be able to efficiently
528accept concatenated strings.</p>
529
530</div>
531
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]532</div>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]533
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]534<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]535<h3>
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]536 <a name="DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt> option</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]537</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]538
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]539<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]540
541<p>Often when working on your pass you will put a bunch of debugging printouts
542and other code into your pass. After you get it working, you want to remove
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]543it, 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]544across).</p>
545
546<p> Naturally, because of this, you don't want to delete the debug printouts,
547but you don't want them to always be noisy. A standard compromise is to comment
548them out, allowing you to enable them if you need them in the future.</p>
549
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]550<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]551file provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to
552this problem. Basically, you can put arbitrary code into the argument of the
553<tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
554tool) is run with the '<tt>-debug</tt>' command line argument:</p>
555
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]556<div class="doc_code">
557<pre>
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]558DEBUG(errs() &lt;&lt; "I am here!\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]559</pre>
560</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]561
562<p>Then you can run your pass like this:</p>
563
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]564<div class="doc_code">
565<pre>
566$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]567<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]568$ opt &lt; a.bc &gt; /dev/null -mypass -debug
569I am here!
570</pre>
571</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]572
573<p>Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you
574to not have to create "yet another" command line option for the debug output for
575your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
576so they do not cause a performance impact at all (for the same reason, they
577should also not contain side-effects!).</p>
578
579<p>One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
580enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
581"<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
582program hasn't been started yet, you can always just run it with
583<tt>-debug</tt>.</p>
584
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]585<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]586<h4>
Chris Lattnerc9151082005-04-26 22:57:07 +0000[diff] [blame]587 <a name="DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt> and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]588 the <tt>-debug-only</tt> option</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]589</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]590
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]591<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]592
593<p>Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
594just turns on <b>too much</b> information (such as when working on the code
595generator). If you want to enable debug information with more fine-grained
596control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
597option as follows:</p>
598
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]599<div class="doc_code">
600<pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]601#undef DEBUG_TYPE
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]602DEBUG(errs() &lt;&lt; "No debug type\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]603#define DEBUG_TYPE "foo"
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]604DEBUG(errs() &lt;&lt; "'foo' debug type\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]605#undef DEBUG_TYPE
606#define DEBUG_TYPE "bar"
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]607DEBUG(errs() &lt;&lt; "'bar' debug type\n"));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]608#undef DEBUG_TYPE
609#define DEBUG_TYPE ""
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]610DEBUG(errs() &lt;&lt; "No debug type (2)\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]611</pre>
612</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]613
614<p>Then you can run your pass like this:</p>
615
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]616<div class="doc_code">
617<pre>
618$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]619<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]620$ opt &lt; a.bc &gt; /dev/null -mypass -debug
621No debug type
622'foo' debug type
623'bar' debug type
624No debug type (2)
625$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=foo
626'foo' debug type
627$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=bar
628'bar' debug type
629</pre>
630</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]631
632<p>Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of
633a 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]634you <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]635<tt>#undef</tt>'s). Also, you should use names more meaningful than "foo" and
636"bar", because there is no system in place to ensure that names do not
637conflict. If two different modules use the same string, they will all be turned
638on when the name is specified. This allows, for example, all debug information
639for instruction scheduling to be enabled with <tt>-debug-type=InstrSched</tt>,
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]640even if the source lives in multiple files.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]641
Daniel Dunbarc3c92392009-08-07 23:48:59 +0000[diff] [blame]642<p>The <tt>DEBUG_WITH_TYPE</tt> macro is also available for situations where you
643would like to set <tt>DEBUG_TYPE</tt>, but only for one specific <tt>DEBUG</tt>
644statement. It takes an additional first parameter, which is the type to use. For
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]645example, the preceding example could be written as:</p>
Daniel Dunbarc3c92392009-08-07 23:48:59 +0000[diff] [blame]646
647
648<div class="doc_code">
649<pre>
650DEBUG_WITH_TYPE("", errs() &lt;&lt; "No debug type\n");
651DEBUG_WITH_TYPE("foo", errs() &lt;&lt; "'foo' debug type\n");
652DEBUG_WITH_TYPE("bar", errs() &lt;&lt; "'bar' debug type\n"));
653DEBUG_WITH_TYPE("", errs() &lt;&lt; "No debug type (2)\n");
654</pre>
655</div>
656
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]657</div>
658
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]659</div>
660
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]661<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]662<h3>
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]663 <a name="Statistic">The <tt>Statistic</tt> class &amp; <tt>-stats</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]664 option</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]665</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]666
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]667<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]668
669<p>The "<tt><a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]670href="/doxygen/Statistic_8h-source.html">llvm/ADT/Statistic.h</a></tt>" file
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]671provides a class named <tt>Statistic</tt> that is used as a unified way to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]672keep track of what the LLVM compiler is doing and how effective various
673optimizations are. It is useful to see what optimizations are contributing to
674making a particular program run faster.</p>
675
676<p>Often you may run your pass on some big program, and you're interested to see
677how many times it makes a certain transformation. Although you can do this with
678hand 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]679for big programs. Using the <tt>Statistic</tt> class makes it very easy to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]680keep track of this information, and the calculated information is presented in a
681uniform manner with the rest of the passes being executed.</p>
682
683<p>There are many examples of <tt>Statistic</tt> uses, but the basics of using
684it are as follows:</p>
685
686<ol>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]687 <li><p>Define your statistic like this:</p>
688
689<div class="doc_code">
690<pre>
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]691#define <a href="#DEBUG_TYPE">DEBUG_TYPE</a> "mypassname" <i>// This goes before any #includes.</i>
692STATISTIC(NumXForms, "The # of times I did stuff");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]693</pre>
694</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]695
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]696 <p>The <tt>STATISTIC</tt> macro defines a static variable, whose name is
697 specified by the first argument. The pass name is taken from the DEBUG_TYPE
698 macro, and the description is taken from the second argument. The variable
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]699 defined ("NumXForms" in this case) acts like an unsigned integer.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]700
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]701 <li><p>Whenever you make a transformation, bump the counter:</p>
702
703<div class="doc_code">
704<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]705++NumXForms; // <i>I did stuff!</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]706</pre>
707</div>
708
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]709 </li>
710 </ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]711
712 <p>That's all you have to do. To get '<tt>opt</tt>' to print out the
713 statistics gathered, use the '<tt>-stats</tt>' option:</p>
714
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]715<div class="doc_code">
716<pre>
717$ opt -stats -mypassname &lt; program.bc &gt; /dev/null
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]718<i>... statistics output ...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]719</pre>
720</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]721
Reid Spencer6b6c73e2007-02-09 16:00:28 +0000[diff] [blame]722 <p> When running <tt>opt</tt> on a C file from the SPEC benchmark
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]723suite, it gives a report that looks like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]724
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]725<div class="doc_code">
726<pre>
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]727 7646 bitcodewriter - Number of normal instructions
728 725 bitcodewriter - Number of oversized instructions
729 129996 bitcodewriter - Number of bitcode bytes written
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]730 2817 raise - Number of insts DCEd or constprop'd
731 3213 raise - Number of cast-of-self removed
732 5046 raise - Number of expression trees converted
733 75 raise - Number of other getelementptr's formed
734 138 raise - Number of load/store peepholes
735 42 deadtypeelim - Number of unused typenames removed from symtab
736 392 funcresolve - Number of varargs functions resolved
737 27 globaldce - Number of global variables removed
738 2 adce - Number of basic blocks removed
739 134 cee - Number of branches revectored
740 49 cee - Number of setcc instruction eliminated
741 532 gcse - Number of loads removed
742 2919 gcse - Number of instructions removed
743 86 indvars - Number of canonical indvars added
744 87 indvars - Number of aux indvars removed
745 25 instcombine - Number of dead inst eliminate
746 434 instcombine - Number of insts combined
747 248 licm - Number of load insts hoisted
748 1298 licm - Number of insts hoisted to a loop pre-header
749 3 licm - Number of insts hoisted to multiple loop preds (bad, no loop pre-header)
750 75 mem2reg - Number of alloca's promoted
751 1444 cfgsimplify - Number of blocks simplified
752</pre>
753</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]754
755<p>Obviously, with so many optimizations, having a unified framework for this
756stuff is very nice. Making your pass fit well into the framework makes it more
757maintainable and useful.</p>
758
759</div>
760
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]761<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]762<h3>
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]763 <a name="ViewGraph">Viewing graphs while debugging code</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]764</h3>
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]765
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]766<div>
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]767
768<p>Several of the important data structures in LLVM are graphs: for example
769CFGs made out of LLVM <a href="#BasicBlock">BasicBlock</a>s, CFGs made out of
770LLVM <a href="CodeGenerator.html#machinebasicblock">MachineBasicBlock</a>s, and
771<a href="CodeGenerator.html#selectiondag_intro">Instruction Selection
772DAGs</a>. In many cases, while debugging various parts of the compiler, it is
773nice to instantly visualize these graphs.</p>
774
775<p>LLVM provides several callbacks that are available in a debug build to do
776exactly that. If you call the <tt>Function::viewCFG()</tt> method, for example,
777the current LLVM tool will pop up a window containing the CFG for the function
778where each basic block is a node in the graph, and each node contains the
779instructions in the block. Similarly, there also exists
780<tt>Function::viewCFGOnly()</tt> (does not include the instructions), the
781<tt>MachineFunction::viewCFG()</tt> and <tt>MachineFunction::viewCFGOnly()</tt>,
782and the <tt>SelectionDAG::viewGraph()</tt> methods. Within GDB, for example,
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]783you can usually use something like <tt>call DAG.viewGraph()</tt> to pop
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]784up a window. Alternatively, you can sprinkle calls to these functions in your
785code in places you want to debug.</p>
786
787<p>Getting this to work requires a small amount of configuration. On Unix
788systems with X11, install the <a href="http://www.graphviz.org">graphviz</a>
789toolkit, and make sure 'dot' and 'gv' are in your path. If you are running on
790Mac OS/X, download and install the Mac OS/X <a
791href="http://www.pixelglow.com/graphviz/">Graphviz program</a>, and add
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]792<tt>/Applications/Graphviz.app/Contents/MacOS/</tt> (or wherever you install
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]793it) to your path. Once in your system and path are set up, rerun the LLVM
794configure script and rebuild LLVM to enable this functionality.</p>
795
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]796<p><tt>SelectionDAG</tt> has been extended to make it easier to locate
797<i>interesting</i> nodes in large complex graphs. From gdb, if you
798<tt>call DAG.setGraphColor(<i>node</i>, "<i>color</i>")</tt>, then the
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]799next <tt>call DAG.viewGraph()</tt> would highlight the node in the
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]800specified color (choices of colors can be found at <a
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]801href="http://www.graphviz.org/doc/info/colors.html">colors</a>.) More
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]802complex node attributes can be provided with <tt>call
803DAG.setGraphAttrs(<i>node</i>, "<i>attributes</i>")</tt> (choices can be
804found at <a href="http://www.graphviz.org/doc/info/attrs.html">Graph
805Attributes</a>.) If you want to restart and clear all the current graph
806attributes, then you can <tt>call DAG.clearGraphAttrs()</tt>. </p>
807
Chris Lattner83f94672011-06-13 15:59:35 +0000[diff] [blame]808<p>Note that graph visualization features are compiled out of Release builds
809to reduce file size. This means that you need a Debug+Asserts or
810Release+Asserts build to use these features.</p>
811
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]812</div>
813
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]814</div>
815
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]816<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]817<h2>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]818 <a name="datastructure">Picking the Right Data Structure for a Task</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]819</h2>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]820<!-- *********************************************************************** -->
821
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]822<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]823
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]824<p>LLVM has a plethora of data structures in the <tt>llvm/ADT/</tt> directory,
825 and we commonly use STL data structures. This section describes the trade-offs
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]826 you should consider when you pick one.</p>
827
828<p>
829The first step is a choose your own adventure: do you want a sequential
830container, a set-like container, or a map-like container? The most important
831thing when choosing a container is the algorithmic properties of how you plan to
832access the container. Based on that, you should use:</p>
833
834<ul>
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]835<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]836 of an value based on another value. Map-like containers also support
837 efficient queries for containment (whether a key is in the map). Map-like
838 containers generally do not support efficient reverse mapping (values to
839 keys). If you need that, use two maps. Some map-like containers also
840 support efficient iteration through the keys in sorted order. Map-like
841 containers are the most expensive sort, only use them if you need one of
842 these capabilities.</li>
843
844<li>a <a href="#ds_set">set-like</a> container if you need to put a bunch of
845 stuff into a container that automatically eliminates duplicates. Some
846 set-like containers support efficient iteration through the elements in
847 sorted order. Set-like containers are more expensive than sequential
848 containers.
849</li>
850
851<li>a <a href="#ds_sequential">sequential</a> container provides
852 the most efficient way to add elements and keeps track of the order they are
853 added to the collection. They permit duplicates and support efficient
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]854 iteration, but do not support efficient look-up based on a key.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]855</li>
856
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]857<li>a <a href="#ds_string">string</a> container is a specialized sequential
858 container or reference structure that is used for character or byte
859 arrays.</li>
860
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]861<li>a <a href="#ds_bit">bit</a> container provides an efficient way to store and
862 perform set operations on sets of numeric id's, while automatically
863 eliminating duplicates. Bit containers require a maximum of 1 bit for each
864 identifier you want to store.
865</li>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]866</ul>
867
868<p>
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]869Once the proper category of container is determined, you can fine tune the
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]870memory use, constant factors, and cache behaviors of access by intelligently
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]871picking a member of the category. Note that constant factors and cache behavior
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]872can be a big deal. If you have a vector that usually only contains a few
873elements (but could contain many), for example, it's much better to use
874<a href="#dss_smallvector">SmallVector</a> than <a href="#dss_vector">vector</a>
875. Doing so avoids (relatively) expensive malloc/free calls, which dwarf the
876cost of adding the elements to the container. </p>
877
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]878<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]879<h3>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]880 <a name="ds_sequential">Sequential Containers (std::vector, std::list, etc)</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]881</h3>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]882
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]883<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]884There are a variety of sequential containers available for you, based on your
885needs. Pick the first in this section that will do what you want.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]886
887<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]888<h4>
Chris Lattner8ae42612011-04-05 23:18:20 +0000[diff] [blame]889 <a name="dss_arrayref">llvm/ADT/ArrayRef.h</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]890</h4>
Chris Lattner8ae42612011-04-05 23:18:20 +0000[diff] [blame]891
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]892<div>
Chris Lattner8ae42612011-04-05 23:18:20 +0000[diff] [blame]893<p>The llvm::ArrayRef class is the preferred class to use in an interface that
894 accepts a sequential list of elements in memory and just reads from them. By
895 taking an ArrayRef, the API can be passed a fixed size array, an std::vector,
896 an llvm::SmallVector and anything else that is contiguous in memory.
897</p>
898</div>
899
900
901
902<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]903<h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]904 <a name="dss_fixedarrays">Fixed Size Arrays</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]905</h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]906
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]907<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]908<p>Fixed size arrays are very simple and very fast. They are good if you know
909exactly how many elements you have, or you have a (low) upper bound on how many
910you have.</p>
911</div>
912
913<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]914<h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]915 <a name="dss_heaparrays">Heap Allocated Arrays</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]916</h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]917
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]918<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]919<p>Heap allocated arrays (new[] + delete[]) are also simple. They are good if
920the number of elements is variable, if you know how many elements you will need
921before the array is allocated, and if the array is usually large (if not,
922consider a <a href="#dss_smallvector">SmallVector</a>). The cost of a heap
923allocated array is the cost of the new/delete (aka malloc/free). Also note that
924if you are allocating an array of a type with a constructor, the constructor and
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]925destructors will be run for every element in the array (re-sizable vectors only
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]926construct those elements actually used).</p>
927</div>
928
929<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]930<h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]931 <a name="dss_smallvector">"llvm/ADT/SmallVector.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]932</h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]933
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]934<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]935<p><tt>SmallVector&lt;Type, N&gt;</tt> is a simple class that looks and smells
936just like <tt>vector&lt;Type&gt;</tt>:
937it supports efficient iteration, lays out elements in memory order (so you can
938do pointer arithmetic between elements), supports efficient push_back/pop_back
939operations, supports efficient random access to its elements, etc.</p>
940
941<p>The advantage of SmallVector is that it allocates space for
942some number of elements (N) <b>in the object itself</b>. Because of this, if
943the SmallVector is dynamically smaller than N, no malloc is performed. This can
944be a big win in cases where the malloc/free call is far more expensive than the
945code that fiddles around with the elements.</p>
946
947<p>This is good for vectors that are "usually small" (e.g. the number of
948predecessors/successors of a block is usually less than 8). On the other hand,
949this makes the size of the SmallVector itself large, so you don't want to
950allocate lots of them (doing so will waste a lot of space). As such,
951SmallVectors are most useful when on the stack.</p>
952
953<p>SmallVector also provides a nice portable and efficient replacement for
954<tt>alloca</tt>.</p>
955
956</div>
957
958<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]959<h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]960 <a name="dss_vector">&lt;vector&gt;</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]961</h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]962
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]963<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]964<p>
965std::vector is well loved and respected. It is useful when SmallVector isn't:
966when the size of the vector is often large (thus the small optimization will
967rarely be a benefit) or if you will be allocating many instances of the vector
968itself (which would waste space for elements that aren't in the container).
969vector is also useful when interfacing with code that expects vectors :).
970</p>
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]971
972<p>One worthwhile note about std::vector: avoid code like this:</p>
973
974<div class="doc_code">
975<pre>
976for ( ... ) {
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]977 std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]978 use V;
979}
980</pre>
981</div>
982
983<p>Instead, write this as:</p>
984
985<div class="doc_code">
986<pre>
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]987std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]988for ( ... ) {
989 use V;
990 V.clear();
991}
992</pre>
993</div>
994
995<p>Doing so will save (at least) one heap allocation and free per iteration of
996the loop.</p>
997
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]998</div>
999
1000<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1001<h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1002 <a name="dss_deque">&lt;deque&gt;</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1003</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1004
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1005<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1006<p>std::deque is, in some senses, a generalized version of std::vector. Like
1007std::vector, it provides constant time random access and other similar
1008properties, but it also provides efficient access to the front of the list. It
1009does not guarantee continuity of elements within memory.</p>
1010
1011<p>In exchange for this extra flexibility, std::deque has significantly higher
1012constant factor costs than std::vector. If possible, use std::vector or
1013something cheaper.</p>
1014</div>
1015
1016<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1017<h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1018 <a name="dss_list">&lt;list&gt;</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1019</h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1020
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1021<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1022<p>std::list is an extremely inefficient class that is rarely useful.
1023It performs a heap allocation for every element inserted into it, thus having an
1024extremely high constant factor, particularly for small data types. std::list
1025also only supports bidirectional iteration, not random access iteration.</p>
1026
1027<p>In exchange for this high cost, std::list supports efficient access to both
1028ends of the list (like std::deque, but unlike std::vector or SmallVector). In
1029addition, the iterator invalidation characteristics of std::list are stronger
1030than that of a vector class: inserting or removing an element into the list does
1031not invalidate iterator or pointers to other elements in the list.</p>
1032</div>
1033
1034<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1035<h4>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1036 <a name="dss_ilist">llvm/ADT/ilist.h</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1037</h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1038
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1039<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1040<p><tt>ilist&lt;T&gt;</tt> implements an 'intrusive' doubly-linked list. It is
1041intrusive, because it requires the element to store and provide access to the
1042prev/next pointers for the list.</p>
1043
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1044<p><tt>ilist</tt> has the same drawbacks as <tt>std::list</tt>, and additionally
1045requires an <tt>ilist_traits</tt> implementation for the element type, but it
1046provides some novel characteristics. In particular, it can efficiently store
1047polymorphic objects, the traits class is informed when an element is inserted or
Gabor Greif0cbcabe2009-03-12 09:47:03 +0000[diff] [blame]1048removed from the list, and <tt>ilist</tt>s are guaranteed to support a
1049constant-time splice operation.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1050
Gabor Greif0cbcabe2009-03-12 09:47:03 +0000[diff] [blame]1051<p>These properties are exactly what we want for things like
1052<tt>Instruction</tt>s and basic blocks, which is why these are implemented with
1053<tt>ilist</tt>s.</p>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1054
1055Related classes of interest are explained in the following subsections:
1056 <ul>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1057 <li><a href="#dss_ilist_traits">ilist_traits</a></li>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1058 <li><a href="#dss_iplist">iplist</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1059 <li><a href="#dss_ilist_node">llvm/ADT/ilist_node.h</a></li>
Gabor Greif6a65f422009-03-12 10:30:31 +0000[diff] [blame]1060 <li><a href="#dss_ilist_sentinel">Sentinels</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1061 </ul>
1062</div>
1063
1064<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1065<h4>
Argyrios Kyrtzidis81df0892011-06-15 19:56:01 +0000[diff] [blame]1066 <a name="dss_packedvector">llvm/ADT/PackedVector.h</a>
1067</h4>
1068
1069<div>
1070<p>
1071Useful for storing a vector of values using only a few number of bits for each
1072value. Apart from the standard operations of a vector-like container, it can
1073also perform an 'or' set operation.
1074</p>
1075
1076<p>For example:</p>
1077
1078<div class="doc_code">
1079<pre>
1080enum State {
1081 None = 0x0,
1082 FirstCondition = 0x1,
1083 SecondCondition = 0x2,
1084 Both = 0x3
1085};
1086
1087State get() {
1088 PackedVector&lt;State, 2&gt; Vec1;
1089 Vec1.push_back(FirstCondition);
1090
1091 PackedVector&lt;State, 2&gt; Vec2;
1092 Vec2.push_back(SecondCondition);
1093
1094 Vec1 |= Vec2;
1095 return Vec1[0]; // returns 'Both'.
1096}
1097</pre>
1098</div>
1099
1100</div>
1101
1102<!-- _______________________________________________________________________ -->
1103<h4>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1104 <a name="dss_ilist_traits">ilist_traits</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1105</h4>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1106
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1107<div>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1108<p><tt>ilist_traits&lt;T&gt;</tt> is <tt>ilist&lt;T&gt;</tt>'s customization
1109mechanism. <tt>iplist&lt;T&gt;</tt> (and consequently <tt>ilist&lt;T&gt;</tt>)
1110publicly derive from this traits class.</p>
1111</div>
1112
1113<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1114<h4>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1115 <a name="dss_iplist">iplist</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1116</h4>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1117
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1118<div>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1119<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]1120supports a slightly narrower interface. Notably, inserters from
1121<tt>T&amp;</tt> are absent.</p>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1122
1123<p><tt>ilist_traits&lt;T&gt;</tt> is a public base of this class and can be
1124used for a wide variety of customizations.</p>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1125</div>
1126
1127<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1128<h4>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1129 <a name="dss_ilist_node">llvm/ADT/ilist_node.h</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1130</h4>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1131
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1132<div>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1133<p><tt>ilist_node&lt;T&gt;</tt> implements a the forward and backward links
1134that are expected by the <tt>ilist&lt;T&gt;</tt> (and analogous containers)
1135in the default manner.</p>
1136
1137<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]1138<tt>T</tt>, usually <tt>T</tt> publicly derives from
1139<tt>ilist_node&lt;T&gt;</tt>.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1140</div>
1141
1142<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1143<h4>
Gabor Greif6a65f422009-03-12 10:30:31 +0000[diff] [blame]1144 <a name="dss_ilist_sentinel">Sentinels</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1145</h4>
Gabor Greif6a65f422009-03-12 10:30:31 +0000[diff] [blame]1146
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1147<div>
Dan Gohmancf0c9bc2010-02-25 23:51:27 +0000[diff] [blame]1148<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]1149citizen in the C++ ecosystem, it needs to support the standard container
1150operations, such as <tt>begin</tt> and <tt>end</tt> iterators, etc. Also, the
1151<tt>operator--</tt> must work correctly on the <tt>end</tt> iterator in the
1152case of non-empty <tt>ilist</tt>s.</p>
1153
1154<p>The only sensible solution to this problem is to allocate a so-called
1155<i>sentinel</i> along with the intrusive list, which serves as the <tt>end</tt>
1156iterator, providing the back-link to the last element. However conforming to the
1157C++ convention it is illegal to <tt>operator++</tt> beyond the sentinel and it
1158also must not be dereferenced.</p>
1159
1160<p>These constraints allow for some implementation freedom to the <tt>ilist</tt>
1161how to allocate and store the sentinel. The corresponding policy is dictated
1162by <tt>ilist_traits&lt;T&gt;</tt>. By default a <tt>T</tt> gets heap-allocated
1163whenever the need for a sentinel arises.</p>
1164
1165<p>While the default policy is sufficient in most cases, it may break down when
1166<tt>T</tt> does not provide a default constructor. Also, in the case of many
1167instances of <tt>ilist</tt>s, the memory overhead of the associated sentinels
1168is wasted. To alleviate the situation with numerous and voluminous
1169<tt>T</tt>-sentinels, sometimes a trick is employed, leading to <i>ghostly
1170sentinels</i>.</p>
1171
1172<p>Ghostly sentinels are obtained by specially-crafted <tt>ilist_traits&lt;T&gt;</tt>
1173which superpose the sentinel with the <tt>ilist</tt> instance in memory. Pointer
1174arithmetic is used to obtain the sentinel, which is relative to the
1175<tt>ilist</tt>'s <tt>this</tt> pointer. The <tt>ilist</tt> is augmented by an
1176extra pointer, which serves as the back-link of the sentinel. This is the only
1177field in the ghostly sentinel which can be legally accessed.</p>
1178</div>
1179
1180<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1181<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1182 <a name="dss_other">Other Sequential Container options</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1183</h4>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1184
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1185<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1186<p>Other STL containers are available, such as std::string.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1187
1188<p>There are also various STL adapter classes such as std::queue,
1189std::priority_queue, std::stack, etc. These provide simplified access to an
1190underlying container but don't affect the cost of the container itself.</p>
1191
1192</div>
1193
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1194</div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1195
1196<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1197<h3>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1198 <a name="ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1199</h3>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1200
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1201<div>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1202
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1203<p>Set-like containers are useful when you need to canonicalize multiple values
1204into a single representation. There are several different choices for how to do
1205this, providing various trade-offs.</p>
1206
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1207<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1208<h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1209 <a name="dss_sortedvectorset">A sorted 'vector'</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1210</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1211
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1212<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1213
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]1214<p>If you intend to insert a lot of elements, then do a lot of queries, a
1215great approach is to use a vector (or other sequential container) with
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1216std::sort+std::unique to remove duplicates. This approach works really well if
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]1217your usage pattern has these two distinct phases (insert then query), and can be
1218coupled with a good choice of <a href="#ds_sequential">sequential container</a>.
1219</p>
1220
1221<p>
1222This combination provides the several nice properties: the result data is
1223contiguous in memory (good for cache locality), has few allocations, is easy to
1224address (iterators in the final vector are just indices or pointers), and can be
1225efficiently queried with a standard binary or radix search.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1226
1227</div>
1228
1229<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1230<h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1231 <a name="dss_smallset">"llvm/ADT/SmallSet.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1232</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1233
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1234<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1235
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1236<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]1237are 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]1238has space for N elements in place (thus, if the set is dynamically smaller than
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1239N, no malloc traffic is required) and accesses them with a simple linear search.
1240When the set grows beyond 'N' elements, it allocates a more expensive representation that
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1241guarantees efficient access (for most types, it falls back to std::set, but for
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1242pointers it uses something far better, <a
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1243href="#dss_smallptrset">SmallPtrSet</a>).</p>
1244
1245<p>The magic of this class is that it handles small sets extremely efficiently,
1246but gracefully handles extremely large sets without loss of efficiency. The
1247drawback is that the interface is quite small: it supports insertion, queries
1248and erasing, but does not support iteration.</p>
1249
1250</div>
1251
1252<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1253<h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1254 <a name="dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1255</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1256
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1257<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1258
Gabor Greif4de73682010-03-26 19:30:47 +0000[diff] [blame]1259<p>SmallPtrSet has all the advantages of <tt>SmallSet</tt> (and a <tt>SmallSet</tt> of pointers is
1260transparently implemented with a <tt>SmallPtrSet</tt>), but also supports iterators. If
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1261more than 'N' insertions are performed, a single quadratically
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1262probed hash table is allocated and grows as needed, providing extremely
1263efficient access (constant time insertion/deleting/queries with low constant
1264factors) and is very stingy with malloc traffic.</p>
1265
Gabor Greif4de73682010-03-26 19:30:47 +0000[diff] [blame]1266<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]1267whenever an insertion occurs. Also, the values visited by the iterators are not
1268visited in sorted order.</p>
1269
1270</div>
1271
1272<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1273<h4>
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]1274 <a name="dss_denseset">"llvm/ADT/DenseSet.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1275</h4>
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]1276
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1277<div>
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]1278
1279<p>
1280DenseSet is a simple quadratically probed hash table. It excels at supporting
1281small values: it uses a single allocation to hold all of the pairs that
1282are currently inserted in the set. DenseSet is a great way to unique small
1283values that are not simple pointers (use <a
1284href="#dss_smallptrset">SmallPtrSet</a> for pointers). Note that DenseSet has
1285the same requirements for the value type that <a
1286href="#dss_densemap">DenseMap</a> has.
1287</p>
1288
1289</div>
1290
1291<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1292<h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1293 <a name="dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1294</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1295
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1296<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1297
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1298<p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1299FoldingSet is an aggregate class that is really good at uniquing
1300expensive-to-create or polymorphic objects. It is a combination of a chained
1301hash table with intrusive links (uniqued objects are required to inherit from
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1302FoldingSetNode) that uses <a href="#dss_smallvector">SmallVector</a> as part of
1303its ID process.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1304
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1305<p>Consider a case where you want to implement a "getOrCreateFoo" method for
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1306a complex object (for example, a node in the code generator). The client has a
1307description of *what* it wants to generate (it knows the opcode and all the
1308operands), 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]1309only to find out it already exists, at which point we would have to delete it
1310and return the node that already exists.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1311</p>
1312
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1313<p>To support this style of client, FoldingSet perform a query with a
1314FoldingSetNodeID (which wraps SmallVector) that can be used to describe the
1315element that we want to query for. The query either returns the element
1316matching the ID or it returns an opaque ID that indicates where insertion should
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1317take place. Construction of the ID usually does not require heap traffic.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1318
1319<p>Because FoldingSet uses intrusive links, it can support polymorphic objects
1320in the set (for example, you can have SDNode instances mixed with LoadSDNodes).
1321Because the elements are individually allocated, pointers to the elements are
1322stable: inserting or removing elements does not invalidate any pointers to other
1323elements.
1324</p>
1325
1326</div>
1327
1328<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1329<h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1330 <a name="dss_set">&lt;set&gt;</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1331</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1332
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1333<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1334
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1335<p><tt>std::set</tt> is a reasonable all-around set class, which is decent at
1336many things but great at nothing. std::set allocates memory for each element
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1337inserted (thus it is very malloc intensive) and typically stores three pointers
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1338per element in the set (thus adding a large amount of per-element space
1339overhead). It offers guaranteed log(n) performance, which is not particularly
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1340fast from a complexity standpoint (particularly if the elements of the set are
1341expensive to compare, like strings), and has extremely high constant factors for
1342lookup, insertion and removal.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1343
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1344<p>The advantages of std::set are that its iterators are stable (deleting or
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1345inserting an element from the set does not affect iterators or pointers to other
1346elements) and that iteration over the set is guaranteed to be in sorted order.
1347If the elements in the set are large, then the relative overhead of the pointers
1348and malloc traffic is not a big deal, but if the elements of the set are small,
1349std::set is almost never a good choice.</p>
1350
1351</div>
1352
1353<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1354<h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1355 <a name="dss_setvector">"llvm/ADT/SetVector.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1356</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1357
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1358<div>
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1359<p>LLVM's SetVector&lt;Type&gt; is an adapter class that combines your choice of
1360a set-like container along with a <a href="#ds_sequential">Sequential
1361Container</a>. The important property
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1362that this provides is efficient insertion with uniquing (duplicate elements are
1363ignored) with iteration support. It implements this by inserting elements into
1364both a set-like container and the sequential container, using the set-like
1365container for uniquing and the sequential container for iteration.
1366</p>
1367
1368<p>The difference between SetVector and other sets is that the order of
1369iteration is guaranteed to match the order of insertion into the SetVector.
1370This property is really important for things like sets of pointers. Because
1371pointer values are non-deterministic (e.g. vary across runs of the program on
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1372different machines), iterating over the pointers in the set will
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1373not be in a well-defined order.</p>
1374
1375<p>
1376The drawback of SetVector is that it requires twice as much space as a normal
1377set and has the sum of constant factors from the set-like container and the
1378sequential container that it uses. Use it *only* if you need to iterate over
1379the elements in a deterministic order. SetVector is also expensive to delete
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1380elements out of (linear time), unless you use it's "pop_back" method, which is
1381faster.
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1382</p>
1383
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1384<p>SetVector is an adapter class that defaults to using std::vector and std::set
1385for the underlying containers, so it is quite expensive. However,
1386<tt>"llvm/ADT/SetVector.h"</tt> also provides a SmallSetVector class, which
1387defaults to using a SmallVector and SmallSet of a specified size. If you use
1388this, and if your sets are dynamically smaller than N, you will save a lot of
1389heap traffic.</p>
1390
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1391</div>
1392
1393<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1394<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1395 <a name="dss_uniquevector">"llvm/ADT/UniqueVector.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1396</h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1397
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1398<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1399
1400<p>
1401UniqueVector is similar to <a href="#dss_setvector">SetVector</a>, but it
1402retains a unique ID for each element inserted into the set. It internally
1403contains a map and a vector, and it assigns a unique ID for each value inserted
1404into the set.</p>
1405
1406<p>UniqueVector is very expensive: its cost is the sum of the cost of
1407maintaining both the map and vector, it has high complexity, high constant
1408factors, and produces a lot of malloc traffic. It should be avoided.</p>
1409
1410</div>
1411
1412
1413<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1414<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1415 <a name="dss_otherset">Other Set-Like Container Options</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1416</h4>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1417
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1418<div>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1419
1420<p>
1421The STL provides several other options, such as std::multiset and the various
Chris Lattnerf1a30822009-03-09 05:20:45 +0000[diff] [blame]1422"hash_set" like containers (whether from C++ TR1 or from the SGI library). We
1423never use hash_set and unordered_set because they are generally very expensive
1424(each insertion requires a malloc) and very non-portable.
1425</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1426
1427<p>std::multiset is useful if you're not interested in elimination of
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1428duplicates, but has all the drawbacks of std::set. A sorted vector (where you
1429don't delete duplicate entries) or some other approach is almost always
1430better.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1431
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1432</div>
1433
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1434</div>
1435
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1436<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1437<h3>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1438 <a name="ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1439</h3>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1440
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1441<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1442Map-like containers are useful when you want to associate data to a key. As
1443usual, there are a lot of different ways to do this. :)
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1444
1445<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1446<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1447 <a name="dss_sortedvectormap">A sorted 'vector'</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1448</h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1449
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1450<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1451
1452<p>
1453If your usage pattern follows a strict insert-then-query approach, you can
1454trivially use the same approach as <a href="#dss_sortedvectorset">sorted vectors
1455for set-like containers</a>. The only difference is that your query function
1456(which uses std::lower_bound to get efficient log(n) lookup) should only compare
1457the key, not both the key and value. This yields the same advantages as sorted
1458vectors for sets.
1459</p>
1460</div>
1461
1462<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1463<h4>
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1464 <a name="dss_stringmap">"llvm/ADT/StringMap.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1465</h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1466
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1467<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1468
1469<p>
1470Strings are commonly used as keys in maps, and they are difficult to support
1471efficiently: they are variable length, inefficient to hash and compare when
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1472long, expensive to copy, etc. StringMap is a specialized container designed to
1473cope with these issues. It supports mapping an arbitrary range of bytes to an
1474arbitrary other object.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1475
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1476<p>The StringMap implementation uses a quadratically-probed hash table, where
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1477the buckets store a pointer to the heap allocated entries (and some other
1478stuff). The entries in the map must be heap allocated because the strings are
1479variable length. The string data (key) and the element object (value) are
1480stored in the same allocation with the string data immediately after the element
1481object. This container guarantees the "<tt>(char*)(&amp;Value+1)</tt>" points
1482to the key string for a value.</p>
1483
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1484<p>The StringMap is very fast for several reasons: quadratic probing is very
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1485cache efficient for lookups, the hash value of strings in buckets is not
Nick Lewycky2a80aca2010-08-01 23:18:45 +0000[diff] [blame]1486recomputed when looking up an element, StringMap rarely has to touch the
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1487memory for unrelated objects when looking up a value (even when hash collisions
1488happen), hash table growth does not recompute the hash values for strings
1489already in the table, and each pair in the map is store in a single allocation
1490(the string data is stored in the same allocation as the Value of a pair).</p>
1491
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1492<p>StringMap also provides query methods that take byte ranges, so it only ever
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1493copies a string if a value is inserted into the table.</p>
1494</div>
1495
1496<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1497<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1498 <a name="dss_indexedmap">"llvm/ADT/IndexedMap.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1499</h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1500
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1501<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1502<p>
1503IndexedMap is a specialized container for mapping small dense integers (or
1504values that can be mapped to small dense integers) to some other type. It is
1505internally implemented as a vector with a mapping function that maps the keys to
1506the dense integer range.
1507</p>
1508
1509<p>
1510This is useful for cases like virtual registers in the LLVM code generator: they
1511have a dense mapping that is offset by a compile-time constant (the first
1512virtual register ID).</p>
1513
1514</div>
1515
1516<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1517<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1518 <a name="dss_densemap">"llvm/ADT/DenseMap.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1519</h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1520
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1521<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1522
1523<p>
1524DenseMap is a simple quadratically probed hash table. It excels at supporting
1525small keys and values: it uses a single allocation to hold all of the pairs that
1526are currently inserted in the map. DenseMap is a great way to map pointers to
1527pointers, or map other small types to each other.
1528</p>
1529
1530<p>
1531There are several aspects of DenseMap that you should be aware of, however. The
1532iterators in a densemap are invalidated whenever an insertion occurs, unlike
1533map. Also, because DenseMap allocates space for a large number of key/value
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1534pairs (it starts with 64 by default), it will waste a lot of space if your keys
1535or values are large. Finally, you must implement a partial specialization of
Chris Lattner76c1b972007-09-17 18:34:04 +0000[diff] [blame]1536DenseMapInfo for the key that you want, if it isn't already supported. This
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1537is required to tell DenseMap about two special marker values (which can never be
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1538inserted into the map) that it needs internally.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1539
1540</div>
1541
1542<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1543<h4>
Jeffrey Yasskin71a5c222009-10-22 22:11:22 +0000[diff] [blame]1544 <a name="dss_valuemap">"llvm/ADT/ValueMap.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1545</h4>
Jeffrey Yasskin71a5c222009-10-22 22:11:22 +0000[diff] [blame]1546
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1547<div>
Jeffrey Yasskin71a5c222009-10-22 22:11:22 +0000[diff] [blame]1548
1549<p>
1550ValueMap is a wrapper around a <a href="#dss_densemap">DenseMap</a> mapping
1551Value*s (or subclasses) to another type. When a Value is deleted or RAUW'ed,
1552ValueMap will update itself so the new version of the key is mapped to the same
1553value, just as if the key were a WeakVH. You can configure exactly how this
1554happens, and what else happens on these two events, by passing
1555a <code>Config</code> parameter to the ValueMap template.</p>
1556
1557</div>
1558
1559<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1560<h4>
Jakob Stoklund Olesenaca0da62010-12-14 00:55:51 +0000[diff] [blame]1561 <a name="dss_intervalmap">"llvm/ADT/IntervalMap.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1562</h4>
Jakob Stoklund Olesenaca0da62010-12-14 00:55:51 +0000[diff] [blame]1563
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1564<div>
Jakob Stoklund Olesenaca0da62010-12-14 00:55:51 +0000[diff] [blame]1565
1566<p> IntervalMap is a compact map for small keys and values. It maps key
1567intervals instead of single keys, and it will automatically coalesce adjacent
1568intervals. When then map only contains a few intervals, they are stored in the
1569map object itself to avoid allocations.</p>
1570
1571<p> The IntervalMap iterators are quite big, so they should not be passed around
1572as STL iterators. The heavyweight iterators allow a smaller data structure.</p>
1573
1574</div>
1575
1576<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1577<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1578 <a name="dss_map">&lt;map&gt;</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1579</h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1580
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1581<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1582
1583<p>
1584std::map has similar characteristics to <a href="#dss_set">std::set</a>: it uses
1585a single allocation per pair inserted into the map, it offers log(n) lookup with
1586an extremely large constant factor, imposes a space penalty of 3 pointers per
1587pair in the map, etc.</p>
1588
1589<p>std::map is most useful when your keys or values are very large, if you need
1590to iterate over the collection in sorted order, or if you need stable iterators
1591into the map (i.e. they don't get invalidated if an insertion or deletion of
1592another element takes place).</p>
1593
1594</div>
1595
1596<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1597<h4>
Jakob Stoklund Olesen4359e5e2011-04-05 20:56:08 +0000[diff] [blame]1598 <a name="dss_inteqclasses">"llvm/ADT/IntEqClasses.h"</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1599</h4>
Jakob Stoklund Olesen4359e5e2011-04-05 20:56:08 +0000[diff] [blame]1600
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1601<div>
Jakob Stoklund Olesen4359e5e2011-04-05 20:56:08 +0000[diff] [blame]1602
1603<p>IntEqClasses provides a compact representation of equivalence classes of
1604small integers. Initially, each integer in the range 0..n-1 has its own
1605equivalence class. Classes can be joined by passing two class representatives to
1606the join(a, b) method. Two integers are in the same class when findLeader()
1607returns the same representative.</p>
1608
1609<p>Once all equivalence classes are formed, the map can be compressed so each
1610integer 0..n-1 maps to an equivalence class number in the range 0..m-1, where m
1611is the total number of equivalence classes. The map must be uncompressed before
1612it can be edited again.</p>
1613
1614</div>
1615
1616<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1617<h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1618 <a name="dss_othermap">Other Map-Like Container Options</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1619</h4>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1620
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1621<div>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1622
1623<p>
1624The STL provides several other options, such as std::multimap and the various
Chris Lattnerf1a30822009-03-09 05:20:45 +0000[diff] [blame]1625"hash_map" like containers (whether from C++ TR1 or from the SGI library). We
1626never use hash_set and unordered_set because they are generally very expensive
1627(each insertion requires a malloc) and very non-portable.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1628
1629<p>std::multimap is useful if you want to map a key to multiple values, but has
1630all the drawbacks of std::map. A sorted vector or some other approach is almost
1631always better.</p>
1632
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1633</div>
1634
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1635</div>
1636
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1637<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1638<h3>
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]1639 <a name="ds_string">String-like containers</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1640</h3>
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]1641
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1642<div>
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]1643
1644<p>
1645TODO: const char* vs stringref vs smallstring vs std::string. Describe twine,
1646xref to #string_apis.
1647</p>
1648
1649</div>
1650
1651<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1652<h3>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1653 <a name="ds_bit">Bit storage containers (BitVector, SparseBitVector)</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1654</h3>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1655
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1656<div>
Chris Lattner7086ce72007-09-25 22:37:50 +0000[diff] [blame]1657<p>Unlike the other containers, there are only two bit storage containers, and
1658choosing when to use each is relatively straightforward.</p>
1659
1660<p>One additional option is
1661<tt>std::vector&lt;bool&gt;</tt>: we discourage its use for two reasons 1) the
1662implementation in many common compilers (e.g. commonly available versions of
1663GCC) is extremely inefficient and 2) the C++ standards committee is likely to
1664deprecate this container and/or change it significantly somehow. In any case,
1665please don't use it.</p>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1666
1667<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1668<h4>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1669 <a name="dss_bitvector">BitVector</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1670</h4>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1671
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1672<div>
Dan Gohman5f7775c2010-01-05 18:24:00 +0000[diff] [blame]1673<p> The BitVector container provides a dynamic size set of bits for manipulation.
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1674It supports individual bit setting/testing, as well as set operations. The set
1675operations take time O(size of bitvector), but operations are performed one word
1676at a time, instead of one bit at a time. This makes the BitVector very fast for
1677set operations compared to other containers. Use the BitVector when you expect
1678the number of set bits to be high (IE a dense set).
1679</p>
1680</div>
1681
1682<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1683<h4>
Dan Gohman5f7775c2010-01-05 18:24:00 +0000[diff] [blame]1684 <a name="dss_smallbitvector">SmallBitVector</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1685</h4>
Dan Gohman5f7775c2010-01-05 18:24:00 +0000[diff] [blame]1686
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1687<div>
Dan Gohman5f7775c2010-01-05 18:24:00 +0000[diff] [blame]1688<p> The SmallBitVector container provides the same interface as BitVector, but
1689it is optimized for the case where only a small number of bits, less than
169025 or so, are needed. It also transparently supports larger bit counts, but
1691slightly less efficiently than a plain BitVector, so SmallBitVector should
1692only be used when larger counts are rare.
1693</p>
1694
1695<p>
1696At this time, SmallBitVector does not support set operations (and, or, xor),
1697and its operator[] does not provide an assignable lvalue.
1698</p>
1699</div>
1700
1701<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1702<h4>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1703 <a name="dss_sparsebitvector">SparseBitVector</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1704</h4>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1705
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1706<div>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1707<p> The SparseBitVector container is much like BitVector, with one major
1708difference: Only the bits that are set, are stored. This makes the
1709SparseBitVector much more space efficient than BitVector when the set is sparse,
1710as well as making set operations O(number of set bits) instead of O(size of
1711universe). 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
1712(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).
1713</p>
1714</div>
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]1715
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1716</div>
1717
1718</div>
1719
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1720<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1721<h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1722 <a name="common">Helpful Hints for Common Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1723</h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1724<!-- *********************************************************************** -->
1725
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1726<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1727
1728<p>This section describes how to perform some very simple transformations of
1729LLVM code. This is meant to give examples of common idioms used, showing the
1730practical side of LLVM transformations. <p> Because this is a "how-to" section,
1731you should also read about the main classes that you will be working with. The
1732<a href="#coreclasses">Core LLVM Class Hierarchy Reference</a> contains details
1733and descriptions of the main classes that you should know about.</p>
1734
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1735<!-- NOTE: this section should be heavy on example code -->
1736<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1737<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1738 <a name="inspection">Basic Inspection and Traversal Routines</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1739</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1740
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1741<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1742
1743<p>The LLVM compiler infrastructure have many different data structures that may
1744be traversed. Following the example of the C++ standard template library, the
1745techniques used to traverse these various data structures are all basically the
1746same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
1747method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
1748function returns an iterator pointing to one past the last valid element of the
1749sequence, and there is some <tt>XXXiterator</tt> data type that is common
1750between the two operations.</p>
1751
1752<p>Because the pattern for iteration is common across many different aspects of
1753the program representation, the standard template library algorithms may be used
1754on them, and it is easier to remember how to iterate. First we show a few common
1755examples of the data structures that need to be traversed. Other data
1756structures are traversed in very similar ways.</p>
1757
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1758<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1759<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1760 <a name="iterate_function">Iterating over the </a><a
1761 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
1762 href="#Function"><tt>Function</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1763</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1764
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1765<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1766
1767<p>It's quite common to have a <tt>Function</tt> instance that you'd like to
1768transform in some way; in particular, you'd like to manipulate its
1769<tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over all of
1770the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>. The following is
1771an example that prints the name of a <tt>BasicBlock</tt> and the number of
1772<tt>Instruction</tt>s it contains:</p>
1773
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1774<div class="doc_code">
1775<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1776// <i>func is a pointer to a Function instance</i>
1777for (Function::iterator i = func-&gt;begin(), e = func-&gt;end(); i != e; ++i)
1778 // <i>Print out the name of the basic block if it has one, and then the</i>
1779 // <i>number of instructions that it contains</i>
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1780 errs() &lt;&lt; "Basic block (name=" &lt;&lt; i-&gt;getName() &lt;&lt; ") has "
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1781 &lt;&lt; i-&gt;size() &lt;&lt; " instructions.\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1782</pre>
1783</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1784
1785<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]1786invoking member functions of the <tt>Instruction</tt> class. This is
1787because the indirection operator is overloaded for the iterator
Chris Lattner7496ec52003-08-05 22:54:23 +0000[diff] [blame]1788classes. In the above code, the expression <tt>i-&gt;size()</tt> is
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1789exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.</p>
1790
1791</div>
1792
1793<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1794<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1795 <a name="iterate_basicblock">Iterating over the </a><a
1796 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1797 href="#BasicBlock"><tt>BasicBlock</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1798</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1799
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1800<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1801
1802<p>Just like when dealing with <tt>BasicBlock</tt>s in <tt>Function</tt>s, it's
1803easy to iterate over the individual instructions that make up
1804<tt>BasicBlock</tt>s. Here's a code snippet that prints out each instruction in
1805a <tt>BasicBlock</tt>:</p>
1806
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1807<div class="doc_code">
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1808<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1809// <i>blk is a pointer to a BasicBlock instance</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1810for (BasicBlock::iterator i = blk-&gt;begin(), e = blk-&gt;end(); i != e; ++i)
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1811 // <i>The next statement works since operator&lt;&lt;(ostream&amp;,...)</i>
1812 // <i>is overloaded for Instruction&amp;</i>
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1813 errs() &lt;&lt; *i &lt;&lt; "\n";
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1814</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1815</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1816
1817<p>However, this isn't really the best way to print out the contents of a
1818<tt>BasicBlock</tt>! Since the ostream operators are overloaded for virtually
1819anything you'll care about, you could have just invoked the print routine on the
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1820basic block itself: <tt>errs() &lt;&lt; *blk &lt;&lt; "\n";</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1821
1822</div>
1823
1824<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1825<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1826 <a name="iterate_institer">Iterating over the </a><a
1827 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1828 href="#Function"><tt>Function</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1829</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1830
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1831<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1832
1833<p>If you're finding that you commonly iterate over a <tt>Function</tt>'s
1834<tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s <tt>Instruction</tt>s,
1835<tt>InstIterator</tt> should be used instead. You'll need to include <a
1836href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>,
1837and then instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1838small 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]1839
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1840<div class="doc_code">
1841<pre>
1842#include "<a href="/doxygen/InstIterator_8h-source.html">llvm/Support/InstIterator.h</a>"
1843
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1844// <i>F is a pointer to a Function instance</i>
Chris Lattnerda021aa2008-06-04 18:20:42 +0000[diff] [blame]1845for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1846 errs() &lt;&lt; *I &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1847</pre>
1848</div>
1849
1850<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]1851work list with its initial contents. For example, if you wanted to
1852initialize a work list to contain all instructions in a <tt>Function</tt>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1853F, all you would need to do is something like:</p>
1854
1855<div class="doc_code">
1856<pre>
1857std::set&lt;Instruction*&gt; worklist;
Chris Lattnerda021aa2008-06-04 18:20:42 +0000[diff] [blame]1858// or better yet, SmallPtrSet&lt;Instruction*, 64&gt; worklist;
1859
1860for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
1861 worklist.insert(&amp;*I);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1862</pre>
1863</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1864
1865<p>The STL set <tt>worklist</tt> would now contain all instructions in the
1866<tt>Function</tt> pointed to by F.</p>
1867
1868</div>
1869
1870<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1871<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1872 <a name="iterate_convert">Turning an iterator into a class pointer (and
1873 vice-versa)</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1874</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1875
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1876<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1877
1878<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]1879instance when all you've got at hand is an iterator. Well, extracting
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1880a reference or a pointer from an iterator is very straight-forward.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1881Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and <tt>j</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1882is a <tt>BasicBlock::const_iterator</tt>:</p>
1883
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1884<div class="doc_code">
1885<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1886Instruction&amp; inst = *i; // <i>Grab reference to instruction reference</i>
1887Instruction* pinst = &amp;*i; // <i>Grab pointer to instruction reference</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1888const Instruction&amp; inst = *j;
1889</pre>
1890</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1891
1892<p>However, the iterators you'll be working with in the LLVM framework are
1893special: they will automatically convert to a ptr-to-instance type whenever they
1894need to. Instead of dereferencing the iterator and then taking the address of
1895the result, you can simply assign the iterator to the proper pointer type and
1896you get the dereference and address-of operation as a result of the assignment
1897(behind the scenes, this is a result of overloading casting mechanisms). Thus
1898the last line of the last example,</p>
1899
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1900<div class="doc_code">
1901<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1902Instruction *pinst = &amp;*i;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1903</pre>
1904</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1905
1906<p>is semantically equivalent to</p>
1907
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1908<div class="doc_code">
1909<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1910Instruction *pinst = i;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1911</pre>
1912</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1913
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1914<p>It's also possible to turn a class pointer into the corresponding iterator,
1915and this is a constant time operation (very efficient). The following code
1916snippet illustrates use of the conversion constructors provided by LLVM
1917iterators. By using these, you can explicitly grab the iterator of something
1918without actually obtaining it via iteration over some structure:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1919
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1920<div class="doc_code">
1921<pre>
1922void printNextInstruction(Instruction* inst) {
1923 BasicBlock::iterator it(inst);
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1924 ++it; // <i>After this line, it refers to the instruction after *inst</i>
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1925 if (it != inst-&gt;getParent()-&gt;end()) errs() &lt;&lt; *it &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1926}
1927</pre>
1928</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1929
Dan Gohman525bf8e2010-03-26 19:39:05 +0000[diff] [blame]1930<p>Unfortunately, these implicit conversions come at a cost; they prevent
1931these iterators from conforming to standard iterator conventions, and thus
Dan Gohman0d91c112010-03-26 19:51:14 +0000[diff] [blame]1932from being usable with standard algorithms and containers. For example, they
1933prevent the following code, where <tt>B</tt> is a <tt>BasicBlock</tt>,
Dan Gohman525bf8e2010-03-26 19:39:05 +0000[diff] [blame]1934from compiling:</p>
1935
1936<div class="doc_code">
1937<pre>
1938 llvm::SmallVector&lt;llvm::Instruction *, 16&gt;(B-&gt;begin(), B-&gt;end());
1939</pre>
1940</div>
1941
1942<p>Because of this, these implicit conversions may be removed some day,
Dan Gohman0d91c112010-03-26 19:51:14 +0000[diff] [blame]1943and <tt>operator*</tt> changed to return a pointer instead of a reference.</p>
Dan Gohman525bf8e2010-03-26 19:39:05 +0000[diff] [blame]1944
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1945</div>
1946
1947<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1948<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1949 <a name="iterate_complex">Finding call sites: a slightly more complex
1950 example</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1951</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1952
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1953<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1954
1955<p>Say that you're writing a FunctionPass and would like to count all the
1956locations in the entire module (that is, across every <tt>Function</tt>) where a
1957certain function (i.e., some <tt>Function</tt>*) is already in scope. As you'll
1958learn 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]1959much more straight-forward manner, but this example will allow us to explore how
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1960you'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]1961is what we want to do:</p>
1962
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1963<div class="doc_code">
1964<pre>
1965initialize callCounter to zero
1966for each Function f in the Module
1967 for each BasicBlock b in f
1968 for each Instruction i in b
1969 if (i is a CallInst and calls the given function)
1970 increment callCounter
1971</pre>
1972</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1973
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1974<p>And the actual code is (remember, because we're writing a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1975<tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply has to
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1976override the <tt>runOnFunction</tt> method):</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1977
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1978<div class="doc_code">
1979<pre>
1980Function* targetFunc = ...;
1981
1982class OurFunctionPass : public FunctionPass {
1983 public:
1984 OurFunctionPass(): callCounter(0) { }
1985
1986 virtual runOnFunction(Function&amp; F) {
1987 for (Function::iterator b = F.begin(), be = F.end(); b != be; ++b) {
Eric Christopher203e71d2008-11-08 08:20:49 +0000[diff] [blame]1988 for (BasicBlock::iterator i = b-&gt;begin(), ie = b-&gt;end(); i != ie; ++i) {
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1989 if (<a href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a>&lt;<a
1990 href="#CallInst">CallInst</a>&gt;(&amp;*i)) {
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1991 // <i>We know we've encountered a call instruction, so we</i>
1992 // <i>need to determine if it's a call to the</i>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1993 // <i>function pointed to by m_func or not.</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1994 if (callInst-&gt;getCalledFunction() == targetFunc)
1995 ++callCounter;
1996 }
1997 }
1998 }
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1999 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2000
2001 private:
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2002 unsigned callCounter;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2003};
2004</pre>
2005</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2006
2007</div>
2008
Brian Gaekef1972c62003-11-07 19:25:45 +0000[diff] [blame]2009<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2010<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2011 <a name="calls_and_invokes">Treating calls and invokes the same way</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2012</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2013
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2014<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2015
2016<p>You may have noticed that the previous example was a bit oversimplified in
2017that it did not deal with call sites generated by 'invoke' instructions. In
2018this, and in other situations, you may find that you want to treat
2019<tt>CallInst</tt>s and <tt>InvokeInst</tt>s the same way, even though their
2020most-specific common base class is <tt>Instruction</tt>, which includes lots of
2021less closely-related things. For these cases, LLVM provides a handy wrapper
2022class called <a
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]2023href="http://llvm.org/doxygen/classllvm_1_1CallSite.html"><tt>CallSite</tt></a>.
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]2024It is essentially a wrapper around an <tt>Instruction</tt> pointer, with some
2025methods that provide functionality common to <tt>CallInst</tt>s and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2026<tt>InvokeInst</tt>s.</p>
2027
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]2028<p>This class has "value semantics": it should be passed by value, not by
2029reference and it should not be dynamically allocated or deallocated using
2030<tt>operator new</tt> or <tt>operator delete</tt>. It is efficiently copyable,
2031assignable and constructable, with costs equivalents to that of a bare pointer.
2032If you look at its definition, it has only a single pointer member.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2033
2034</div>
2035
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]2036<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2037<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2038 <a name="iterate_chains">Iterating over def-use &amp; use-def chains</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2039</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2040
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2041<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2042
2043<p>Frequently, we might have an instance of the <a
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]2044href="/doxygen/classllvm_1_1Value.html">Value Class</a> and we want to
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]2045determine which <tt>User</tt>s use the <tt>Value</tt>. The list of all
2046<tt>User</tt>s of a particular <tt>Value</tt> is called a <i>def-use</i> chain.
2047For example, let's say we have a <tt>Function*</tt> named <tt>F</tt> to a
2048particular function <tt>foo</tt>. Finding all of the instructions that
2049<i>use</i> <tt>foo</tt> is as simple as iterating over the <i>def-use</i> chain
2050of <tt>F</tt>:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2051
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2052<div class="doc_code">
2053<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2054Function *F = ...;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2055
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2056for (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]2057 if (Instruction *Inst = dyn_cast&lt;Instruction&gt;(*i)) {
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]2058 errs() &lt;&lt; "F is used in instruction:\n";
2059 errs() &lt;&lt; *Inst &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2060 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2061</pre>
Gabor Greif394fdfb2010-03-26 19:35:48 +0000[diff] [blame]2062</div>
2063
Gabor Greifce94319532010-03-26 19:40:38 +0000[diff] [blame]2064<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]2065operation. Instead of performing <tt>*i</tt> above several times, consider
Gabor Greifce94319532010-03-26 19:40:38 +0000[diff] [blame]2066doing it only once in the loop body and reusing its result.</p>
Gabor Greif4de73682010-03-26 19:30:47 +0000[diff] [blame]2067
Gabor Greif6091ff32010-03-26 19:04:42 +0000[diff] [blame]2068<p>Alternatively, it's common to have an instance of the <a
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]2069href="/doxygen/classllvm_1_1User.html">User Class</a> and need to know what
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2070<tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used by a
2071<tt>User</tt> is known as a <i>use-def</i> chain. Instances of class
2072<tt>Instruction</tt> are common <tt>User</tt>s, so we might want to iterate over
2073all of the values that a particular instruction uses (that is, the operands of
2074the particular <tt>Instruction</tt>):</p>
2075
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2076<div class="doc_code">
2077<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2078Instruction *pi = ...;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2079
2080for (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]2081 Value *v = *i;
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2082 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2083}
2084</pre>
2085</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2086
Gabor Greif4de73682010-03-26 19:30:47 +0000[diff] [blame]2087<p>Declaring objects as <tt>const</tt> is an important tool of enforcing
Gabor Greifce94319532010-03-26 19:40:38 +0000[diff] [blame]2088mutation free algorithms (such as analyses, etc.). For this purpose above
Gabor Greif4de73682010-03-26 19:30:47 +0000[diff] [blame]2089iterators come in constant flavors as <tt>Value::const_use_iterator</tt>
2090and <tt>Value::const_op_iterator</tt>. They automatically arise when
2091calling <tt>use/op_begin()</tt> on <tt>const Value*</tt>s or
2092<tt>const User*</tt>s respectively. Upon dereferencing, they return
Gabor Greifce94319532010-03-26 19:40:38 +0000[diff] [blame]2093<tt>const Use*</tt>s. Otherwise the above patterns remain unchanged.</p>
2094
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2095</div>
2096
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2097<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2098<h4>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2099 <a name="iterate_preds">Iterating over predecessors &amp;
2100successors of blocks</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2101</h4>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2102
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2103<div>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2104
2105<p>Iterating over the predecessors and successors of a block is quite easy
2106with the routines defined in <tt>"llvm/Support/CFG.h"</tt>. Just use code like
2107this to iterate over all predecessors of BB:</p>
2108
2109<div class="doc_code">
2110<pre>
2111#include "llvm/Support/CFG.h"
2112BasicBlock *BB = ...;
2113
2114for (pred_iterator PI = pred_begin(BB), E = pred_end(BB); PI != E; ++PI) {
2115 BasicBlock *Pred = *PI;
2116 // <i>...</i>
2117}
2118</pre>
2119</div>
2120
2121<p>Similarly, to iterate over successors use
2122succ_iterator/succ_begin/succ_end.</p>
2123
2124</div>
2125
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2126</div>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]2127
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2128<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2129<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2130 <a name="simplechanges">Making simple changes</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2131</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2132
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2133<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2134
2135<p>There are some primitive transformation operations present in the LLVM
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]2136infrastructure that are worth knowing about. When performing
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2137transformations, it's fairly common to manipulate the contents of basic
2138blocks. This section describes some of the common methods for doing so
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2139and gives example code.</p>
2140
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2141<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2142<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2143 <a name="schanges_creating">Creating and inserting new
2144 <tt>Instruction</tt>s</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2145</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2146
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2147<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2148
2149<p><i>Instantiating Instructions</i></p>
2150
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]2151<p>Creation of <tt>Instruction</tt>s is straight-forward: simply call the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2152constructor for the kind of instruction to instantiate and provide the necessary
2153parameters. For example, an <tt>AllocaInst</tt> only <i>requires</i> a
2154(const-ptr-to) <tt>Type</tt>. Thus:</p>
2155
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2156<div class="doc_code">
2157<pre>
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2158AllocaInst* ai = new AllocaInst(Type::Int32Ty);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2159</pre>
2160</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2161
2162<p>will create an <tt>AllocaInst</tt> instance that represents the allocation of
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]2163one integer in the current stack frame, at run time. Each <tt>Instruction</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2164subclass is likely to have varying default parameters which change the semantics
2165of the instruction, so refer to the <a
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]2166href="/doxygen/classllvm_1_1Instruction.html">doxygen documentation for the subclass of
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2167Instruction</a> that you're interested in instantiating.</p>
2168
2169<p><i>Naming values</i></p>
2170
2171<p>It is very useful to name the values of instructions when you're able to, as
2172this facilitates the debugging of your transformations. If you end up looking
2173at generated LLVM machine code, you definitely want to have logical names
2174associated with the results of instructions! By supplying a value for the
2175<tt>Name</tt> (default) parameter of the <tt>Instruction</tt> constructor, you
2176associate a logical name with the result of the instruction's execution at
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]2177run time. For example, say that I'm writing a transformation that dynamically
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2178allocates space for an integer on the stack, and that integer is going to be
2179used as some kind of index by some other code. To accomplish this, I place an
2180<tt>AllocaInst</tt> at the first point in the first <tt>BasicBlock</tt> of some
2181<tt>Function</tt>, and I'm intending to use it within the same
2182<tt>Function</tt>. I might do:</p>
2183
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2184<div class="doc_code">
2185<pre>
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2186AllocaInst* pa = new AllocaInst(Type::Int32Ty, 0, "indexLoc");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2187</pre>
2188</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2189
2190<p>where <tt>indexLoc</tt> is now the logical name of the instruction's
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]2191execution value, which is a pointer to an integer on the run time stack.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2192
2193<p><i>Inserting instructions</i></p>
2194
2195<p>There are essentially two ways to insert an <tt>Instruction</tt>
2196into an existing sequence of instructions that form a <tt>BasicBlock</tt>:</p>
2197
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]2198<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2199 <li>Insertion into an explicit instruction list
2200
2201 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within that
2202 <tt>BasicBlock</tt>, and a newly-created instruction we wish to insert
2203 before <tt>*pi</tt>, we do the following: </p>
2204
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2205<div class="doc_code">
2206<pre>
2207BasicBlock *pb = ...;
2208Instruction *pi = ...;
2209Instruction *newInst = new Instruction(...);
2210
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2211pb-&gt;getInstList().insert(pi, newInst); // <i>Inserts newInst before pi in pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2212</pre>
2213</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]2214
2215 <p>Appending to the end of a <tt>BasicBlock</tt> is so common that
2216 the <tt>Instruction</tt> class and <tt>Instruction</tt>-derived
2217 classes provide constructors which take a pointer to a
2218 <tt>BasicBlock</tt> to be appended to. For example code that
2219 looked like: </p>
2220
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2221<div class="doc_code">
2222<pre>
2223BasicBlock *pb = ...;
2224Instruction *newInst = new Instruction(...);
2225
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2226pb-&gt;getInstList().push_back(newInst); // <i>Appends newInst to pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2227</pre>
2228</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]2229
2230 <p>becomes: </p>
2231
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2232<div class="doc_code">
2233<pre>
2234BasicBlock *pb = ...;
2235Instruction *newInst = new Instruction(..., pb);
2236</pre>
2237</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]2238
2239 <p>which is much cleaner, especially if you are creating
2240 long instruction streams.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2241
2242 <li>Insertion into an implicit instruction list
2243
2244 <p><tt>Instruction</tt> instances that are already in <tt>BasicBlock</tt>s
2245 are implicitly associated with an existing instruction list: the instruction
2246 list of the enclosing basic block. Thus, we could have accomplished the same
2247 thing as the above code without being given a <tt>BasicBlock</tt> by doing:
2248 </p>
2249
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2250<div class="doc_code">
2251<pre>
2252Instruction *pi = ...;
2253Instruction *newInst = new Instruction(...);
2254
2255pi-&gt;getParent()-&gt;getInstList().insert(pi, newInst);
2256</pre>
2257</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2258
2259 <p>In fact, this sequence of steps occurs so frequently that the
2260 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes provide
2261 constructors which take (as a default parameter) a pointer to an
2262 <tt>Instruction</tt> which the newly-created <tt>Instruction</tt> should
2263 precede. That is, <tt>Instruction</tt> constructors are capable of
2264 inserting the newly-created instance into the <tt>BasicBlock</tt> of a
2265 provided instruction, immediately before that instruction. Using an
2266 <tt>Instruction</tt> constructor with a <tt>insertBefore</tt> (default)
2267 parameter, the above code becomes:</p>
2268
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2269<div class="doc_code">
2270<pre>
2271Instruction* pi = ...;
2272Instruction* newInst = new Instruction(..., pi);
2273</pre>
2274</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2275
2276 <p>which is much cleaner, especially if you're creating a lot of
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2277 instructions and adding them to <tt>BasicBlock</tt>s.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2278</ul>
2279
2280</div>
2281
2282<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2283<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2284 <a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2285</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2286
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2287<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2288
2289<p>Deleting an instruction from an existing sequence of instructions that form a
Chris Lattner49e0ccf2011-03-24 16:13:31 +0000[diff] [blame]2290<a href="#BasicBlock"><tt>BasicBlock</tt></a> is very straight-forward: just
2291call the instruction's eraseFromParent() method. For example:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2292
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2293<div class="doc_code">
2294<pre>
2295<a href="#Instruction">Instruction</a> *I = .. ;
Chris Lattner9f8ec252008-02-15 22:57:17 +0000[diff] [blame]2296I-&gt;eraseFromParent();
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2297</pre>
2298</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2299
Chris Lattner49e0ccf2011-03-24 16:13:31 +0000[diff] [blame]2300<p>This unlinks the instruction from its containing basic block and deletes
2301it. If you'd just like to unlink the instruction from its containing basic
2302block but not delete it, you can use the <tt>removeFromParent()</tt> method.</p>
2303
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2304</div>
2305
2306<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2307<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2308 <a name="schanges_replacing">Replacing an <tt>Instruction</tt> with another
2309 <tt>Value</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2310</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2311
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2312<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2313
2314<p><i>Replacing individual instructions</i></p>
2315
2316<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]2317permits use of two very useful replace functions: <tt>ReplaceInstWithValue</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2318and <tt>ReplaceInstWithInst</tt>.</p>
2319
NAKAMURA Takumi06c6d9a2011-04-18 01:17:51 +0000[diff] [blame]2320<h5><a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a></h5>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2321
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2322<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2323 <li><tt>ReplaceInstWithValue</tt>
2324
Nick Lewyckyb6d1f392008-09-15 06:31:52 +0000[diff] [blame]2325 <p>This function replaces all uses of a given instruction with a value,
2326 and then removes the original instruction. The following example
2327 illustrates the replacement of the result of a particular
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]2328 <tt>AllocaInst</tt> that allocates memory for a single integer with a null
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2329 pointer to an integer.</p>
2330
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2331<div class="doc_code">
2332<pre>
2333AllocaInst* instToReplace = ...;
2334BasicBlock::iterator ii(instToReplace);
2335
2336ReplaceInstWithValue(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
Daniel Dunbar58c2ac02008-10-03 22:17:25 +0000[diff] [blame]2337 Constant::getNullValue(PointerType::getUnqual(Type::Int32Ty)));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2338</pre></div></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2339
2340 <li><tt>ReplaceInstWithInst</tt>
2341
2342 <p>This function replaces a particular instruction with another
Nick Lewyckyb6d1f392008-09-15 06:31:52 +0000[diff] [blame]2343 instruction, inserting the new instruction into the basic block at the
2344 location where the old instruction was, and replacing any uses of the old
2345 instruction with the new instruction. The following example illustrates
2346 the replacement of one <tt>AllocaInst</tt> with another.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2347
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2348<div class="doc_code">
2349<pre>
2350AllocaInst* instToReplace = ...;
2351BasicBlock::iterator ii(instToReplace);
2352
2353ReplaceInstWithInst(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2354 new AllocaInst(Type::Int32Ty, 0, "ptrToReplacedInt"));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2355</pre></div></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2356</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2357
2358<p><i>Replacing multiple uses of <tt>User</tt>s and <tt>Value</tt>s</i></p>
2359
2360<p>You can use <tt>Value::replaceAllUsesWith</tt> and
2361<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]2362doxygen documentation for the <a href="/doxygen/classllvm_1_1Value.html">Value Class</a>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]2363and <a href="/doxygen/classllvm_1_1User.html">User Class</a>, respectively, for more
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2364information.</p>
2365
2366<!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
2367include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
2368ReplaceInstWithValue, ReplaceInstWithInst -->
2369
2370</div>
2371
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2372<!--_______________________________________________________________________-->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2373<h4>
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2374 <a name="schanges_deletingGV">Deleting <tt>GlobalVariable</tt>s</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2375</h4>
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2376
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2377<div>
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2378
Tanya Lattnerc5dfcdb2007-06-20 20:46:37 +0000[diff] [blame]2379<p>Deleting a global variable from a module is just as easy as deleting an
2380Instruction. First, you must have a pointer to the global variable that you wish
2381 to delete. You use this pointer to erase it from its parent, the module.
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2382 For example:</p>
2383
2384<div class="doc_code">
2385<pre>
2386<a href="#GlobalVariable">GlobalVariable</a> *GV = .. ;
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2387
Tanya Lattnerc5dfcdb2007-06-20 20:46:37 +0000[diff] [blame]2388GV-&gt;eraseFromParent();
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2389</pre>
2390</div>
2391
2392</div>
2393
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2394</div>
2395
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2396<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2397<h3>
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2398 <a name="create_types">How to Create Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2399</h3>
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2400
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2401<div>
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2402
2403<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]2404statically, you can use <tt>TypeBuilder&lt;...&gt;::get()</tt>, defined
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2405in <tt>llvm/Support/TypeBuilder.h</tt>, to retrieve them. <tt>TypeBuilder</tt>
2406has two forms depending on whether you're building types for cross-compilation
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2407or native library use. <tt>TypeBuilder&lt;T, true&gt;</tt> requires
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2408that <tt>T</tt> be independent of the host environment, meaning that it's built
2409out of types from
2410the <a href="/doxygen/namespacellvm_1_1types.html"><tt>llvm::types</tt></a>
2411namespace and pointers, functions, arrays, etc. built of
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2412those. <tt>TypeBuilder&lt;T, false&gt;</tt> additionally allows native C types
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2413whose size may depend on the host compiler. For example,</p>
2414
2415<div class="doc_code">
2416<pre>
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2417FunctionType *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]2418</pre>
2419</div>
2420
2421<p>is easier to read and write than the equivalent</p>
2422
2423<div class="doc_code">
2424<pre>
Owen Anderson5e8c50e2009-06-16 17:40:28 +0000[diff] [blame]2425std::vector&lt;const Type*&gt; params;
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2426params.push_back(PointerType::getUnqual(Type::Int32Ty));
2427FunctionType *ft = FunctionType::get(Type::Int8Ty, params, false);
2428</pre>
2429</div>
2430
2431<p>See the <a href="/doxygen/TypeBuilder_8h-source.html#l00001">class
2432comment</a> for more details.</p>
2433
2434</div>
2435
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2436</div>
2437
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]2438<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2439<h2>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2440 <a name="threading">Threads and LLVM</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2441</h2>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2442<!-- *********************************************************************** -->
2443
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2444<div>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2445<p>
2446This section describes the interaction of the LLVM APIs with multithreading,
2447both on the part of client applications, and in the JIT, in the hosted
2448application.
2449</p>
2450
2451<p>
2452Note that LLVM's support for multithreading is still relatively young. Up
2453through version 2.5, the execution of threaded hosted applications was
2454supported, but not threaded client access to the APIs. While this use case is
2455now supported, clients <em>must</em> adhere to the guidelines specified below to
2456ensure proper operation in multithreaded mode.
2457</p>
2458
2459<p>
2460Note that, on Unix-like platforms, LLVM requires the presence of GCC's atomic
2461intrinsics in order to support threaded operation. If you need a
2462multhreading-capable LLVM on a platform without a suitably modern system
2463compiler, consider compiling LLVM and LLVM-GCC in single-threaded mode, and
2464using the resultant compiler to build a copy of LLVM with multithreading
2465support.
2466</p>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2467
2468<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2469<h3>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2470 <a name="startmultithreaded">Entering and Exiting Multithreaded Mode</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2471</h3>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2472
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2473<div>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2474
2475<p>
2476In order to properly protect its internal data structures while avoiding
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2477excessive locking overhead in the single-threaded case, the LLVM must intialize
2478certain data structures necessary to provide guards around its internals. To do
2479so, the client program must invoke <tt>llvm_start_multithreaded()</tt> before
2480making any concurrent LLVM API calls. To subsequently tear down these
2481structures, use the <tt>llvm_stop_multithreaded()</tt> call. You can also use
2482the <tt>llvm_is_multithreaded()</tt> call to check the status of multithreaded
2483mode.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2484</p>
2485
2486<p>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2487Note that both of these calls must be made <em>in isolation</em>. That is to
2488say that no other LLVM API calls may be executing at any time during the
2489execution of <tt>llvm_start_multithreaded()</tt> or <tt>llvm_stop_multithreaded
2490</tt>. It's is the client's responsibility to enforce this isolation.
2491</p>
2492
2493<p>
2494The return value of <tt>llvm_start_multithreaded()</tt> indicates the success or
2495failure of the initialization. Failure typically indicates that your copy of
2496LLVM was built without multithreading support, typically because GCC atomic
2497intrinsics were not found in your system compiler. In this case, the LLVM API
2498will not be safe for concurrent calls. However, it <em>will</em> be safe for
Jeffrey Yasskin01eba392010-01-29 19:10:38 +0000[diff] [blame]2499hosting threaded applications in the JIT, though <a href="#jitthreading">care
2500must be taken</a> to ensure that side exits and the like do not accidentally
2501result in concurrent LLVM API calls.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2502</p>
2503</div>
2504
2505<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2506<h3>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2507 <a name="shutdown">Ending Execution with <tt>llvm_shutdown()</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2508</h3>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2509
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2510<div>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2511<p>
2512When you are done using the LLVM APIs, you should call <tt>llvm_shutdown()</tt>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2513to deallocate memory used for internal structures. This will also invoke
2514<tt>llvm_stop_multithreaded()</tt> if LLVM is operating in multithreaded mode.
2515As such, <tt>llvm_shutdown()</tt> requires the same isolation guarantees as
2516<tt>llvm_stop_multithreaded()</tt>.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2517</p>
2518
2519<p>
2520Note that, if you use scope-based shutdown, you can use the
2521<tt>llvm_shutdown_obj</tt> class, which calls <tt>llvm_shutdown()</tt> in its
2522destructor.
2523</div>
2524
2525<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2526<h3>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2527 <a name="managedstatic">Lazy Initialization with <tt>ManagedStatic</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2528</h3>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2529
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2530<div>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2531<p>
2532<tt>ManagedStatic</tt> is a utility class in LLVM used to implement static
2533initialization of static resources, such as the global type tables. Before the
2534invocation of <tt>llvm_shutdown()</tt>, it implements a simple lazy
2535initialization scheme. Once <tt>llvm_start_multithreaded()</tt> returns,
2536however, it uses double-checked locking to implement thread-safe lazy
2537initialization.
2538</p>
2539
2540<p>
2541Note that, because no other threads are allowed to issue LLVM API calls before
2542<tt>llvm_start_multithreaded()</tt> returns, it is possible to have
2543<tt>ManagedStatic</tt>s of <tt>llvm::sys::Mutex</tt>s.
2544</p>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2545
2546<p>
2547The <tt>llvm_acquire_global_lock()</tt> and <tt>llvm_release_global_lock</tt>
2548APIs provide access to the global lock used to implement the double-checked
2549locking for lazy initialization. These should only be used internally to LLVM,
2550and only if you know what you're doing!
2551</p>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2552</div>
2553
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]2554<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2555<h3>
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]2556 <a name="llvmcontext">Achieving Isolation with <tt>LLVMContext</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2557</h3>
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]2558
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2559<div>
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]2560<p>
2561<tt>LLVMContext</tt> is an opaque class in the LLVM API which clients can use
2562to operate multiple, isolated instances of LLVM concurrently within the same
2563address space. For instance, in a hypothetical compile-server, the compilation
2564of an individual translation unit is conceptually independent from all the
2565others, and it would be desirable to be able to compile incoming translation
2566units concurrently on independent server threads. Fortunately,
2567<tt>LLVMContext</tt> exists to enable just this kind of scenario!
2568</p>
2569
2570<p>
2571Conceptually, <tt>LLVMContext</tt> provides isolation. Every LLVM entity
2572(<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]2573in LLVM's in-memory IR belongs to an <tt>LLVMContext</tt>. Entities in
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]2574different contexts <em>cannot</em> interact with each other: <tt>Module</tt>s in
2575different contexts cannot be linked together, <tt>Function</tt>s cannot be added
2576to <tt>Module</tt>s in different contexts, etc. What this means is that is is
2577safe to compile on multiple threads simultaneously, as long as no two threads
2578operate on entities within the same context.
2579</p>
2580
2581<p>
2582In practice, very few places in the API require the explicit specification of a
2583<tt>LLVMContext</tt>, other than the <tt>Type</tt> creation/lookup APIs.
2584Because every <tt>Type</tt> carries a reference to its owning context, most
2585other entities can determine what context they belong to by looking at their
2586own <tt>Type</tt>. If you are adding new entities to LLVM IR, please try to
2587maintain this interface design.
2588</p>
2589
2590<p>
2591For clients that do <em>not</em> require the benefits of isolation, LLVM
2592provides a convenience API <tt>getGlobalContext()</tt>. This returns a global,
2593lazily initialized <tt>LLVMContext</tt> that may be used in situations where
2594isolation is not a concern.
2595</p>
2596</div>
2597
Jeffrey Yasskin01eba392010-01-29 19:10:38 +0000[diff] [blame]2598<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2599<h3>
Jeffrey Yasskin01eba392010-01-29 19:10:38 +0000[diff] [blame]2600 <a name="jitthreading">Threads and the JIT</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2601</h3>
Jeffrey Yasskin01eba392010-01-29 19:10:38 +0000[diff] [blame]2602
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2603<div>
Jeffrey Yasskin01eba392010-01-29 19:10:38 +0000[diff] [blame]2604<p>
2605LLVM's "eager" JIT compiler is safe to use in threaded programs. Multiple
2606threads can call <tt>ExecutionEngine::getPointerToFunction()</tt> or
2607<tt>ExecutionEngine::runFunction()</tt> concurrently, and multiple threads can
2608run code output by the JIT concurrently. The user must still ensure that only
2609one thread accesses IR in a given <tt>LLVMContext</tt> while another thread
2610might be modifying it. One way to do that is to always hold the JIT lock while
2611accessing IR outside the JIT (the JIT <em>modifies</em> the IR by adding
2612<tt>CallbackVH</tt>s). Another way is to only
2613call <tt>getPointerToFunction()</tt> from the <tt>LLVMContext</tt>'s thread.
2614</p>
2615
2616<p>When the JIT is configured to compile lazily (using
2617<tt>ExecutionEngine::DisableLazyCompilation(false)</tt>), there is currently a
2618<a href="http://llvm.org/bugs/show_bug.cgi?id=5184">race condition</a> in
2619updating call sites after a function is lazily-jitted. It's still possible to
2620use the lazy JIT in a threaded program if you ensure that only one thread at a
2621time can call any particular lazy stub and that the JIT lock guards any IR
2622access, but we suggest using only the eager JIT in threaded programs.
2623</p>
2624</div>
2625
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2626</div>
2627
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2628<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2629<h2>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2630 <a name="advanced">Advanced Topics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2631</h2>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2632<!-- *********************************************************************** -->
2633
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2634<div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2635<p>
2636This section describes some of the advanced or obscure API's that most clients
2637do not need to be aware of. These API's tend manage the inner workings of the
2638LLVM system, and only need to be accessed in unusual circumstances.
2639</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2640
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2641
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2642<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2643<h3>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2644 <a name="SymbolTable">The <tt>ValueSymbolTable</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2645</h3>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2646
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2647<div>
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2648<p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1ValueSymbolTable.html">
2649ValueSymbolTable</a></tt> class provides a symbol table that the <a
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2650href="#Function"><tt>Function</tt></a> and <a href="#Module">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2651<tt>Module</tt></a> classes use for naming value definitions. The symbol table
2652can provide a name for any <a href="#Value"><tt>Value</tt></a>.
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2653</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2654
Reid Spencera6362242007-01-07 00:41:39 +0000[diff] [blame]2655<p>Note that the <tt>SymbolTable</tt> class should not be directly accessed
2656by most clients. It should only be used when iteration over the symbol table
2657names themselves are required, which is very special purpose. Note that not
2658all LLVM
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2659<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]2660an empty name) do not exist in the symbol table.
2661</p>
2662
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2663<p>Symbol tables support iteration over the values in the symbol
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2664table with <tt>begin/end/iterator</tt> and supports querying to see if a
2665specific name is in the symbol table (with <tt>lookup</tt>). The
2666<tt>ValueSymbolTable</tt> class exposes no public mutator methods, instead,
2667simply call <tt>setName</tt> on a value, which will autoinsert it into the
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2668appropriate symbol table.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2669
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2670</div>
2671
2672
2673
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2674<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2675<h3>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2676 <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]2677</h3>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2678
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2679<div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2680<p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1User.html">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2681User</a></tt> class provides a basis for expressing the ownership of <tt>User</tt>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2682towards other <tt><a href="http://llvm.org/doxygen/classllvm_1_1Value.html">
2683Value</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]2684Use</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]2685addition and removal.</p>
2686
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2687<!-- ______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2688<h4>
2689 <a name="Use2User">
2690 Interaction and relationship between <tt>User</tt> and <tt>Use</tt> objects
2691 </a>
2692</h4>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2693
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2694<div>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2695<p>
2696A subclass of <tt>User</tt> can choose between incorporating its <tt>Use</tt> objects
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2697or refer to them out-of-line by means of a pointer. A mixed variant
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2698(some <tt>Use</tt>s inline others hung off) is impractical and breaks the invariant
2699that the <tt>Use</tt> objects belonging to the same <tt>User</tt> form a contiguous array.
2700</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2701
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2702<p>
2703We have 2 different layouts in the <tt>User</tt> (sub)classes:
2704<ul>
2705<li><p>Layout a)
2706The <tt>Use</tt> object(s) are inside (resp. at fixed offset) of the <tt>User</tt>
2707object and there are a fixed number of them.</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2708
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2709<li><p>Layout b)
2710The <tt>Use</tt> object(s) are referenced by a pointer to an
2711array from the <tt>User</tt> object and there may be a variable
2712number of them.</p>
2713</ul>
2714<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2715As of v2.4 each layout still possesses a direct pointer to the
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2716start of the array of <tt>Use</tt>s. Though not mandatory for layout a),
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2717we stick to this redundancy for the sake of simplicity.
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2718The <tt>User</tt> object also stores the number of <tt>Use</tt> objects it
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2719has. (Theoretically this information can also be calculated
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2720given the scheme presented below.)</p>
2721<p>
2722Special forms of allocation operators (<tt>operator new</tt>)
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2723enforce the following memory layouts:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2724
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2725<ul>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2726<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]2727
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2728<pre>
2729...---.---.---.---.-------...
2730 | P | P | P | P | User
2731'''---'---'---'---'-------'''
2732</pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2733
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2734<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]2735<pre>
2736.-------...
2737| User
2738'-------'''
2739 |
2740 v
2741 .---.---.---.---...
2742 | P | P | P | P |
2743 '---'---'---'---'''
2744</pre>
2745</ul>
2746<i>(In the above figures '<tt>P</tt>' stands for the <tt>Use**</tt> that
2747 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]2748
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2749</div>
2750
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2751<!-- ______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2752<h4>
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2753 <a name="Waymarking">The waymarking algorithm</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2754</h4>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2755
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2756<div>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2757<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2758Since the <tt>Use</tt> objects are deprived of the direct (back)pointer to
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2759their <tt>User</tt> objects, there must be a fast and exact method to
2760recover it. This is accomplished by the following scheme:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2761
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2762A 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]2763start of the <tt>User</tt> object:
2764<ul>
2765<li><tt>00</tt> &mdash;&gt; binary digit 0</li>
2766<li><tt>01</tt> &mdash;&gt; binary digit 1</li>
2767<li><tt>10</tt> &mdash;&gt; stop and calculate (<tt>s</tt>)</li>
2768<li><tt>11</tt> &mdash;&gt; full stop (<tt>S</tt>)</li>
2769</ul>
2770<p>
2771Given a <tt>Use*</tt>, all we have to do is to walk till we get
2772a stop and we either have a <tt>User</tt> immediately behind or
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2773we have to walk to the next stop picking up digits
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2774and calculating the offset:</p>
2775<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2776.---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.----------------
2777| 1 | s | 1 | 0 | 1 | 0 | s | 1 | 1 | 0 | s | 1 | 1 | s | 1 | S | User (or User*)
2778'---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'----------------
2779 |+15 |+10 |+6 |+3 |+1
2780 | | | | |__>
2781 | | | |__________>
2782 | | |______________________>
2783 | |______________________________________>
2784 |__________________________________________________________>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2785</pre>
2786<p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2787Only the significant number of bits need to be stored between the
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2788stops, so that the <i>worst case is 20 memory accesses</i> when there are
27891000 <tt>Use</tt> objects associated with a <tt>User</tt>.</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2790
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2791</div>
2792
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2793<!-- ______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2794<h4>
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2795 <a name="ReferenceImpl">Reference implementation</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2796</h4>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2797
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2798<div>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2799<p>
2800The following literate Haskell fragment demonstrates the concept:</p>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2801
2802<div class="doc_code">
2803<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2804> import Test.QuickCheck
2805>
2806> digits :: Int -> [Char] -> [Char]
2807> digits 0 acc = '0' : acc
2808> digits 1 acc = '1' : acc
2809> digits n acc = digits (n `div` 2) $ digits (n `mod` 2) acc
2810>
2811> dist :: Int -> [Char] -> [Char]
2812> dist 0 [] = ['S']
2813> dist 0 acc = acc
2814> dist 1 acc = let r = dist 0 acc in 's' : digits (length r) r
2815> dist n acc = dist (n - 1) $ dist 1 acc
2816>
2817> takeLast n ss = reverse $ take n $ reverse ss
2818>
2819> test = takeLast 40 $ dist 20 []
2820>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2821</pre>
2822</div>
2823<p>
2824Printing &lt;test&gt; gives: <tt>"1s100000s11010s10100s1111s1010s110s11s1S"</tt></p>
2825<p>
2826The reverse algorithm computes the length of the string just by examining
2827a certain prefix:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2828
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2829<div class="doc_code">
2830<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2831> pref :: [Char] -> Int
2832> pref "S" = 1
2833> pref ('s':'1':rest) = decode 2 1 rest
2834> pref (_:rest) = 1 + pref rest
2835>
2836> decode walk acc ('0':rest) = decode (walk + 1) (acc * 2) rest
2837> decode walk acc ('1':rest) = decode (walk + 1) (acc * 2 + 1) rest
2838> decode walk acc _ = walk + acc
2839>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2840</pre>
2841</div>
2842<p>
2843Now, as expected, printing &lt;pref test&gt; gives <tt>40</tt>.</p>
2844<p>
2845We can <i>quickCheck</i> this with following property:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2846
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2847<div class="doc_code">
2848<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2849> testcase = dist 2000 []
2850> testcaseLength = length testcase
2851>
2852> identityProp n = n > 0 && n <= testcaseLength ==> length arr == pref arr
2853> where arr = takeLast n testcase
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2854>
2855</pre>
2856</div>
2857<p>
2858As expected &lt;quickCheck identityProp&gt; gives:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2859
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2860<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2861*Main> quickCheck identityProp
2862OK, passed 100 tests.
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2863</pre>
2864<p>
2865Let's be a bit more exhaustive:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2866
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2867<div class="doc_code">
2868<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2869>
2870> deepCheck p = check (defaultConfig { configMaxTest = 500 }) p
2871>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2872</pre>
2873</div>
2874<p>
2875And here is the result of &lt;deepCheck identityProp&gt;:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2876
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2877<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2878*Main> deepCheck identityProp
2879OK, passed 500 tests.
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2880</pre>
2881
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2882</div>
2883
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2884<!-- ______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2885<h4>
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2886 <a name="Tagging">Tagging considerations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2887</h4>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2888
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2889<div>
2890
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2891<p>
2892To maintain the invariant that the 2 LSBits of each <tt>Use**</tt> in <tt>Use</tt>
2893never change after being set up, setters of <tt>Use::Prev</tt> must re-tag the
2894new <tt>Use**</tt> on every modification. Accordingly getters must strip the
2895tag bits.</p>
2896<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2897For layout b) instead of the <tt>User</tt> we find a pointer (<tt>User*</tt> with LSBit set).
2898Following this pointer brings us to the <tt>User</tt>. A portable trick ensures
2899that the first bytes of <tt>User</tt> (if interpreted as a pointer) never has
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2900the LSBit set. (Portability is relying on the fact that all known compilers place the
2901<tt>vptr</tt> in the first word of the instances.)</p>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2902
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2903</div>
2904
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2905</div>
2906
2907</div>
2908
2909<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2910<h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2911 <a name="coreclasses">The Core LLVM Class Hierarchy Reference </a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2912</h2>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2913<!-- *********************************************************************** -->
2914
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2915<div>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2916<p><tt>#include "<a href="/doxygen/Type_8h-source.html">llvm/Type.h</a>"</tt>
2917<br>doxygen info: <a href="/doxygen/classllvm_1_1Type.html">Type Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2918
2919<p>The Core LLVM classes are the primary means of representing the program
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2920being inspected or transformed. The core LLVM classes are defined in
2921header files in the <tt>include/llvm/</tt> directory, and implemented in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2922the <tt>lib/VMCore</tt> directory.</p>
2923
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2924<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2925<h3>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2926 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2927</h3>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2928
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2929<div>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2930
2931 <p><tt>Type</tt> is a superclass of all type classes. Every <tt>Value</tt> has
2932 a <tt>Type</tt>. <tt>Type</tt> cannot be instantiated directly but only
2933 through its subclasses. Certain primitive types (<tt>VoidType</tt>,
2934 <tt>LabelType</tt>, <tt>FloatType</tt> and <tt>DoubleType</tt>) have hidden
2935 subclasses. They are hidden because they offer no useful functionality beyond
2936 what the <tt>Type</tt> class offers except to distinguish themselves from
2937 other subclasses of <tt>Type</tt>.</p>
2938 <p>All other types are subclasses of <tt>DerivedType</tt>. Types can be
2939 named, but this is not a requirement. There exists exactly
2940 one instance of a given shape at any one time. This allows type equality to
2941 be performed with address equality of the Type Instance. That is, given two
2942 <tt>Type*</tt> values, the types are identical if the pointers are identical.
2943 </p>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2944
2945<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2946<h4>
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2947 <a name="m_Type">Important Public Methods</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2948</h4>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2949
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2950<div>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2951
2952<ul>
Duncan Sandsb0bc6c32010-02-15 16:12:20 +0000[diff] [blame]2953 <li><tt>bool isIntegerTy() const</tt>: Returns true for any integer type.</li>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2954
Duncan Sandsb0bc6c32010-02-15 16:12:20 +0000[diff] [blame]2955 <li><tt>bool isFloatingPointTy()</tt>: Return true if this is one of the five
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2956 floating point types.</li>
2957
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2958 <li><tt>bool isSized()</tt>: Return true if the type has known size. Things
2959 that don't have a size are abstract types, labels and void.</li>
2960
2961</ul>
2962</div>
2963
2964<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2965<h4>
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2966 <a name="derivedtypes">Important Derived Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2967</h4>
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2968<div>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2969<dl>
2970 <dt><tt>IntegerType</tt></dt>
2971 <dd>Subclass of DerivedType that represents integer types of any bit width.
2972 Any bit width between <tt>IntegerType::MIN_INT_BITS</tt> (1) and
2973 <tt>IntegerType::MAX_INT_BITS</tt> (~8 million) can be represented.
2974 <ul>
2975 <li><tt>static const IntegerType* get(unsigned NumBits)</tt>: get an integer
2976 type of a specific bit width.</li>
2977 <li><tt>unsigned getBitWidth() const</tt>: Get the bit width of an integer
2978 type.</li>
2979 </ul>
2980 </dd>
2981 <dt><tt>SequentialType</tt></dt>
2982 <dd>This is subclassed by ArrayType and PointerType
2983 <ul>
2984 <li><tt>const Type * getElementType() const</tt>: Returns the type of each
2985 of the elements in the sequential type. </li>
2986 </ul>
2987 </dd>
2988 <dt><tt>ArrayType</tt></dt>
2989 <dd>This is a subclass of SequentialType and defines the interface for array
2990 types.
2991 <ul>
2992 <li><tt>unsigned getNumElements() const</tt>: Returns the number of
2993 elements in the array. </li>
2994 </ul>
2995 </dd>
2996 <dt><tt>PointerType</tt></dt>
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]2997 <dd>Subclass of SequentialType for pointer types.</dd>
Reid Spencer9d6565a2007-02-15 02:26:10 +0000[diff] [blame]2998 <dt><tt>VectorType</tt></dt>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2999 <dd>Subclass of SequentialType for vector types. A
3000 vector type is similar to an ArrayType but is distinguished because it is
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3001 a first class type whereas ArrayType is not. Vector types are used for
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3002 vector operations and are usually small vectors of of an integer or floating
3003 point type.</dd>
3004 <dt><tt>StructType</tt></dt>
3005 <dd>Subclass of DerivedTypes for struct types.</dd>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3006 <dt><tt><a name="FunctionType">FunctionType</a></tt></dt>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3007 <dd>Subclass of DerivedTypes for function types.
3008 <ul>
Dan Gohman4bb31bf2010-03-30 20:04:57 +0000[diff] [blame]3009 <li><tt>bool isVarArg() const</tt>: Returns true if it's a vararg
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3010 function</li>
3011 <li><tt> const Type * getReturnType() const</tt>: Returns the
3012 return type of the function.</li>
3013 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns
3014 the type of the ith parameter.</li>
3015 <li><tt> const unsigned getNumParams() const</tt>: Returns the
3016 number of formal parameters.</li>
3017 </ul>
3018 </dd>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3019</dl>
3020</div>
3021
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3022</div>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3023
3024<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3025<h3>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3026 <a name="Module">The <tt>Module</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3027</h3>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3028
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3029<div>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3030
3031<p><tt>#include "<a
3032href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt><br> doxygen info:
3033<a href="/doxygen/classllvm_1_1Module.html">Module Class</a></p>
3034
3035<p>The <tt>Module</tt> class represents the top level structure present in LLVM
3036programs. An LLVM module is effectively either a translation unit of the
3037original program or a combination of several translation units merged by the
3038linker. The <tt>Module</tt> class keeps track of a list of <a
3039href="#Function"><tt>Function</tt></a>s, a list of <a
3040href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
3041href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
3042helpful member functions that try to make common operations easy.</p>
3043
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3044<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3045<h4>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3046 <a name="m_Module">Important Public Members of the <tt>Module</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3047</h4>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3048
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3049<div>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3050
3051<ul>
3052 <li><tt>Module::Module(std::string name = "")</tt></li>
3053</ul>
3054
3055<p>Constructing a <a href="#Module">Module</a> is easy. You can optionally
3056provide a name for it (probably based on the name of the translation unit).</p>
3057
3058<ul>
3059 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
3060 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
3061
3062 <tt>begin()</tt>, <tt>end()</tt>
3063 <tt>size()</tt>, <tt>empty()</tt>
3064
3065 <p>These are forwarding methods that make it easy to access the contents of
3066 a <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
3067 list.</p></li>
3068
3069 <li><tt>Module::FunctionListType &amp;getFunctionList()</tt>
3070
3071 <p> Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
3072 necessary to use when you need to update the list or perform a complex
3073 action that doesn't have a forwarding method.</p>
3074
3075 <p><!-- Global Variable --></p></li>
3076</ul>
3077
3078<hr>
3079
3080<ul>
3081 <li><tt>Module::global_iterator</tt> - Typedef for global variable list iterator<br>
3082
3083 <tt>Module::const_global_iterator</tt> - Typedef for const_iterator.<br>
3084
3085 <tt>global_begin()</tt>, <tt>global_end()</tt>
3086 <tt>global_size()</tt>, <tt>global_empty()</tt>
3087
3088 <p> These are forwarding methods that make it easy to access the contents of
3089 a <tt>Module</tt> object's <a
3090 href="#GlobalVariable"><tt>GlobalVariable</tt></a> list.</p></li>
3091
3092 <li><tt>Module::GlobalListType &amp;getGlobalList()</tt>
3093
3094 <p>Returns the list of <a
3095 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s. This is necessary to
3096 use when you need to update the list or perform a complex action that
3097 doesn't have a forwarding method.</p>
3098
3099 <p><!-- Symbol table stuff --> </p></li>
3100</ul>
3101
3102<hr>
3103
3104<ul>
3105 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
3106
3107 <p>Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
3108 for this <tt>Module</tt>.</p>
3109
3110 <p><!-- Convenience methods --></p></li>
3111</ul>
3112
3113<hr>
3114
3115<ul>
3116 <li><tt><a href="#Function">Function</a> *getFunction(const std::string
3117 &amp;Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt>
3118
3119 <p>Look up the specified function in the <tt>Module</tt> <a
3120 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
3121 <tt>null</tt>.</p></li>
3122
3123 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const
3124 std::string &amp;Name, const <a href="#FunctionType">FunctionType</a> *T)</tt>
3125
3126 <p>Look up the specified function in the <tt>Module</tt> <a
3127 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
3128 external declaration for the function and return it.</p></li>
3129
3130 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt>
3131
3132 <p>If there is at least one entry in the <a
3133 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
3134 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
3135 string.</p></li>
3136
3137 <li><tt>bool addTypeName(const std::string &amp;Name, const <a
3138 href="#Type">Type</a> *Ty)</tt>
3139
3140 <p>Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
3141 mapping <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this
3142 name, true is returned and the <a
3143 href="#SymbolTable"><tt>SymbolTable</tt></a> is not modified.</p></li>
3144</ul>
3145
3146</div>
3147
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3148</div>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3149
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3150<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3151<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3152 <a name="Value">The <tt>Value</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3153</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3154
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3155<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3156
3157<p><tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt>
3158<br>
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]3159doxygen info: <a href="/doxygen/classllvm_1_1Value.html">Value Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3160
3161<p>The <tt>Value</tt> class is the most important class in the LLVM Source
3162base. It represents a typed value that may be used (among other things) as an
3163operand to an instruction. There are many different types of <tt>Value</tt>s,
3164such as <a href="#Constant"><tt>Constant</tt></a>s,<a
3165href="#Argument"><tt>Argument</tt></a>s. Even <a
3166href="#Instruction"><tt>Instruction</tt></a>s and <a
3167href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.</p>
3168
3169<p>A particular <tt>Value</tt> may be used many times in the LLVM representation
3170for a program. For example, an incoming argument to a function (represented
3171with an instance of the <a href="#Argument">Argument</a> class) is "used" by
3172every instruction in the function that references the argument. To keep track
3173of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
3174href="#User"><tt>User</tt></a>s that is using it (the <a
3175href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
3176graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
3177def-use information in the program, and is accessible through the <tt>use_</tt>*
3178methods, shown below.</p>
3179
3180<p>Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed,
3181and this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
3182method. In addition, all LLVM values can be named. The "name" of the
3183<tt>Value</tt> is a symbolic string printed in the LLVM code:</p>
3184
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3185<div class="doc_code">
3186<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]3187%<b>foo</b> = add i32 1, 2
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3188</pre>
3189</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3190
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3191<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]3192that the name of any value may be missing (an empty string), so names should
3193<b>ONLY</b> be used for debugging (making the source code easier to read,
3194debugging printouts), they should not be used to keep track of values or map
3195between them. For this purpose, use a <tt>std::map</tt> of pointers to the
3196<tt>Value</tt> itself instead.</p>
3197
3198<p>One important aspect of LLVM is that there is no distinction between an SSA
3199variable and the operation that produces it. Because of this, any reference to
3200the value produced by an instruction (or the value available as an incoming
Chris Lattnerd5fc4fc2004-03-18 14:58:55 +0000[diff] [blame]3201argument, for example) is represented as a direct pointer to the instance of
3202the class that
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3203represents this value. Although this may take some getting used to, it
3204simplifies the representation and makes it easier to manipulate.</p>
3205
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3206<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3207<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3208 <a name="m_Value">Important Public Members of the <tt>Value</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3209</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3210
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3211<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3212
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3213<ul>
3214 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the
3215use-list<br>
Gabor Greifbbbf9a22010-03-26 19:59:25 +0000[diff] [blame]3216 <tt>Value::const_use_iterator</tt> - Typedef for const_iterator over
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3217the use-list<br>
3218 <tt>unsigned use_size()</tt> - Returns the number of users of the
3219value.<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3220 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3221 <tt>use_iterator use_begin()</tt> - Get an iterator to the start of
3222the use-list.<br>
3223 <tt>use_iterator use_end()</tt> - Get an iterator to the end of the
3224use-list.<br>
3225 <tt><a href="#User">User</a> *use_back()</tt> - Returns the last
3226element in the list.
3227 <p> These methods are the interface to access the def-use
3228information in LLVM. As with all other iterators in LLVM, the naming
3229conventions follow the conventions defined by the <a href="#stl">STL</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3230 </li>
3231 <li><tt><a href="#Type">Type</a> *getType() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3232 <p>This method returns the Type of the Value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3233 </li>
3234 <li><tt>bool hasName() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3235 <tt>std::string getName() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3236 <tt>void setName(const std::string &amp;Name)</tt>
3237 <p> This family of methods is used to access and assign a name to a <tt>Value</tt>,
3238be aware of the <a href="#nameWarning">precaution above</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3239 </li>
3240 <li><tt>void replaceAllUsesWith(Value *V)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3241
3242 <p>This method traverses the use list of a <tt>Value</tt> changing all <a
3243 href="#User"><tt>User</tt>s</a> of the current value to refer to
3244 "<tt>V</tt>" instead. For example, if you detect that an instruction always
3245 produces a constant value (for example through constant folding), you can
3246 replace all uses of the instruction with the constant like this:</p>
3247
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3248<div class="doc_code">
3249<pre>
3250Inst-&gt;replaceAllUsesWith(ConstVal);
3251</pre>
3252</div>
3253
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3254</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3255
3256</div>
3257
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3258</div>
3259
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3260<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3261<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3262 <a name="User">The <tt>User</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3263</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3264
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3265<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3266
3267<p>
3268<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]3269doxygen info: <a href="/doxygen/classllvm_1_1User.html">User Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3270Superclass: <a href="#Value"><tt>Value</tt></a></p>
3271
3272<p>The <tt>User</tt> class is the common base class of all LLVM nodes that may
3273refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
3274that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
3275referring to. The <tt>User</tt> class itself is a subclass of
3276<tt>Value</tt>.</p>
3277
3278<p>The operands of a <tt>User</tt> point directly to the LLVM <a
3279href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
3280Single Assignment (SSA) form, there can only be one definition referred to,
3281allowing this direct connection. This connection provides the use-def
3282information in LLVM.</p>
3283
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3284<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3285<h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3286 <a name="m_User">Important Public Members of the <tt>User</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3287</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3288
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3289<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3290
3291<p>The <tt>User</tt> class exposes the operand list in two ways: through
3292an index access interface and through an iterator based interface.</p>
3293
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3294<ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3295 <li><tt>Value *getOperand(unsigned i)</tt><br>
3296 <tt>unsigned getNumOperands()</tt>
3297 <p> These two methods expose the operands of the <tt>User</tt> in a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3298convenient form for direct access.</p></li>
3299
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3300 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand
3301list<br>
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]3302 <tt>op_iterator op_begin()</tt> - Get an iterator to the start of
3303the operand list.<br>
3304 <tt>op_iterator op_end()</tt> - Get an iterator to the end of the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3305operand list.
3306 <p> Together, these methods make up the iterator based interface to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3307the operands of a <tt>User</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3308</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3309
3310</div>
3311
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3312</div>
3313
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3314<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3315<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3316 <a name="Instruction">The <tt>Instruction</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3317</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3318
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3319<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3320
3321<p><tt>#include "</tt><tt><a
3322href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt><br>
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3323doxygen info: <a href="/doxygen/classllvm_1_1Instruction.html">Instruction Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3324Superclasses: <a href="#User"><tt>User</tt></a>, <a
3325href="#Value"><tt>Value</tt></a></p>
3326
3327<p>The <tt>Instruction</tt> class is the common base class for all LLVM
3328instructions. It provides only a few methods, but is a very commonly used
3329class. The primary data tracked by the <tt>Instruction</tt> class itself is the
3330opcode (instruction type) and the parent <a
3331href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
3332into. To represent a specific type of instruction, one of many subclasses of
3333<tt>Instruction</tt> are used.</p>
3334
3335<p> Because the <tt>Instruction</tt> class subclasses the <a
3336href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
3337way as for other <a href="#User"><tt>User</tt></a>s (with the
3338<tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
3339<tt>op_begin()</tt>/<tt>op_end()</tt> methods).</p> <p> An important file for
3340the <tt>Instruction</tt> class is the <tt>llvm/Instruction.def</tt> file. This
3341file contains some meta-data about the various different types of instructions
3342in LLVM. It describes the enum values that are used as opcodes (for example
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3343<tt>Instruction::Add</tt> and <tt>Instruction::ICmp</tt>), as well as the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3344concrete sub-classes of <tt>Instruction</tt> that implement the instruction (for
3345example <tt><a href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3346href="#CmpInst">CmpInst</a></tt>). Unfortunately, the use of macros in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3347this file confuses doxygen, so these enum values don't show up correctly in the
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3348<a href="/doxygen/classllvm_1_1Instruction.html">doxygen output</a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3349
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3350<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3351<h4>
3352 <a name="s_Instruction">
3353 Important Subclasses of the <tt>Instruction</tt> class
3354 </a>
3355</h4>
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3356<div>
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3357 <ul>
3358 <li><tt><a name="BinaryOperator">BinaryOperator</a></tt>
3359 <p>This subclasses represents all two operand instructions whose operands
3360 must be the same type, except for the comparison instructions.</p></li>
3361 <li><tt><a name="CastInst">CastInst</a></tt>
3362 <p>This subclass is the parent of the 12 casting instructions. It provides
3363 common operations on cast instructions.</p>
3364 <li><tt><a name="CmpInst">CmpInst</a></tt>
3365 <p>This subclass respresents the two comparison instructions,
3366 <a href="LangRef.html#i_icmp">ICmpInst</a> (integer opreands), and
3367 <a href="LangRef.html#i_fcmp">FCmpInst</a> (floating point operands).</p>
3368 <li><tt><a name="TerminatorInst">TerminatorInst</a></tt>
3369 <p>This subclass is the parent of all terminator instructions (those which
3370 can terminate a block).</p>
3371 </ul>
3372 </div>
3373
3374<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3375<h4>
3376 <a name="m_Instruction">
3377 Important Public Members of the <tt>Instruction</tt> class
3378 </a>
3379</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3380
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3381<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3382
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3383<ul>
3384 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3385 <p>Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that
3386this <tt>Instruction</tt> is embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3387 <li><tt>bool mayWriteToMemory()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3388 <p>Returns true if the instruction writes to memory, i.e. it is a
3389 <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]3390 <li><tt>unsigned getOpcode()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3391 <p>Returns the opcode for the <tt>Instruction</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3392 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3393 <p>Returns another instance of the specified instruction, identical
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3394in all ways to the original except that the instruction has no parent
3395(ie it's not embedded into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>),
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3396and it has no name</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3397</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3398
3399</div>
3400
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3401</div>
3402
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3403<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3404<h3>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3405 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3406</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3407
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3408<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3409
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3410<p>Constant represents a base class for different types of constants. It
3411is subclassed by ConstantInt, ConstantArray, etc. for representing
3412the various types of Constants. <a href="#GlobalValue">GlobalValue</a> is also
3413a subclass, which represents the address of a global variable or function.
3414</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3415
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3416<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3417<h4>Important Subclasses of Constant</h4>
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3418<div>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3419<ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3420 <li>ConstantInt : This subclass of Constant represents an integer constant of
3421 any width.
3422 <ul>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]3423 <li><tt>const APInt&amp; getValue() const</tt>: Returns the underlying
3424 value of this constant, an APInt value.</li>
3425 <li><tt>int64_t getSExtValue() const</tt>: Converts the underlying APInt
3426 value to an int64_t via sign extension. If the value (not the bit width)
3427 of the APInt is too large to fit in an int64_t, an assertion will result.
3428 For this reason, use of this method is discouraged.</li>
3429 <li><tt>uint64_t getZExtValue() const</tt>: Converts the underlying APInt
3430 value to a uint64_t via zero extension. IF the value (not the bit width)
3431 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]3432 For this reason, use of this method is discouraged.</li>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]3433 <li><tt>static ConstantInt* get(const APInt&amp; Val)</tt>: Returns the
3434 ConstantInt object that represents the value provided by <tt>Val</tt>.
3435 The type is implied as the IntegerType that corresponds to the bit width
3436 of <tt>Val</tt>.</li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3437 <li><tt>static ConstantInt* get(const Type *Ty, uint64_t Val)</tt>:
3438 Returns the ConstantInt object that represents the value provided by
3439 <tt>Val</tt> for integer type <tt>Ty</tt>.</li>
3440 </ul>
3441 </li>
3442 <li>ConstantFP : This class represents a floating point constant.
3443 <ul>
3444 <li><tt>double getValue() const</tt>: Returns the underlying value of
3445 this constant. </li>
3446 </ul>
3447 </li>
3448 <li>ConstantArray : This represents a constant array.
3449 <ul>
3450 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
3451 a vector of component constants that makeup this array. </li>
3452 </ul>
3453 </li>
3454 <li>ConstantStruct : This represents a constant struct.
3455 <ul>
3456 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
3457 a vector of component constants that makeup this array. </li>
3458 </ul>
3459 </li>
3460 <li>GlobalValue : This represents either a global variable or a function. In
3461 either case, the value is a constant fixed address (after linking).
3462 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3463</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3464</div>
3465
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3466</div>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3467
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3468<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3469<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3470 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3471</h3>
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><tt>#include "<a
3476href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt><br>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]3477doxygen info: <a href="/doxygen/classllvm_1_1GlobalValue.html">GlobalValue
3478Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3479Superclasses: <a href="#Constant"><tt>Constant</tt></a>,
3480<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]3481
3482<p>Global values (<a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
3483href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
3484visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
3485Because they are visible at global scope, they are also subject to linking with
3486other globals defined in different translation units. To control the linking
3487process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
3488<tt>GlobalValue</tt>s know whether they have internal or external linkage, as
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]3489defined by the <tt>LinkageTypes</tt> enumeration.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3490
3491<p>If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
3492<tt>static</tt> in C), it is not visible to code outside the current translation
3493unit, and does not participate in linking. If it has external linkage, it is
3494visible to external code, and does participate in linking. In addition to
3495linkage information, <tt>GlobalValue</tt>s keep track of which <a
3496href="#Module"><tt>Module</tt></a> they are currently part of.</p>
3497
3498<p>Because <tt>GlobalValue</tt>s are memory objects, they are always referred to
3499by their <b>address</b>. As such, the <a href="#Type"><tt>Type</tt></a> of a
3500global is always a pointer to its contents. It is important to remember this
3501when using the <tt>GetElementPtrInst</tt> instruction because this pointer must
3502be dereferenced first. For example, if you have a <tt>GlobalVariable</tt> (a
3503subclass 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]3504i32]</tt>, then the <tt>GlobalVariable</tt> is a pointer to that array. Although
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3505the address of the first element of this array and the value of the
3506<tt>GlobalVariable</tt> are the same, they have different types. The
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]3507<tt>GlobalVariable</tt>'s type is <tt>[24 x i32]</tt>. The first element's type
3508is <tt>i32.</tt> Because of this, accessing a global value requires you to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3509dereference the pointer with <tt>GetElementPtrInst</tt> first, then its elements
3510can be accessed. This is explained in the <a href="LangRef.html#globalvars">LLVM
3511Language Reference Manual</a>.</p>
3512
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3513<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3514<h4>
3515 <a name="m_GlobalValue">
3516 Important Public Members of the <tt>GlobalValue</tt> class
3517 </a>
3518</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3519
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3520<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3521
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3522<ul>
3523 <li><tt>bool hasInternalLinkage() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3524 <tt>bool hasExternalLinkage() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3525 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt>
3526 <p> These methods manipulate the linkage characteristics of the <tt>GlobalValue</tt>.</p>
3527 <p> </p>
3528 </li>
3529 <li><tt><a href="#Module">Module</a> *getParent()</tt>
3530 <p> This returns the <a href="#Module"><tt>Module</tt></a> that the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3531GlobalValue is currently embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3532</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3533
3534</div>
3535
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3536</div>
3537
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3538<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3539<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3540 <a name="Function">The <tt>Function</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3541</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3542
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3543<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3544
3545<p><tt>#include "<a
3546href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt><br> doxygen
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3547info: <a href="/doxygen/classllvm_1_1Function.html">Function Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3548Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
3549<a href="#Constant"><tt>Constant</tt></a>,
3550<a href="#User"><tt>User</tt></a>,
3551<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3552
3553<p>The <tt>Function</tt> class represents a single procedure in LLVM. It is
Torok Edwin87469292009-10-12 13:37:29 +0000[diff] [blame]3554actually one of the more complex classes in the LLVM hierarchy because it must
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3555keep track of a large amount of data. The <tt>Function</tt> class keeps track
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3556of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal
3557<a href="#Argument"><tt>Argument</tt></a>s, and a
3558<a href="#SymbolTable"><tt>SymbolTable</tt></a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3559
3560<p>The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most
3561commonly used part of <tt>Function</tt> objects. The list imposes an implicit
3562ordering of the blocks in the function, which indicate how the code will be
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3563laid out by the backend. Additionally, the first <a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3564href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
3565<tt>Function</tt>. It is not legal in LLVM to explicitly branch to this initial
3566block. There are no implicit exit nodes, and in fact there may be multiple exit
3567nodes from a single <tt>Function</tt>. If the <a
3568href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
3569the <tt>Function</tt> is actually a function declaration: the actual body of the
3570function hasn't been linked in yet.</p>
3571
3572<p>In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
3573<tt>Function</tt> class also keeps track of the list of formal <a
3574href="#Argument"><tt>Argument</tt></a>s that the function receives. This
3575container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
3576nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
3577the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.</p>
3578
3579<p>The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used
3580LLVM feature that is only used when you have to look up a value by name. Aside
3581from that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used
3582internally to make sure that there are not conflicts between the names of <a
3583href="#Instruction"><tt>Instruction</tt></a>s, <a
3584href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
3585href="#Argument"><tt>Argument</tt></a>s in the function body.</p>
3586
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]3587<p>Note that <tt>Function</tt> is a <a href="#GlobalValue">GlobalValue</a>
3588and therefore also a <a href="#Constant">Constant</a>. The value of the function
3589is its address (after linking) which is guaranteed to be constant.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3590
3591<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3592<h4>
3593 <a name="m_Function">
3594 Important Public Members of the <tt>Function</tt> class
3595 </a>
3596</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3597
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3598<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3599
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3600<ul>
3601 <li><tt>Function(const </tt><tt><a href="#FunctionType">FunctionType</a>
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]3602 *Ty, LinkageTypes Linkage, const std::string &amp;N = "", Module* Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3603
3604 <p>Constructor used when you need to create new <tt>Function</tt>s to add
3605 the the program. The constructor must specify the type of the function to
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]3606 create and what type of linkage the function should have. The <a
3607 href="#FunctionType"><tt>FunctionType</tt></a> argument
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3608 specifies the formal arguments and return value for the function. The same
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3609 <a href="#FunctionType"><tt>FunctionType</tt></a> value can be used to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3610 create multiple functions. The <tt>Parent</tt> argument specifies the Module
3611 in which the function is defined. If this argument is provided, the function
3612 will automatically be inserted into that module's list of
3613 functions.</p></li>
3614
Chris Lattner62810e32008-11-25 18:34:50 +0000[diff] [blame]3615 <li><tt>bool isDeclaration()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3616
3617 <p>Return whether or not the <tt>Function</tt> has a body defined. If the
3618 function is "external", it does not have a body, and thus must be resolved
3619 by linking with a function defined in a different translation unit.</p></li>
3620
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3621 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3622 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3623
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]3624 <tt>begin()</tt>, <tt>end()</tt>
3625 <tt>size()</tt>, <tt>empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3626
3627 <p>These are forwarding methods that make it easy to access the contents of
3628 a <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
3629 list.</p></li>
3630
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3631 <li><tt>Function::BasicBlockListType &amp;getBasicBlockList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3632
3633 <p>Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This
3634 is necessary to use when you need to update the list or perform a complex
3635 action that doesn't have a forwarding method.</p></li>
3636
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3637 <li><tt>Function::arg_iterator</tt> - Typedef for the argument list
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3638iterator<br>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3639 <tt>Function::const_arg_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3640
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]3641 <tt>arg_begin()</tt>, <tt>arg_end()</tt>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3642 <tt>arg_size()</tt>, <tt>arg_empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3643
3644 <p>These are forwarding methods that make it easy to access the contents of
3645 a <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a>
3646 list.</p></li>
3647
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3648 <li><tt>Function::ArgumentListType &amp;getArgumentList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3649
3650 <p>Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
3651 necessary to use when you need to update the list or perform a complex
3652 action that doesn't have a forwarding method.</p></li>
3653
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3654 <li><tt><a href="#BasicBlock">BasicBlock</a> &amp;getEntryBlock()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3655
3656 <p>Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
3657 function. Because the entry block for the function is always the first
3658 block, this returns the first block of the <tt>Function</tt>.</p></li>
3659
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3660 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
3661 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3662
3663 <p>This traverses the <a href="#Type"><tt>Type</tt></a> of the
3664 <tt>Function</tt> and returns the return type of the function, or the <a
3665 href="#FunctionType"><tt>FunctionType</tt></a> of the actual
3666 function.</p></li>
3667
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3668 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3669
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3670 <p> Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3671 for this <tt>Function</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3672</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3673
3674</div>
3675
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3676</div>
3677
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3678<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3679<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3680 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3681</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3682
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3683<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3684
3685<p><tt>#include "<a
3686href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt>
3687<br>
Tanya Lattnera3da7772004-06-22 08:02:25 +0000[diff] [blame]3688doxygen info: <a href="/doxygen/classllvm_1_1GlobalVariable.html">GlobalVariable
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3689 Class</a><br>
3690Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
3691<a href="#Constant"><tt>Constant</tt></a>,
3692<a href="#User"><tt>User</tt></a>,
3693<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3694
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3695<p>Global variables are represented with the (surprise surprise)
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3696<tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are also
3697subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such are
3698always referenced by their address (global values must live in memory, so their
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3699"name" refers to their constant address). See
3700<a href="#GlobalValue"><tt>GlobalValue</tt></a> for more on this. Global
3701variables may have an initial value (which must be a
3702<a href="#Constant"><tt>Constant</tt></a>), and if they have an initializer,
3703they may be marked as "constant" themselves (indicating that their contents
3704never change at runtime).</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3705
3706<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3707<h4>
3708 <a name="m_GlobalVariable">
3709 Important Public Members of the <tt>GlobalVariable</tt> class
3710 </a>
3711</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3712
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3713<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3714
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3715<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3716 <li><tt>GlobalVariable(const </tt><tt><a href="#Type">Type</a> *Ty, bool
3717 isConstant, LinkageTypes&amp; Linkage, <a href="#Constant">Constant</a>
3718 *Initializer = 0, const std::string &amp;Name = "", Module* Parent = 0)</tt>
3719
3720 <p>Create a new global variable of the specified type. If
3721 <tt>isConstant</tt> is true then the global variable will be marked as
3722 unchanging for the program. The Linkage parameter specifies the type of
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]3723 linkage (internal, external, weak, linkonce, appending) for the variable.
3724 If the linkage is InternalLinkage, WeakAnyLinkage, WeakODRLinkage,
3725 LinkOnceAnyLinkage or LinkOnceODRLinkage,&nbsp; then the resultant
3726 global variable will have internal linkage. AppendingLinkage concatenates
3727 together all instances (in different translation units) of the variable
3728 into a single variable but is only applicable to arrays. &nbsp;See
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3729 the <a href="LangRef.html#modulestructure">LLVM Language Reference</a> for
3730 further details on linkage types. Optionally an initializer, a name, and the
3731 module to put the variable into may be specified for the global variable as
3732 well.</p></li>
3733
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3734 <li><tt>bool isConstant() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3735
3736 <p>Returns true if this is a global variable that is known not to
3737 be modified at runtime.</p></li>
3738
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3739 <li><tt>bool hasInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3740
3741 <p>Returns true if this <tt>GlobalVariable</tt> has an intializer.</p></li>
3742
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3743 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3744
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3745 <p>Returns the initial value for a <tt>GlobalVariable</tt>. It is not legal
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3746 to call this method if there is no initializer.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3747</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3748
3749</div>
3750
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3751</div>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3752
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3753<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3754<h3>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3755 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3756</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3757
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3758<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3759
3760<p><tt>#include "<a
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3761href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt><br>
Stefanus Du Toit24e04112009-06-17 21:12:26 +0000[diff] [blame]3762doxygen info: <a href="/doxygen/classllvm_1_1BasicBlock.html">BasicBlock
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3763Class</a><br>
3764Superclass: <a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3765
Nick Lewyckyccd279d2011-02-17 02:19:22 +0000[diff] [blame]3766<p>This class represents a single entry single exit section of the code,
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3767commonly known as a basic block by the compiler community. The
3768<tt>BasicBlock</tt> class maintains a list of <a
3769href="#Instruction"><tt>Instruction</tt></a>s, which form the body of the block.
3770Matching the language definition, the last element of this list of instructions
3771is always a terminator instruction (a subclass of the <a
3772href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).</p>
3773
3774<p>In addition to tracking the list of instructions that make up the block, the
3775<tt>BasicBlock</tt> class also keeps track of the <a
3776href="#Function"><tt>Function</tt></a> that it is embedded into.</p>
3777
3778<p>Note that <tt>BasicBlock</tt>s themselves are <a
3779href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
3780like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
3781<tt>label</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3782
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3783<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3784<h4>
3785 <a name="m_BasicBlock">
3786 Important Public Members of the <tt>BasicBlock</tt> class
3787 </a>
3788</h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3789
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3790<div>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3791<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3792
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3793<li><tt>BasicBlock(const std::string &amp;Name = "", </tt><tt><a
3794 href="#Function">Function</a> *Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3795
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3796<p>The <tt>BasicBlock</tt> constructor is used to create new basic blocks for
3797insertion into a function. The constructor optionally takes a name for the new
3798block, and a <a href="#Function"><tt>Function</tt></a> to insert it into. If
3799the <tt>Parent</tt> parameter is specified, the new <tt>BasicBlock</tt> is
3800automatically inserted at the end of the specified <a
3801href="#Function"><tt>Function</tt></a>, if not specified, the BasicBlock must be
3802manually inserted into the <a href="#Function"><tt>Function</tt></a>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3803
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3804<li><tt>BasicBlock::iterator</tt> - Typedef for instruction list iterator<br>
3805<tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
3806<tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
3807<tt>size()</tt>, <tt>empty()</tt>
3808STL-style functions for accessing the instruction list.
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3809
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3810<p>These methods and typedefs are forwarding functions that have the same
3811semantics as the standard library methods of the same names. These methods
3812expose the underlying instruction list of a basic block in a way that is easy to
3813manipulate. To get the full complement of container operations (including
3814operations to update the list), you must use the <tt>getInstList()</tt>
3815method.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3816
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3817<li><tt>BasicBlock::InstListType &amp;getInstList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3818
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3819<p>This method is used to get access to the underlying container that actually
3820holds the Instructions. This method must be used when there isn't a forwarding
3821function in the <tt>BasicBlock</tt> class for the operation that you would like
3822to perform. Because there are no forwarding functions for "updating"
3823operations, you need to use this if you want to update the contents of a
3824<tt>BasicBlock</tt>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3825
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3826<li><tt><a href="#Function">Function</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3827
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3828<p> Returns a pointer to <a href="#Function"><tt>Function</tt></a> the block is
3829embedded into, or a null pointer if it is homeless.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3830
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3831<li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3832
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3833<p> Returns a pointer to the terminator instruction that appears at the end of
3834the <tt>BasicBlock</tt>. If there is no terminator instruction, or if the last
3835instruction in the block is not a terminator, then a null pointer is
3836returned.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3837
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3838</ul>
3839
3840</div>
3841
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3842</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3843
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3844<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3845<h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3846 <a name="Argument">The <tt>Argument</tt> class</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3847</h3>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3848
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3849<div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3850
3851<p>This subclass of Value defines the interface for incoming formal
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]3852arguments to a function. A Function maintains a list of its formal
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3853arguments. An argument has a pointer to the parent Function.</p>
3854
3855</div>
3856
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3857</div>
3858
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3859<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3860<hr>
3861<address>
3862 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]3863 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3864 <a href="http://validator.w3.org/check/referer"><img
Gabor Greifa9c0f2b2008-06-18 14:05:31 +0000[diff] [blame]3865 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]3866
3867 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
3868 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
NAKAMURA Takumib9a33632011-04-09 02:13:37 +0000[diff] [blame]3869 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3870 Last modified: $Date$
3871</address>
3872
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3873</body>
3874</html>