Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 583b6dbcc586359e6ea7c3e4ac46bdb708779dce / . / docs / ProgrammersManual.html
blob: 83251e5102dcd0f3183530e2e177cd2d3ea83351 [file] [log] [blame]
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3<html>
4<head>
Owen Anderson5e8c50e2009-06-16 17:40:28 +0000[diff] [blame]5 <meta http-equiv="Content-type" content="text/html;charset=UTF-8">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]6 <title>LLVM Programmer's Manual</title>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]7 <link rel="stylesheet" href="llvm.css" type="text/css">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]8</head>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]9<body>
10
11<div class="doc_title">
12 LLVM Programmer's Manual
13</div>
14
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]15<ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]16 <li><a href="#introduction">Introduction</a></li>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]17 <li><a href="#general">General Information</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]18 <ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]19 <li><a href="#stl">The C++ Standard Template Library</a></li>
20<!--
21 <li>The <tt>-time-passes</tt> option</li>
22 <li>How to use the LLVM Makefile system</li>
23 <li>How to write a regression test</li>
Chris Lattner61db4652004-12-08 19:05:44 +0000[diff] [blame]24
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]25-->
Chris Lattner84b7f8d2003-08-01 22:20:59 +0000[diff] [blame]26 </ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]27 </li>
28 <li><a href="#apis">Important and useful LLVM APIs</a>
29 <ul>
30 <li><a href="#isa">The <tt>isa&lt;&gt;</tt>, <tt>cast&lt;&gt;</tt>
31and <tt>dyn_cast&lt;&gt;</tt> templates</a> </li>
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>
59 <li><a href="#dss_fixedarrays">Fixed Size Arrays</a></li>
60 <li><a href="#dss_heaparrays">Heap Allocated Arrays</a></li>
61 <li><a href="#dss_smallvector">"llvm/ADT/SmallVector.h"</a></li>
62 <li><a href="#dss_vector">&lt;vector&gt;</a></li>
63 <li><a href="#dss_deque">&lt;deque&gt;</a></li>
64 <li><a href="#dss_list">&lt;list&gt;</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]65 <li><a href="#dss_ilist">llvm/ADT/ilist.h</a></li>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]66 <li><a href="#dss_other">Other Sequential Container Options</a></li>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]67 </ul></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]68 <li><a href="#ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
69 <ul>
70 <li><a href="#dss_sortedvectorset">A sorted 'vector'</a></li>
71 <li><a href="#dss_smallset">"llvm/ADT/SmallSet.h"</a></li>
72 <li><a href="#dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a></li>
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]73 <li><a href="#dss_denseset">"llvm/ADT/DenseSet.h"</a></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]74 <li><a href="#dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a></li>
75 <li><a href="#dss_set">&lt;set&gt;</a></li>
76 <li><a href="#dss_setvector">"llvm/ADT/SetVector.h"</a></li>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]77 <li><a href="#dss_uniquevector">"llvm/ADT/UniqueVector.h"</a></li>
78 <li><a href="#dss_otherset">Other Set-Like ContainerOptions</a></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]79 </ul></li>
Chris Lattnerf3692522007-02-03 19:51:56 +0000[diff] [blame]80 <li><a href="#ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
81 <ul>
82 <li><a href="#dss_sortedvectormap">A sorted 'vector'</a></li>
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]83 <li><a href="#dss_stringmap">"llvm/ADT/StringMap.h"</a></li>
Chris Lattnerf3692522007-02-03 19:51:56 +0000[diff] [blame]84 <li><a href="#dss_indexedmap">"llvm/ADT/IndexedMap.h"</a></li>
85 <li><a href="#dss_densemap">"llvm/ADT/DenseMap.h"</a></li>
86 <li><a href="#dss_map">&lt;map&gt;</a></li>
87 <li><a href="#dss_othermap">Other Map-Like Container Options</a></li>
88 </ul></li>
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]89 <li><a href="#ds_string">String-like containers</a>
Benjamin Kramere15192b2009-08-05 15:42:44 +0000[diff] [blame]90 <!--<ul>
91 todo
92 </ul>--></li>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]93 <li><a href="#ds_bit">BitVector-like containers</a>
94 <ul>
95 <li><a href="#dss_bitvector">A dense bitvector</a></li>
96 <li><a href="#dss_sparsebitvector">A sparse bitvector</a></li>
97 </ul></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]98 </ul>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]99 </li>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]100 <li><a href="#common">Helpful Hints for Common Operations</a>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]101 <ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]102 <li><a href="#inspection">Basic Inspection and Traversal Routines</a>
103 <ul>
104 <li><a href="#iterate_function">Iterating over the <tt>BasicBlock</tt>s
105in a <tt>Function</tt></a> </li>
106 <li><a href="#iterate_basicblock">Iterating over the <tt>Instruction</tt>s
107in a <tt>BasicBlock</tt></a> </li>
108 <li><a href="#iterate_institer">Iterating over the <tt>Instruction</tt>s
109in a <tt>Function</tt></a> </li>
110 <li><a href="#iterate_convert">Turning an iterator into a
111class pointer</a> </li>
112 <li><a href="#iterate_complex">Finding call sites: a more
113complex example</a> </li>
114 <li><a href="#calls_and_invokes">Treating calls and invokes
115the same way</a> </li>
116 <li><a href="#iterate_chains">Iterating over def-use &amp;
117use-def chains</a> </li>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]118 <li><a href="#iterate_preds">Iterating over predecessors &amp;
119successors of blocks</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]120 </ul>
121 </li>
122 <li><a href="#simplechanges">Making simple changes</a>
123 <ul>
124 <li><a href="#schanges_creating">Creating and inserting new
125 <tt>Instruction</tt>s</a> </li>
126 <li><a href="#schanges_deleting">Deleting <tt>Instruction</tt>s</a> </li>
127 <li><a href="#schanges_replacing">Replacing an <tt>Instruction</tt>
128with another <tt>Value</tt></a> </li>
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]129 <li><a href="#schanges_deletingGV">Deleting <tt>GlobalVariable</tt>s</a> </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]130 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]131 </li>
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]132 <li><a href="#create_types">How to Create Types</a></li>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]133<!--
134 <li>Working with the Control Flow Graph
135 <ul>
136 <li>Accessing predecessors and successors of a <tt>BasicBlock</tt>
137 <li>
138 <li>
139 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]140-->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]141 </ul>
142 </li>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]143
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]144 <li><a href="#threading">Threads and LLVM</a>
145 <ul>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]146 <li><a href="#startmultithreaded">Entering and Exiting Multithreaded Mode
147 </a></li>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]148 <li><a href="#shutdown">Ending execution with <tt>llvm_shutdown()</tt></a></li>
149 <li><a href="#managedstatic">Lazy initialization with <tt>ManagedStatic</tt></a></li>
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]150 <li><a href="#llvmcontext">Achieving Isolation with <tt>LLVMContext</tt></a></li>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]151 </ul>
152 </li>
153
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]154 <li><a href="#advanced">Advanced Topics</a>
155 <ul>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]156 <li><a href="#TypeResolve">LLVM Type Resolution</a>
157 <ul>
158 <li><a href="#BuildRecType">Basic Recursive Type Construction</a></li>
159 <li><a href="#refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a></li>
160 <li><a href="#PATypeHolder">The PATypeHolder Class</a></li>
161 <li><a href="#AbstractTypeUser">The AbstractTypeUser Class</a></li>
162 </ul></li>
163
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]164 <li><a href="#SymbolTable">The <tt>ValueSymbolTable</tt> and <tt>TypeSymbolTable</tt> classes</a></li>
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<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]207<div class="doc_section">
208 <a name="introduction">Introduction </a>
209</div>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]210<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]211
212<div class="doc_text">
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<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]239<div class="doc_section">
240 <a name="general">General Information</a>
241</div>
242<!-- *********************************************************************** -->
243
244<div class="doc_text">
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
249</div>
250
251<!-- ======================================================================= -->
252<div class="doc_subsection">
253 <a name="stl">The C++ Standard Template Library</a>
254</div>
255
256<div class="doc_text">
257
258<p>LLVM makes heavy use of the C++ Standard Template Library (STL),
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]259perhaps much more than you are used to, or have seen before. Because of
260this, you might want to do a little background reading in the
261techniques used and capabilities of the library. There are many good
262pages that discuss the STL, and several books on the subject that you
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]263can get, so it will not be discussed in this document.</p>
264
265<p>Here are some useful links:</p>
266
267<ol>
268
269<li><a href="http://www.dinkumware.com/refxcpp.html">Dinkumware C++ Library
270reference</a> - an excellent reference for the STL and other parts of the
271standard C++ library.</li>
272
273<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]274O'Reilly book in the making. It has a decent Standard Library
275Reference that rivals Dinkumware's, and is unfortunately no longer free since the
276book has been published.</li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]277
278<li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
279Questions</a></li>
280
281<li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
282Contains a useful <a
283href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
284STL</a>.</li>
285
286<li><a href="http://www.research.att.com/%7Ebs/C++.html">Bjarne Stroustrup's C++
287Page</a></li>
288
Tanya Lattner79445ba2004-12-08 18:34:56 +0000[diff] [blame]289<li><a href="http://64.78.49.204/">
Reid Spencer096603a2004-05-26 08:41:35 +0000[diff] [blame]290Bruce Eckel's Thinking in C++, 2nd ed. Volume 2 Revision 4.0 (even better, get
291the book).</a></li>
292
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]293</ol>
294
295<p>You are also encouraged to take a look at the <a
296href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
297to write maintainable code more than where to put your curly braces.</p>
298
299</div>
300
301<!-- ======================================================================= -->
302<div class="doc_subsection">
303 <a name="stl">Other useful references</a>
304</div>
305
306<div class="doc_text">
307
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]308<ol>
309<li><a href="http://www.psc.edu/%7Esemke/cvs_branches.html">CVS
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]310Branch and Tag Primer</a></li>
Misha Brukmana0f71e42004-06-18 18:39:00 +0000[diff] [blame]311<li><a href="http://www.fortran-2000.com/ArnaudRecipes/sharedlib.html">Using
312static and shared libraries across platforms</a></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]313</ol>
314
315</div>
316
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]317<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]318<div class="doc_section">
319 <a name="apis">Important and useful LLVM APIs</a>
320</div>
321<!-- *********************************************************************** -->
322
323<div class="doc_text">
324
325<p>Here we highlight some LLVM APIs that are generally useful and good to
326know about when writing transformations.</p>
327
328</div>
329
330<!-- ======================================================================= -->
331<div class="doc_subsection">
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]332 <a name="isa">The <tt>isa&lt;&gt;</tt>, <tt>cast&lt;&gt;</tt> and
333 <tt>dyn_cast&lt;&gt;</tt> templates</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]334</div>
335
336<div class="doc_text">
337
338<p>The LLVM source-base makes extensive use of a custom form of RTTI.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]339These templates have many similarities to the C++ <tt>dynamic_cast&lt;&gt;</tt>
340operator, but they don't have some drawbacks (primarily stemming from
341the fact that <tt>dynamic_cast&lt;&gt;</tt> only works on classes that
342have a v-table). Because they are used so often, you must know what they
343do and how they work. All of these templates are defined in the <a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]344 href="/doxygen/Casting_8h-source.html"><tt>llvm/Support/Casting.h</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]345file (note that you very rarely have to include this file directly).</p>
346
347<dl>
348 <dt><tt>isa&lt;&gt;</tt>: </dt>
349
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]350 <dd><p>The <tt>isa&lt;&gt;</tt> operator works exactly like the Java
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]351 "<tt>instanceof</tt>" operator. It returns true or false depending on whether
352 a reference or pointer points to an instance of the specified class. This can
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]353 be very useful for constraint checking of various sorts (example below).</p>
354 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]355
356 <dt><tt>cast&lt;&gt;</tt>: </dt>
357
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]358 <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]359 converts a pointer or reference from a base class to a derived class, causing
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]360 an assertion failure if it is not really an instance of the right type. This
361 should be used in cases where you have some information that makes you believe
362 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]363 and <tt>cast&lt;&gt;</tt> template is:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]364
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]365<div class="doc_code">
366<pre>
367static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
368 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))
369 return true;
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]370
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]371 // <i>Otherwise, it must be an instruction...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]372 return !L-&gt;contains(cast&lt;<a href="#Instruction">Instruction</a>&gt;(V)-&gt;getParent());
373}
374</pre>
375</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]376
377 <p>Note that you should <b>not</b> use an <tt>isa&lt;&gt;</tt> test followed
378 by a <tt>cast&lt;&gt;</tt>, for that use the <tt>dyn_cast&lt;&gt;</tt>
379 operator.</p>
380
381 </dd>
382
383 <dt><tt>dyn_cast&lt;&gt;</tt>:</dt>
384
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]385 <dd><p>The <tt>dyn_cast&lt;&gt;</tt> operator is a "checking cast" operation.
386 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]387 pointer to it (this operator does not work with references). If the operand is
388 not of the correct type, a null pointer is returned. Thus, this works very
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]389 much like the <tt>dynamic_cast&lt;&gt;</tt> operator in C++, and should be
390 used in the same circumstances. Typically, the <tt>dyn_cast&lt;&gt;</tt>
391 operator is used in an <tt>if</tt> statement or some other flow control
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]392 statement like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]393
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]394<div class="doc_code">
395<pre>
396if (<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]397 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]398}
399</pre>
400</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]401
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]402 <p>This form of the <tt>if</tt> statement effectively combines together a call
403 to <tt>isa&lt;&gt;</tt> and a call to <tt>cast&lt;&gt;</tt> into one
404 statement, which is very convenient.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]405
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]406 <p>Note that the <tt>dyn_cast&lt;&gt;</tt> operator, like C++'s
407 <tt>dynamic_cast&lt;&gt;</tt> or Java's <tt>instanceof</tt> operator, can be
408 abused. In particular, you should not use big chained <tt>if/then/else</tt>
409 blocks to check for lots of different variants of classes. If you find
410 yourself wanting to do this, it is much cleaner and more efficient to use the
411 <tt>InstVisitor</tt> class to dispatch over the instruction type directly.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]412
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]413 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]414
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]415 <dt><tt>cast_or_null&lt;&gt;</tt>: </dt>
416
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]417 <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]418 <tt>cast&lt;&gt;</tt> operator, except that it allows for a null pointer as an
419 argument (which it then propagates). This can sometimes be useful, allowing
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]420 you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]421
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]422 <dt><tt>dyn_cast_or_null&lt;&gt;</tt>: </dt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]423
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]424 <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]425 <tt>dyn_cast&lt;&gt;</tt> operator, except that it allows for a null pointer
426 as an argument (which it then propagates). This can sometimes be useful,
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]427 allowing you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]428
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]429</dl>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]430
431<p>These five templates can be used with any classes, whether they have a
432v-table or not. To add support for these templates, you simply need to add
433<tt>classof</tt> static methods to the class you are interested casting
434to. Describing this is currently outside the scope of this document, but there
435are lots of examples in the LLVM source base.</p>
436
437</div>
438
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]439
440<!-- ======================================================================= -->
441<div class="doc_subsection">
442 <a name="string_apis">Passing strings (the <tt>StringRef</tt>
443and <tt>Twine</tt> classes)</a>
444</div>
445
446<div class="doc_text">
447
448<p>Although LLVM generally does not do much string manipulation, we do have
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]449several important APIs which take strings. Two important examples are the
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]450Value class -- which has names for instructions, functions, etc. -- and the
451StringMap class which is used extensively in LLVM and Clang.</p>
452
453<p>These are generic classes, and they need to be able to accept strings which
454may have embedded null characters. Therefore, they cannot simply take
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]455a <tt>const char *</tt>, and taking a <tt>const std::string&amp;</tt> requires
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]456clients to perform a heap allocation which is usually unnecessary. Instead,
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]457many LLVM APIs use a <tt>const StringRef&amp;</tt> or a <tt>const
458Twine&amp;</tt> for passing strings efficiently.</p>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]459
460</div>
461
462<!-- _______________________________________________________________________ -->
463<div class="doc_subsubsection">
464 <a name="StringRef">The <tt>StringRef</tt> class</a>
465</div>
466
467<div class="doc_text">
468
469<p>The <tt>StringRef</tt> data type represents a reference to a constant string
470(a character array and a length) and supports the common operations available
471on <tt>std:string</tt>, but does not require heap allocation.</p>
472
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]473<p>It can be implicitly constructed using a C style null-terminated string,
474an <tt>std::string</tt>, or explicitly with a character pointer and length.
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]475For example, the <tt>StringRef</tt> find function is declared as:</p>
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]476
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]477<div class="doc_code">
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]478 iterator find(const StringRef &amp;Key);
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]479</div>
480
481<p>and clients can call it using any one of:</p>
482
483<div class="doc_code">
484<pre>
485 Map.find("foo"); <i>// Lookup "foo"</i>
486 Map.find(std::string("bar")); <i>// Lookup "bar"</i>
487 Map.find(StringRef("\0baz", 4)); <i>// Lookup "\0baz"</i>
488</pre>
489</div>
490
491<p>Similarly, APIs which need to return a string may return a <tt>StringRef</tt>
492instance, which can be used directly or converted to an <tt>std::string</tt>
493using the <tt>str</tt> member function. See
494"<tt><a href="/doxygen/classllvm_1_1StringRef_8h-source.html">llvm/ADT/StringRef.h</a></tt>"
495for more information.</p>
496
497<p>You should rarely use the <tt>StringRef</tt> class directly, because it contains
498pointers to external memory it is not generally safe to store an instance of the
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]499class (unless you know that the external storage will not be freed).</p>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]500
501</div>
502
503<!-- _______________________________________________________________________ -->
504<div class="doc_subsubsection">
505 <a name="Twine">The <tt>Twine</tt> class</a>
506</div>
507
508<div class="doc_text">
509
510<p>The <tt>Twine</tt> class is an efficient way for APIs to accept concatenated
511strings. For example, a common LLVM paradigm is to name one instruction based on
512the name of another instruction with a suffix, for example:</p>
513
514<div class="doc_code">
515<pre>
516 New = CmpInst::Create(<i>...</i>, SO->getName() + ".cmp");
517</pre>
518</div>
519
520<p>The <tt>Twine</tt> class is effectively a
521lightweight <a href="http://en.wikipedia.org/wiki/Rope_(computer_science)">rope</a>
522which points to temporary (stack allocated) objects. Twines can be implicitly
523constructed as the result of the plus operator applied to strings (i.e., a C
524strings, an <tt>std::string</tt>, or a <tt>StringRef</tt>). The twine delays the
525actual concatentation of strings until it is actually required, at which point
526it can be efficiently rendered directly into a character array. This avoids
527unnecessary heap allocation involved in constructing the temporary results of
528string concatenation. See
529"<tt><a href="/doxygen/classllvm_1_1Twine_8h-source.html">llvm/ADT/Twine.h</a></tt>"
Benjamin Kramere15192b2009-08-05 15:42:44 +0000[diff] [blame]530for more information.</p>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]531
532<p>As with a <tt>StringRef</tt>, <tt>Twine</tt> objects point to external memory
533and should almost never be stored or mentioned directly. They are intended
534solely for use when defining a function which should be able to efficiently
535accept concatenated strings.</p>
536
537</div>
538
539
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]540<!-- ======================================================================= -->
541<div class="doc_subsection">
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]542 <a name="DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt> option</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]543</div>
544
545<div class="doc_text">
546
547<p>Often when working on your pass you will put a bunch of debugging printouts
548and other code into your pass. After you get it working, you want to remove
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]549it, 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]550across).</p>
551
552<p> Naturally, because of this, you don't want to delete the debug printouts,
553but you don't want them to always be noisy. A standard compromise is to comment
554them out, allowing you to enable them if you need them in the future.</p>
555
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]556<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]557file provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to
558this problem. Basically, you can put arbitrary code into the argument of the
559<tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
560tool) is run with the '<tt>-debug</tt>' command line argument:</p>
561
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]562<div class="doc_code">
563<pre>
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]564DEBUG(errs() &lt;&lt; "I am here!\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]565</pre>
566</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]567
568<p>Then you can run your pass like this:</p>
569
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]570<div class="doc_code">
571<pre>
572$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]573<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]574$ opt &lt; a.bc &gt; /dev/null -mypass -debug
575I am here!
576</pre>
577</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]578
579<p>Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you
580to not have to create "yet another" command line option for the debug output for
581your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
582so they do not cause a performance impact at all (for the same reason, they
583should also not contain side-effects!).</p>
584
585<p>One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
586enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
587"<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
588program hasn't been started yet, you can always just run it with
589<tt>-debug</tt>.</p>
590
591</div>
592
593<!-- _______________________________________________________________________ -->
594<div class="doc_subsubsection">
Chris Lattnerc9151082005-04-26 22:57:07 +0000[diff] [blame]595 <a name="DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt> and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]596 the <tt>-debug-only</tt> option</a>
597</div>
598
599<div class="doc_text">
600
601<p>Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
602just turns on <b>too much</b> information (such as when working on the code
603generator). If you want to enable debug information with more fine-grained
604control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
605option as follows:</p>
606
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]607<div class="doc_code">
608<pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]609#undef DEBUG_TYPE
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]610DEBUG(errs() &lt;&lt; "No debug type\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]611#define DEBUG_TYPE "foo"
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]612DEBUG(errs() &lt;&lt; "'foo' debug type\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]613#undef DEBUG_TYPE
614#define DEBUG_TYPE "bar"
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]615DEBUG(errs() &lt;&lt; "'bar' debug type\n"));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]616#undef DEBUG_TYPE
617#define DEBUG_TYPE ""
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]618DEBUG(errs() &lt;&lt; "No debug type (2)\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]619</pre>
620</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]621
622<p>Then you can run your pass like this:</p>
623
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]624<div class="doc_code">
625<pre>
626$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]627<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]628$ opt &lt; a.bc &gt; /dev/null -mypass -debug
629No debug type
630'foo' debug type
631'bar' debug type
632No debug type (2)
633$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=foo
634'foo' debug type
635$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=bar
636'bar' debug type
637</pre>
638</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]639
640<p>Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of
641a 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]642you <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]643<tt>#undef</tt>'s). Also, you should use names more meaningful than "foo" and
644"bar", because there is no system in place to ensure that names do not
645conflict. If two different modules use the same string, they will all be turned
646on when the name is specified. This allows, for example, all debug information
647for instruction scheduling to be enabled with <tt>-debug-type=InstrSched</tt>,
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]648even if the source lives in multiple files.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]649
Daniel Dunbarc3c92392009-08-07 23:48:59 +0000[diff] [blame]650<p>The <tt>DEBUG_WITH_TYPE</tt> macro is also available for situations where you
651would like to set <tt>DEBUG_TYPE</tt>, but only for one specific <tt>DEBUG</tt>
652statement. It takes an additional first parameter, which is the type to use. For
653example, the preceeding example could be written as:</p>
654
655
656<div class="doc_code">
657<pre>
658DEBUG_WITH_TYPE("", errs() &lt;&lt; "No debug type\n");
659DEBUG_WITH_TYPE("foo", errs() &lt;&lt; "'foo' debug type\n");
660DEBUG_WITH_TYPE("bar", errs() &lt;&lt; "'bar' debug type\n"));
661DEBUG_WITH_TYPE("", errs() &lt;&lt; "No debug type (2)\n");
662</pre>
663</div>
664
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]665</div>
666
667<!-- ======================================================================= -->
668<div class="doc_subsection">
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]669 <a name="Statistic">The <tt>Statistic</tt> class &amp; <tt>-stats</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]670 option</a>
671</div>
672
673<div class="doc_text">
674
675<p>The "<tt><a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]676href="/doxygen/Statistic_8h-source.html">llvm/ADT/Statistic.h</a></tt>" file
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]677provides a class named <tt>Statistic</tt> that is used as a unified way to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]678keep track of what the LLVM compiler is doing and how effective various
679optimizations are. It is useful to see what optimizations are contributing to
680making a particular program run faster.</p>
681
682<p>Often you may run your pass on some big program, and you're interested to see
683how many times it makes a certain transformation. Although you can do this with
684hand 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]685for big programs. Using the <tt>Statistic</tt> class makes it very easy to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]686keep track of this information, and the calculated information is presented in a
687uniform manner with the rest of the passes being executed.</p>
688
689<p>There are many examples of <tt>Statistic</tt> uses, but the basics of using
690it are as follows:</p>
691
692<ol>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]693 <li><p>Define your statistic like this:</p>
694
695<div class="doc_code">
696<pre>
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]697#define <a href="#DEBUG_TYPE">DEBUG_TYPE</a> "mypassname" <i>// This goes before any #includes.</i>
698STATISTIC(NumXForms, "The # of times I did stuff");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]699</pre>
700</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]701
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]702 <p>The <tt>STATISTIC</tt> macro defines a static variable, whose name is
703 specified by the first argument. The pass name is taken from the DEBUG_TYPE
704 macro, and the description is taken from the second argument. The variable
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]705 defined ("NumXForms" in this case) acts like an unsigned integer.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]706
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]707 <li><p>Whenever you make a transformation, bump the counter:</p>
708
709<div class="doc_code">
710<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]711++NumXForms; // <i>I did stuff!</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]712</pre>
713</div>
714
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]715 </li>
716 </ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]717
718 <p>That's all you have to do. To get '<tt>opt</tt>' to print out the
719 statistics gathered, use the '<tt>-stats</tt>' option:</p>
720
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]721<div class="doc_code">
722<pre>
723$ opt -stats -mypassname &lt; program.bc &gt; /dev/null
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]724<i>... statistics output ...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]725</pre>
726</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]727
Reid Spencer6b6c73e2007-02-09 16:00:28 +0000[diff] [blame]728 <p> When running <tt>opt</tt> on a C file from the SPEC benchmark
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]729suite, it gives a report that looks like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]730
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]731<div class="doc_code">
732<pre>
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]733 7646 bitcodewriter - Number of normal instructions
734 725 bitcodewriter - Number of oversized instructions
735 129996 bitcodewriter - Number of bitcode bytes written
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]736 2817 raise - Number of insts DCEd or constprop'd
737 3213 raise - Number of cast-of-self removed
738 5046 raise - Number of expression trees converted
739 75 raise - Number of other getelementptr's formed
740 138 raise - Number of load/store peepholes
741 42 deadtypeelim - Number of unused typenames removed from symtab
742 392 funcresolve - Number of varargs functions resolved
743 27 globaldce - Number of global variables removed
744 2 adce - Number of basic blocks removed
745 134 cee - Number of branches revectored
746 49 cee - Number of setcc instruction eliminated
747 532 gcse - Number of loads removed
748 2919 gcse - Number of instructions removed
749 86 indvars - Number of canonical indvars added
750 87 indvars - Number of aux indvars removed
751 25 instcombine - Number of dead inst eliminate
752 434 instcombine - Number of insts combined
753 248 licm - Number of load insts hoisted
754 1298 licm - Number of insts hoisted to a loop pre-header
755 3 licm - Number of insts hoisted to multiple loop preds (bad, no loop pre-header)
756 75 mem2reg - Number of alloca's promoted
757 1444 cfgsimplify - Number of blocks simplified
758</pre>
759</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]760
761<p>Obviously, with so many optimizations, having a unified framework for this
762stuff is very nice. Making your pass fit well into the framework makes it more
763maintainable and useful.</p>
764
765</div>
766
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]767<!-- ======================================================================= -->
768<div class="doc_subsection">
769 <a name="ViewGraph">Viewing graphs while debugging code</a>
770</div>
771
772<div class="doc_text">
773
774<p>Several of the important data structures in LLVM are graphs: for example
775CFGs made out of LLVM <a href="#BasicBlock">BasicBlock</a>s, CFGs made out of
776LLVM <a href="CodeGenerator.html#machinebasicblock">MachineBasicBlock</a>s, and
777<a href="CodeGenerator.html#selectiondag_intro">Instruction Selection
778DAGs</a>. In many cases, while debugging various parts of the compiler, it is
779nice to instantly visualize these graphs.</p>
780
781<p>LLVM provides several callbacks that are available in a debug build to do
782exactly that. If you call the <tt>Function::viewCFG()</tt> method, for example,
783the current LLVM tool will pop up a window containing the CFG for the function
784where each basic block is a node in the graph, and each node contains the
785instructions in the block. Similarly, there also exists
786<tt>Function::viewCFGOnly()</tt> (does not include the instructions), the
787<tt>MachineFunction::viewCFG()</tt> and <tt>MachineFunction::viewCFGOnly()</tt>,
788and the <tt>SelectionDAG::viewGraph()</tt> methods. Within GDB, for example,
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]789you can usually use something like <tt>call DAG.viewGraph()</tt> to pop
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]790up a window. Alternatively, you can sprinkle calls to these functions in your
791code in places you want to debug.</p>
792
793<p>Getting this to work requires a small amount of configuration. On Unix
794systems with X11, install the <a href="http://www.graphviz.org">graphviz</a>
795toolkit, and make sure 'dot' and 'gv' are in your path. If you are running on
796Mac OS/X, download and install the Mac OS/X <a
797href="http://www.pixelglow.com/graphviz/">Graphviz program</a>, and add
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]798<tt>/Applications/Graphviz.app/Contents/MacOS/</tt> (or wherever you install
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]799it) to your path. Once in your system and path are set up, rerun the LLVM
800configure script and rebuild LLVM to enable this functionality.</p>
801
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]802<p><tt>SelectionDAG</tt> has been extended to make it easier to locate
803<i>interesting</i> nodes in large complex graphs. From gdb, if you
804<tt>call DAG.setGraphColor(<i>node</i>, "<i>color</i>")</tt>, then the
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]805next <tt>call DAG.viewGraph()</tt> would highlight the node in the
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]806specified color (choices of colors can be found at <a
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]807href="http://www.graphviz.org/doc/info/colors.html">colors</a>.) More
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]808complex node attributes can be provided with <tt>call
809DAG.setGraphAttrs(<i>node</i>, "<i>attributes</i>")</tt> (choices can be
810found at <a href="http://www.graphviz.org/doc/info/attrs.html">Graph
811Attributes</a>.) If you want to restart and clear all the current graph
812attributes, then you can <tt>call DAG.clearGraphAttrs()</tt>. </p>
813
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]814</div>
815
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]816<!-- *********************************************************************** -->
817<div class="doc_section">
818 <a name="datastructure">Picking the Right Data Structure for a Task</a>
819</div>
820<!-- *********************************************************************** -->
821
822<div class="doc_text">
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
878</div>
879
880<!-- ======================================================================= -->
881<div class="doc_subsection">
882 <a name="ds_sequential">Sequential Containers (std::vector, std::list, etc)</a>
883</div>
884
885<div class="doc_text">
886There are a variety of sequential containers available for you, based on your
887needs. Pick the first in this section that will do what you want.
888</div>
889
890<!-- _______________________________________________________________________ -->
891<div class="doc_subsubsection">
892 <a name="dss_fixedarrays">Fixed Size Arrays</a>
893</div>
894
895<div class="doc_text">
896<p>Fixed size arrays are very simple and very fast. They are good if you know
897exactly how many elements you have, or you have a (low) upper bound on how many
898you have.</p>
899</div>
900
901<!-- _______________________________________________________________________ -->
902<div class="doc_subsubsection">
903 <a name="dss_heaparrays">Heap Allocated Arrays</a>
904</div>
905
906<div class="doc_text">
907<p>Heap allocated arrays (new[] + delete[]) are also simple. They are good if
908the number of elements is variable, if you know how many elements you will need
909before the array is allocated, and if the array is usually large (if not,
910consider a <a href="#dss_smallvector">SmallVector</a>). The cost of a heap
911allocated array is the cost of the new/delete (aka malloc/free). Also note that
912if you are allocating an array of a type with a constructor, the constructor and
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]913destructors will be run for every element in the array (re-sizable vectors only
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]914construct those elements actually used).</p>
915</div>
916
917<!-- _______________________________________________________________________ -->
918<div class="doc_subsubsection">
919 <a name="dss_smallvector">"llvm/ADT/SmallVector.h"</a>
920</div>
921
922<div class="doc_text">
923<p><tt>SmallVector&lt;Type, N&gt;</tt> is a simple class that looks and smells
924just like <tt>vector&lt;Type&gt;</tt>:
925it supports efficient iteration, lays out elements in memory order (so you can
926do pointer arithmetic between elements), supports efficient push_back/pop_back
927operations, supports efficient random access to its elements, etc.</p>
928
929<p>The advantage of SmallVector is that it allocates space for
930some number of elements (N) <b>in the object itself</b>. Because of this, if
931the SmallVector is dynamically smaller than N, no malloc is performed. This can
932be a big win in cases where the malloc/free call is far more expensive than the
933code that fiddles around with the elements.</p>
934
935<p>This is good for vectors that are "usually small" (e.g. the number of
936predecessors/successors of a block is usually less than 8). On the other hand,
937this makes the size of the SmallVector itself large, so you don't want to
938allocate lots of them (doing so will waste a lot of space). As such,
939SmallVectors are most useful when on the stack.</p>
940
941<p>SmallVector also provides a nice portable and efficient replacement for
942<tt>alloca</tt>.</p>
943
944</div>
945
946<!-- _______________________________________________________________________ -->
947<div class="doc_subsubsection">
948 <a name="dss_vector">&lt;vector&gt;</a>
949</div>
950
951<div class="doc_text">
952<p>
953std::vector is well loved and respected. It is useful when SmallVector isn't:
954when the size of the vector is often large (thus the small optimization will
955rarely be a benefit) or if you will be allocating many instances of the vector
956itself (which would waste space for elements that aren't in the container).
957vector is also useful when interfacing with code that expects vectors :).
958</p>
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]959
960<p>One worthwhile note about std::vector: avoid code like this:</p>
961
962<div class="doc_code">
963<pre>
964for ( ... ) {
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]965 std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]966 use V;
967}
968</pre>
969</div>
970
971<p>Instead, write this as:</p>
972
973<div class="doc_code">
974<pre>
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]975std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]976for ( ... ) {
977 use V;
978 V.clear();
979}
980</pre>
981</div>
982
983<p>Doing so will save (at least) one heap allocation and free per iteration of
984the loop.</p>
985
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]986</div>
987
988<!-- _______________________________________________________________________ -->
989<div class="doc_subsubsection">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]990 <a name="dss_deque">&lt;deque&gt;</a>
991</div>
992
993<div class="doc_text">
994<p>std::deque is, in some senses, a generalized version of std::vector. Like
995std::vector, it provides constant time random access and other similar
996properties, but it also provides efficient access to the front of the list. It
997does not guarantee continuity of elements within memory.</p>
998
999<p>In exchange for this extra flexibility, std::deque has significantly higher
1000constant factor costs than std::vector. If possible, use std::vector or
1001something cheaper.</p>
1002</div>
1003
1004<!-- _______________________________________________________________________ -->
1005<div class="doc_subsubsection">
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1006 <a name="dss_list">&lt;list&gt;</a>
1007</div>
1008
1009<div class="doc_text">
1010<p>std::list is an extremely inefficient class that is rarely useful.
1011It performs a heap allocation for every element inserted into it, thus having an
1012extremely high constant factor, particularly for small data types. std::list
1013also only supports bidirectional iteration, not random access iteration.</p>
1014
1015<p>In exchange for this high cost, std::list supports efficient access to both
1016ends of the list (like std::deque, but unlike std::vector or SmallVector). In
1017addition, the iterator invalidation characteristics of std::list are stronger
1018than that of a vector class: inserting or removing an element into the list does
1019not invalidate iterator or pointers to other elements in the list.</p>
1020</div>
1021
1022<!-- _______________________________________________________________________ -->
1023<div class="doc_subsubsection">
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1024 <a name="dss_ilist">llvm/ADT/ilist.h</a>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1025</div>
1026
1027<div class="doc_text">
1028<p><tt>ilist&lt;T&gt;</tt> implements an 'intrusive' doubly-linked list. It is
1029intrusive, because it requires the element to store and provide access to the
1030prev/next pointers for the list.</p>
1031
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1032<p><tt>ilist</tt> has the same drawbacks as <tt>std::list</tt>, and additionally
1033requires an <tt>ilist_traits</tt> implementation for the element type, but it
1034provides some novel characteristics. In particular, it can efficiently store
1035polymorphic objects, the traits class is informed when an element is inserted or
Gabor Greif0cbcabe2009-03-12 09:47:03 +0000[diff] [blame]1036removed from the list, and <tt>ilist</tt>s are guaranteed to support a
1037constant-time splice operation.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1038
Gabor Greif0cbcabe2009-03-12 09:47:03 +0000[diff] [blame]1039<p>These properties are exactly what we want for things like
1040<tt>Instruction</tt>s and basic blocks, which is why these are implemented with
1041<tt>ilist</tt>s.</p>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1042
1043Related classes of interest are explained in the following subsections:
1044 <ul>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1045 <li><a href="#dss_ilist_traits">ilist_traits</a></li>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1046 <li><a href="#dss_iplist">iplist</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1047 <li><a href="#dss_ilist_node">llvm/ADT/ilist_node.h</a></li>
Gabor Greif6a65f422009-03-12 10:30:31 +0000[diff] [blame]1048 <li><a href="#dss_ilist_sentinel">Sentinels</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1049 </ul>
1050</div>
1051
1052<!-- _______________________________________________________________________ -->
1053<div class="doc_subsubsection">
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1054 <a name="dss_ilist_traits">ilist_traits</a>
1055</div>
1056
1057<div class="doc_text">
1058<p><tt>ilist_traits&lt;T&gt;</tt> is <tt>ilist&lt;T&gt;</tt>'s customization
1059mechanism. <tt>iplist&lt;T&gt;</tt> (and consequently <tt>ilist&lt;T&gt;</tt>)
1060publicly derive from this traits class.</p>
1061</div>
1062
1063<!-- _______________________________________________________________________ -->
1064<div class="doc_subsubsection">
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1065 <a name="dss_iplist">iplist</a>
1066</div>
1067
1068<div class="doc_text">
1069<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]1070supports a slightly narrower interface. Notably, inserters from
1071<tt>T&amp;</tt> are absent.</p>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1072
1073<p><tt>ilist_traits&lt;T&gt;</tt> is a public base of this class and can be
1074used for a wide variety of customizations.</p>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1075</div>
1076
1077<!-- _______________________________________________________________________ -->
1078<div class="doc_subsubsection">
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1079 <a name="dss_ilist_node">llvm/ADT/ilist_node.h</a>
1080</div>
1081
1082<div class="doc_text">
1083<p><tt>ilist_node&lt;T&gt;</tt> implements a the forward and backward links
1084that are expected by the <tt>ilist&lt;T&gt;</tt> (and analogous containers)
1085in the default manner.</p>
1086
1087<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]1088<tt>T</tt>, usually <tt>T</tt> publicly derives from
1089<tt>ilist_node&lt;T&gt;</tt>.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1090</div>
1091
1092<!-- _______________________________________________________________________ -->
1093<div class="doc_subsubsection">
Gabor Greif6a65f422009-03-12 10:30:31 +0000[diff] [blame]1094 <a name="dss_ilist_sentinel">Sentinels</a>
1095</div>
1096
1097<div class="doc_text">
1098<p><tt>ilist</tt>s have another speciality that must be considered. To be a good
1099citizen in the C++ ecosystem, it needs to support the standard container
1100operations, such as <tt>begin</tt> and <tt>end</tt> iterators, etc. Also, the
1101<tt>operator--</tt> must work correctly on the <tt>end</tt> iterator in the
1102case of non-empty <tt>ilist</tt>s.</p>
1103
1104<p>The only sensible solution to this problem is to allocate a so-called
1105<i>sentinel</i> along with the intrusive list, which serves as the <tt>end</tt>
1106iterator, providing the back-link to the last element. However conforming to the
1107C++ convention it is illegal to <tt>operator++</tt> beyond the sentinel and it
1108also must not be dereferenced.</p>
1109
1110<p>These constraints allow for some implementation freedom to the <tt>ilist</tt>
1111how to allocate and store the sentinel. The corresponding policy is dictated
1112by <tt>ilist_traits&lt;T&gt;</tt>. By default a <tt>T</tt> gets heap-allocated
1113whenever the need for a sentinel arises.</p>
1114
1115<p>While the default policy is sufficient in most cases, it may break down when
1116<tt>T</tt> does not provide a default constructor. Also, in the case of many
1117instances of <tt>ilist</tt>s, the memory overhead of the associated sentinels
1118is wasted. To alleviate the situation with numerous and voluminous
1119<tt>T</tt>-sentinels, sometimes a trick is employed, leading to <i>ghostly
1120sentinels</i>.</p>
1121
1122<p>Ghostly sentinels are obtained by specially-crafted <tt>ilist_traits&lt;T&gt;</tt>
1123which superpose the sentinel with the <tt>ilist</tt> instance in memory. Pointer
1124arithmetic is used to obtain the sentinel, which is relative to the
1125<tt>ilist</tt>'s <tt>this</tt> pointer. The <tt>ilist</tt> is augmented by an
1126extra pointer, which serves as the back-link of the sentinel. This is the only
1127field in the ghostly sentinel which can be legally accessed.</p>
1128</div>
1129
1130<!-- _______________________________________________________________________ -->
1131<div class="doc_subsubsection">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1132 <a name="dss_other">Other Sequential Container options</a>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1133</div>
1134
1135<div class="doc_text">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1136<p>Other STL containers are available, such as std::string.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1137
1138<p>There are also various STL adapter classes such as std::queue,
1139std::priority_queue, std::stack, etc. These provide simplified access to an
1140underlying container but don't affect the cost of the container itself.</p>
1141
1142</div>
1143
1144
1145<!-- ======================================================================= -->
1146<div class="doc_subsection">
1147 <a name="ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
1148</div>
1149
1150<div class="doc_text">
1151
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1152<p>Set-like containers are useful when you need to canonicalize multiple values
1153into a single representation. There are several different choices for how to do
1154this, providing various trade-offs.</p>
1155
1156</div>
1157
1158
1159<!-- _______________________________________________________________________ -->
1160<div class="doc_subsubsection">
1161 <a name="dss_sortedvectorset">A sorted 'vector'</a>
1162</div>
1163
1164<div class="doc_text">
1165
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]1166<p>If you intend to insert a lot of elements, then do a lot of queries, a
1167great approach is to use a vector (or other sequential container) with
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1168std::sort+std::unique to remove duplicates. This approach works really well if
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]1169your usage pattern has these two distinct phases (insert then query), and can be
1170coupled with a good choice of <a href="#ds_sequential">sequential container</a>.
1171</p>
1172
1173<p>
1174This combination provides the several nice properties: the result data is
1175contiguous in memory (good for cache locality), has few allocations, is easy to
1176address (iterators in the final vector are just indices or pointers), and can be
1177efficiently queried with a standard binary or radix search.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1178
1179</div>
1180
1181<!-- _______________________________________________________________________ -->
1182<div class="doc_subsubsection">
1183 <a name="dss_smallset">"llvm/ADT/SmallSet.h"</a>
1184</div>
1185
1186<div class="doc_text">
1187
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1188<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]1189are 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]1190has space for N elements in place (thus, if the set is dynamically smaller than
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1191N, no malloc traffic is required) and accesses them with a simple linear search.
1192When the set grows beyond 'N' elements, it allocates a more expensive representation that
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1193guarantees efficient access (for most types, it falls back to std::set, but for
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1194pointers it uses something far better, <a
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1195href="#dss_smallptrset">SmallPtrSet</a>).</p>
1196
1197<p>The magic of this class is that it handles small sets extremely efficiently,
1198but gracefully handles extremely large sets without loss of efficiency. The
1199drawback is that the interface is quite small: it supports insertion, queries
1200and erasing, but does not support iteration.</p>
1201
1202</div>
1203
1204<!-- _______________________________________________________________________ -->
1205<div class="doc_subsubsection">
1206 <a name="dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a>
1207</div>
1208
1209<div class="doc_text">
1210
1211<p>SmallPtrSet has all the advantages of SmallSet (and a SmallSet of pointers is
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1212transparently implemented with a SmallPtrSet), but also supports iterators. If
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1213more than 'N' insertions are performed, a single quadratically
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1214probed hash table is allocated and grows as needed, providing extremely
1215efficient access (constant time insertion/deleting/queries with low constant
1216factors) and is very stingy with malloc traffic.</p>
1217
1218<p>Note that, unlike std::set, the iterators of SmallPtrSet are invalidated
1219whenever an insertion occurs. Also, the values visited by the iterators are not
1220visited in sorted order.</p>
1221
1222</div>
1223
1224<!-- _______________________________________________________________________ -->
1225<div class="doc_subsubsection">
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]1226 <a name="dss_denseset">"llvm/ADT/DenseSet.h"</a>
1227</div>
1228
1229<div class="doc_text">
1230
1231<p>
1232DenseSet is a simple quadratically probed hash table. It excels at supporting
1233small values: it uses a single allocation to hold all of the pairs that
1234are currently inserted in the set. DenseSet is a great way to unique small
1235values that are not simple pointers (use <a
1236href="#dss_smallptrset">SmallPtrSet</a> for pointers). Note that DenseSet has
1237the same requirements for the value type that <a
1238href="#dss_densemap">DenseMap</a> has.
1239</p>
1240
1241</div>
1242
1243<!-- _______________________________________________________________________ -->
1244<div class="doc_subsubsection">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1245 <a name="dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a>
1246</div>
1247
1248<div class="doc_text">
1249
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1250<p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1251FoldingSet is an aggregate class that is really good at uniquing
1252expensive-to-create or polymorphic objects. It is a combination of a chained
1253hash table with intrusive links (uniqued objects are required to inherit from
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1254FoldingSetNode) that uses <a href="#dss_smallvector">SmallVector</a> as part of
1255its ID process.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1256
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1257<p>Consider a case where you want to implement a "getOrCreateFoo" method for
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1258a complex object (for example, a node in the code generator). The client has a
1259description of *what* it wants to generate (it knows the opcode and all the
1260operands), 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]1261only to find out it already exists, at which point we would have to delete it
1262and return the node that already exists.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1263</p>
1264
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1265<p>To support this style of client, FoldingSet perform a query with a
1266FoldingSetNodeID (which wraps SmallVector) that can be used to describe the
1267element that we want to query for. The query either returns the element
1268matching the ID or it returns an opaque ID that indicates where insertion should
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1269take place. Construction of the ID usually does not require heap traffic.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1270
1271<p>Because FoldingSet uses intrusive links, it can support polymorphic objects
1272in the set (for example, you can have SDNode instances mixed with LoadSDNodes).
1273Because the elements are individually allocated, pointers to the elements are
1274stable: inserting or removing elements does not invalidate any pointers to other
1275elements.
1276</p>
1277
1278</div>
1279
1280<!-- _______________________________________________________________________ -->
1281<div class="doc_subsubsection">
1282 <a name="dss_set">&lt;set&gt;</a>
1283</div>
1284
1285<div class="doc_text">
1286
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1287<p><tt>std::set</tt> is a reasonable all-around set class, which is decent at
1288many things but great at nothing. std::set allocates memory for each element
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1289inserted (thus it is very malloc intensive) and typically stores three pointers
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1290per element in the set (thus adding a large amount of per-element space
1291overhead). It offers guaranteed log(n) performance, which is not particularly
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1292fast from a complexity standpoint (particularly if the elements of the set are
1293expensive to compare, like strings), and has extremely high constant factors for
1294lookup, insertion and removal.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1295
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1296<p>The advantages of std::set are that its iterators are stable (deleting or
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1297inserting an element from the set does not affect iterators or pointers to other
1298elements) and that iteration over the set is guaranteed to be in sorted order.
1299If the elements in the set are large, then the relative overhead of the pointers
1300and malloc traffic is not a big deal, but if the elements of the set are small,
1301std::set is almost never a good choice.</p>
1302
1303</div>
1304
1305<!-- _______________________________________________________________________ -->
1306<div class="doc_subsubsection">
1307 <a name="dss_setvector">"llvm/ADT/SetVector.h"</a>
1308</div>
1309
1310<div class="doc_text">
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1311<p>LLVM's SetVector&lt;Type&gt; is an adapter class that combines your choice of
1312a set-like container along with a <a href="#ds_sequential">Sequential
1313Container</a>. The important property
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1314that this provides is efficient insertion with uniquing (duplicate elements are
1315ignored) with iteration support. It implements this by inserting elements into
1316both a set-like container and the sequential container, using the set-like
1317container for uniquing and the sequential container for iteration.
1318</p>
1319
1320<p>The difference between SetVector and other sets is that the order of
1321iteration is guaranteed to match the order of insertion into the SetVector.
1322This property is really important for things like sets of pointers. Because
1323pointer values are non-deterministic (e.g. vary across runs of the program on
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1324different machines), iterating over the pointers in the set will
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1325not be in a well-defined order.</p>
1326
1327<p>
1328The drawback of SetVector is that it requires twice as much space as a normal
1329set and has the sum of constant factors from the set-like container and the
1330sequential container that it uses. Use it *only* if you need to iterate over
1331the elements in a deterministic order. SetVector is also expensive to delete
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1332elements out of (linear time), unless you use it's "pop_back" method, which is
1333faster.
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1334</p>
1335
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1336<p>SetVector is an adapter class that defaults to using std::vector and std::set
1337for the underlying containers, so it is quite expensive. However,
1338<tt>"llvm/ADT/SetVector.h"</tt> also provides a SmallSetVector class, which
1339defaults to using a SmallVector and SmallSet of a specified size. If you use
1340this, and if your sets are dynamically smaller than N, you will save a lot of
1341heap traffic.</p>
1342
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1343</div>
1344
1345<!-- _______________________________________________________________________ -->
1346<div class="doc_subsubsection">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1347 <a name="dss_uniquevector">"llvm/ADT/UniqueVector.h"</a>
1348</div>
1349
1350<div class="doc_text">
1351
1352<p>
1353UniqueVector is similar to <a href="#dss_setvector">SetVector</a>, but it
1354retains a unique ID for each element inserted into the set. It internally
1355contains a map and a vector, and it assigns a unique ID for each value inserted
1356into the set.</p>
1357
1358<p>UniqueVector is very expensive: its cost is the sum of the cost of
1359maintaining both the map and vector, it has high complexity, high constant
1360factors, and produces a lot of malloc traffic. It should be avoided.</p>
1361
1362</div>
1363
1364
1365<!-- _______________________________________________________________________ -->
1366<div class="doc_subsubsection">
1367 <a name="dss_otherset">Other Set-Like Container Options</a>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1368</div>
1369
1370<div class="doc_text">
1371
1372<p>
1373The STL provides several other options, such as std::multiset and the various
Chris Lattnerf1a30822009-03-09 05:20:45 +0000[diff] [blame]1374"hash_set" like containers (whether from C++ TR1 or from the SGI library). We
1375never use hash_set and unordered_set because they are generally very expensive
1376(each insertion requires a malloc) and very non-portable.
1377</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1378
1379<p>std::multiset is useful if you're not interested in elimination of
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1380duplicates, but has all the drawbacks of std::set. A sorted vector (where you
1381don't delete duplicate entries) or some other approach is almost always
1382better.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1383
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1384</div>
1385
1386<!-- ======================================================================= -->
1387<div class="doc_subsection">
1388 <a name="ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
1389</div>
1390
1391<div class="doc_text">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1392Map-like containers are useful when you want to associate data to a key. As
1393usual, there are a lot of different ways to do this. :)
1394</div>
1395
1396<!-- _______________________________________________________________________ -->
1397<div class="doc_subsubsection">
1398 <a name="dss_sortedvectormap">A sorted 'vector'</a>
1399</div>
1400
1401<div class="doc_text">
1402
1403<p>
1404If your usage pattern follows a strict insert-then-query approach, you can
1405trivially use the same approach as <a href="#dss_sortedvectorset">sorted vectors
1406for set-like containers</a>. The only difference is that your query function
1407(which uses std::lower_bound to get efficient log(n) lookup) should only compare
1408the key, not both the key and value. This yields the same advantages as sorted
1409vectors for sets.
1410</p>
1411</div>
1412
1413<!-- _______________________________________________________________________ -->
1414<div class="doc_subsubsection">
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1415 <a name="dss_stringmap">"llvm/ADT/StringMap.h"</a>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1416</div>
1417
1418<div class="doc_text">
1419
1420<p>
1421Strings are commonly used as keys in maps, and they are difficult to support
1422efficiently: they are variable length, inefficient to hash and compare when
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1423long, expensive to copy, etc. StringMap is a specialized container designed to
1424cope with these issues. It supports mapping an arbitrary range of bytes to an
1425arbitrary other object.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1426
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1427<p>The StringMap implementation uses a quadratically-probed hash table, where
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1428the buckets store a pointer to the heap allocated entries (and some other
1429stuff). The entries in the map must be heap allocated because the strings are
1430variable length. The string data (key) and the element object (value) are
1431stored in the same allocation with the string data immediately after the element
1432object. This container guarantees the "<tt>(char*)(&amp;Value+1)</tt>" points
1433to the key string for a value.</p>
1434
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1435<p>The StringMap is very fast for several reasons: quadratic probing is very
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1436cache efficient for lookups, the hash value of strings in buckets is not
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1437recomputed when lookup up an element, StringMap rarely has to touch the
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1438memory for unrelated objects when looking up a value (even when hash collisions
1439happen), hash table growth does not recompute the hash values for strings
1440already in the table, and each pair in the map is store in a single allocation
1441(the string data is stored in the same allocation as the Value of a pair).</p>
1442
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1443<p>StringMap also provides query methods that take byte ranges, so it only ever
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1444copies a string if a value is inserted into the table.</p>
1445</div>
1446
1447<!-- _______________________________________________________________________ -->
1448<div class="doc_subsubsection">
1449 <a name="dss_indexedmap">"llvm/ADT/IndexedMap.h"</a>
1450</div>
1451
1452<div class="doc_text">
1453<p>
1454IndexedMap is a specialized container for mapping small dense integers (or
1455values that can be mapped to small dense integers) to some other type. It is
1456internally implemented as a vector with a mapping function that maps the keys to
1457the dense integer range.
1458</p>
1459
1460<p>
1461This is useful for cases like virtual registers in the LLVM code generator: they
1462have a dense mapping that is offset by a compile-time constant (the first
1463virtual register ID).</p>
1464
1465</div>
1466
1467<!-- _______________________________________________________________________ -->
1468<div class="doc_subsubsection">
1469 <a name="dss_densemap">"llvm/ADT/DenseMap.h"</a>
1470</div>
1471
1472<div class="doc_text">
1473
1474<p>
1475DenseMap is a simple quadratically probed hash table. It excels at supporting
1476small keys and values: it uses a single allocation to hold all of the pairs that
1477are currently inserted in the map. DenseMap is a great way to map pointers to
1478pointers, or map other small types to each other.
1479</p>
1480
1481<p>
1482There are several aspects of DenseMap that you should be aware of, however. The
1483iterators in a densemap are invalidated whenever an insertion occurs, unlike
1484map. Also, because DenseMap allocates space for a large number of key/value
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1485pairs (it starts with 64 by default), it will waste a lot of space if your keys
1486or values are large. Finally, you must implement a partial specialization of
Chris Lattner76c1b972007-09-17 18:34:04 +0000[diff] [blame]1487DenseMapInfo for the key that you want, if it isn't already supported. This
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1488is required to tell DenseMap about two special marker values (which can never be
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1489inserted into the map) that it needs internally.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1490
1491</div>
1492
1493<!-- _______________________________________________________________________ -->
1494<div class="doc_subsubsection">
1495 <a name="dss_map">&lt;map&gt;</a>
1496</div>
1497
1498<div class="doc_text">
1499
1500<p>
1501std::map has similar characteristics to <a href="#dss_set">std::set</a>: it uses
1502a single allocation per pair inserted into the map, it offers log(n) lookup with
1503an extremely large constant factor, imposes a space penalty of 3 pointers per
1504pair in the map, etc.</p>
1505
1506<p>std::map is most useful when your keys or values are very large, if you need
1507to iterate over the collection in sorted order, or if you need stable iterators
1508into the map (i.e. they don't get invalidated if an insertion or deletion of
1509another element takes place).</p>
1510
1511</div>
1512
1513<!-- _______________________________________________________________________ -->
1514<div class="doc_subsubsection">
1515 <a name="dss_othermap">Other Map-Like Container Options</a>
1516</div>
1517
1518<div class="doc_text">
1519
1520<p>
1521The STL provides several other options, such as std::multimap and the various
Chris Lattnerf1a30822009-03-09 05:20:45 +0000[diff] [blame]1522"hash_map" like containers (whether from C++ TR1 or from the SGI library). We
1523never use hash_set and unordered_set because they are generally very expensive
1524(each insertion requires a malloc) and very non-portable.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1525
1526<p>std::multimap is useful if you want to map a key to multiple values, but has
1527all the drawbacks of std::map. A sorted vector or some other approach is almost
1528always better.</p>
1529
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1530</div>
1531
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1532<!-- ======================================================================= -->
1533<div class="doc_subsection">
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]1534 <a name="ds_string">String-like containers</a>
1535</div>
1536
1537<div class="doc_text">
1538
1539<p>
1540TODO: const char* vs stringref vs smallstring vs std::string. Describe twine,
1541xref to #string_apis.
1542</p>
1543
1544</div>
1545
1546<!-- ======================================================================= -->
1547<div class="doc_subsection">
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1548 <a name="ds_bit">Bit storage containers (BitVector, SparseBitVector)</a>
1549</div>
1550
1551<div class="doc_text">
Chris Lattner7086ce72007-09-25 22:37:50 +0000[diff] [blame]1552<p>Unlike the other containers, there are only two bit storage containers, and
1553choosing when to use each is relatively straightforward.</p>
1554
1555<p>One additional option is
1556<tt>std::vector&lt;bool&gt;</tt>: we discourage its use for two reasons 1) the
1557implementation in many common compilers (e.g. commonly available versions of
1558GCC) is extremely inefficient and 2) the C++ standards committee is likely to
1559deprecate this container and/or change it significantly somehow. In any case,
1560please don't use it.</p>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1561</div>
1562
1563<!-- _______________________________________________________________________ -->
1564<div class="doc_subsubsection">
1565 <a name="dss_bitvector">BitVector</a>
1566</div>
1567
1568<div class="doc_text">
1569<p> The BitVector container provides a fixed size set of bits for manipulation.
1570It supports individual bit setting/testing, as well as set operations. The set
1571operations take time O(size of bitvector), but operations are performed one word
1572at a time, instead of one bit at a time. This makes the BitVector very fast for
1573set operations compared to other containers. Use the BitVector when you expect
1574the number of set bits to be high (IE a dense set).
1575</p>
1576</div>
1577
1578<!-- _______________________________________________________________________ -->
1579<div class="doc_subsubsection">
1580 <a name="dss_sparsebitvector">SparseBitVector</a>
1581</div>
1582
1583<div class="doc_text">
1584<p> The SparseBitVector container is much like BitVector, with one major
1585difference: Only the bits that are set, are stored. This makes the
1586SparseBitVector much more space efficient than BitVector when the set is sparse,
1587as well as making set operations O(number of set bits) instead of O(size of
1588universe). 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
1589(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).
1590</p>
1591</div>
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]1592
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1593<!-- *********************************************************************** -->
1594<div class="doc_section">
1595 <a name="common">Helpful Hints for Common Operations</a>
1596</div>
1597<!-- *********************************************************************** -->
1598
1599<div class="doc_text">
1600
1601<p>This section describes how to perform some very simple transformations of
1602LLVM code. This is meant to give examples of common idioms used, showing the
1603practical side of LLVM transformations. <p> Because this is a "how-to" section,
1604you should also read about the main classes that you will be working with. The
1605<a href="#coreclasses">Core LLVM Class Hierarchy Reference</a> contains details
1606and descriptions of the main classes that you should know about.</p>
1607
1608</div>
1609
1610<!-- NOTE: this section should be heavy on example code -->
1611<!-- ======================================================================= -->
1612<div class="doc_subsection">
1613 <a name="inspection">Basic Inspection and Traversal Routines</a>
1614</div>
1615
1616<div class="doc_text">
1617
1618<p>The LLVM compiler infrastructure have many different data structures that may
1619be traversed. Following the example of the C++ standard template library, the
1620techniques used to traverse these various data structures are all basically the
1621same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
1622method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
1623function returns an iterator pointing to one past the last valid element of the
1624sequence, and there is some <tt>XXXiterator</tt> data type that is common
1625between the two operations.</p>
1626
1627<p>Because the pattern for iteration is common across many different aspects of
1628the program representation, the standard template library algorithms may be used
1629on them, and it is easier to remember how to iterate. First we show a few common
1630examples of the data structures that need to be traversed. Other data
1631structures are traversed in very similar ways.</p>
1632
1633</div>
1634
1635<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1636<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1637 <a name="iterate_function">Iterating over the </a><a
1638 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
1639 href="#Function"><tt>Function</tt></a>
1640</div>
1641
1642<div class="doc_text">
1643
1644<p>It's quite common to have a <tt>Function</tt> instance that you'd like to
1645transform in some way; in particular, you'd like to manipulate its
1646<tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over all of
1647the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>. The following is
1648an example that prints the name of a <tt>BasicBlock</tt> and the number of
1649<tt>Instruction</tt>s it contains:</p>
1650
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1651<div class="doc_code">
1652<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1653// <i>func is a pointer to a Function instance</i>
1654for (Function::iterator i = func-&gt;begin(), e = func-&gt;end(); i != e; ++i)
1655 // <i>Print out the name of the basic block if it has one, and then the</i>
1656 // <i>number of instructions that it contains</i>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1657 llvm::cerr &lt;&lt; "Basic block (name=" &lt;&lt; i-&gt;getName() &lt;&lt; ") has "
1658 &lt;&lt; i-&gt;size() &lt;&lt; " instructions.\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1659</pre>
1660</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1661
1662<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]1663invoking member functions of the <tt>Instruction</tt> class. This is
1664because the indirection operator is overloaded for the iterator
Chris Lattner7496ec52003-08-05 22:54:23 +0000[diff] [blame]1665classes. In the above code, the expression <tt>i-&gt;size()</tt> is
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1666exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.</p>
1667
1668</div>
1669
1670<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1671<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1672 <a name="iterate_basicblock">Iterating over the </a><a
1673 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1674 href="#BasicBlock"><tt>BasicBlock</tt></a>
1675</div>
1676
1677<div class="doc_text">
1678
1679<p>Just like when dealing with <tt>BasicBlock</tt>s in <tt>Function</tt>s, it's
1680easy to iterate over the individual instructions that make up
1681<tt>BasicBlock</tt>s. Here's a code snippet that prints out each instruction in
1682a <tt>BasicBlock</tt>:</p>
1683
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1684<div class="doc_code">
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1685<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1686// <i>blk is a pointer to a BasicBlock instance</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1687for (BasicBlock::iterator i = blk-&gt;begin(), e = blk-&gt;end(); i != e; ++i)
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1688 // <i>The next statement works since operator&lt;&lt;(ostream&amp;,...)</i>
1689 // <i>is overloaded for Instruction&amp;</i>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1690 llvm::cerr &lt;&lt; *i &lt;&lt; "\n";
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1691</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1692</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1693
1694<p>However, this isn't really the best way to print out the contents of a
1695<tt>BasicBlock</tt>! Since the ostream operators are overloaded for virtually
1696anything you'll care about, you could have just invoked the print routine on the
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1697basic block itself: <tt>llvm::cerr &lt;&lt; *blk &lt;&lt; "\n";</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1698
1699</div>
1700
1701<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1702<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1703 <a name="iterate_institer">Iterating over the </a><a
1704 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1705 href="#Function"><tt>Function</tt></a>
1706</div>
1707
1708<div class="doc_text">
1709
1710<p>If you're finding that you commonly iterate over a <tt>Function</tt>'s
1711<tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s <tt>Instruction</tt>s,
1712<tt>InstIterator</tt> should be used instead. You'll need to include <a
1713href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>,
1714and then instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1715small 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]1716
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1717<div class="doc_code">
1718<pre>
1719#include "<a href="/doxygen/InstIterator_8h-source.html">llvm/Support/InstIterator.h</a>"
1720
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1721// <i>F is a pointer to a Function instance</i>
Chris Lattnerda021aa2008-06-04 18:20:42 +0000[diff] [blame]1722for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
1723 llvm::cerr &lt;&lt; *I &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1724</pre>
1725</div>
1726
1727<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]1728work list with its initial contents. For example, if you wanted to
1729initialize a work list to contain all instructions in a <tt>Function</tt>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1730F, all you would need to do is something like:</p>
1731
1732<div class="doc_code">
1733<pre>
1734std::set&lt;Instruction*&gt; worklist;
Chris Lattnerda021aa2008-06-04 18:20:42 +0000[diff] [blame]1735// or better yet, SmallPtrSet&lt;Instruction*, 64&gt; worklist;
1736
1737for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
1738 worklist.insert(&amp;*I);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1739</pre>
1740</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1741
1742<p>The STL set <tt>worklist</tt> would now contain all instructions in the
1743<tt>Function</tt> pointed to by F.</p>
1744
1745</div>
1746
1747<!-- _______________________________________________________________________ -->
1748<div class="doc_subsubsection">
1749 <a name="iterate_convert">Turning an iterator into a class pointer (and
1750 vice-versa)</a>
1751</div>
1752
1753<div class="doc_text">
1754
1755<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]1756instance when all you've got at hand is an iterator. Well, extracting
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1757a reference or a pointer from an iterator is very straight-forward.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1758Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and <tt>j</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1759is a <tt>BasicBlock::const_iterator</tt>:</p>
1760
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1761<div class="doc_code">
1762<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1763Instruction&amp; inst = *i; // <i>Grab reference to instruction reference</i>
1764Instruction* pinst = &amp;*i; // <i>Grab pointer to instruction reference</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1765const Instruction&amp; inst = *j;
1766</pre>
1767</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1768
1769<p>However, the iterators you'll be working with in the LLVM framework are
1770special: they will automatically convert to a ptr-to-instance type whenever they
1771need to. Instead of dereferencing the iterator and then taking the address of
1772the result, you can simply assign the iterator to the proper pointer type and
1773you get the dereference and address-of operation as a result of the assignment
1774(behind the scenes, this is a result of overloading casting mechanisms). Thus
1775the last line of the last example,</p>
1776
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1777<div class="doc_code">
1778<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1779Instruction *pinst = &amp;*i;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1780</pre>
1781</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1782
1783<p>is semantically equivalent to</p>
1784
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1785<div class="doc_code">
1786<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1787Instruction *pinst = i;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1788</pre>
1789</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1790
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1791<p>It's also possible to turn a class pointer into the corresponding iterator,
1792and this is a constant time operation (very efficient). The following code
1793snippet illustrates use of the conversion constructors provided by LLVM
1794iterators. By using these, you can explicitly grab the iterator of something
1795without actually obtaining it via iteration over some structure:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1796
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1797<div class="doc_code">
1798<pre>
1799void printNextInstruction(Instruction* inst) {
1800 BasicBlock::iterator it(inst);
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1801 ++it; // <i>After this line, it refers to the instruction after *inst</i>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1802 if (it != inst-&gt;getParent()-&gt;end()) llvm::cerr &lt;&lt; *it &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1803}
1804</pre>
1805</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1806
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1807</div>
1808
1809<!--_______________________________________________________________________-->
1810<div class="doc_subsubsection">
1811 <a name="iterate_complex">Finding call sites: a slightly more complex
1812 example</a>
1813</div>
1814
1815<div class="doc_text">
1816
1817<p>Say that you're writing a FunctionPass and would like to count all the
1818locations in the entire module (that is, across every <tt>Function</tt>) where a
1819certain function (i.e., some <tt>Function</tt>*) is already in scope. As you'll
1820learn 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]1821much more straight-forward manner, but this example will allow us to explore how
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1822you'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]1823is what we want to do:</p>
1824
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1825<div class="doc_code">
1826<pre>
1827initialize callCounter to zero
1828for each Function f in the Module
1829 for each BasicBlock b in f
1830 for each Instruction i in b
1831 if (i is a CallInst and calls the given function)
1832 increment callCounter
1833</pre>
1834</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1835
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1836<p>And the actual code is (remember, because we're writing a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1837<tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply has to
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1838override the <tt>runOnFunction</tt> method):</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>
1842Function* targetFunc = ...;
1843
1844class OurFunctionPass : public FunctionPass {
1845 public:
1846 OurFunctionPass(): callCounter(0) { }
1847
1848 virtual runOnFunction(Function&amp; F) {
1849 for (Function::iterator b = F.begin(), be = F.end(); b != be; ++b) {
Eric Christopher203e71d2008-11-08 08:20:49 +0000[diff] [blame]1850 for (BasicBlock::iterator i = b-&gt;begin(), ie = b-&gt;end(); i != ie; ++i) {
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1851 if (<a href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a>&lt;<a
1852 href="#CallInst">CallInst</a>&gt;(&amp;*i)) {
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1853 // <i>We know we've encountered a call instruction, so we</i>
1854 // <i>need to determine if it's a call to the</i>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1855 // <i>function pointed to by m_func or not.</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1856 if (callInst-&gt;getCalledFunction() == targetFunc)
1857 ++callCounter;
1858 }
1859 }
1860 }
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1861 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1862
1863 private:
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1864 unsigned callCounter;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1865};
1866</pre>
1867</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1868
1869</div>
1870
Brian Gaekef1972c62003-11-07 19:25:45 +0000[diff] [blame]1871<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1872<div class="doc_subsubsection">
1873 <a name="calls_and_invokes">Treating calls and invokes the same way</a>
1874</div>
1875
1876<div class="doc_text">
1877
1878<p>You may have noticed that the previous example was a bit oversimplified in
1879that it did not deal with call sites generated by 'invoke' instructions. In
1880this, and in other situations, you may find that you want to treat
1881<tt>CallInst</tt>s and <tt>InvokeInst</tt>s the same way, even though their
1882most-specific common base class is <tt>Instruction</tt>, which includes lots of
1883less closely-related things. For these cases, LLVM provides a handy wrapper
1884class called <a
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]1885href="http://llvm.org/doxygen/classllvm_1_1CallSite.html"><tt>CallSite</tt></a>.
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1886It is essentially a wrapper around an <tt>Instruction</tt> pointer, with some
1887methods that provide functionality common to <tt>CallInst</tt>s and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1888<tt>InvokeInst</tt>s.</p>
1889
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1890<p>This class has "value semantics": it should be passed by value, not by
1891reference and it should not be dynamically allocated or deallocated using
1892<tt>operator new</tt> or <tt>operator delete</tt>. It is efficiently copyable,
1893assignable and constructable, with costs equivalents to that of a bare pointer.
1894If you look at its definition, it has only a single pointer member.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1895
1896</div>
1897
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]1898<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1899<div class="doc_subsubsection">
1900 <a name="iterate_chains">Iterating over def-use &amp; use-def chains</a>
1901</div>
1902
1903<div class="doc_text">
1904
1905<p>Frequently, we might have an instance of the <a
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]1906href="/doxygen/classllvm_1_1Value.html">Value Class</a> and we want to
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]1907determine which <tt>User</tt>s use the <tt>Value</tt>. The list of all
1908<tt>User</tt>s of a particular <tt>Value</tt> is called a <i>def-use</i> chain.
1909For example, let's say we have a <tt>Function*</tt> named <tt>F</tt> to a
1910particular function <tt>foo</tt>. Finding all of the instructions that
1911<i>use</i> <tt>foo</tt> is as simple as iterating over the <i>def-use</i> chain
1912of <tt>F</tt>:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1913
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1914<div class="doc_code">
1915<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1916Function *F = ...;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1917
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1918for (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]1919 if (Instruction *Inst = dyn_cast&lt;Instruction&gt;(*i)) {
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1920 llvm::cerr &lt;&lt; "F is used in instruction:\n";
1921 llvm::cerr &lt;&lt; *Inst &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1922 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1923</pre>
1924</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1925
1926<p>Alternately, it's common to have an instance of the <a
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]1927href="/doxygen/classllvm_1_1User.html">User Class</a> and need to know what
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1928<tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used by a
1929<tt>User</tt> is known as a <i>use-def</i> chain. Instances of class
1930<tt>Instruction</tt> are common <tt>User</tt>s, so we might want to iterate over
1931all of the values that a particular instruction uses (that is, the operands of
1932the particular <tt>Instruction</tt>):</p>
1933
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1934<div class="doc_code">
1935<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1936Instruction *pi = ...;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1937
1938for (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]1939 Value *v = *i;
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1940 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1941}
1942</pre>
1943</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1944
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]1945<!--
1946 def-use chains ("finding all users of"): Value::use_begin/use_end
1947 use-def chains ("finding all values used"): User::op_begin/op_end [op=operand]
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1948-->
1949
1950</div>
1951
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1952<!--_______________________________________________________________________-->
1953<div class="doc_subsubsection">
1954 <a name="iterate_preds">Iterating over predecessors &amp;
1955successors of blocks</a>
1956</div>
1957
1958<div class="doc_text">
1959
1960<p>Iterating over the predecessors and successors of a block is quite easy
1961with the routines defined in <tt>"llvm/Support/CFG.h"</tt>. Just use code like
1962this to iterate over all predecessors of BB:</p>
1963
1964<div class="doc_code">
1965<pre>
1966#include "llvm/Support/CFG.h"
1967BasicBlock *BB = ...;
1968
1969for (pred_iterator PI = pred_begin(BB), E = pred_end(BB); PI != E; ++PI) {
1970 BasicBlock *Pred = *PI;
1971 // <i>...</i>
1972}
1973</pre>
1974</div>
1975
1976<p>Similarly, to iterate over successors use
1977succ_iterator/succ_begin/succ_end.</p>
1978
1979</div>
1980
1981
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1982<!-- ======================================================================= -->
1983<div class="doc_subsection">
1984 <a name="simplechanges">Making simple changes</a>
1985</div>
1986
1987<div class="doc_text">
1988
1989<p>There are some primitive transformation operations present in the LLVM
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]1990infrastructure that are worth knowing about. When performing
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1991transformations, it's fairly common to manipulate the contents of basic
1992blocks. This section describes some of the common methods for doing so
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1993and gives example code.</p>
1994
1995</div>
1996
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1997<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1998<div class="doc_subsubsection">
1999 <a name="schanges_creating">Creating and inserting new
2000 <tt>Instruction</tt>s</a>
2001</div>
2002
2003<div class="doc_text">
2004
2005<p><i>Instantiating Instructions</i></p>
2006
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]2007<p>Creation of <tt>Instruction</tt>s is straight-forward: simply call the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2008constructor for the kind of instruction to instantiate and provide the necessary
2009parameters. For example, an <tt>AllocaInst</tt> only <i>requires</i> a
2010(const-ptr-to) <tt>Type</tt>. Thus:</p>
2011
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2012<div class="doc_code">
2013<pre>
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2014AllocaInst* ai = new AllocaInst(Type::Int32Ty);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2015</pre>
2016</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2017
2018<p>will create an <tt>AllocaInst</tt> instance that represents the allocation of
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]2019one integer in the current stack frame, at run time. Each <tt>Instruction</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2020subclass is likely to have varying default parameters which change the semantics
2021of the instruction, so refer to the <a
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]2022href="/doxygen/classllvm_1_1Instruction.html">doxygen documentation for the subclass of
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2023Instruction</a> that you're interested in instantiating.</p>
2024
2025<p><i>Naming values</i></p>
2026
2027<p>It is very useful to name the values of instructions when you're able to, as
2028this facilitates the debugging of your transformations. If you end up looking
2029at generated LLVM machine code, you definitely want to have logical names
2030associated with the results of instructions! By supplying a value for the
2031<tt>Name</tt> (default) parameter of the <tt>Instruction</tt> constructor, you
2032associate a logical name with the result of the instruction's execution at
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]2033run time. For example, say that I'm writing a transformation that dynamically
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2034allocates space for an integer on the stack, and that integer is going to be
2035used as some kind of index by some other code. To accomplish this, I place an
2036<tt>AllocaInst</tt> at the first point in the first <tt>BasicBlock</tt> of some
2037<tt>Function</tt>, and I'm intending to use it within the same
2038<tt>Function</tt>. I might do:</p>
2039
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2040<div class="doc_code">
2041<pre>
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2042AllocaInst* pa = new AllocaInst(Type::Int32Ty, 0, "indexLoc");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2043</pre>
2044</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2045
2046<p>where <tt>indexLoc</tt> is now the logical name of the instruction's
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]2047execution value, which is a pointer to an integer on the run time stack.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2048
2049<p><i>Inserting instructions</i></p>
2050
2051<p>There are essentially two ways to insert an <tt>Instruction</tt>
2052into an existing sequence of instructions that form a <tt>BasicBlock</tt>:</p>
2053
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]2054<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2055 <li>Insertion into an explicit instruction list
2056
2057 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within that
2058 <tt>BasicBlock</tt>, and a newly-created instruction we wish to insert
2059 before <tt>*pi</tt>, we do the following: </p>
2060
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2061<div class="doc_code">
2062<pre>
2063BasicBlock *pb = ...;
2064Instruction *pi = ...;
2065Instruction *newInst = new Instruction(...);
2066
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2067pb-&gt;getInstList().insert(pi, newInst); // <i>Inserts newInst before pi in pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2068</pre>
2069</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]2070
2071 <p>Appending to the end of a <tt>BasicBlock</tt> is so common that
2072 the <tt>Instruction</tt> class and <tt>Instruction</tt>-derived
2073 classes provide constructors which take a pointer to a
2074 <tt>BasicBlock</tt> to be appended to. For example code that
2075 looked like: </p>
2076
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2077<div class="doc_code">
2078<pre>
2079BasicBlock *pb = ...;
2080Instruction *newInst = new Instruction(...);
2081
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2082pb-&gt;getInstList().push_back(newInst); // <i>Appends newInst to pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2083</pre>
2084</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]2085
2086 <p>becomes: </p>
2087
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2088<div class="doc_code">
2089<pre>
2090BasicBlock *pb = ...;
2091Instruction *newInst = new Instruction(..., pb);
2092</pre>
2093</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]2094
2095 <p>which is much cleaner, especially if you are creating
2096 long instruction streams.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2097
2098 <li>Insertion into an implicit instruction list
2099
2100 <p><tt>Instruction</tt> instances that are already in <tt>BasicBlock</tt>s
2101 are implicitly associated with an existing instruction list: the instruction
2102 list of the enclosing basic block. Thus, we could have accomplished the same
2103 thing as the above code without being given a <tt>BasicBlock</tt> by doing:
2104 </p>
2105
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2106<div class="doc_code">
2107<pre>
2108Instruction *pi = ...;
2109Instruction *newInst = new Instruction(...);
2110
2111pi-&gt;getParent()-&gt;getInstList().insert(pi, newInst);
2112</pre>
2113</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2114
2115 <p>In fact, this sequence of steps occurs so frequently that the
2116 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes provide
2117 constructors which take (as a default parameter) a pointer to an
2118 <tt>Instruction</tt> which the newly-created <tt>Instruction</tt> should
2119 precede. That is, <tt>Instruction</tt> constructors are capable of
2120 inserting the newly-created instance into the <tt>BasicBlock</tt> of a
2121 provided instruction, immediately before that instruction. Using an
2122 <tt>Instruction</tt> constructor with a <tt>insertBefore</tt> (default)
2123 parameter, the above code becomes:</p>
2124
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2125<div class="doc_code">
2126<pre>
2127Instruction* pi = ...;
2128Instruction* newInst = new Instruction(..., pi);
2129</pre>
2130</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2131
2132 <p>which is much cleaner, especially if you're creating a lot of
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2133 instructions and adding them to <tt>BasicBlock</tt>s.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2134</ul>
2135
2136</div>
2137
2138<!--_______________________________________________________________________-->
2139<div class="doc_subsubsection">
2140 <a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a>
2141</div>
2142
2143<div class="doc_text">
2144
2145<p>Deleting an instruction from an existing sequence of instructions that form a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]2146<a href="#BasicBlock"><tt>BasicBlock</tt></a> is very straight-forward. First,
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2147you must have a pointer to the instruction that you wish to delete. Second, you
2148need to obtain the pointer to that instruction's basic block. You use the
2149pointer to the basic block to get its list of instructions and then use the
2150erase function to remove your instruction. For example:</p>
2151
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2152<div class="doc_code">
2153<pre>
2154<a href="#Instruction">Instruction</a> *I = .. ;
Chris Lattner9f8ec252008-02-15 22:57:17 +0000[diff] [blame]2155I-&gt;eraseFromParent();
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2156</pre>
2157</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2158
2159</div>
2160
2161<!--_______________________________________________________________________-->
2162<div class="doc_subsubsection">
2163 <a name="schanges_replacing">Replacing an <tt>Instruction</tt> with another
2164 <tt>Value</tt></a>
2165</div>
2166
2167<div class="doc_text">
2168
2169<p><i>Replacing individual instructions</i></p>
2170
2171<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]2172permits use of two very useful replace functions: <tt>ReplaceInstWithValue</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2173and <tt>ReplaceInstWithInst</tt>.</p>
2174
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2175<h4><a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a></h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2176
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2177<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2178 <li><tt>ReplaceInstWithValue</tt>
2179
Nick Lewyckyb6d1f392008-09-15 06:31:52 +0000[diff] [blame]2180 <p>This function replaces all uses of a given instruction with a value,
2181 and then removes the original instruction. The following example
2182 illustrates the replacement of the result of a particular
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]2183 <tt>AllocaInst</tt> that allocates memory for a single integer with a null
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2184 pointer to an integer.</p>
2185
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2186<div class="doc_code">
2187<pre>
2188AllocaInst* instToReplace = ...;
2189BasicBlock::iterator ii(instToReplace);
2190
2191ReplaceInstWithValue(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
Daniel Dunbar58c2ac02008-10-03 22:17:25 +0000[diff] [blame]2192 Constant::getNullValue(PointerType::getUnqual(Type::Int32Ty)));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2193</pre></div></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2194
2195 <li><tt>ReplaceInstWithInst</tt>
2196
2197 <p>This function replaces a particular instruction with another
Nick Lewyckyb6d1f392008-09-15 06:31:52 +0000[diff] [blame]2198 instruction, inserting the new instruction into the basic block at the
2199 location where the old instruction was, and replacing any uses of the old
2200 instruction with the new instruction. The following example illustrates
2201 the replacement of one <tt>AllocaInst</tt> with another.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2202
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2203<div class="doc_code">
2204<pre>
2205AllocaInst* instToReplace = ...;
2206BasicBlock::iterator ii(instToReplace);
2207
2208ReplaceInstWithInst(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2209 new AllocaInst(Type::Int32Ty, 0, "ptrToReplacedInt"));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2210</pre></div></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2211</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2212
2213<p><i>Replacing multiple uses of <tt>User</tt>s and <tt>Value</tt>s</i></p>
2214
2215<p>You can use <tt>Value::replaceAllUsesWith</tt> and
2216<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]2217doxygen documentation for the <a href="/doxygen/classllvm_1_1Value.html">Value Class</a>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]2218and <a href="/doxygen/classllvm_1_1User.html">User Class</a>, respectively, for more
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2219information.</p>
2220
2221<!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
2222include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
2223ReplaceInstWithValue, ReplaceInstWithInst -->
2224
2225</div>
2226
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2227<!--_______________________________________________________________________-->
2228<div class="doc_subsubsection">
2229 <a name="schanges_deletingGV">Deleting <tt>GlobalVariable</tt>s</a>
2230</div>
2231
2232<div class="doc_text">
2233
Tanya Lattnerc5dfcdb2007-06-20 20:46:37 +0000[diff] [blame]2234<p>Deleting a global variable from a module is just as easy as deleting an
2235Instruction. First, you must have a pointer to the global variable that you wish
2236 to delete. You use this pointer to erase it from its parent, the module.
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2237 For example:</p>
2238
2239<div class="doc_code">
2240<pre>
2241<a href="#GlobalVariable">GlobalVariable</a> *GV = .. ;
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2242
Tanya Lattnerc5dfcdb2007-06-20 20:46:37 +0000[diff] [blame]2243GV-&gt;eraseFromParent();
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2244</pre>
2245</div>
2246
2247</div>
2248
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2249<!-- ======================================================================= -->
2250<div class="doc_subsection">
2251 <a name="create_types">How to Create Types</a>
2252</div>
2253
2254<div class="doc_text">
2255
2256<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]2257statically, you can use <tt>TypeBuilder&lt;...&gt;::get()</tt>, defined
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2258in <tt>llvm/Support/TypeBuilder.h</tt>, to retrieve them. <tt>TypeBuilder</tt>
2259has two forms depending on whether you're building types for cross-compilation
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2260or native library use. <tt>TypeBuilder&lt;T, true&gt;</tt> requires
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2261that <tt>T</tt> be independent of the host environment, meaning that it's built
2262out of types from
2263the <a href="/doxygen/namespacellvm_1_1types.html"><tt>llvm::types</tt></a>
2264namespace and pointers, functions, arrays, etc. built of
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2265those. <tt>TypeBuilder&lt;T, false&gt;</tt> additionally allows native C types
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2266whose size may depend on the host compiler. For example,</p>
2267
2268<div class="doc_code">
2269<pre>
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2270FunctionType *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]2271</pre>
2272</div>
2273
2274<p>is easier to read and write than the equivalent</p>
2275
2276<div class="doc_code">
2277<pre>
Owen Anderson5e8c50e2009-06-16 17:40:28 +0000[diff] [blame]2278std::vector&lt;const Type*&gt; params;
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2279params.push_back(PointerType::getUnqual(Type::Int32Ty));
2280FunctionType *ft = FunctionType::get(Type::Int8Ty, params, false);
2281</pre>
2282</div>
2283
2284<p>See the <a href="/doxygen/TypeBuilder_8h-source.html#l00001">class
2285comment</a> for more details.</p>
2286
2287</div>
2288
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]2289<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2290<div class="doc_section">
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2291 <a name="threading">Threads and LLVM</a>
2292</div>
2293<!-- *********************************************************************** -->
2294
2295<div class="doc_text">
2296<p>
2297This section describes the interaction of the LLVM APIs with multithreading,
2298both on the part of client applications, and in the JIT, in the hosted
2299application.
2300</p>
2301
2302<p>
2303Note that LLVM's support for multithreading is still relatively young. Up
2304through version 2.5, the execution of threaded hosted applications was
2305supported, but not threaded client access to the APIs. While this use case is
2306now supported, clients <em>must</em> adhere to the guidelines specified below to
2307ensure proper operation in multithreaded mode.
2308</p>
2309
2310<p>
2311Note that, on Unix-like platforms, LLVM requires the presence of GCC's atomic
2312intrinsics in order to support threaded operation. If you need a
2313multhreading-capable LLVM on a platform without a suitably modern system
2314compiler, consider compiling LLVM and LLVM-GCC in single-threaded mode, and
2315using the resultant compiler to build a copy of LLVM with multithreading
2316support.
2317</p>
2318</div>
2319
2320<!-- ======================================================================= -->
2321<div class="doc_subsection">
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2322 <a name="startmultithreaded">Entering and Exiting Multithreaded Mode</a>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2323</div>
2324
2325<div class="doc_text">
2326
2327<p>
2328In order to properly protect its internal data structures while avoiding
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2329excessive locking overhead in the single-threaded case, the LLVM must intialize
2330certain data structures necessary to provide guards around its internals. To do
2331so, the client program must invoke <tt>llvm_start_multithreaded()</tt> before
2332making any concurrent LLVM API calls. To subsequently tear down these
2333structures, use the <tt>llvm_stop_multithreaded()</tt> call. You can also use
2334the <tt>llvm_is_multithreaded()</tt> call to check the status of multithreaded
2335mode.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2336</p>
2337
2338<p>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2339Note that both of these calls must be made <em>in isolation</em>. That is to
2340say that no other LLVM API calls may be executing at any time during the
2341execution of <tt>llvm_start_multithreaded()</tt> or <tt>llvm_stop_multithreaded
2342</tt>. It's is the client's responsibility to enforce this isolation.
2343</p>
2344
2345<p>
2346The return value of <tt>llvm_start_multithreaded()</tt> indicates the success or
2347failure of the initialization. Failure typically indicates that your copy of
2348LLVM was built without multithreading support, typically because GCC atomic
2349intrinsics were not found in your system compiler. In this case, the LLVM API
2350will not be safe for concurrent calls. However, it <em>will</em> be safe for
2351hosting threaded applications in the JIT, though care must be taken to ensure
2352that side exits and the like do not accidentally result in concurrent LLVM API
2353calls.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2354</p>
2355</div>
2356
2357<!-- ======================================================================= -->
2358<div class="doc_subsection">
2359 <a name="shutdown">Ending Execution with <tt>llvm_shutdown()</tt></a>
2360</div>
2361
2362<div class="doc_text">
2363<p>
2364When you are done using the LLVM APIs, you should call <tt>llvm_shutdown()</tt>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2365to deallocate memory used for internal structures. This will also invoke
2366<tt>llvm_stop_multithreaded()</tt> if LLVM is operating in multithreaded mode.
2367As such, <tt>llvm_shutdown()</tt> requires the same isolation guarantees as
2368<tt>llvm_stop_multithreaded()</tt>.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2369</p>
2370
2371<p>
2372Note that, if you use scope-based shutdown, you can use the
2373<tt>llvm_shutdown_obj</tt> class, which calls <tt>llvm_shutdown()</tt> in its
2374destructor.
2375</div>
2376
2377<!-- ======================================================================= -->
2378<div class="doc_subsection">
2379 <a name="managedstatic">Lazy Initialization with <tt>ManagedStatic</tt></a>
2380</div>
2381
2382<div class="doc_text">
2383<p>
2384<tt>ManagedStatic</tt> is a utility class in LLVM used to implement static
2385initialization of static resources, such as the global type tables. Before the
2386invocation of <tt>llvm_shutdown()</tt>, it implements a simple lazy
2387initialization scheme. Once <tt>llvm_start_multithreaded()</tt> returns,
2388however, it uses double-checked locking to implement thread-safe lazy
2389initialization.
2390</p>
2391
2392<p>
2393Note that, because no other threads are allowed to issue LLVM API calls before
2394<tt>llvm_start_multithreaded()</tt> returns, it is possible to have
2395<tt>ManagedStatic</tt>s of <tt>llvm::sys::Mutex</tt>s.
2396</p>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2397
2398<p>
2399The <tt>llvm_acquire_global_lock()</tt> and <tt>llvm_release_global_lock</tt>
2400APIs provide access to the global lock used to implement the double-checked
2401locking for lazy initialization. These should only be used internally to LLVM,
2402and only if you know what you're doing!
2403</p>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2404</div>
2405
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]2406<!-- ======================================================================= -->
2407<div class="doc_subsection">
2408 <a name="llvmcontext">Achieving Isolation with <tt>LLVMContext</tt></a>
2409</div>
2410
2411<div class="doc_text">
2412<p>
2413<tt>LLVMContext</tt> is an opaque class in the LLVM API which clients can use
2414to operate multiple, isolated instances of LLVM concurrently within the same
2415address space. For instance, in a hypothetical compile-server, the compilation
2416of an individual translation unit is conceptually independent from all the
2417others, and it would be desirable to be able to compile incoming translation
2418units concurrently on independent server threads. Fortunately,
2419<tt>LLVMContext</tt> exists to enable just this kind of scenario!
2420</p>
2421
2422<p>
2423Conceptually, <tt>LLVMContext</tt> provides isolation. Every LLVM entity
2424(<tt>Module</tt>s, <tt>Value</tt>s, <tt>Type</tt>s, <tt>Constant</tt>s, etc.)
Chris Lattnerf636b962009-08-20 00:29:24 +0000[diff] [blame]2425in LLVM's in-memory IR belong to an <tt>LLVMContext</tt>. Entities in
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]2426different contexts <em>cannot</em> interact with each other: <tt>Module</tt>s in
2427different contexts cannot be linked together, <tt>Function</tt>s cannot be added
2428to <tt>Module</tt>s in different contexts, etc. What this means is that is is
2429safe to compile on multiple threads simultaneously, as long as no two threads
2430operate on entities within the same context.
2431</p>
2432
2433<p>
2434In practice, very few places in the API require the explicit specification of a
2435<tt>LLVMContext</tt>, other than the <tt>Type</tt> creation/lookup APIs.
2436Because every <tt>Type</tt> carries a reference to its owning context, most
2437other entities can determine what context they belong to by looking at their
2438own <tt>Type</tt>. If you are adding new entities to LLVM IR, please try to
2439maintain this interface design.
2440</p>
2441
2442<p>
2443For clients that do <em>not</em> require the benefits of isolation, LLVM
2444provides a convenience API <tt>getGlobalContext()</tt>. This returns a global,
2445lazily initialized <tt>LLVMContext</tt> that may be used in situations where
2446isolation is not a concern.
2447</p>
2448</div>
2449
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2450<!-- *********************************************************************** -->
2451<div class="doc_section">
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2452 <a name="advanced">Advanced Topics</a>
2453</div>
2454<!-- *********************************************************************** -->
2455
2456<div class="doc_text">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2457<p>
2458This section describes some of the advanced or obscure API's that most clients
2459do not need to be aware of. These API's tend manage the inner workings of the
2460LLVM system, and only need to be accessed in unusual circumstances.
2461</p>
2462</div>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2463
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2464<!-- ======================================================================= -->
2465<div class="doc_subsection">
2466 <a name="TypeResolve">LLVM Type Resolution</a>
2467</div>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2468
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2469<div class="doc_text">
2470
2471<p>
2472The LLVM type system has a very simple goal: allow clients to compare types for
2473structural equality with a simple pointer comparison (aka a shallow compare).
2474This goal makes clients much simpler and faster, and is used throughout the LLVM
2475system.
2476</p>
2477
2478<p>
2479Unfortunately achieving this goal is not a simple matter. In particular,
2480recursive types and late resolution of opaque types makes the situation very
2481difficult to handle. Fortunately, for the most part, our implementation makes
2482most clients able to be completely unaware of the nasty internal details. The
2483primary case where clients are exposed to the inner workings of it are when
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]2484building a recursive type. In addition to this case, the LLVM bitcode reader,
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2485assembly parser, and linker also have to be aware of the inner workings of this
2486system.
2487</p>
2488
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]2489<p>
2490For our purposes below, we need three concepts. First, an "Opaque Type" is
2491exactly as defined in the <a href="LangRef.html#t_opaque">language
2492reference</a>. Second an "Abstract Type" is any type which includes an
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2493opaque type as part of its type graph (for example "<tt>{ opaque, i32 }</tt>").
2494Third, a concrete type is a type that is not an abstract type (e.g. "<tt>{ i32,
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]2495float }</tt>").
2496</p>
2497
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2498</div>
2499
2500<!-- ______________________________________________________________________ -->
2501<div class="doc_subsubsection">
2502 <a name="BuildRecType">Basic Recursive Type Construction</a>
2503</div>
2504
2505<div class="doc_text">
2506
2507<p>
2508Because the most common question is "how do I build a recursive type with LLVM",
2509we answer it now and explain it as we go. Here we include enough to cause this
2510to be emitted to an output .ll file:
2511</p>
2512
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2513<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2514<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2515%mylist = type { %mylist*, i32 }
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2516</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2517</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2518
2519<p>
2520To build this, use the following LLVM APIs:
2521</p>
2522
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2523<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2524<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2525// <i>Create the initial outer struct</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2526<a href="#PATypeHolder">PATypeHolder</a> StructTy = OpaqueType::get();
2527std::vector&lt;const Type*&gt; Elts;
Daniel Dunbar58c2ac02008-10-03 22:17:25 +0000[diff] [blame]2528Elts.push_back(PointerType::getUnqual(StructTy));
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2529Elts.push_back(Type::Int32Ty);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2530StructType *NewSTy = StructType::get(Elts);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2531
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2532// <i>At this point, NewSTy = "{ opaque*, i32 }". Tell VMCore that</i>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2533// <i>the struct and the opaque type are actually the same.</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2534cast&lt;OpaqueType&gt;(StructTy.get())-&gt;<a href="#refineAbstractTypeTo">refineAbstractTypeTo</a>(NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2535
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2536// <i>NewSTy is potentially invalidated, but StructTy (a <a href="#PATypeHolder">PATypeHolder</a>) is</i>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2537// <i>kept up-to-date</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2538NewSTy = cast&lt;StructType&gt;(StructTy.get());
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2539
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2540// <i>Add a name for the type to the module symbol table (optional)</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2541MyModule-&gt;addTypeName("mylist", NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2542</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2543</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2544
2545<p>
2546This code shows the basic approach used to build recursive types: build a
2547non-recursive type using 'opaque', then use type unification to close the cycle.
2548The type unification step is performed by the <tt><a
Chris Lattneraff26d12007-02-03 03:06:52 +0000[diff] [blame]2549href="#refineAbstractTypeTo">refineAbstractTypeTo</a></tt> method, which is
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2550described next. After that, we describe the <a
2551href="#PATypeHolder">PATypeHolder class</a>.
2552</p>
2553
2554</div>
2555
2556<!-- ______________________________________________________________________ -->
2557<div class="doc_subsubsection">
2558 <a name="refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a>
2559</div>
2560
2561<div class="doc_text">
2562<p>
2563The <tt>refineAbstractTypeTo</tt> method starts the type unification process.
2564While this method is actually a member of the DerivedType class, it is most
2565often used on OpaqueType instances. Type unification is actually a recursive
2566process. After unification, types can become structurally isomorphic to
2567existing types, and all duplicates are deleted (to preserve pointer equality).
2568</p>
2569
2570<p>
2571In the example above, the OpaqueType object is definitely deleted.
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2572Additionally, if there is an "{ \2*, i32}" type already created in the system,
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2573the pointer and struct type created are <b>also</b> deleted. Obviously whenever
2574a type is deleted, any "Type*" pointers in the program are invalidated. As
2575such, it is safest to avoid having <i>any</i> "Type*" pointers to abstract types
2576live across a call to <tt>refineAbstractTypeTo</tt> (note that non-abstract
2577types can never move or be deleted). To deal with this, the <a
2578href="#PATypeHolder">PATypeHolder</a> class is used to maintain a stable
2579reference to a possibly refined type, and the <a
2580href="#AbstractTypeUser">AbstractTypeUser</a> class is used to update more
2581complex datastructures.
2582</p>
2583
2584</div>
2585
2586<!-- ______________________________________________________________________ -->
2587<div class="doc_subsubsection">
2588 <a name="PATypeHolder">The PATypeHolder Class</a>
2589</div>
2590
2591<div class="doc_text">
2592<p>
2593PATypeHolder is a form of a "smart pointer" for Type objects. When VMCore
2594happily goes about nuking types that become isomorphic to existing types, it
2595automatically updates all PATypeHolder objects to point to the new type. In the
2596example above, this allows the code to maintain a pointer to the resultant
2597resolved recursive type, even though the Type*'s are potentially invalidated.
2598</p>
2599
2600<p>
2601PATypeHolder is an extremely light-weight object that uses a lazy union-find
2602implementation to update pointers. For example the pointer from a Value to its
2603Type is maintained by PATypeHolder objects.
2604</p>
2605
2606</div>
2607
2608<!-- ______________________________________________________________________ -->
2609<div class="doc_subsubsection">
2610 <a name="AbstractTypeUser">The AbstractTypeUser Class</a>
2611</div>
2612
2613<div class="doc_text">
2614
2615<p>
2616Some data structures need more to perform more complex updates when types get
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2617resolved. To support this, a class can derive from the AbstractTypeUser class.
2618This class
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2619allows it to get callbacks when certain types are resolved. To register to get
2620callbacks for a particular type, the DerivedType::{add/remove}AbstractTypeUser
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]2621methods can be called on a type. Note that these methods only work for <i>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2622 abstract</i> types. Concrete types (those that do not include any opaque
2623objects) can never be refined.
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2624</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2625</div>
2626
2627
2628<!-- ======================================================================= -->
2629<div class="doc_subsection">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2630 <a name="SymbolTable">The <tt>ValueSymbolTable</tt> and
2631 <tt>TypeSymbolTable</tt> classes</a>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2632</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2633
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2634<div class="doc_text">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2635<p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1ValueSymbolTable.html">
2636ValueSymbolTable</a></tt> class provides a symbol table that the <a
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2637href="#Function"><tt>Function</tt></a> and <a href="#Module">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2638<tt>Module</tt></a> classes use for naming value definitions. The symbol table
2639can provide a name for any <a href="#Value"><tt>Value</tt></a>.
2640The <tt><a href="http://llvm.org/doxygen/classllvm_1_1TypeSymbolTable.html">
2641TypeSymbolTable</a></tt> class is used by the <tt>Module</tt> class to store
2642names for types.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2643
Reid Spencera6362242007-01-07 00:41:39 +0000[diff] [blame]2644<p>Note that the <tt>SymbolTable</tt> class should not be directly accessed
2645by most clients. It should only be used when iteration over the symbol table
2646names themselves are required, which is very special purpose. Note that not
2647all LLVM
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2648<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]2649an empty name) do not exist in the symbol table.
2650</p>
2651
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2652<p>These symbol tables support iteration over the values/types in the symbol
2653table with <tt>begin/end/iterator</tt> and supports querying to see if a
2654specific name is in the symbol table (with <tt>lookup</tt>). The
2655<tt>ValueSymbolTable</tt> class exposes no public mutator methods, instead,
2656simply call <tt>setName</tt> on a value, which will autoinsert it into the
2657appropriate symbol table. For types, use the Module::addTypeName method to
2658insert entries into the symbol table.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2659
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2660</div>
2661
2662
2663
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2664<!-- ======================================================================= -->
2665<div class="doc_subsection">
2666 <a name="UserLayout">The <tt>User</tt> and owned <tt>Use</tt> classes' memory layout</a>
2667</div>
2668
2669<div class="doc_text">
2670<p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1User.html">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2671User</a></tt> class provides a basis for expressing the ownership of <tt>User</tt>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2672towards other <tt><a href="http://llvm.org/doxygen/classllvm_1_1Value.html">
2673Value</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]2674Use</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]2675addition and removal.</p>
2676
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2677<!-- ______________________________________________________________________ -->
2678<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2679 <a name="Use2User">Interaction and relationship between <tt>User</tt> and <tt>Use</tt> objects</a>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2680</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2681
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2682<div class="doc_text">
2683<p>
2684A subclass of <tt>User</tt> can choose between incorporating its <tt>Use</tt> objects
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2685or refer to them out-of-line by means of a pointer. A mixed variant
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2686(some <tt>Use</tt>s inline others hung off) is impractical and breaks the invariant
2687that the <tt>Use</tt> objects belonging to the same <tt>User</tt> form a contiguous array.
2688</p>
2689</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2690
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2691<p>
2692We have 2 different layouts in the <tt>User</tt> (sub)classes:
2693<ul>
2694<li><p>Layout a)
2695The <tt>Use</tt> object(s) are inside (resp. at fixed offset) of the <tt>User</tt>
2696object and there are a fixed number of them.</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2697
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2698<li><p>Layout b)
2699The <tt>Use</tt> object(s) are referenced by a pointer to an
2700array from the <tt>User</tt> object and there may be a variable
2701number of them.</p>
2702</ul>
2703<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2704As of v2.4 each layout still possesses a direct pointer to the
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2705start of the array of <tt>Use</tt>s. Though not mandatory for layout a),
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2706we stick to this redundancy for the sake of simplicity.
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2707The <tt>User</tt> object also stores the number of <tt>Use</tt> objects it
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2708has. (Theoretically this information can also be calculated
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2709given the scheme presented below.)</p>
2710<p>
2711Special forms of allocation operators (<tt>operator new</tt>)
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2712enforce the following memory layouts:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2713
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2714<ul>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2715<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]2716
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2717<pre>
2718...---.---.---.---.-------...
2719 | P | P | P | P | User
2720'''---'---'---'---'-------'''
2721</pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2722
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2723<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]2724<pre>
2725.-------...
2726| User
2727'-------'''
2728 |
2729 v
2730 .---.---.---.---...
2731 | P | P | P | P |
2732 '---'---'---'---'''
2733</pre>
2734</ul>
2735<i>(In the above figures '<tt>P</tt>' stands for the <tt>Use**</tt> that
2736 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]2737
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2738<!-- ______________________________________________________________________ -->
2739<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2740 <a name="Waymarking">The waymarking algorithm</a>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2741</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2742
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2743<div class="doc_text">
2744<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2745Since the <tt>Use</tt> objects are deprived of the direct (back)pointer to
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2746their <tt>User</tt> objects, there must be a fast and exact method to
2747recover it. This is accomplished by the following scheme:</p>
2748</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2749
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2750A 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]2751start of the <tt>User</tt> object:
2752<ul>
2753<li><tt>00</tt> &mdash;&gt; binary digit 0</li>
2754<li><tt>01</tt> &mdash;&gt; binary digit 1</li>
2755<li><tt>10</tt> &mdash;&gt; stop and calculate (<tt>s</tt>)</li>
2756<li><tt>11</tt> &mdash;&gt; full stop (<tt>S</tt>)</li>
2757</ul>
2758<p>
2759Given a <tt>Use*</tt>, all we have to do is to walk till we get
2760a stop and we either have a <tt>User</tt> immediately behind or
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2761we have to walk to the next stop picking up digits
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2762and calculating the offset:</p>
2763<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2764.---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.----------------
2765| 1 | s | 1 | 0 | 1 | 0 | s | 1 | 1 | 0 | s | 1 | 1 | s | 1 | S | User (or User*)
2766'---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'----------------
2767 |+15 |+10 |+6 |+3 |+1
2768 | | | | |__>
2769 | | | |__________>
2770 | | |______________________>
2771 | |______________________________________>
2772 |__________________________________________________________>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2773</pre>
2774<p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2775Only the significant number of bits need to be stored between the
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2776stops, so that the <i>worst case is 20 memory accesses</i> when there are
27771000 <tt>Use</tt> objects associated with a <tt>User</tt>.</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2778
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2779<!-- ______________________________________________________________________ -->
2780<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2781 <a name="ReferenceImpl">Reference implementation</a>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2782</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2783
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2784<div class="doc_text">
2785<p>
2786The following literate Haskell fragment demonstrates the concept:</p>
2787</div>
2788
2789<div class="doc_code">
2790<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2791> import Test.QuickCheck
2792>
2793> digits :: Int -> [Char] -> [Char]
2794> digits 0 acc = '0' : acc
2795> digits 1 acc = '1' : acc
2796> digits n acc = digits (n `div` 2) $ digits (n `mod` 2) acc
2797>
2798> dist :: Int -> [Char] -> [Char]
2799> dist 0 [] = ['S']
2800> dist 0 acc = acc
2801> dist 1 acc = let r = dist 0 acc in 's' : digits (length r) r
2802> dist n acc = dist (n - 1) $ dist 1 acc
2803>
2804> takeLast n ss = reverse $ take n $ reverse ss
2805>
2806> test = takeLast 40 $ dist 20 []
2807>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2808</pre>
2809</div>
2810<p>
2811Printing &lt;test&gt; gives: <tt>"1s100000s11010s10100s1111s1010s110s11s1S"</tt></p>
2812<p>
2813The reverse algorithm computes the length of the string just by examining
2814a certain prefix:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2815
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2816<div class="doc_code">
2817<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2818> pref :: [Char] -> Int
2819> pref "S" = 1
2820> pref ('s':'1':rest) = decode 2 1 rest
2821> pref (_:rest) = 1 + pref rest
2822>
2823> decode walk acc ('0':rest) = decode (walk + 1) (acc * 2) rest
2824> decode walk acc ('1':rest) = decode (walk + 1) (acc * 2 + 1) rest
2825> decode walk acc _ = walk + acc
2826>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2827</pre>
2828</div>
2829<p>
2830Now, as expected, printing &lt;pref test&gt; gives <tt>40</tt>.</p>
2831<p>
2832We can <i>quickCheck</i> this with following property:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2833
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2834<div class="doc_code">
2835<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2836> testcase = dist 2000 []
2837> testcaseLength = length testcase
2838>
2839> identityProp n = n > 0 && n <= testcaseLength ==> length arr == pref arr
2840> where arr = takeLast n testcase
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2841>
2842</pre>
2843</div>
2844<p>
2845As expected &lt;quickCheck identityProp&gt; gives:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2846
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2847<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2848*Main> quickCheck identityProp
2849OK, passed 100 tests.
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2850</pre>
2851<p>
2852Let's be a bit more exhaustive:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2853
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2854<div class="doc_code">
2855<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2856>
2857> deepCheck p = check (defaultConfig { configMaxTest = 500 }) p
2858>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2859</pre>
2860</div>
2861<p>
2862And here is the result of &lt;deepCheck identityProp&gt;:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2863
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2864<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2865*Main> deepCheck identityProp
2866OK, passed 500 tests.
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2867</pre>
2868
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2869<!-- ______________________________________________________________________ -->
2870<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2871 <a name="Tagging">Tagging considerations</a>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2872</div>
2873
2874<p>
2875To maintain the invariant that the 2 LSBits of each <tt>Use**</tt> in <tt>Use</tt>
2876never change after being set up, setters of <tt>Use::Prev</tt> must re-tag the
2877new <tt>Use**</tt> on every modification. Accordingly getters must strip the
2878tag bits.</p>
2879<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2880For layout b) instead of the <tt>User</tt> we find a pointer (<tt>User*</tt> with LSBit set).
2881Following this pointer brings us to the <tt>User</tt>. A portable trick ensures
2882that the first bytes of <tt>User</tt> (if interpreted as a pointer) never has
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2883the LSBit set. (Portability is relying on the fact that all known compilers place the
2884<tt>vptr</tt> in the first word of the instances.)</p>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2885
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2886</div>
2887
2888 <!-- *********************************************************************** -->
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2889<div class="doc_section">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2890 <a name="coreclasses">The Core LLVM Class Hierarchy Reference </a>
2891</div>
2892<!-- *********************************************************************** -->
2893
2894<div class="doc_text">
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2895<p><tt>#include "<a href="/doxygen/Type_8h-source.html">llvm/Type.h</a>"</tt>
2896<br>doxygen info: <a href="/doxygen/classllvm_1_1Type.html">Type Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2897
2898<p>The Core LLVM classes are the primary means of representing the program
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2899being inspected or transformed. The core LLVM classes are defined in
2900header files in the <tt>include/llvm/</tt> directory, and implemented in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2901the <tt>lib/VMCore</tt> directory.</p>
2902
2903</div>
2904
2905<!-- ======================================================================= -->
2906<div class="doc_subsection">
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2907 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
2908</div>
2909
2910<div class="doc_text">
2911
2912 <p><tt>Type</tt> is a superclass of all type classes. Every <tt>Value</tt> has
2913 a <tt>Type</tt>. <tt>Type</tt> cannot be instantiated directly but only
2914 through its subclasses. Certain primitive types (<tt>VoidType</tt>,
2915 <tt>LabelType</tt>, <tt>FloatType</tt> and <tt>DoubleType</tt>) have hidden
2916 subclasses. They are hidden because they offer no useful functionality beyond
2917 what the <tt>Type</tt> class offers except to distinguish themselves from
2918 other subclasses of <tt>Type</tt>.</p>
2919 <p>All other types are subclasses of <tt>DerivedType</tt>. Types can be
2920 named, but this is not a requirement. There exists exactly
2921 one instance of a given shape at any one time. This allows type equality to
2922 be performed with address equality of the Type Instance. That is, given two
2923 <tt>Type*</tt> values, the types are identical if the pointers are identical.
2924 </p>
2925</div>
2926
2927<!-- _______________________________________________________________________ -->
2928<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2929 <a name="m_Type">Important Public Methods</a>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2930</div>
2931
2932<div class="doc_text">
2933
2934<ul>
Chris Lattner8f79df32007-01-15 01:55:32 +0000[diff] [blame]2935 <li><tt>bool isInteger() const</tt>: Returns true for any integer type.</li>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2936
2937 <li><tt>bool isFloatingPoint()</tt>: Return true if this is one of the two
2938 floating point types.</li>
2939
2940 <li><tt>bool isAbstract()</tt>: Return true if the type is abstract (contains
2941 an OpaqueType anywhere in its definition).</li>
2942
2943 <li><tt>bool isSized()</tt>: Return true if the type has known size. Things
2944 that don't have a size are abstract types, labels and void.</li>
2945
2946</ul>
2947</div>
2948
2949<!-- _______________________________________________________________________ -->
2950<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2951 <a name="derivedtypes">Important Derived Types</a>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2952</div>
2953<div class="doc_text">
2954<dl>
2955 <dt><tt>IntegerType</tt></dt>
2956 <dd>Subclass of DerivedType that represents integer types of any bit width.
2957 Any bit width between <tt>IntegerType::MIN_INT_BITS</tt> (1) and
2958 <tt>IntegerType::MAX_INT_BITS</tt> (~8 million) can be represented.
2959 <ul>
2960 <li><tt>static const IntegerType* get(unsigned NumBits)</tt>: get an integer
2961 type of a specific bit width.</li>
2962 <li><tt>unsigned getBitWidth() const</tt>: Get the bit width of an integer
2963 type.</li>
2964 </ul>
2965 </dd>
2966 <dt><tt>SequentialType</tt></dt>
2967 <dd>This is subclassed by ArrayType and PointerType
2968 <ul>
2969 <li><tt>const Type * getElementType() const</tt>: Returns the type of each
2970 of the elements in the sequential type. </li>
2971 </ul>
2972 </dd>
2973 <dt><tt>ArrayType</tt></dt>
2974 <dd>This is a subclass of SequentialType and defines the interface for array
2975 types.
2976 <ul>
2977 <li><tt>unsigned getNumElements() const</tt>: Returns the number of
2978 elements in the array. </li>
2979 </ul>
2980 </dd>
2981 <dt><tt>PointerType</tt></dt>
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]2982 <dd>Subclass of SequentialType for pointer types.</dd>
Reid Spencer9d6565a2007-02-15 02:26:10 +0000[diff] [blame]2983 <dt><tt>VectorType</tt></dt>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2984 <dd>Subclass of SequentialType for vector types. A
2985 vector type is similar to an ArrayType but is distinguished because it is
2986 a first class type wherease ArrayType is not. Vector types are used for
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2987 vector operations and are usually small vectors of of an integer or floating
2988 point type.</dd>
2989 <dt><tt>StructType</tt></dt>
2990 <dd>Subclass of DerivedTypes for struct types.</dd>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]2991 <dt><tt><a name="FunctionType">FunctionType</a></tt></dt>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2992 <dd>Subclass of DerivedTypes for function types.
2993 <ul>
2994 <li><tt>bool isVarArg() const</tt>: Returns true if its a vararg
2995 function</li>
2996 <li><tt> const Type * getReturnType() const</tt>: Returns the
2997 return type of the function.</li>
2998 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns
2999 the type of the ith parameter.</li>
3000 <li><tt> const unsigned getNumParams() const</tt>: Returns the
3001 number of formal parameters.</li>
3002 </ul>
3003 </dd>
3004 <dt><tt>OpaqueType</tt></dt>
3005 <dd>Sublcass of DerivedType for abstract types. This class
3006 defines no content and is used as a placeholder for some other type. Note
3007 that OpaqueType is used (temporarily) during type resolution for forward
3008 references of types. Once the referenced type is resolved, the OpaqueType
3009 is replaced with the actual type. OpaqueType can also be used for data
3010 abstraction. At link time opaque types can be resolved to actual types
3011 of the same name.</dd>
3012</dl>
3013</div>
3014
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3015
3016
3017<!-- ======================================================================= -->
3018<div class="doc_subsection">
3019 <a name="Module">The <tt>Module</tt> class</a>
3020</div>
3021
3022<div class="doc_text">
3023
3024<p><tt>#include "<a
3025href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt><br> doxygen info:
3026<a href="/doxygen/classllvm_1_1Module.html">Module Class</a></p>
3027
3028<p>The <tt>Module</tt> class represents the top level structure present in LLVM
3029programs. An LLVM module is effectively either a translation unit of the
3030original program or a combination of several translation units merged by the
3031linker. The <tt>Module</tt> class keeps track of a list of <a
3032href="#Function"><tt>Function</tt></a>s, a list of <a
3033href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
3034href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
3035helpful member functions that try to make common operations easy.</p>
3036
3037</div>
3038
3039<!-- _______________________________________________________________________ -->
3040<div class="doc_subsubsection">
3041 <a name="m_Module">Important Public Members of the <tt>Module</tt> class</a>
3042</div>
3043
3044<div class="doc_text">
3045
3046<ul>
3047 <li><tt>Module::Module(std::string name = "")</tt></li>
3048</ul>
3049
3050<p>Constructing a <a href="#Module">Module</a> is easy. You can optionally
3051provide a name for it (probably based on the name of the translation unit).</p>
3052
3053<ul>
3054 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
3055 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
3056
3057 <tt>begin()</tt>, <tt>end()</tt>
3058 <tt>size()</tt>, <tt>empty()</tt>
3059
3060 <p>These are forwarding methods that make it easy to access the contents of
3061 a <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
3062 list.</p></li>
3063
3064 <li><tt>Module::FunctionListType &amp;getFunctionList()</tt>
3065
3066 <p> Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
3067 necessary to use when you need to update the list or perform a complex
3068 action that doesn't have a forwarding method.</p>
3069
3070 <p><!-- Global Variable --></p></li>
3071</ul>
3072
3073<hr>
3074
3075<ul>
3076 <li><tt>Module::global_iterator</tt> - Typedef for global variable list iterator<br>
3077
3078 <tt>Module::const_global_iterator</tt> - Typedef for const_iterator.<br>
3079
3080 <tt>global_begin()</tt>, <tt>global_end()</tt>
3081 <tt>global_size()</tt>, <tt>global_empty()</tt>
3082
3083 <p> These are forwarding methods that make it easy to access the contents of
3084 a <tt>Module</tt> object's <a
3085 href="#GlobalVariable"><tt>GlobalVariable</tt></a> list.</p></li>
3086
3087 <li><tt>Module::GlobalListType &amp;getGlobalList()</tt>
3088
3089 <p>Returns the list of <a
3090 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s. This is necessary to
3091 use when you need to update the list or perform a complex action that
3092 doesn't have a forwarding method.</p>
3093
3094 <p><!-- Symbol table stuff --> </p></li>
3095</ul>
3096
3097<hr>
3098
3099<ul>
3100 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
3101
3102 <p>Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
3103 for this <tt>Module</tt>.</p>
3104
3105 <p><!-- Convenience methods --></p></li>
3106</ul>
3107
3108<hr>
3109
3110<ul>
3111 <li><tt><a href="#Function">Function</a> *getFunction(const std::string
3112 &amp;Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt>
3113
3114 <p>Look up the specified function in the <tt>Module</tt> <a
3115 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
3116 <tt>null</tt>.</p></li>
3117
3118 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const
3119 std::string &amp;Name, const <a href="#FunctionType">FunctionType</a> *T)</tt>
3120
3121 <p>Look up the specified function in the <tt>Module</tt> <a
3122 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
3123 external declaration for the function and return it.</p></li>
3124
3125 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt>
3126
3127 <p>If there is at least one entry in the <a
3128 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
3129 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
3130 string.</p></li>
3131
3132 <li><tt>bool addTypeName(const std::string &amp;Name, const <a
3133 href="#Type">Type</a> *Ty)</tt>
3134
3135 <p>Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
3136 mapping <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this
3137 name, true is returned and the <a
3138 href="#SymbolTable"><tt>SymbolTable</tt></a> is not modified.</p></li>
3139</ul>
3140
3141</div>
3142
3143
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3144<!-- ======================================================================= -->
3145<div class="doc_subsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3146 <a name="Value">The <tt>Value</tt> class</a>
3147</div>
3148
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3149<div class="doc_text">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3150
3151<p><tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt>
3152<br>
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]3153doxygen info: <a href="/doxygen/classllvm_1_1Value.html">Value Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3154
3155<p>The <tt>Value</tt> class is the most important class in the LLVM Source
3156base. It represents a typed value that may be used (among other things) as an
3157operand to an instruction. There are many different types of <tt>Value</tt>s,
3158such as <a href="#Constant"><tt>Constant</tt></a>s,<a
3159href="#Argument"><tt>Argument</tt></a>s. Even <a
3160href="#Instruction"><tt>Instruction</tt></a>s and <a
3161href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.</p>
3162
3163<p>A particular <tt>Value</tt> may be used many times in the LLVM representation
3164for a program. For example, an incoming argument to a function (represented
3165with an instance of the <a href="#Argument">Argument</a> class) is "used" by
3166every instruction in the function that references the argument. To keep track
3167of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
3168href="#User"><tt>User</tt></a>s that is using it (the <a
3169href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
3170graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
3171def-use information in the program, and is accessible through the <tt>use_</tt>*
3172methods, shown below.</p>
3173
3174<p>Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed,
3175and this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
3176method. In addition, all LLVM values can be named. The "name" of the
3177<tt>Value</tt> is a symbolic string printed in the LLVM code:</p>
3178
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3179<div class="doc_code">
3180<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]3181%<b>foo</b> = add i32 1, 2
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3182</pre>
3183</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3184
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3185<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]3186that the name of any value may be missing (an empty string), so names should
3187<b>ONLY</b> be used for debugging (making the source code easier to read,
3188debugging printouts), they should not be used to keep track of values or map
3189between them. For this purpose, use a <tt>std::map</tt> of pointers to the
3190<tt>Value</tt> itself instead.</p>
3191
3192<p>One important aspect of LLVM is that there is no distinction between an SSA
3193variable and the operation that produces it. Because of this, any reference to
3194the value produced by an instruction (or the value available as an incoming
Chris Lattnerd5fc4fc2004-03-18 14:58:55 +0000[diff] [blame]3195argument, for example) is represented as a direct pointer to the instance of
3196the class that
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3197represents this value. Although this may take some getting used to, it
3198simplifies the representation and makes it easier to manipulate.</p>
3199
3200</div>
3201
3202<!-- _______________________________________________________________________ -->
3203<div class="doc_subsubsection">
3204 <a name="m_Value">Important Public Members of the <tt>Value</tt> class</a>
3205</div>
3206
3207<div class="doc_text">
3208
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3209<ul>
3210 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the
3211use-list<br>
3212 <tt>Value::use_const_iterator</tt> - Typedef for const_iterator over
3213the use-list<br>
3214 <tt>unsigned use_size()</tt> - Returns the number of users of the
3215value.<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3216 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3217 <tt>use_iterator use_begin()</tt> - Get an iterator to the start of
3218the use-list.<br>
3219 <tt>use_iterator use_end()</tt> - Get an iterator to the end of the
3220use-list.<br>
3221 <tt><a href="#User">User</a> *use_back()</tt> - Returns the last
3222element in the list.
3223 <p> These methods are the interface to access the def-use
3224information in LLVM. As with all other iterators in LLVM, the naming
3225conventions follow the conventions defined by the <a href="#stl">STL</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3226 </li>
3227 <li><tt><a href="#Type">Type</a> *getType() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3228 <p>This method returns the Type of the Value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3229 </li>
3230 <li><tt>bool hasName() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3231 <tt>std::string getName() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3232 <tt>void setName(const std::string &amp;Name)</tt>
3233 <p> This family of methods is used to access and assign a name to a <tt>Value</tt>,
3234be aware of the <a href="#nameWarning">precaution above</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3235 </li>
3236 <li><tt>void replaceAllUsesWith(Value *V)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3237
3238 <p>This method traverses the use list of a <tt>Value</tt> changing all <a
3239 href="#User"><tt>User</tt>s</a> of the current value to refer to
3240 "<tt>V</tt>" instead. For example, if you detect that an instruction always
3241 produces a constant value (for example through constant folding), you can
3242 replace all uses of the instruction with the constant like this:</p>
3243
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3244<div class="doc_code">
3245<pre>
3246Inst-&gt;replaceAllUsesWith(ConstVal);
3247</pre>
3248</div>
3249
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3250</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3251
3252</div>
3253
3254<!-- ======================================================================= -->
3255<div class="doc_subsection">
3256 <a name="User">The <tt>User</tt> class</a>
3257</div>
3258
3259<div class="doc_text">
3260
3261<p>
3262<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]3263doxygen info: <a href="/doxygen/classllvm_1_1User.html">User Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3264Superclass: <a href="#Value"><tt>Value</tt></a></p>
3265
3266<p>The <tt>User</tt> class is the common base class of all LLVM nodes that may
3267refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
3268that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
3269referring to. The <tt>User</tt> class itself is a subclass of
3270<tt>Value</tt>.</p>
3271
3272<p>The operands of a <tt>User</tt> point directly to the LLVM <a
3273href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
3274Single Assignment (SSA) form, there can only be one definition referred to,
3275allowing this direct connection. This connection provides the use-def
3276information in LLVM.</p>
3277
3278</div>
3279
3280<!-- _______________________________________________________________________ -->
3281<div class="doc_subsubsection">
3282 <a name="m_User">Important Public Members of the <tt>User</tt> class</a>
3283</div>
3284
3285<div class="doc_text">
3286
3287<p>The <tt>User</tt> class exposes the operand list in two ways: through
3288an index access interface and through an iterator based interface.</p>
3289
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3290<ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3291 <li><tt>Value *getOperand(unsigned i)</tt><br>
3292 <tt>unsigned getNumOperands()</tt>
3293 <p> These two methods expose the operands of the <tt>User</tt> in a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3294convenient form for direct access.</p></li>
3295
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3296 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand
3297list<br>
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]3298 <tt>op_iterator op_begin()</tt> - Get an iterator to the start of
3299the operand list.<br>
3300 <tt>op_iterator op_end()</tt> - Get an iterator to the end of the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3301operand list.
3302 <p> Together, these methods make up the iterator based interface to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3303the operands of a <tt>User</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3304</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3305
3306</div>
3307
3308<!-- ======================================================================= -->
3309<div class="doc_subsection">
3310 <a name="Instruction">The <tt>Instruction</tt> class</a>
3311</div>
3312
3313<div class="doc_text">
3314
3315<p><tt>#include "</tt><tt><a
3316href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt><br>
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3317doxygen info: <a href="/doxygen/classllvm_1_1Instruction.html">Instruction Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3318Superclasses: <a href="#User"><tt>User</tt></a>, <a
3319href="#Value"><tt>Value</tt></a></p>
3320
3321<p>The <tt>Instruction</tt> class is the common base class for all LLVM
3322instructions. It provides only a few methods, but is a very commonly used
3323class. The primary data tracked by the <tt>Instruction</tt> class itself is the
3324opcode (instruction type) and the parent <a
3325href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
3326into. To represent a specific type of instruction, one of many subclasses of
3327<tt>Instruction</tt> are used.</p>
3328
3329<p> Because the <tt>Instruction</tt> class subclasses the <a
3330href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
3331way as for other <a href="#User"><tt>User</tt></a>s (with the
3332<tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
3333<tt>op_begin()</tt>/<tt>op_end()</tt> methods).</p> <p> An important file for
3334the <tt>Instruction</tt> class is the <tt>llvm/Instruction.def</tt> file. This
3335file contains some meta-data about the various different types of instructions
3336in LLVM. It describes the enum values that are used as opcodes (for example
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3337<tt>Instruction::Add</tt> and <tt>Instruction::ICmp</tt>), as well as the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3338concrete sub-classes of <tt>Instruction</tt> that implement the instruction (for
3339example <tt><a href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3340href="#CmpInst">CmpInst</a></tt>). Unfortunately, the use of macros in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3341this file confuses doxygen, so these enum values don't show up correctly in the
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3342<a href="/doxygen/classllvm_1_1Instruction.html">doxygen output</a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3343
3344</div>
3345
3346<!-- _______________________________________________________________________ -->
3347<div class="doc_subsubsection">
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3348 <a name="s_Instruction">Important Subclasses of the <tt>Instruction</tt>
3349 class</a>
3350</div>
3351<div class="doc_text">
3352 <ul>
3353 <li><tt><a name="BinaryOperator">BinaryOperator</a></tt>
3354 <p>This subclasses represents all two operand instructions whose operands
3355 must be the same type, except for the comparison instructions.</p></li>
3356 <li><tt><a name="CastInst">CastInst</a></tt>
3357 <p>This subclass is the parent of the 12 casting instructions. It provides
3358 common operations on cast instructions.</p>
3359 <li><tt><a name="CmpInst">CmpInst</a></tt>
3360 <p>This subclass respresents the two comparison instructions,
3361 <a href="LangRef.html#i_icmp">ICmpInst</a> (integer opreands), and
3362 <a href="LangRef.html#i_fcmp">FCmpInst</a> (floating point operands).</p>
3363 <li><tt><a name="TerminatorInst">TerminatorInst</a></tt>
3364 <p>This subclass is the parent of all terminator instructions (those which
3365 can terminate a block).</p>
3366 </ul>
3367 </div>
3368
3369<!-- _______________________________________________________________________ -->
3370<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3371 <a name="m_Instruction">Important Public Members of the <tt>Instruction</tt>
3372 class</a>
3373</div>
3374
3375<div class="doc_text">
3376
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3377<ul>
3378 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3379 <p>Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that
3380this <tt>Instruction</tt> is embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3381 <li><tt>bool mayWriteToMemory()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3382 <p>Returns true if the instruction writes to memory, i.e. it is a
3383 <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]3384 <li><tt>unsigned getOpcode()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3385 <p>Returns the opcode for the <tt>Instruction</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3386 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3387 <p>Returns another instance of the specified instruction, identical
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3388in all ways to the original except that the instruction has no parent
3389(ie it's not embedded into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>),
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3390and it has no name</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3391</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3392
3393</div>
3394
3395<!-- ======================================================================= -->
3396<div class="doc_subsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3397 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3398</div>
3399
3400<div class="doc_text">
3401
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3402<p>Constant represents a base class for different types of constants. It
3403is subclassed by ConstantInt, ConstantArray, etc. for representing
3404the various types of Constants. <a href="#GlobalValue">GlobalValue</a> is also
3405a subclass, which represents the address of a global variable or function.
3406</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3407
3408</div>
3409
3410<!-- _______________________________________________________________________ -->
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3411<div class="doc_subsubsection">Important Subclasses of Constant </div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3412<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3413<ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3414 <li>ConstantInt : This subclass of Constant represents an integer constant of
3415 any width.
3416 <ul>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]3417 <li><tt>const APInt&amp; getValue() const</tt>: Returns the underlying
3418 value of this constant, an APInt value.</li>
3419 <li><tt>int64_t getSExtValue() const</tt>: Converts the underlying APInt
3420 value to an int64_t via sign extension. If the value (not the bit width)
3421 of the APInt is too large to fit in an int64_t, an assertion will result.
3422 For this reason, use of this method is discouraged.</li>
3423 <li><tt>uint64_t getZExtValue() const</tt>: Converts the underlying APInt
3424 value to a uint64_t via zero extension. IF the value (not the bit width)
3425 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]3426 For this reason, use of this method is discouraged.</li>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]3427 <li><tt>static ConstantInt* get(const APInt&amp; Val)</tt>: Returns the
3428 ConstantInt object that represents the value provided by <tt>Val</tt>.
3429 The type is implied as the IntegerType that corresponds to the bit width
3430 of <tt>Val</tt>.</li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3431 <li><tt>static ConstantInt* get(const Type *Ty, uint64_t Val)</tt>:
3432 Returns the ConstantInt object that represents the value provided by
3433 <tt>Val</tt> for integer type <tt>Ty</tt>.</li>
3434 </ul>
3435 </li>
3436 <li>ConstantFP : This class represents a floating point constant.
3437 <ul>
3438 <li><tt>double getValue() const</tt>: Returns the underlying value of
3439 this constant. </li>
3440 </ul>
3441 </li>
3442 <li>ConstantArray : This represents a constant array.
3443 <ul>
3444 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
3445 a vector of component constants that makeup this array. </li>
3446 </ul>
3447 </li>
3448 <li>ConstantStruct : This represents a constant struct.
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>GlobalValue : This represents either a global variable or a function. In
3455 either case, the value is a constant fixed address (after linking).
3456 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3457</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3458</div>
3459
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3460
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3461<!-- ======================================================================= -->
3462<div class="doc_subsection">
3463 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
3464</div>
3465
3466<div class="doc_text">
3467
3468<p><tt>#include "<a
3469href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt><br>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]3470doxygen info: <a href="/doxygen/classllvm_1_1GlobalValue.html">GlobalValue
3471Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3472Superclasses: <a href="#Constant"><tt>Constant</tt></a>,
3473<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]3474
3475<p>Global values (<a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
3476href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
3477visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
3478Because they are visible at global scope, they are also subject to linking with
3479other globals defined in different translation units. To control the linking
3480process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
3481<tt>GlobalValue</tt>s know whether they have internal or external linkage, as
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]3482defined by the <tt>LinkageTypes</tt> enumeration.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3483
3484<p>If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
3485<tt>static</tt> in C), it is not visible to code outside the current translation
3486unit, and does not participate in linking. If it has external linkage, it is
3487visible to external code, and does participate in linking. In addition to
3488linkage information, <tt>GlobalValue</tt>s keep track of which <a
3489href="#Module"><tt>Module</tt></a> they are currently part of.</p>
3490
3491<p>Because <tt>GlobalValue</tt>s are memory objects, they are always referred to
3492by their <b>address</b>. As such, the <a href="#Type"><tt>Type</tt></a> of a
3493global is always a pointer to its contents. It is important to remember this
3494when using the <tt>GetElementPtrInst</tt> instruction because this pointer must
3495be dereferenced first. For example, if you have a <tt>GlobalVariable</tt> (a
3496subclass 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]3497i32]</tt>, then the <tt>GlobalVariable</tt> is a pointer to that array. Although
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3498the address of the first element of this array and the value of the
3499<tt>GlobalVariable</tt> are the same, they have different types. The
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]3500<tt>GlobalVariable</tt>'s type is <tt>[24 x i32]</tt>. The first element's type
3501is <tt>i32.</tt> Because of this, accessing a global value requires you to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3502dereference the pointer with <tt>GetElementPtrInst</tt> first, then its elements
3503can be accessed. This is explained in the <a href="LangRef.html#globalvars">LLVM
3504Language Reference Manual</a>.</p>
3505
3506</div>
3507
3508<!-- _______________________________________________________________________ -->
3509<div class="doc_subsubsection">
3510 <a name="m_GlobalValue">Important Public Members of the <tt>GlobalValue</tt>
3511 class</a>
3512</div>
3513
3514<div class="doc_text">
3515
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3516<ul>
3517 <li><tt>bool hasInternalLinkage() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3518 <tt>bool hasExternalLinkage() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3519 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt>
3520 <p> These methods manipulate the linkage characteristics of the <tt>GlobalValue</tt>.</p>
3521 <p> </p>
3522 </li>
3523 <li><tt><a href="#Module">Module</a> *getParent()</tt>
3524 <p> This returns the <a href="#Module"><tt>Module</tt></a> that the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3525GlobalValue is currently embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3526</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3527
3528</div>
3529
3530<!-- ======================================================================= -->
3531<div class="doc_subsection">
3532 <a name="Function">The <tt>Function</tt> class</a>
3533</div>
3534
3535<div class="doc_text">
3536
3537<p><tt>#include "<a
3538href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt><br> doxygen
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3539info: <a href="/doxygen/classllvm_1_1Function.html">Function Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3540Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
3541<a href="#Constant"><tt>Constant</tt></a>,
3542<a href="#User"><tt>User</tt></a>,
3543<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3544
3545<p>The <tt>Function</tt> class represents a single procedure in LLVM. It is
3546actually one of the more complex classes in the LLVM heirarchy because it must
3547keep track of a large amount of data. The <tt>Function</tt> class keeps track
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3548of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal
3549<a href="#Argument"><tt>Argument</tt></a>s, and a
3550<a href="#SymbolTable"><tt>SymbolTable</tt></a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3551
3552<p>The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most
3553commonly used part of <tt>Function</tt> objects. The list imposes an implicit
3554ordering of the blocks in the function, which indicate how the code will be
3555layed out by the backend. Additionally, the first <a
3556href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
3557<tt>Function</tt>. It is not legal in LLVM to explicitly branch to this initial
3558block. There are no implicit exit nodes, and in fact there may be multiple exit
3559nodes from a single <tt>Function</tt>. If the <a
3560href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
3561the <tt>Function</tt> is actually a function declaration: the actual body of the
3562function hasn't been linked in yet.</p>
3563
3564<p>In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
3565<tt>Function</tt> class also keeps track of the list of formal <a
3566href="#Argument"><tt>Argument</tt></a>s that the function receives. This
3567container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
3568nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
3569the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.</p>
3570
3571<p>The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used
3572LLVM feature that is only used when you have to look up a value by name. Aside
3573from that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used
3574internally to make sure that there are not conflicts between the names of <a
3575href="#Instruction"><tt>Instruction</tt></a>s, <a
3576href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
3577href="#Argument"><tt>Argument</tt></a>s in the function body.</p>
3578
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]3579<p>Note that <tt>Function</tt> is a <a href="#GlobalValue">GlobalValue</a>
3580and therefore also a <a href="#Constant">Constant</a>. The value of the function
3581is its address (after linking) which is guaranteed to be constant.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3582</div>
3583
3584<!-- _______________________________________________________________________ -->
3585<div class="doc_subsubsection">
3586 <a name="m_Function">Important Public Members of the <tt>Function</tt>
3587 class</a>
3588</div>
3589
3590<div class="doc_text">
3591
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3592<ul>
3593 <li><tt>Function(const </tt><tt><a href="#FunctionType">FunctionType</a>
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]3594 *Ty, LinkageTypes Linkage, const std::string &amp;N = "", Module* Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3595
3596 <p>Constructor used when you need to create new <tt>Function</tt>s to add
3597 the the program. The constructor must specify the type of the function to
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]3598 create and what type of linkage the function should have. The <a
3599 href="#FunctionType"><tt>FunctionType</tt></a> argument
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3600 specifies the formal arguments and return value for the function. The same
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3601 <a href="#FunctionType"><tt>FunctionType</tt></a> value can be used to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3602 create multiple functions. The <tt>Parent</tt> argument specifies the Module
3603 in which the function is defined. If this argument is provided, the function
3604 will automatically be inserted into that module's list of
3605 functions.</p></li>
3606
Chris Lattner62810e32008-11-25 18:34:50 +0000[diff] [blame]3607 <li><tt>bool isDeclaration()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3608
3609 <p>Return whether or not the <tt>Function</tt> has a body defined. If the
3610 function is "external", it does not have a body, and thus must be resolved
3611 by linking with a function defined in a different translation unit.</p></li>
3612
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3613 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3614 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3615
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]3616 <tt>begin()</tt>, <tt>end()</tt>
3617 <tt>size()</tt>, <tt>empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3618
3619 <p>These are forwarding methods that make it easy to access the contents of
3620 a <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
3621 list.</p></li>
3622
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3623 <li><tt>Function::BasicBlockListType &amp;getBasicBlockList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3624
3625 <p>Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This
3626 is necessary to use when you need to update the list or perform a complex
3627 action that doesn't have a forwarding method.</p></li>
3628
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3629 <li><tt>Function::arg_iterator</tt> - Typedef for the argument list
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3630iterator<br>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3631 <tt>Function::const_arg_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3632
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]3633 <tt>arg_begin()</tt>, <tt>arg_end()</tt>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3634 <tt>arg_size()</tt>, <tt>arg_empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3635
3636 <p>These are forwarding methods that make it easy to access the contents of
3637 a <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a>
3638 list.</p></li>
3639
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3640 <li><tt>Function::ArgumentListType &amp;getArgumentList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3641
3642 <p>Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
3643 necessary to use when you need to update the list or perform a complex
3644 action that doesn't have a forwarding method.</p></li>
3645
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3646 <li><tt><a href="#BasicBlock">BasicBlock</a> &amp;getEntryBlock()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3647
3648 <p>Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
3649 function. Because the entry block for the function is always the first
3650 block, this returns the first block of the <tt>Function</tt>.</p></li>
3651
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3652 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
3653 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3654
3655 <p>This traverses the <a href="#Type"><tt>Type</tt></a> of the
3656 <tt>Function</tt> and returns the return type of the function, or the <a
3657 href="#FunctionType"><tt>FunctionType</tt></a> of the actual
3658 function.</p></li>
3659
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3660 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3661
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3662 <p> Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3663 for this <tt>Function</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3664</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3665
3666</div>
3667
3668<!-- ======================================================================= -->
3669<div class="doc_subsection">
3670 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
3671</div>
3672
3673<div class="doc_text">
3674
3675<p><tt>#include "<a
3676href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt>
3677<br>
Tanya Lattnera3da7772004-06-22 08:02:25 +0000[diff] [blame]3678doxygen info: <a href="/doxygen/classllvm_1_1GlobalVariable.html">GlobalVariable
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3679 Class</a><br>
3680Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
3681<a href="#Constant"><tt>Constant</tt></a>,
3682<a href="#User"><tt>User</tt></a>,
3683<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3684
3685<p>Global variables are represented with the (suprise suprise)
3686<tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are also
3687subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such are
3688always referenced by their address (global values must live in memory, so their
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3689"name" refers to their constant address). See
3690<a href="#GlobalValue"><tt>GlobalValue</tt></a> for more on this. Global
3691variables may have an initial value (which must be a
3692<a href="#Constant"><tt>Constant</tt></a>), and if they have an initializer,
3693they may be marked as "constant" themselves (indicating that their contents
3694never change at runtime).</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3695</div>
3696
3697<!-- _______________________________________________________________________ -->
3698<div class="doc_subsubsection">
3699 <a name="m_GlobalVariable">Important Public Members of the
3700 <tt>GlobalVariable</tt> class</a>
3701</div>
3702
3703<div class="doc_text">
3704
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3705<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3706 <li><tt>GlobalVariable(const </tt><tt><a href="#Type">Type</a> *Ty, bool
3707 isConstant, LinkageTypes&amp; Linkage, <a href="#Constant">Constant</a>
3708 *Initializer = 0, const std::string &amp;Name = "", Module* Parent = 0)</tt>
3709
3710 <p>Create a new global variable of the specified type. If
3711 <tt>isConstant</tt> is true then the global variable will be marked as
3712 unchanging for the program. The Linkage parameter specifies the type of
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]3713 linkage (internal, external, weak, linkonce, appending) for the variable.
3714 If the linkage is InternalLinkage, WeakAnyLinkage, WeakODRLinkage,
3715 LinkOnceAnyLinkage or LinkOnceODRLinkage,&nbsp; then the resultant
3716 global variable will have internal linkage. AppendingLinkage concatenates
3717 together all instances (in different translation units) of the variable
3718 into a single variable but is only applicable to arrays. &nbsp;See
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3719 the <a href="LangRef.html#modulestructure">LLVM Language Reference</a> for
3720 further details on linkage types. Optionally an initializer, a name, and the
3721 module to put the variable into may be specified for the global variable as
3722 well.</p></li>
3723
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3724 <li><tt>bool isConstant() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3725
3726 <p>Returns true if this is a global variable that is known not to
3727 be modified at runtime.</p></li>
3728
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3729 <li><tt>bool hasInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3730
3731 <p>Returns true if this <tt>GlobalVariable</tt> has an intializer.</p></li>
3732
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3733 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3734
3735 <p>Returns the intial value for a <tt>GlobalVariable</tt>. It is not legal
3736 to call this method if there is no initializer.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3737</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3738
3739</div>
3740
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3741
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3742<!-- ======================================================================= -->
3743<div class="doc_subsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3744 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3745</div>
3746
3747<div class="doc_text">
3748
3749<p><tt>#include "<a
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3750href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt><br>
Stefanus Du Toit24e04112009-06-17 21:12:26 +0000[diff] [blame]3751doxygen info: <a href="/doxygen/classllvm_1_1BasicBlock.html">BasicBlock
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3752Class</a><br>
3753Superclass: <a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3754
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3755<p>This class represents a single entry multiple exit section of the code,
3756commonly known as a basic block by the compiler community. The
3757<tt>BasicBlock</tt> class maintains a list of <a
3758href="#Instruction"><tt>Instruction</tt></a>s, which form the body of the block.
3759Matching the language definition, the last element of this list of instructions
3760is always a terminator instruction (a subclass of the <a
3761href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).</p>
3762
3763<p>In addition to tracking the list of instructions that make up the block, the
3764<tt>BasicBlock</tt> class also keeps track of the <a
3765href="#Function"><tt>Function</tt></a> that it is embedded into.</p>
3766
3767<p>Note that <tt>BasicBlock</tt>s themselves are <a
3768href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
3769like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
3770<tt>label</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3771
3772</div>
3773
3774<!-- _______________________________________________________________________ -->
3775<div class="doc_subsubsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3776 <a name="m_BasicBlock">Important Public Members of the <tt>BasicBlock</tt>
3777 class</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3778</div>
3779
3780<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3781<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3782
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3783<li><tt>BasicBlock(const std::string &amp;Name = "", </tt><tt><a
3784 href="#Function">Function</a> *Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3785
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3786<p>The <tt>BasicBlock</tt> constructor is used to create new basic blocks for
3787insertion into a function. The constructor optionally takes a name for the new
3788block, and a <a href="#Function"><tt>Function</tt></a> to insert it into. If
3789the <tt>Parent</tt> parameter is specified, the new <tt>BasicBlock</tt> is
3790automatically inserted at the end of the specified <a
3791href="#Function"><tt>Function</tt></a>, if not specified, the BasicBlock must be
3792manually inserted into the <a href="#Function"><tt>Function</tt></a>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3793
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3794<li><tt>BasicBlock::iterator</tt> - Typedef for instruction list iterator<br>
3795<tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
3796<tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
3797<tt>size()</tt>, <tt>empty()</tt>
3798STL-style functions for accessing the instruction list.
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3799
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3800<p>These methods and typedefs are forwarding functions that have the same
3801semantics as the standard library methods of the same names. These methods
3802expose the underlying instruction list of a basic block in a way that is easy to
3803manipulate. To get the full complement of container operations (including
3804operations to update the list), you must use the <tt>getInstList()</tt>
3805method.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3806
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3807<li><tt>BasicBlock::InstListType &amp;getInstList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3808
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3809<p>This method is used to get access to the underlying container that actually
3810holds the Instructions. This method must be used when there isn't a forwarding
3811function in the <tt>BasicBlock</tt> class for the operation that you would like
3812to perform. Because there are no forwarding functions for "updating"
3813operations, you need to use this if you want to update the contents of a
3814<tt>BasicBlock</tt>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3815
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3816<li><tt><a href="#Function">Function</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3817
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3818<p> Returns a pointer to <a href="#Function"><tt>Function</tt></a> the block is
3819embedded into, or a null pointer if it is homeless.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3820
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3821<li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3822
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3823<p> Returns a pointer to the terminator instruction that appears at the end of
3824the <tt>BasicBlock</tt>. If there is no terminator instruction, or if the last
3825instruction in the block is not a terminator, then a null pointer is
3826returned.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3827
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3828</ul>
3829
3830</div>
3831
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3832
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3833<!-- ======================================================================= -->
3834<div class="doc_subsection">
3835 <a name="Argument">The <tt>Argument</tt> class</a>
3836</div>
3837
3838<div class="doc_text">
3839
3840<p>This subclass of Value defines the interface for incoming formal
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]3841arguments to a function. A Function maintains a list of its formal
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3842arguments. An argument has a pointer to the parent Function.</p>
3843
3844</div>
3845
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3846<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3847<hr>
3848<address>
3849 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]3850 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3851 <a href="http://validator.w3.org/check/referer"><img
Gabor Greifa9c0f2b2008-06-18 14:05:31 +0000[diff] [blame]3852 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]3853
3854 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
3855 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]3856 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3857 Last modified: $Date$
3858</address>
3859
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3860</body>
3861</html>