Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 39a1fab856acec40e486d075a8017ed824c8bc25 / . / docs / ProgrammersManual.html
blob: e4e3dc29c30b21c08b6c54348377a895a8743d37 [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>
Jeffrey Yasskin71a5c222009-10-22 22:11:22 +0000[diff] [blame]86 <li><a href="#dss_valuemap">"llvm/ADT/ValueMap.h"</a></li>
Chris Lattnerf3692522007-02-03 19:51:56 +0000[diff] [blame]87 <li><a href="#dss_map">&lt;map&gt;</a></li>
88 <li><a href="#dss_othermap">Other Map-Like Container Options</a></li>
89 </ul></li>
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]90 <li><a href="#ds_string">String-like containers</a>
Benjamin Kramere15192b2009-08-05 15:42:44 +0000[diff] [blame]91 <!--<ul>
92 todo
93 </ul>--></li>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]94 <li><a href="#ds_bit">BitVector-like containers</a>
95 <ul>
96 <li><a href="#dss_bitvector">A dense bitvector</a></li>
97 <li><a href="#dss_sparsebitvector">A sparse bitvector</a></li>
98 </ul></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]99 </ul>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]100 </li>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]101 <li><a href="#common">Helpful Hints for Common Operations</a>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]102 <ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]103 <li><a href="#inspection">Basic Inspection and Traversal Routines</a>
104 <ul>
105 <li><a href="#iterate_function">Iterating over the <tt>BasicBlock</tt>s
106in a <tt>Function</tt></a> </li>
107 <li><a href="#iterate_basicblock">Iterating over the <tt>Instruction</tt>s
108in a <tt>BasicBlock</tt></a> </li>
109 <li><a href="#iterate_institer">Iterating over the <tt>Instruction</tt>s
110in a <tt>Function</tt></a> </li>
111 <li><a href="#iterate_convert">Turning an iterator into a
112class pointer</a> </li>
113 <li><a href="#iterate_complex">Finding call sites: a more
114complex example</a> </li>
115 <li><a href="#calls_and_invokes">Treating calls and invokes
116the same way</a> </li>
117 <li><a href="#iterate_chains">Iterating over def-use &amp;
118use-def chains</a> </li>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]119 <li><a href="#iterate_preds">Iterating over predecessors &amp;
120successors of blocks</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]121 </ul>
122 </li>
123 <li><a href="#simplechanges">Making simple changes</a>
124 <ul>
125 <li><a href="#schanges_creating">Creating and inserting new
126 <tt>Instruction</tt>s</a> </li>
127 <li><a href="#schanges_deleting">Deleting <tt>Instruction</tt>s</a> </li>
128 <li><a href="#schanges_replacing">Replacing an <tt>Instruction</tt>
129with another <tt>Value</tt></a> </li>
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]130 <li><a href="#schanges_deletingGV">Deleting <tt>GlobalVariable</tt>s</a> </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]131 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]132 </li>
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]133 <li><a href="#create_types">How to Create Types</a></li>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]134<!--
135 <li>Working with the Control Flow Graph
136 <ul>
137 <li>Accessing predecessors and successors of a <tt>BasicBlock</tt>
138 <li>
139 <li>
140 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]141-->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]142 </ul>
143 </li>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]144
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]145 <li><a href="#threading">Threads and LLVM</a>
146 <ul>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]147 <li><a href="#startmultithreaded">Entering and Exiting Multithreaded Mode
148 </a></li>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]149 <li><a href="#shutdown">Ending execution with <tt>llvm_shutdown()</tt></a></li>
150 <li><a href="#managedstatic">Lazy initialization with <tt>ManagedStatic</tt></a></li>
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]151 <li><a href="#llvmcontext">Achieving Isolation with <tt>LLVMContext</tt></a></li>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]152 </ul>
153 </li>
154
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]155 <li><a href="#advanced">Advanced Topics</a>
156 <ul>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]157 <li><a href="#TypeResolve">LLVM Type Resolution</a>
158 <ul>
159 <li><a href="#BuildRecType">Basic Recursive Type Construction</a></li>
160 <li><a href="#refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a></li>
161 <li><a href="#PATypeHolder">The PATypeHolder Class</a></li>
162 <li><a href="#AbstractTypeUser">The AbstractTypeUser Class</a></li>
163 </ul></li>
164
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]165 <li><a href="#SymbolTable">The <tt>ValueSymbolTable</tt> and <tt>TypeSymbolTable</tt> classes</a></li>
166 <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]167 </ul></li>
168
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]169 <li><a href="#coreclasses">The Core LLVM Class Hierarchy Reference</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]170 <ul>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]171 <li><a href="#Type">The <tt>Type</tt> class</a> </li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]172 <li><a href="#Module">The <tt>Module</tt> class</a></li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]173 <li><a href="#Value">The <tt>Value</tt> class</a>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]174 <ul>
175 <li><a href="#User">The <tt>User</tt> class</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]176 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]177 <li><a href="#Instruction">The <tt>Instruction</tt> class</a></li>
178 <li><a href="#Constant">The <tt>Constant</tt> class</a>
179 <ul>
180 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]181 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]182 <li><a href="#Function">The <tt>Function</tt> class</a></li>
183 <li><a href="#GlobalVariable">The <tt>GlobalVariable</tt> class</a></li>
184 </ul>
185 </li>
186 </ul>
187 </li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]188 </ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]189 </li>
190 <li><a href="#BasicBlock">The <tt>BasicBlock</tt> class</a></li>
191 <li><a href="#Argument">The <tt>Argument</tt> class</a></li>
192 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]193 </li>
194 </ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]195 </li>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]196</ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]197
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]198<div class="doc_author">
199 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
Chris Lattner94c43592004-05-26 16:52:55 +0000[diff] [blame]200 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>,
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]201 <a href="mailto:ggreif@gmail.com">Gabor Greif</a>,
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]202 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a>,
203 <a href="mailto:rspencer@x10sys.com">Reid Spencer</a> and
204 <a href="mailto:owen@apple.com">Owen Anderson</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]205</div>
206
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]207<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]208<div class="doc_section">
209 <a name="introduction">Introduction </a>
210</div>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]211<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]212
213<div class="doc_text">
214
215<p>This document is meant to highlight some of the important classes and
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]216interfaces available in the LLVM source-base. This manual is not
217intended to explain what LLVM is, how it works, and what LLVM code looks
218like. It assumes that you know the basics of LLVM and are interested
219in writing transformations or otherwise analyzing or manipulating the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]220code.</p>
221
222<p>This document should get you oriented so that you can find your
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]223way in the continuously growing source code that makes up the LLVM
224infrastructure. Note that this manual is not intended to serve as a
225replacement for reading the source code, so if you think there should be
226a method in one of these classes to do something, but it's not listed,
227check the source. Links to the <a href="/doxygen/">doxygen</a> sources
228are provided to make this as easy as possible.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]229
230<p>The first section of this document describes general information that is
231useful to know when working in the LLVM infrastructure, and the second describes
232the Core LLVM classes. In the future this manual will be extended with
233information describing how to use extension libraries, such as dominator
234information, CFG traversal routines, and useful utilities like the <tt><a
235href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.</p>
236
237</div>
238
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]239<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]240<div class="doc_section">
241 <a name="general">General Information</a>
242</div>
243<!-- *********************************************************************** -->
244
245<div class="doc_text">
246
247<p>This section contains general information that is useful if you are working
248in the LLVM source-base, but that isn't specific to any particular API.</p>
249
250</div>
251
252<!-- ======================================================================= -->
253<div class="doc_subsection">
254 <a name="stl">The C++ Standard Template Library</a>
255</div>
256
257<div class="doc_text">
258
259<p>LLVM makes heavy use of the C++ Standard Template Library (STL),
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]260perhaps much more than you are used to, or have seen before. Because of
261this, you might want to do a little background reading in the
262techniques used and capabilities of the library. There are many good
263pages that discuss the STL, and several books on the subject that you
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]264can get, so it will not be discussed in this document.</p>
265
266<p>Here are some useful links:</p>
267
268<ol>
269
270<li><a href="http://www.dinkumware.com/refxcpp.html">Dinkumware C++ Library
271reference</a> - an excellent reference for the STL and other parts of the
272standard C++ library.</li>
273
274<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]275O'Reilly book in the making. It has a decent Standard Library
276Reference that rivals Dinkumware's, and is unfortunately no longer free since the
277book has been published.</li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]278
279<li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
280Questions</a></li>
281
282<li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
283Contains a useful <a
284href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
285STL</a>.</li>
286
287<li><a href="http://www.research.att.com/%7Ebs/C++.html">Bjarne Stroustrup's C++
288Page</a></li>
289
Tanya Lattner79445ba2004-12-08 18:34:56 +0000[diff] [blame]290<li><a href="http://64.78.49.204/">
Reid Spencer096603a2004-05-26 08:41:35 +0000[diff] [blame]291Bruce Eckel's Thinking in C++, 2nd ed. Volume 2 Revision 4.0 (even better, get
292the book).</a></li>
293
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]294</ol>
295
296<p>You are also encouraged to take a look at the <a
297href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
298to write maintainable code more than where to put your curly braces.</p>
299
300</div>
301
302<!-- ======================================================================= -->
303<div class="doc_subsection">
304 <a name="stl">Other useful references</a>
305</div>
306
307<div class="doc_text">
308
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]309<ol>
310<li><a href="http://www.psc.edu/%7Esemke/cvs_branches.html">CVS
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]311Branch and Tag Primer</a></li>
Misha Brukmana0f71e42004-06-18 18:39:00 +0000[diff] [blame]312<li><a href="http://www.fortran-2000.com/ArnaudRecipes/sharedlib.html">Using
313static and shared libraries across platforms</a></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]314</ol>
315
316</div>
317
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]318<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]319<div class="doc_section">
320 <a name="apis">Important and useful LLVM APIs</a>
321</div>
322<!-- *********************************************************************** -->
323
324<div class="doc_text">
325
326<p>Here we highlight some LLVM APIs that are generally useful and good to
327know about when writing transformations.</p>
328
329</div>
330
331<!-- ======================================================================= -->
332<div class="doc_subsection">
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]333 <a name="isa">The <tt>isa&lt;&gt;</tt>, <tt>cast&lt;&gt;</tt> and
334 <tt>dyn_cast&lt;&gt;</tt> templates</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]335</div>
336
337<div class="doc_text">
338
339<p>The LLVM source-base makes extensive use of a custom form of RTTI.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]340These templates have many similarities to the C++ <tt>dynamic_cast&lt;&gt;</tt>
341operator, but they don't have some drawbacks (primarily stemming from
342the fact that <tt>dynamic_cast&lt;&gt;</tt> only works on classes that
343have a v-table). Because they are used so often, you must know what they
344do and how they work. All of these templates are defined in the <a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]345 href="/doxygen/Casting_8h-source.html"><tt>llvm/Support/Casting.h</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]346file (note that you very rarely have to include this file directly).</p>
347
348<dl>
349 <dt><tt>isa&lt;&gt;</tt>: </dt>
350
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]351 <dd><p>The <tt>isa&lt;&gt;</tt> operator works exactly like the Java
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]352 "<tt>instanceof</tt>" operator. It returns true or false depending on whether
353 a reference or pointer points to an instance of the specified class. This can
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]354 be very useful for constraint checking of various sorts (example below).</p>
355 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]356
357 <dt><tt>cast&lt;&gt;</tt>: </dt>
358
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]359 <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]360 converts a pointer or reference from a base class to a derived class, causing
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]361 an assertion failure if it is not really an instance of the right type. This
362 should be used in cases where you have some information that makes you believe
363 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]364 and <tt>cast&lt;&gt;</tt> template is:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]365
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]366<div class="doc_code">
367<pre>
368static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
369 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))
370 return true;
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]371
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]372 // <i>Otherwise, it must be an instruction...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]373 return !L-&gt;contains(cast&lt;<a href="#Instruction">Instruction</a>&gt;(V)-&gt;getParent());
374}
375</pre>
376</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]377
378 <p>Note that you should <b>not</b> use an <tt>isa&lt;&gt;</tt> test followed
379 by a <tt>cast&lt;&gt;</tt>, for that use the <tt>dyn_cast&lt;&gt;</tt>
380 operator.</p>
381
382 </dd>
383
384 <dt><tt>dyn_cast&lt;&gt;</tt>:</dt>
385
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]386 <dd><p>The <tt>dyn_cast&lt;&gt;</tt> operator is a "checking cast" operation.
387 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]388 pointer to it (this operator does not work with references). If the operand is
389 not of the correct type, a null pointer is returned. Thus, this works very
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]390 much like the <tt>dynamic_cast&lt;&gt;</tt> operator in C++, and should be
391 used in the same circumstances. Typically, the <tt>dyn_cast&lt;&gt;</tt>
392 operator is used in an <tt>if</tt> statement or some other flow control
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]393 statement like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]394
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]395<div class="doc_code">
396<pre>
397if (<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]398 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]399}
400</pre>
401</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]402
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]403 <p>This form of the <tt>if</tt> statement effectively combines together a call
404 to <tt>isa&lt;&gt;</tt> and a call to <tt>cast&lt;&gt;</tt> into one
405 statement, which is very convenient.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]406
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]407 <p>Note that the <tt>dyn_cast&lt;&gt;</tt> operator, like C++'s
408 <tt>dynamic_cast&lt;&gt;</tt> or Java's <tt>instanceof</tt> operator, can be
409 abused. In particular, you should not use big chained <tt>if/then/else</tt>
410 blocks to check for lots of different variants of classes. If you find
411 yourself wanting to do this, it is much cleaner and more efficient to use the
412 <tt>InstVisitor</tt> class to dispatch over the instruction type directly.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]413
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]414 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]415
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]416 <dt><tt>cast_or_null&lt;&gt;</tt>: </dt>
417
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]418 <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]419 <tt>cast&lt;&gt;</tt> operator, except that it allows for a null pointer as an
420 argument (which it then propagates). This can sometimes be useful, allowing
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]421 you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]422
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]423 <dt><tt>dyn_cast_or_null&lt;&gt;</tt>: </dt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]424
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]425 <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]426 <tt>dyn_cast&lt;&gt;</tt> operator, except that it allows for a null pointer
427 as an argument (which it then propagates). This can sometimes be useful,
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]428 allowing you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]429
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]430</dl>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]431
432<p>These five templates can be used with any classes, whether they have a
433v-table or not. To add support for these templates, you simply need to add
434<tt>classof</tt> static methods to the class you are interested casting
435to. Describing this is currently outside the scope of this document, but there
436are lots of examples in the LLVM source base.</p>
437
438</div>
439
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]440
441<!-- ======================================================================= -->
442<div class="doc_subsection">
443 <a name="string_apis">Passing strings (the <tt>StringRef</tt>
444and <tt>Twine</tt> classes)</a>
445</div>
446
447<div class="doc_text">
448
449<p>Although LLVM generally does not do much string manipulation, we do have
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]450several important APIs which take strings. Two important examples are the
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]451Value class -- which has names for instructions, functions, etc. -- and the
452StringMap class which is used extensively in LLVM and Clang.</p>
453
454<p>These are generic classes, and they need to be able to accept strings which
455may have embedded null characters. Therefore, they cannot simply take
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]456a <tt>const char *</tt>, and taking a <tt>const std::string&amp;</tt> requires
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]457clients to perform a heap allocation which is usually unnecessary. Instead,
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]458many LLVM APIs use a <tt>const StringRef&amp;</tt> or a <tt>const
459Twine&amp;</tt> for passing strings efficiently.</p>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]460
461</div>
462
463<!-- _______________________________________________________________________ -->
464<div class="doc_subsubsection">
465 <a name="StringRef">The <tt>StringRef</tt> class</a>
466</div>
467
468<div class="doc_text">
469
470<p>The <tt>StringRef</tt> data type represents a reference to a constant string
471(a character array and a length) and supports the common operations available
472on <tt>std:string</tt>, but does not require heap allocation.</p>
473
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]474<p>It can be implicitly constructed using a C style null-terminated string,
475an <tt>std::string</tt>, or explicitly with a character pointer and length.
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]476For example, the <tt>StringRef</tt> find function is declared as:</p>
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]477
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]478<div class="doc_code">
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]479 iterator find(const StringRef &amp;Key);
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]480</div>
481
482<p>and clients can call it using any one of:</p>
483
484<div class="doc_code">
485<pre>
486 Map.find("foo"); <i>// Lookup "foo"</i>
487 Map.find(std::string("bar")); <i>// Lookup "bar"</i>
488 Map.find(StringRef("\0baz", 4)); <i>// Lookup "\0baz"</i>
489</pre>
490</div>
491
492<p>Similarly, APIs which need to return a string may return a <tt>StringRef</tt>
493instance, which can be used directly or converted to an <tt>std::string</tt>
494using the <tt>str</tt> member function. See
495"<tt><a href="/doxygen/classllvm_1_1StringRef_8h-source.html">llvm/ADT/StringRef.h</a></tt>"
496for more information.</p>
497
498<p>You should rarely use the <tt>StringRef</tt> class directly, because it contains
499pointers to external memory it is not generally safe to store an instance of the
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]500class (unless you know that the external storage will not be freed).</p>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]501
502</div>
503
504<!-- _______________________________________________________________________ -->
505<div class="doc_subsubsection">
506 <a name="Twine">The <tt>Twine</tt> class</a>
507</div>
508
509<div class="doc_text">
510
511<p>The <tt>Twine</tt> class is an efficient way for APIs to accept concatenated
512strings. For example, a common LLVM paradigm is to name one instruction based on
513the name of another instruction with a suffix, for example:</p>
514
515<div class="doc_code">
516<pre>
517 New = CmpInst::Create(<i>...</i>, SO->getName() + ".cmp");
518</pre>
519</div>
520
521<p>The <tt>Twine</tt> class is effectively a
522lightweight <a href="http://en.wikipedia.org/wiki/Rope_(computer_science)">rope</a>
523which points to temporary (stack allocated) objects. Twines can be implicitly
524constructed as the result of the plus operator applied to strings (i.e., a C
525strings, an <tt>std::string</tt>, or a <tt>StringRef</tt>). The twine delays the
526actual concatentation of strings until it is actually required, at which point
527it can be efficiently rendered directly into a character array. This avoids
528unnecessary heap allocation involved in constructing the temporary results of
529string concatenation. See
530"<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]531for more information.</p>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]532
533<p>As with a <tt>StringRef</tt>, <tt>Twine</tt> objects point to external memory
534and should almost never be stored or mentioned directly. They are intended
535solely for use when defining a function which should be able to efficiently
536accept concatenated strings.</p>
537
538</div>
539
540
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]541<!-- ======================================================================= -->
542<div class="doc_subsection">
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]543 <a name="DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt> option</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]544</div>
545
546<div class="doc_text">
547
548<p>Often when working on your pass you will put a bunch of debugging printouts
549and other code into your pass. After you get it working, you want to remove
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]550it, but you may need it again in the future (to work out new bugs that you run
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]551across).</p>
552
553<p> Naturally, because of this, you don't want to delete the debug printouts,
554but you don't want them to always be noisy. A standard compromise is to comment
555them out, allowing you to enable them if you need them in the future.</p>
556
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]557<p>The "<tt><a href="/doxygen/Debug_8h-source.html">llvm/Support/Debug.h</a></tt>"
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]558file provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to
559this problem. Basically, you can put arbitrary code into the argument of the
560<tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
561tool) is run with the '<tt>-debug</tt>' command line argument:</p>
562
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]563<div class="doc_code">
564<pre>
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]565DEBUG(errs() &lt;&lt; "I am here!\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]566</pre>
567</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]568
569<p>Then you can run your pass like this:</p>
570
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]571<div class="doc_code">
572<pre>
573$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]574<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]575$ opt &lt; a.bc &gt; /dev/null -mypass -debug
576I am here!
577</pre>
578</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]579
580<p>Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you
581to not have to create "yet another" command line option for the debug output for
582your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
583so they do not cause a performance impact at all (for the same reason, they
584should also not contain side-effects!).</p>
585
586<p>One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
587enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
588"<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
589program hasn't been started yet, you can always just run it with
590<tt>-debug</tt>.</p>
591
592</div>
593
594<!-- _______________________________________________________________________ -->
595<div class="doc_subsubsection">
Chris Lattnerc9151082005-04-26 22:57:07 +0000[diff] [blame]596 <a name="DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt> and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]597 the <tt>-debug-only</tt> option</a>
598</div>
599
600<div class="doc_text">
601
602<p>Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
603just turns on <b>too much</b> information (such as when working on the code
604generator). If you want to enable debug information with more fine-grained
605control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
606option as follows:</p>
607
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]608<div class="doc_code">
609<pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]610#undef DEBUG_TYPE
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]611DEBUG(errs() &lt;&lt; "No debug type\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]612#define DEBUG_TYPE "foo"
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]613DEBUG(errs() &lt;&lt; "'foo' debug type\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]614#undef DEBUG_TYPE
615#define DEBUG_TYPE "bar"
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]616DEBUG(errs() &lt;&lt; "'bar' debug type\n"));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]617#undef DEBUG_TYPE
618#define DEBUG_TYPE ""
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]619DEBUG(errs() &lt;&lt; "No debug type (2)\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]620</pre>
621</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]622
623<p>Then you can run your pass like this:</p>
624
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]625<div class="doc_code">
626<pre>
627$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]628<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]629$ opt &lt; a.bc &gt; /dev/null -mypass -debug
630No debug type
631'foo' debug type
632'bar' debug type
633No debug type (2)
634$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=foo
635'foo' debug type
636$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=bar
637'bar' debug type
638</pre>
639</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]640
641<p>Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of
642a 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]643you <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]644<tt>#undef</tt>'s). Also, you should use names more meaningful than "foo" and
645"bar", because there is no system in place to ensure that names do not
646conflict. If two different modules use the same string, they will all be turned
647on when the name is specified. This allows, for example, all debug information
648for instruction scheduling to be enabled with <tt>-debug-type=InstrSched</tt>,
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]649even if the source lives in multiple files.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]650
Daniel Dunbarc3c92392009-08-07 23:48:59 +0000[diff] [blame]651<p>The <tt>DEBUG_WITH_TYPE</tt> macro is also available for situations where you
652would like to set <tt>DEBUG_TYPE</tt>, but only for one specific <tt>DEBUG</tt>
653statement. It takes an additional first parameter, which is the type to use. For
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]654example, the preceding example could be written as:</p>
Daniel Dunbarc3c92392009-08-07 23:48:59 +0000[diff] [blame]655
656
657<div class="doc_code">
658<pre>
659DEBUG_WITH_TYPE("", errs() &lt;&lt; "No debug type\n");
660DEBUG_WITH_TYPE("foo", errs() &lt;&lt; "'foo' debug type\n");
661DEBUG_WITH_TYPE("bar", errs() &lt;&lt; "'bar' debug type\n"));
662DEBUG_WITH_TYPE("", errs() &lt;&lt; "No debug type (2)\n");
663</pre>
664</div>
665
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]666</div>
667
668<!-- ======================================================================= -->
669<div class="doc_subsection">
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]670 <a name="Statistic">The <tt>Statistic</tt> class &amp; <tt>-stats</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]671 option</a>
672</div>
673
674<div class="doc_text">
675
676<p>The "<tt><a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]677href="/doxygen/Statistic_8h-source.html">llvm/ADT/Statistic.h</a></tt>" file
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]678provides a class named <tt>Statistic</tt> that is used as a unified way to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]679keep track of what the LLVM compiler is doing and how effective various
680optimizations are. It is useful to see what optimizations are contributing to
681making a particular program run faster.</p>
682
683<p>Often you may run your pass on some big program, and you're interested to see
684how many times it makes a certain transformation. Although you can do this with
685hand inspection, or some ad-hoc method, this is a real pain and not very useful
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]686for big programs. Using the <tt>Statistic</tt> class makes it very easy to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]687keep track of this information, and the calculated information is presented in a
688uniform manner with the rest of the passes being executed.</p>
689
690<p>There are many examples of <tt>Statistic</tt> uses, but the basics of using
691it are as follows:</p>
692
693<ol>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]694 <li><p>Define your statistic like this:</p>
695
696<div class="doc_code">
697<pre>
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]698#define <a href="#DEBUG_TYPE">DEBUG_TYPE</a> "mypassname" <i>// This goes before any #includes.</i>
699STATISTIC(NumXForms, "The # of times I did stuff");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]700</pre>
701</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]702
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]703 <p>The <tt>STATISTIC</tt> macro defines a static variable, whose name is
704 specified by the first argument. The pass name is taken from the DEBUG_TYPE
705 macro, and the description is taken from the second argument. The variable
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]706 defined ("NumXForms" in this case) acts like an unsigned integer.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]707
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]708 <li><p>Whenever you make a transformation, bump the counter:</p>
709
710<div class="doc_code">
711<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]712++NumXForms; // <i>I did stuff!</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]713</pre>
714</div>
715
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]716 </li>
717 </ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]718
719 <p>That's all you have to do. To get '<tt>opt</tt>' to print out the
720 statistics gathered, use the '<tt>-stats</tt>' option:</p>
721
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]722<div class="doc_code">
723<pre>
724$ opt -stats -mypassname &lt; program.bc &gt; /dev/null
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]725<i>... statistics output ...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]726</pre>
727</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]728
Reid Spencer6b6c73e2007-02-09 16:00:28 +0000[diff] [blame]729 <p> When running <tt>opt</tt> on a C file from the SPEC benchmark
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]730suite, it gives a report that looks like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]731
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]732<div class="doc_code">
733<pre>
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]734 7646 bitcodewriter - Number of normal instructions
735 725 bitcodewriter - Number of oversized instructions
736 129996 bitcodewriter - Number of bitcode bytes written
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]737 2817 raise - Number of insts DCEd or constprop'd
738 3213 raise - Number of cast-of-self removed
739 5046 raise - Number of expression trees converted
740 75 raise - Number of other getelementptr's formed
741 138 raise - Number of load/store peepholes
742 42 deadtypeelim - Number of unused typenames removed from symtab
743 392 funcresolve - Number of varargs functions resolved
744 27 globaldce - Number of global variables removed
745 2 adce - Number of basic blocks removed
746 134 cee - Number of branches revectored
747 49 cee - Number of setcc instruction eliminated
748 532 gcse - Number of loads removed
749 2919 gcse - Number of instructions removed
750 86 indvars - Number of canonical indvars added
751 87 indvars - Number of aux indvars removed
752 25 instcombine - Number of dead inst eliminate
753 434 instcombine - Number of insts combined
754 248 licm - Number of load insts hoisted
755 1298 licm - Number of insts hoisted to a loop pre-header
756 3 licm - Number of insts hoisted to multiple loop preds (bad, no loop pre-header)
757 75 mem2reg - Number of alloca's promoted
758 1444 cfgsimplify - Number of blocks simplified
759</pre>
760</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]761
762<p>Obviously, with so many optimizations, having a unified framework for this
763stuff is very nice. Making your pass fit well into the framework makes it more
764maintainable and useful.</p>
765
766</div>
767
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]768<!-- ======================================================================= -->
769<div class="doc_subsection">
770 <a name="ViewGraph">Viewing graphs while debugging code</a>
771</div>
772
773<div class="doc_text">
774
775<p>Several of the important data structures in LLVM are graphs: for example
776CFGs made out of LLVM <a href="#BasicBlock">BasicBlock</a>s, CFGs made out of
777LLVM <a href="CodeGenerator.html#machinebasicblock">MachineBasicBlock</a>s, and
778<a href="CodeGenerator.html#selectiondag_intro">Instruction Selection
779DAGs</a>. In many cases, while debugging various parts of the compiler, it is
780nice to instantly visualize these graphs.</p>
781
782<p>LLVM provides several callbacks that are available in a debug build to do
783exactly that. If you call the <tt>Function::viewCFG()</tt> method, for example,
784the current LLVM tool will pop up a window containing the CFG for the function
785where each basic block is a node in the graph, and each node contains the
786instructions in the block. Similarly, there also exists
787<tt>Function::viewCFGOnly()</tt> (does not include the instructions), the
788<tt>MachineFunction::viewCFG()</tt> and <tt>MachineFunction::viewCFGOnly()</tt>,
789and the <tt>SelectionDAG::viewGraph()</tt> methods. Within GDB, for example,
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]790you can usually use something like <tt>call DAG.viewGraph()</tt> to pop
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]791up a window. Alternatively, you can sprinkle calls to these functions in your
792code in places you want to debug.</p>
793
794<p>Getting this to work requires a small amount of configuration. On Unix
795systems with X11, install the <a href="http://www.graphviz.org">graphviz</a>
796toolkit, and make sure 'dot' and 'gv' are in your path. If you are running on
797Mac OS/X, download and install the Mac OS/X <a
798href="http://www.pixelglow.com/graphviz/">Graphviz program</a>, and add
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]799<tt>/Applications/Graphviz.app/Contents/MacOS/</tt> (or wherever you install
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]800it) to your path. Once in your system and path are set up, rerun the LLVM
801configure script and rebuild LLVM to enable this functionality.</p>
802
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]803<p><tt>SelectionDAG</tt> has been extended to make it easier to locate
804<i>interesting</i> nodes in large complex graphs. From gdb, if you
805<tt>call DAG.setGraphColor(<i>node</i>, "<i>color</i>")</tt>, then the
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]806next <tt>call DAG.viewGraph()</tt> would highlight the node in the
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]807specified color (choices of colors can be found at <a
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]808href="http://www.graphviz.org/doc/info/colors.html">colors</a>.) More
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]809complex node attributes can be provided with <tt>call
810DAG.setGraphAttrs(<i>node</i>, "<i>attributes</i>")</tt> (choices can be
811found at <a href="http://www.graphviz.org/doc/info/attrs.html">Graph
812Attributes</a>.) If you want to restart and clear all the current graph
813attributes, then you can <tt>call DAG.clearGraphAttrs()</tt>. </p>
814
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]815</div>
816
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]817<!-- *********************************************************************** -->
818<div class="doc_section">
819 <a name="datastructure">Picking the Right Data Structure for a Task</a>
820</div>
821<!-- *********************************************************************** -->
822
823<div class="doc_text">
824
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]825<p>LLVM has a plethora of data structures in the <tt>llvm/ADT/</tt> directory,
826 and we commonly use STL data structures. This section describes the trade-offs
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]827 you should consider when you pick one.</p>
828
829<p>
830The first step is a choose your own adventure: do you want a sequential
831container, a set-like container, or a map-like container? The most important
832thing when choosing a container is the algorithmic properties of how you plan to
833access the container. Based on that, you should use:</p>
834
835<ul>
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]836<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]837 of an value based on another value. Map-like containers also support
838 efficient queries for containment (whether a key is in the map). Map-like
839 containers generally do not support efficient reverse mapping (values to
840 keys). If you need that, use two maps. Some map-like containers also
841 support efficient iteration through the keys in sorted order. Map-like
842 containers are the most expensive sort, only use them if you need one of
843 these capabilities.</li>
844
845<li>a <a href="#ds_set">set-like</a> container if you need to put a bunch of
846 stuff into a container that automatically eliminates duplicates. Some
847 set-like containers support efficient iteration through the elements in
848 sorted order. Set-like containers are more expensive than sequential
849 containers.
850</li>
851
852<li>a <a href="#ds_sequential">sequential</a> container provides
853 the most efficient way to add elements and keeps track of the order they are
854 added to the collection. They permit duplicates and support efficient
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]855 iteration, but do not support efficient look-up based on a key.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]856</li>
857
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]858<li>a <a href="#ds_string">string</a> container is a specialized sequential
859 container or reference structure that is used for character or byte
860 arrays.</li>
861
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]862<li>a <a href="#ds_bit">bit</a> container provides an efficient way to store and
863 perform set operations on sets of numeric id's, while automatically
864 eliminating duplicates. Bit containers require a maximum of 1 bit for each
865 identifier you want to store.
866</li>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]867</ul>
868
869<p>
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]870Once the proper category of container is determined, you can fine tune the
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]871memory use, constant factors, and cache behaviors of access by intelligently
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]872picking a member of the category. Note that constant factors and cache behavior
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]873can be a big deal. If you have a vector that usually only contains a few
874elements (but could contain many), for example, it's much better to use
875<a href="#dss_smallvector">SmallVector</a> than <a href="#dss_vector">vector</a>
876. Doing so avoids (relatively) expensive malloc/free calls, which dwarf the
877cost of adding the elements to the container. </p>
878
879</div>
880
881<!-- ======================================================================= -->
882<div class="doc_subsection">
883 <a name="ds_sequential">Sequential Containers (std::vector, std::list, etc)</a>
884</div>
885
886<div class="doc_text">
887There are a variety of sequential containers available for you, based on your
888needs. Pick the first in this section that will do what you want.
889</div>
890
891<!-- _______________________________________________________________________ -->
892<div class="doc_subsubsection">
893 <a name="dss_fixedarrays">Fixed Size Arrays</a>
894</div>
895
896<div class="doc_text">
897<p>Fixed size arrays are very simple and very fast. They are good if you know
898exactly how many elements you have, or you have a (low) upper bound on how many
899you have.</p>
900</div>
901
902<!-- _______________________________________________________________________ -->
903<div class="doc_subsubsection">
904 <a name="dss_heaparrays">Heap Allocated Arrays</a>
905</div>
906
907<div class="doc_text">
908<p>Heap allocated arrays (new[] + delete[]) are also simple. They are good if
909the number of elements is variable, if you know how many elements you will need
910before the array is allocated, and if the array is usually large (if not,
911consider a <a href="#dss_smallvector">SmallVector</a>). The cost of a heap
912allocated array is the cost of the new/delete (aka malloc/free). Also note that
913if you are allocating an array of a type with a constructor, the constructor and
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]914destructors will be run for every element in the array (re-sizable vectors only
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]915construct those elements actually used).</p>
916</div>
917
918<!-- _______________________________________________________________________ -->
919<div class="doc_subsubsection">
920 <a name="dss_smallvector">"llvm/ADT/SmallVector.h"</a>
921</div>
922
923<div class="doc_text">
924<p><tt>SmallVector&lt;Type, N&gt;</tt> is a simple class that looks and smells
925just like <tt>vector&lt;Type&gt;</tt>:
926it supports efficient iteration, lays out elements in memory order (so you can
927do pointer arithmetic between elements), supports efficient push_back/pop_back
928operations, supports efficient random access to its elements, etc.</p>
929
930<p>The advantage of SmallVector is that it allocates space for
931some number of elements (N) <b>in the object itself</b>. Because of this, if
932the SmallVector is dynamically smaller than N, no malloc is performed. This can
933be a big win in cases where the malloc/free call is far more expensive than the
934code that fiddles around with the elements.</p>
935
936<p>This is good for vectors that are "usually small" (e.g. the number of
937predecessors/successors of a block is usually less than 8). On the other hand,
938this makes the size of the SmallVector itself large, so you don't want to
939allocate lots of them (doing so will waste a lot of space). As such,
940SmallVectors are most useful when on the stack.</p>
941
942<p>SmallVector also provides a nice portable and efficient replacement for
943<tt>alloca</tt>.</p>
944
945</div>
946
947<!-- _______________________________________________________________________ -->
948<div class="doc_subsubsection">
949 <a name="dss_vector">&lt;vector&gt;</a>
950</div>
951
952<div class="doc_text">
953<p>
954std::vector is well loved and respected. It is useful when SmallVector isn't:
955when the size of the vector is often large (thus the small optimization will
956rarely be a benefit) or if you will be allocating many instances of the vector
957itself (which would waste space for elements that aren't in the container).
958vector is also useful when interfacing with code that expects vectors :).
959</p>
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]960
961<p>One worthwhile note about std::vector: avoid code like this:</p>
962
963<div class="doc_code">
964<pre>
965for ( ... ) {
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]966 std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]967 use V;
968}
969</pre>
970</div>
971
972<p>Instead, write this as:</p>
973
974<div class="doc_code">
975<pre>
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]976std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]977for ( ... ) {
978 use V;
979 V.clear();
980}
981</pre>
982</div>
983
984<p>Doing so will save (at least) one heap allocation and free per iteration of
985the loop.</p>
986
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]987</div>
988
989<!-- _______________________________________________________________________ -->
990<div class="doc_subsubsection">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]991 <a name="dss_deque">&lt;deque&gt;</a>
992</div>
993
994<div class="doc_text">
995<p>std::deque is, in some senses, a generalized version of std::vector. Like
996std::vector, it provides constant time random access and other similar
997properties, but it also provides efficient access to the front of the list. It
998does not guarantee continuity of elements within memory.</p>
999
1000<p>In exchange for this extra flexibility, std::deque has significantly higher
1001constant factor costs than std::vector. If possible, use std::vector or
1002something cheaper.</p>
1003</div>
1004
1005<!-- _______________________________________________________________________ -->
1006<div class="doc_subsubsection">
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1007 <a name="dss_list">&lt;list&gt;</a>
1008</div>
1009
1010<div class="doc_text">
1011<p>std::list is an extremely inefficient class that is rarely useful.
1012It performs a heap allocation for every element inserted into it, thus having an
1013extremely high constant factor, particularly for small data types. std::list
1014also only supports bidirectional iteration, not random access iteration.</p>
1015
1016<p>In exchange for this high cost, std::list supports efficient access to both
1017ends of the list (like std::deque, but unlike std::vector or SmallVector). In
1018addition, the iterator invalidation characteristics of std::list are stronger
1019than that of a vector class: inserting or removing an element into the list does
1020not invalidate iterator or pointers to other elements in the list.</p>
1021</div>
1022
1023<!-- _______________________________________________________________________ -->
1024<div class="doc_subsubsection">
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1025 <a name="dss_ilist">llvm/ADT/ilist.h</a>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1026</div>
1027
1028<div class="doc_text">
1029<p><tt>ilist&lt;T&gt;</tt> implements an 'intrusive' doubly-linked list. It is
1030intrusive, because it requires the element to store and provide access to the
1031prev/next pointers for the list.</p>
1032
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1033<p><tt>ilist</tt> has the same drawbacks as <tt>std::list</tt>, and additionally
1034requires an <tt>ilist_traits</tt> implementation for the element type, but it
1035provides some novel characteristics. In particular, it can efficiently store
1036polymorphic objects, the traits class is informed when an element is inserted or
Gabor Greif0cbcabe2009-03-12 09:47:03 +0000[diff] [blame]1037removed from the list, and <tt>ilist</tt>s are guaranteed to support a
1038constant-time splice operation.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1039
Gabor Greif0cbcabe2009-03-12 09:47:03 +0000[diff] [blame]1040<p>These properties are exactly what we want for things like
1041<tt>Instruction</tt>s and basic blocks, which is why these are implemented with
1042<tt>ilist</tt>s.</p>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1043
1044Related classes of interest are explained in the following subsections:
1045 <ul>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1046 <li><a href="#dss_ilist_traits">ilist_traits</a></li>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1047 <li><a href="#dss_iplist">iplist</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1048 <li><a href="#dss_ilist_node">llvm/ADT/ilist_node.h</a></li>
Gabor Greif6a65f422009-03-12 10:30:31 +0000[diff] [blame]1049 <li><a href="#dss_ilist_sentinel">Sentinels</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1050 </ul>
1051</div>
1052
1053<!-- _______________________________________________________________________ -->
1054<div class="doc_subsubsection">
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1055 <a name="dss_ilist_traits">ilist_traits</a>
1056</div>
1057
1058<div class="doc_text">
1059<p><tt>ilist_traits&lt;T&gt;</tt> is <tt>ilist&lt;T&gt;</tt>'s customization
1060mechanism. <tt>iplist&lt;T&gt;</tt> (and consequently <tt>ilist&lt;T&gt;</tt>)
1061publicly derive from this traits class.</p>
1062</div>
1063
1064<!-- _______________________________________________________________________ -->
1065<div class="doc_subsubsection">
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1066 <a name="dss_iplist">iplist</a>
1067</div>
1068
1069<div class="doc_text">
1070<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]1071supports a slightly narrower interface. Notably, inserters from
1072<tt>T&amp;</tt> are absent.</p>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1073
1074<p><tt>ilist_traits&lt;T&gt;</tt> is a public base of this class and can be
1075used for a wide variety of customizations.</p>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1076</div>
1077
1078<!-- _______________________________________________________________________ -->
1079<div class="doc_subsubsection">
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1080 <a name="dss_ilist_node">llvm/ADT/ilist_node.h</a>
1081</div>
1082
1083<div class="doc_text">
1084<p><tt>ilist_node&lt;T&gt;</tt> implements a the forward and backward links
1085that are expected by the <tt>ilist&lt;T&gt;</tt> (and analogous containers)
1086in the default manner.</p>
1087
1088<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]1089<tt>T</tt>, usually <tt>T</tt> publicly derives from
1090<tt>ilist_node&lt;T&gt;</tt>.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1091</div>
1092
1093<!-- _______________________________________________________________________ -->
1094<div class="doc_subsubsection">
Gabor Greif6a65f422009-03-12 10:30:31 +0000[diff] [blame]1095 <a name="dss_ilist_sentinel">Sentinels</a>
1096</div>
1097
1098<div class="doc_text">
1099<p><tt>ilist</tt>s have another speciality that must be considered. To be a good
1100citizen in the C++ ecosystem, it needs to support the standard container
1101operations, such as <tt>begin</tt> and <tt>end</tt> iterators, etc. Also, the
1102<tt>operator--</tt> must work correctly on the <tt>end</tt> iterator in the
1103case of non-empty <tt>ilist</tt>s.</p>
1104
1105<p>The only sensible solution to this problem is to allocate a so-called
1106<i>sentinel</i> along with the intrusive list, which serves as the <tt>end</tt>
1107iterator, providing the back-link to the last element. However conforming to the
1108C++ convention it is illegal to <tt>operator++</tt> beyond the sentinel and it
1109also must not be dereferenced.</p>
1110
1111<p>These constraints allow for some implementation freedom to the <tt>ilist</tt>
1112how to allocate and store the sentinel. The corresponding policy is dictated
1113by <tt>ilist_traits&lt;T&gt;</tt>. By default a <tt>T</tt> gets heap-allocated
1114whenever the need for a sentinel arises.</p>
1115
1116<p>While the default policy is sufficient in most cases, it may break down when
1117<tt>T</tt> does not provide a default constructor. Also, in the case of many
1118instances of <tt>ilist</tt>s, the memory overhead of the associated sentinels
1119is wasted. To alleviate the situation with numerous and voluminous
1120<tt>T</tt>-sentinels, sometimes a trick is employed, leading to <i>ghostly
1121sentinels</i>.</p>
1122
1123<p>Ghostly sentinels are obtained by specially-crafted <tt>ilist_traits&lt;T&gt;</tt>
1124which superpose the sentinel with the <tt>ilist</tt> instance in memory. Pointer
1125arithmetic is used to obtain the sentinel, which is relative to the
1126<tt>ilist</tt>'s <tt>this</tt> pointer. The <tt>ilist</tt> is augmented by an
1127extra pointer, which serves as the back-link of the sentinel. This is the only
1128field in the ghostly sentinel which can be legally accessed.</p>
1129</div>
1130
1131<!-- _______________________________________________________________________ -->
1132<div class="doc_subsubsection">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1133 <a name="dss_other">Other Sequential Container options</a>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1134</div>
1135
1136<div class="doc_text">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1137<p>Other STL containers are available, such as std::string.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1138
1139<p>There are also various STL adapter classes such as std::queue,
1140std::priority_queue, std::stack, etc. These provide simplified access to an
1141underlying container but don't affect the cost of the container itself.</p>
1142
1143</div>
1144
1145
1146<!-- ======================================================================= -->
1147<div class="doc_subsection">
1148 <a name="ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
1149</div>
1150
1151<div class="doc_text">
1152
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1153<p>Set-like containers are useful when you need to canonicalize multiple values
1154into a single representation. There are several different choices for how to do
1155this, providing various trade-offs.</p>
1156
1157</div>
1158
1159
1160<!-- _______________________________________________________________________ -->
1161<div class="doc_subsubsection">
1162 <a name="dss_sortedvectorset">A sorted 'vector'</a>
1163</div>
1164
1165<div class="doc_text">
1166
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]1167<p>If you intend to insert a lot of elements, then do a lot of queries, a
1168great approach is to use a vector (or other sequential container) with
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1169std::sort+std::unique to remove duplicates. This approach works really well if
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]1170your usage pattern has these two distinct phases (insert then query), and can be
1171coupled with a good choice of <a href="#ds_sequential">sequential container</a>.
1172</p>
1173
1174<p>
1175This combination provides the several nice properties: the result data is
1176contiguous in memory (good for cache locality), has few allocations, is easy to
1177address (iterators in the final vector are just indices or pointers), and can be
1178efficiently queried with a standard binary or radix search.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1179
1180</div>
1181
1182<!-- _______________________________________________________________________ -->
1183<div class="doc_subsubsection">
1184 <a name="dss_smallset">"llvm/ADT/SmallSet.h"</a>
1185</div>
1186
1187<div class="doc_text">
1188
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1189<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]1190are 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]1191has space for N elements in place (thus, if the set is dynamically smaller than
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1192N, no malloc traffic is required) and accesses them with a simple linear search.
1193When the set grows beyond 'N' elements, it allocates a more expensive representation that
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1194guarantees efficient access (for most types, it falls back to std::set, but for
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1195pointers it uses something far better, <a
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1196href="#dss_smallptrset">SmallPtrSet</a>).</p>
1197
1198<p>The magic of this class is that it handles small sets extremely efficiently,
1199but gracefully handles extremely large sets without loss of efficiency. The
1200drawback is that the interface is quite small: it supports insertion, queries
1201and erasing, but does not support iteration.</p>
1202
1203</div>
1204
1205<!-- _______________________________________________________________________ -->
1206<div class="doc_subsubsection">
1207 <a name="dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a>
1208</div>
1209
1210<div class="doc_text">
1211
1212<p>SmallPtrSet has all the advantages of SmallSet (and a SmallSet of pointers is
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1213transparently implemented with a SmallPtrSet), but also supports iterators. If
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1214more than 'N' insertions are performed, a single quadratically
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1215probed hash table is allocated and grows as needed, providing extremely
1216efficient access (constant time insertion/deleting/queries with low constant
1217factors) and is very stingy with malloc traffic.</p>
1218
1219<p>Note that, unlike std::set, the iterators of SmallPtrSet are invalidated
1220whenever an insertion occurs. Also, the values visited by the iterators are not
1221visited in sorted order.</p>
1222
1223</div>
1224
1225<!-- _______________________________________________________________________ -->
1226<div class="doc_subsubsection">
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]1227 <a name="dss_denseset">"llvm/ADT/DenseSet.h"</a>
1228</div>
1229
1230<div class="doc_text">
1231
1232<p>
1233DenseSet is a simple quadratically probed hash table. It excels at supporting
1234small values: it uses a single allocation to hold all of the pairs that
1235are currently inserted in the set. DenseSet is a great way to unique small
1236values that are not simple pointers (use <a
1237href="#dss_smallptrset">SmallPtrSet</a> for pointers). Note that DenseSet has
1238the same requirements for the value type that <a
1239href="#dss_densemap">DenseMap</a> has.
1240</p>
1241
1242</div>
1243
1244<!-- _______________________________________________________________________ -->
1245<div class="doc_subsubsection">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1246 <a name="dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a>
1247</div>
1248
1249<div class="doc_text">
1250
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1251<p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1252FoldingSet is an aggregate class that is really good at uniquing
1253expensive-to-create or polymorphic objects. It is a combination of a chained
1254hash table with intrusive links (uniqued objects are required to inherit from
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1255FoldingSetNode) that uses <a href="#dss_smallvector">SmallVector</a> as part of
1256its ID process.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1257
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1258<p>Consider a case where you want to implement a "getOrCreateFoo" method for
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1259a complex object (for example, a node in the code generator). The client has a
1260description of *what* it wants to generate (it knows the opcode and all the
1261operands), 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]1262only to find out it already exists, at which point we would have to delete it
1263and return the node that already exists.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1264</p>
1265
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1266<p>To support this style of client, FoldingSet perform a query with a
1267FoldingSetNodeID (which wraps SmallVector) that can be used to describe the
1268element that we want to query for. The query either returns the element
1269matching the ID or it returns an opaque ID that indicates where insertion should
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1270take place. Construction of the ID usually does not require heap traffic.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1271
1272<p>Because FoldingSet uses intrusive links, it can support polymorphic objects
1273in the set (for example, you can have SDNode instances mixed with LoadSDNodes).
1274Because the elements are individually allocated, pointers to the elements are
1275stable: inserting or removing elements does not invalidate any pointers to other
1276elements.
1277</p>
1278
1279</div>
1280
1281<!-- _______________________________________________________________________ -->
1282<div class="doc_subsubsection">
1283 <a name="dss_set">&lt;set&gt;</a>
1284</div>
1285
1286<div class="doc_text">
1287
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1288<p><tt>std::set</tt> is a reasonable all-around set class, which is decent at
1289many things but great at nothing. std::set allocates memory for each element
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1290inserted (thus it is very malloc intensive) and typically stores three pointers
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1291per element in the set (thus adding a large amount of per-element space
1292overhead). It offers guaranteed log(n) performance, which is not particularly
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1293fast from a complexity standpoint (particularly if the elements of the set are
1294expensive to compare, like strings), and has extremely high constant factors for
1295lookup, insertion and removal.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1296
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1297<p>The advantages of std::set are that its iterators are stable (deleting or
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1298inserting an element from the set does not affect iterators or pointers to other
1299elements) and that iteration over the set is guaranteed to be in sorted order.
1300If the elements in the set are large, then the relative overhead of the pointers
1301and malloc traffic is not a big deal, but if the elements of the set are small,
1302std::set is almost never a good choice.</p>
1303
1304</div>
1305
1306<!-- _______________________________________________________________________ -->
1307<div class="doc_subsubsection">
1308 <a name="dss_setvector">"llvm/ADT/SetVector.h"</a>
1309</div>
1310
1311<div class="doc_text">
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1312<p>LLVM's SetVector&lt;Type&gt; is an adapter class that combines your choice of
1313a set-like container along with a <a href="#ds_sequential">Sequential
1314Container</a>. The important property
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1315that this provides is efficient insertion with uniquing (duplicate elements are
1316ignored) with iteration support. It implements this by inserting elements into
1317both a set-like container and the sequential container, using the set-like
1318container for uniquing and the sequential container for iteration.
1319</p>
1320
1321<p>The difference between SetVector and other sets is that the order of
1322iteration is guaranteed to match the order of insertion into the SetVector.
1323This property is really important for things like sets of pointers. Because
1324pointer values are non-deterministic (e.g. vary across runs of the program on
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1325different machines), iterating over the pointers in the set will
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1326not be in a well-defined order.</p>
1327
1328<p>
1329The drawback of SetVector is that it requires twice as much space as a normal
1330set and has the sum of constant factors from the set-like container and the
1331sequential container that it uses. Use it *only* if you need to iterate over
1332the elements in a deterministic order. SetVector is also expensive to delete
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1333elements out of (linear time), unless you use it's "pop_back" method, which is
1334faster.
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1335</p>
1336
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1337<p>SetVector is an adapter class that defaults to using std::vector and std::set
1338for the underlying containers, so it is quite expensive. However,
1339<tt>"llvm/ADT/SetVector.h"</tt> also provides a SmallSetVector class, which
1340defaults to using a SmallVector and SmallSet of a specified size. If you use
1341this, and if your sets are dynamically smaller than N, you will save a lot of
1342heap traffic.</p>
1343
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1344</div>
1345
1346<!-- _______________________________________________________________________ -->
1347<div class="doc_subsubsection">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1348 <a name="dss_uniquevector">"llvm/ADT/UniqueVector.h"</a>
1349</div>
1350
1351<div class="doc_text">
1352
1353<p>
1354UniqueVector is similar to <a href="#dss_setvector">SetVector</a>, but it
1355retains a unique ID for each element inserted into the set. It internally
1356contains a map and a vector, and it assigns a unique ID for each value inserted
1357into the set.</p>
1358
1359<p>UniqueVector is very expensive: its cost is the sum of the cost of
1360maintaining both the map and vector, it has high complexity, high constant
1361factors, and produces a lot of malloc traffic. It should be avoided.</p>
1362
1363</div>
1364
1365
1366<!-- _______________________________________________________________________ -->
1367<div class="doc_subsubsection">
1368 <a name="dss_otherset">Other Set-Like Container Options</a>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1369</div>
1370
1371<div class="doc_text">
1372
1373<p>
1374The STL provides several other options, such as std::multiset and the various
Chris Lattnerf1a30822009-03-09 05:20:45 +0000[diff] [blame]1375"hash_set" like containers (whether from C++ TR1 or from the SGI library). We
1376never use hash_set and unordered_set because they are generally very expensive
1377(each insertion requires a malloc) and very non-portable.
1378</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1379
1380<p>std::multiset is useful if you're not interested in elimination of
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1381duplicates, but has all the drawbacks of std::set. A sorted vector (where you
1382don't delete duplicate entries) or some other approach is almost always
1383better.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1384
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1385</div>
1386
1387<!-- ======================================================================= -->
1388<div class="doc_subsection">
1389 <a name="ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
1390</div>
1391
1392<div class="doc_text">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1393Map-like containers are useful when you want to associate data to a key. As
1394usual, there are a lot of different ways to do this. :)
1395</div>
1396
1397<!-- _______________________________________________________________________ -->
1398<div class="doc_subsubsection">
1399 <a name="dss_sortedvectormap">A sorted 'vector'</a>
1400</div>
1401
1402<div class="doc_text">
1403
1404<p>
1405If your usage pattern follows a strict insert-then-query approach, you can
1406trivially use the same approach as <a href="#dss_sortedvectorset">sorted vectors
1407for set-like containers</a>. The only difference is that your query function
1408(which uses std::lower_bound to get efficient log(n) lookup) should only compare
1409the key, not both the key and value. This yields the same advantages as sorted
1410vectors for sets.
1411</p>
1412</div>
1413
1414<!-- _______________________________________________________________________ -->
1415<div class="doc_subsubsection">
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1416 <a name="dss_stringmap">"llvm/ADT/StringMap.h"</a>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1417</div>
1418
1419<div class="doc_text">
1420
1421<p>
1422Strings are commonly used as keys in maps, and they are difficult to support
1423efficiently: they are variable length, inefficient to hash and compare when
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1424long, expensive to copy, etc. StringMap is a specialized container designed to
1425cope with these issues. It supports mapping an arbitrary range of bytes to an
1426arbitrary other object.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1427
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1428<p>The StringMap implementation uses a quadratically-probed hash table, where
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1429the buckets store a pointer to the heap allocated entries (and some other
1430stuff). The entries in the map must be heap allocated because the strings are
1431variable length. The string data (key) and the element object (value) are
1432stored in the same allocation with the string data immediately after the element
1433object. This container guarantees the "<tt>(char*)(&amp;Value+1)</tt>" points
1434to the key string for a value.</p>
1435
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1436<p>The StringMap is very fast for several reasons: quadratic probing is very
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1437cache efficient for lookups, the hash value of strings in buckets is not
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1438recomputed when lookup up an element, StringMap rarely has to touch the
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1439memory for unrelated objects when looking up a value (even when hash collisions
1440happen), hash table growth does not recompute the hash values for strings
1441already in the table, and each pair in the map is store in a single allocation
1442(the string data is stored in the same allocation as the Value of a pair).</p>
1443
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1444<p>StringMap also provides query methods that take byte ranges, so it only ever
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1445copies a string if a value is inserted into the table.</p>
1446</div>
1447
1448<!-- _______________________________________________________________________ -->
1449<div class="doc_subsubsection">
1450 <a name="dss_indexedmap">"llvm/ADT/IndexedMap.h"</a>
1451</div>
1452
1453<div class="doc_text">
1454<p>
1455IndexedMap is a specialized container for mapping small dense integers (or
1456values that can be mapped to small dense integers) to some other type. It is
1457internally implemented as a vector with a mapping function that maps the keys to
1458the dense integer range.
1459</p>
1460
1461<p>
1462This is useful for cases like virtual registers in the LLVM code generator: they
1463have a dense mapping that is offset by a compile-time constant (the first
1464virtual register ID).</p>
1465
1466</div>
1467
1468<!-- _______________________________________________________________________ -->
1469<div class="doc_subsubsection">
1470 <a name="dss_densemap">"llvm/ADT/DenseMap.h"</a>
1471</div>
1472
1473<div class="doc_text">
1474
1475<p>
1476DenseMap is a simple quadratically probed hash table. It excels at supporting
1477small keys and values: it uses a single allocation to hold all of the pairs that
1478are currently inserted in the map. DenseMap is a great way to map pointers to
1479pointers, or map other small types to each other.
1480</p>
1481
1482<p>
1483There are several aspects of DenseMap that you should be aware of, however. The
1484iterators in a densemap are invalidated whenever an insertion occurs, unlike
1485map. Also, because DenseMap allocates space for a large number of key/value
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1486pairs (it starts with 64 by default), it will waste a lot of space if your keys
1487or values are large. Finally, you must implement a partial specialization of
Chris Lattner76c1b972007-09-17 18:34:04 +0000[diff] [blame]1488DenseMapInfo for the key that you want, if it isn't already supported. This
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1489is required to tell DenseMap about two special marker values (which can never be
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1490inserted into the map) that it needs internally.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1491
1492</div>
1493
1494<!-- _______________________________________________________________________ -->
1495<div class="doc_subsubsection">
Jeffrey Yasskin71a5c222009-10-22 22:11:22 +0000[diff] [blame]1496 <a name="dss_valuemap">"llvm/ADT/ValueMap.h"</a>
1497</div>
1498
1499<div class="doc_text">
1500
1501<p>
1502ValueMap is a wrapper around a <a href="#dss_densemap">DenseMap</a> mapping
1503Value*s (or subclasses) to another type. When a Value is deleted or RAUW'ed,
1504ValueMap will update itself so the new version of the key is mapped to the same
1505value, just as if the key were a WeakVH. You can configure exactly how this
1506happens, and what else happens on these two events, by passing
1507a <code>Config</code> parameter to the ValueMap template.</p>
1508
1509</div>
1510
1511<!-- _______________________________________________________________________ -->
1512<div class="doc_subsubsection">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1513 <a name="dss_map">&lt;map&gt;</a>
1514</div>
1515
1516<div class="doc_text">
1517
1518<p>
1519std::map has similar characteristics to <a href="#dss_set">std::set</a>: it uses
1520a single allocation per pair inserted into the map, it offers log(n) lookup with
1521an extremely large constant factor, imposes a space penalty of 3 pointers per
1522pair in the map, etc.</p>
1523
1524<p>std::map is most useful when your keys or values are very large, if you need
1525to iterate over the collection in sorted order, or if you need stable iterators
1526into the map (i.e. they don't get invalidated if an insertion or deletion of
1527another element takes place).</p>
1528
1529</div>
1530
1531<!-- _______________________________________________________________________ -->
1532<div class="doc_subsubsection">
1533 <a name="dss_othermap">Other Map-Like Container Options</a>
1534</div>
1535
1536<div class="doc_text">
1537
1538<p>
1539The STL provides several other options, such as std::multimap and the various
Chris Lattnerf1a30822009-03-09 05:20:45 +0000[diff] [blame]1540"hash_map" like containers (whether from C++ TR1 or from the SGI library). We
1541never use hash_set and unordered_set because they are generally very expensive
1542(each insertion requires a malloc) and very non-portable.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1543
1544<p>std::multimap is useful if you want to map a key to multiple values, but has
1545all the drawbacks of std::map. A sorted vector or some other approach is almost
1546always better.</p>
1547
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1548</div>
1549
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1550<!-- ======================================================================= -->
1551<div class="doc_subsection">
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]1552 <a name="ds_string">String-like containers</a>
1553</div>
1554
1555<div class="doc_text">
1556
1557<p>
1558TODO: const char* vs stringref vs smallstring vs std::string. Describe twine,
1559xref to #string_apis.
1560</p>
1561
1562</div>
1563
1564<!-- ======================================================================= -->
1565<div class="doc_subsection">
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1566 <a name="ds_bit">Bit storage containers (BitVector, SparseBitVector)</a>
1567</div>
1568
1569<div class="doc_text">
Chris Lattner7086ce72007-09-25 22:37:50 +0000[diff] [blame]1570<p>Unlike the other containers, there are only two bit storage containers, and
1571choosing when to use each is relatively straightforward.</p>
1572
1573<p>One additional option is
1574<tt>std::vector&lt;bool&gt;</tt>: we discourage its use for two reasons 1) the
1575implementation in many common compilers (e.g. commonly available versions of
1576GCC) is extremely inefficient and 2) the C++ standards committee is likely to
1577deprecate this container and/or change it significantly somehow. In any case,
1578please don't use it.</p>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1579</div>
1580
1581<!-- _______________________________________________________________________ -->
1582<div class="doc_subsubsection">
1583 <a name="dss_bitvector">BitVector</a>
1584</div>
1585
1586<div class="doc_text">
1587<p> The BitVector container provides a fixed size set of bits for manipulation.
1588It supports individual bit setting/testing, as well as set operations. The set
1589operations take time O(size of bitvector), but operations are performed one word
1590at a time, instead of one bit at a time. This makes the BitVector very fast for
1591set operations compared to other containers. Use the BitVector when you expect
1592the number of set bits to be high (IE a dense set).
1593</p>
1594</div>
1595
1596<!-- _______________________________________________________________________ -->
1597<div class="doc_subsubsection">
1598 <a name="dss_sparsebitvector">SparseBitVector</a>
1599</div>
1600
1601<div class="doc_text">
1602<p> The SparseBitVector container is much like BitVector, with one major
1603difference: Only the bits that are set, are stored. This makes the
1604SparseBitVector much more space efficient than BitVector when the set is sparse,
1605as well as making set operations O(number of set bits) instead of O(size of
1606universe). 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
1607(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).
1608</p>
1609</div>
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]1610
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1611<!-- *********************************************************************** -->
1612<div class="doc_section">
1613 <a name="common">Helpful Hints for Common Operations</a>
1614</div>
1615<!-- *********************************************************************** -->
1616
1617<div class="doc_text">
1618
1619<p>This section describes how to perform some very simple transformations of
1620LLVM code. This is meant to give examples of common idioms used, showing the
1621practical side of LLVM transformations. <p> Because this is a "how-to" section,
1622you should also read about the main classes that you will be working with. The
1623<a href="#coreclasses">Core LLVM Class Hierarchy Reference</a> contains details
1624and descriptions of the main classes that you should know about.</p>
1625
1626</div>
1627
1628<!-- NOTE: this section should be heavy on example code -->
1629<!-- ======================================================================= -->
1630<div class="doc_subsection">
1631 <a name="inspection">Basic Inspection and Traversal Routines</a>
1632</div>
1633
1634<div class="doc_text">
1635
1636<p>The LLVM compiler infrastructure have many different data structures that may
1637be traversed. Following the example of the C++ standard template library, the
1638techniques used to traverse these various data structures are all basically the
1639same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
1640method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
1641function returns an iterator pointing to one past the last valid element of the
1642sequence, and there is some <tt>XXXiterator</tt> data type that is common
1643between the two operations.</p>
1644
1645<p>Because the pattern for iteration is common across many different aspects of
1646the program representation, the standard template library algorithms may be used
1647on them, and it is easier to remember how to iterate. First we show a few common
1648examples of the data structures that need to be traversed. Other data
1649structures are traversed in very similar ways.</p>
1650
1651</div>
1652
1653<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1654<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1655 <a name="iterate_function">Iterating over the </a><a
1656 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
1657 href="#Function"><tt>Function</tt></a>
1658</div>
1659
1660<div class="doc_text">
1661
1662<p>It's quite common to have a <tt>Function</tt> instance that you'd like to
1663transform in some way; in particular, you'd like to manipulate its
1664<tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over all of
1665the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>. The following is
1666an example that prints the name of a <tt>BasicBlock</tt> and the number of
1667<tt>Instruction</tt>s it contains:</p>
1668
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1669<div class="doc_code">
1670<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1671// <i>func is a pointer to a Function instance</i>
1672for (Function::iterator i = func-&gt;begin(), e = func-&gt;end(); i != e; ++i)
1673 // <i>Print out the name of the basic block if it has one, and then the</i>
1674 // <i>number of instructions that it contains</i>
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1675 errs() &lt;&lt; "Basic block (name=" &lt;&lt; i-&gt;getName() &lt;&lt; ") has "
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1676 &lt;&lt; i-&gt;size() &lt;&lt; " instructions.\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1677</pre>
1678</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1679
1680<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]1681invoking member functions of the <tt>Instruction</tt> class. This is
1682because the indirection operator is overloaded for the iterator
Chris Lattner7496ec52003-08-05 22:54:23 +0000[diff] [blame]1683classes. In the above code, the expression <tt>i-&gt;size()</tt> is
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1684exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.</p>
1685
1686</div>
1687
1688<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1689<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1690 <a name="iterate_basicblock">Iterating over the </a><a
1691 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1692 href="#BasicBlock"><tt>BasicBlock</tt></a>
1693</div>
1694
1695<div class="doc_text">
1696
1697<p>Just like when dealing with <tt>BasicBlock</tt>s in <tt>Function</tt>s, it's
1698easy to iterate over the individual instructions that make up
1699<tt>BasicBlock</tt>s. Here's a code snippet that prints out each instruction in
1700a <tt>BasicBlock</tt>:</p>
1701
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1702<div class="doc_code">
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1703<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1704// <i>blk is a pointer to a BasicBlock instance</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1705for (BasicBlock::iterator i = blk-&gt;begin(), e = blk-&gt;end(); i != e; ++i)
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1706 // <i>The next statement works since operator&lt;&lt;(ostream&amp;,...)</i>
1707 // <i>is overloaded for Instruction&amp;</i>
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1708 errs() &lt;&lt; *i &lt;&lt; "\n";
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1709</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1710</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1711
1712<p>However, this isn't really the best way to print out the contents of a
1713<tt>BasicBlock</tt>! Since the ostream operators are overloaded for virtually
1714anything you'll care about, you could have just invoked the print routine on the
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1715basic block itself: <tt>errs() &lt;&lt; *blk &lt;&lt; "\n";</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1716
1717</div>
1718
1719<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1720<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1721 <a name="iterate_institer">Iterating over the </a><a
1722 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1723 href="#Function"><tt>Function</tt></a>
1724</div>
1725
1726<div class="doc_text">
1727
1728<p>If you're finding that you commonly iterate over a <tt>Function</tt>'s
1729<tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s <tt>Instruction</tt>s,
1730<tt>InstIterator</tt> should be used instead. You'll need to include <a
1731href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>,
1732and then instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1733small 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]1734
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1735<div class="doc_code">
1736<pre>
1737#include "<a href="/doxygen/InstIterator_8h-source.html">llvm/Support/InstIterator.h</a>"
1738
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1739// <i>F is a pointer to a Function instance</i>
Chris Lattnerda021aa2008-06-04 18:20:42 +0000[diff] [blame]1740for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1741 errs() &lt;&lt; *I &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1742</pre>
1743</div>
1744
1745<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]1746work list with its initial contents. For example, if you wanted to
1747initialize a work list to contain all instructions in a <tt>Function</tt>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1748F, all you would need to do is something like:</p>
1749
1750<div class="doc_code">
1751<pre>
1752std::set&lt;Instruction*&gt; worklist;
Chris Lattnerda021aa2008-06-04 18:20:42 +0000[diff] [blame]1753// or better yet, SmallPtrSet&lt;Instruction*, 64&gt; worklist;
1754
1755for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
1756 worklist.insert(&amp;*I);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1757</pre>
1758</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1759
1760<p>The STL set <tt>worklist</tt> would now contain all instructions in the
1761<tt>Function</tt> pointed to by F.</p>
1762
1763</div>
1764
1765<!-- _______________________________________________________________________ -->
1766<div class="doc_subsubsection">
1767 <a name="iterate_convert">Turning an iterator into a class pointer (and
1768 vice-versa)</a>
1769</div>
1770
1771<div class="doc_text">
1772
1773<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]1774instance when all you've got at hand is an iterator. Well, extracting
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1775a reference or a pointer from an iterator is very straight-forward.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1776Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and <tt>j</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1777is a <tt>BasicBlock::const_iterator</tt>:</p>
1778
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1779<div class="doc_code">
1780<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1781Instruction&amp; inst = *i; // <i>Grab reference to instruction reference</i>
1782Instruction* pinst = &amp;*i; // <i>Grab pointer to instruction reference</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1783const Instruction&amp; inst = *j;
1784</pre>
1785</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1786
1787<p>However, the iterators you'll be working with in the LLVM framework are
1788special: they will automatically convert to a ptr-to-instance type whenever they
1789need to. Instead of dereferencing the iterator and then taking the address of
1790the result, you can simply assign the iterator to the proper pointer type and
1791you get the dereference and address-of operation as a result of the assignment
1792(behind the scenes, this is a result of overloading casting mechanisms). Thus
1793the last line of the last example,</p>
1794
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1795<div class="doc_code">
1796<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1797Instruction *pinst = &amp;*i;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1798</pre>
1799</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1800
1801<p>is semantically equivalent to</p>
1802
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1803<div class="doc_code">
1804<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1805Instruction *pinst = i;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1806</pre>
1807</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1808
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1809<p>It's also possible to turn a class pointer into the corresponding iterator,
1810and this is a constant time operation (very efficient). The following code
1811snippet illustrates use of the conversion constructors provided by LLVM
1812iterators. By using these, you can explicitly grab the iterator of something
1813without actually obtaining it via iteration over some structure:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1814
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1815<div class="doc_code">
1816<pre>
1817void printNextInstruction(Instruction* inst) {
1818 BasicBlock::iterator it(inst);
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1819 ++it; // <i>After this line, it refers to the instruction after *inst</i>
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1820 if (it != inst-&gt;getParent()-&gt;end()) errs() &lt;&lt; *it &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1821}
1822</pre>
1823</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1824
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1825</div>
1826
1827<!--_______________________________________________________________________-->
1828<div class="doc_subsubsection">
1829 <a name="iterate_complex">Finding call sites: a slightly more complex
1830 example</a>
1831</div>
1832
1833<div class="doc_text">
1834
1835<p>Say that you're writing a FunctionPass and would like to count all the
1836locations in the entire module (that is, across every <tt>Function</tt>) where a
1837certain function (i.e., some <tt>Function</tt>*) is already in scope. As you'll
1838learn 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]1839much more straight-forward manner, but this example will allow us to explore how
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1840you'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]1841is what we want to do:</p>
1842
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1843<div class="doc_code">
1844<pre>
1845initialize callCounter to zero
1846for each Function f in the Module
1847 for each BasicBlock b in f
1848 for each Instruction i in b
1849 if (i is a CallInst and calls the given function)
1850 increment callCounter
1851</pre>
1852</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1853
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1854<p>And the actual code is (remember, because we're writing a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1855<tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply has to
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1856override the <tt>runOnFunction</tt> method):</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1857
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1858<div class="doc_code">
1859<pre>
1860Function* targetFunc = ...;
1861
1862class OurFunctionPass : public FunctionPass {
1863 public:
1864 OurFunctionPass(): callCounter(0) { }
1865
1866 virtual runOnFunction(Function&amp; F) {
1867 for (Function::iterator b = F.begin(), be = F.end(); b != be; ++b) {
Eric Christopher203e71d2008-11-08 08:20:49 +0000[diff] [blame]1868 for (BasicBlock::iterator i = b-&gt;begin(), ie = b-&gt;end(); i != ie; ++i) {
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1869 if (<a href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a>&lt;<a
1870 href="#CallInst">CallInst</a>&gt;(&amp;*i)) {
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1871 // <i>We know we've encountered a call instruction, so we</i>
1872 // <i>need to determine if it's a call to the</i>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1873 // <i>function pointed to by m_func or not.</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1874 if (callInst-&gt;getCalledFunction() == targetFunc)
1875 ++callCounter;
1876 }
1877 }
1878 }
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1879 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1880
1881 private:
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1882 unsigned callCounter;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1883};
1884</pre>
1885</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1886
1887</div>
1888
Brian Gaekef1972c62003-11-07 19:25:45 +0000[diff] [blame]1889<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1890<div class="doc_subsubsection">
1891 <a name="calls_and_invokes">Treating calls and invokes the same way</a>
1892</div>
1893
1894<div class="doc_text">
1895
1896<p>You may have noticed that the previous example was a bit oversimplified in
1897that it did not deal with call sites generated by 'invoke' instructions. In
1898this, and in other situations, you may find that you want to treat
1899<tt>CallInst</tt>s and <tt>InvokeInst</tt>s the same way, even though their
1900most-specific common base class is <tt>Instruction</tt>, which includes lots of
1901less closely-related things. For these cases, LLVM provides a handy wrapper
1902class called <a
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]1903href="http://llvm.org/doxygen/classllvm_1_1CallSite.html"><tt>CallSite</tt></a>.
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1904It is essentially a wrapper around an <tt>Instruction</tt> pointer, with some
1905methods that provide functionality common to <tt>CallInst</tt>s and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1906<tt>InvokeInst</tt>s.</p>
1907
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1908<p>This class has "value semantics": it should be passed by value, not by
1909reference and it should not be dynamically allocated or deallocated using
1910<tt>operator new</tt> or <tt>operator delete</tt>. It is efficiently copyable,
1911assignable and constructable, with costs equivalents to that of a bare pointer.
1912If you look at its definition, it has only a single pointer member.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1913
1914</div>
1915
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]1916<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1917<div class="doc_subsubsection">
1918 <a name="iterate_chains">Iterating over def-use &amp; use-def chains</a>
1919</div>
1920
1921<div class="doc_text">
1922
1923<p>Frequently, we might have an instance of the <a
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]1924href="/doxygen/classllvm_1_1Value.html">Value Class</a> and we want to
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]1925determine which <tt>User</tt>s use the <tt>Value</tt>. The list of all
1926<tt>User</tt>s of a particular <tt>Value</tt> is called a <i>def-use</i> chain.
1927For example, let's say we have a <tt>Function*</tt> named <tt>F</tt> to a
1928particular function <tt>foo</tt>. Finding all of the instructions that
1929<i>use</i> <tt>foo</tt> is as simple as iterating over the <i>def-use</i> chain
1930of <tt>F</tt>:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1931
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1932<div class="doc_code">
1933<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1934Function *F = ...;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1935
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1936for (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]1937 if (Instruction *Inst = dyn_cast&lt;Instruction&gt;(*i)) {
Chris Lattner3fee6ed2009-09-08 05:15:50 +0000[diff] [blame]1938 errs() &lt;&lt; "F is used in instruction:\n";
1939 errs() &lt;&lt; *Inst &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1940 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1941</pre>
1942</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1943
1944<p>Alternately, it's common to have an instance of the <a
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]1945href="/doxygen/classllvm_1_1User.html">User Class</a> and need to know what
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1946<tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used by a
1947<tt>User</tt> is known as a <i>use-def</i> chain. Instances of class
1948<tt>Instruction</tt> are common <tt>User</tt>s, so we might want to iterate over
1949all of the values that a particular instruction uses (that is, the operands of
1950the particular <tt>Instruction</tt>):</p>
1951
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1952<div class="doc_code">
1953<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1954Instruction *pi = ...;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1955
1956for (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]1957 Value *v = *i;
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1958 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1959}
1960</pre>
1961</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1962
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]1963<!--
1964 def-use chains ("finding all users of"): Value::use_begin/use_end
1965 use-def chains ("finding all values used"): User::op_begin/op_end [op=operand]
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1966-->
1967
1968</div>
1969
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1970<!--_______________________________________________________________________-->
1971<div class="doc_subsubsection">
1972 <a name="iterate_preds">Iterating over predecessors &amp;
1973successors of blocks</a>
1974</div>
1975
1976<div class="doc_text">
1977
1978<p>Iterating over the predecessors and successors of a block is quite easy
1979with the routines defined in <tt>"llvm/Support/CFG.h"</tt>. Just use code like
1980this to iterate over all predecessors of BB:</p>
1981
1982<div class="doc_code">
1983<pre>
1984#include "llvm/Support/CFG.h"
1985BasicBlock *BB = ...;
1986
1987for (pred_iterator PI = pred_begin(BB), E = pred_end(BB); PI != E; ++PI) {
1988 BasicBlock *Pred = *PI;
1989 // <i>...</i>
1990}
1991</pre>
1992</div>
1993
1994<p>Similarly, to iterate over successors use
1995succ_iterator/succ_begin/succ_end.</p>
1996
1997</div>
1998
1999
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2000<!-- ======================================================================= -->
2001<div class="doc_subsection">
2002 <a name="simplechanges">Making simple changes</a>
2003</div>
2004
2005<div class="doc_text">
2006
2007<p>There are some primitive transformation operations present in the LLVM
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]2008infrastructure that are worth knowing about. When performing
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2009transformations, it's fairly common to manipulate the contents of basic
2010blocks. This section describes some of the common methods for doing so
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2011and gives example code.</p>
2012
2013</div>
2014
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2015<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2016<div class="doc_subsubsection">
2017 <a name="schanges_creating">Creating and inserting new
2018 <tt>Instruction</tt>s</a>
2019</div>
2020
2021<div class="doc_text">
2022
2023<p><i>Instantiating Instructions</i></p>
2024
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]2025<p>Creation of <tt>Instruction</tt>s is straight-forward: simply call the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2026constructor for the kind of instruction to instantiate and provide the necessary
2027parameters. For example, an <tt>AllocaInst</tt> only <i>requires</i> a
2028(const-ptr-to) <tt>Type</tt>. Thus:</p>
2029
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2030<div class="doc_code">
2031<pre>
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2032AllocaInst* ai = new AllocaInst(Type::Int32Ty);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2033</pre>
2034</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2035
2036<p>will create an <tt>AllocaInst</tt> instance that represents the allocation of
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]2037one integer in the current stack frame, at run time. Each <tt>Instruction</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2038subclass is likely to have varying default parameters which change the semantics
2039of the instruction, so refer to the <a
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]2040href="/doxygen/classllvm_1_1Instruction.html">doxygen documentation for the subclass of
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2041Instruction</a> that you're interested in instantiating.</p>
2042
2043<p><i>Naming values</i></p>
2044
2045<p>It is very useful to name the values of instructions when you're able to, as
2046this facilitates the debugging of your transformations. If you end up looking
2047at generated LLVM machine code, you definitely want to have logical names
2048associated with the results of instructions! By supplying a value for the
2049<tt>Name</tt> (default) parameter of the <tt>Instruction</tt> constructor, you
2050associate a logical name with the result of the instruction's execution at
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]2051run time. For example, say that I'm writing a transformation that dynamically
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2052allocates space for an integer on the stack, and that integer is going to be
2053used as some kind of index by some other code. To accomplish this, I place an
2054<tt>AllocaInst</tt> at the first point in the first <tt>BasicBlock</tt> of some
2055<tt>Function</tt>, and I'm intending to use it within the same
2056<tt>Function</tt>. I might do:</p>
2057
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2058<div class="doc_code">
2059<pre>
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2060AllocaInst* pa = new AllocaInst(Type::Int32Ty, 0, "indexLoc");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2061</pre>
2062</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2063
2064<p>where <tt>indexLoc</tt> is now the logical name of the instruction's
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]2065execution value, which is a pointer to an integer on the run time stack.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2066
2067<p><i>Inserting instructions</i></p>
2068
2069<p>There are essentially two ways to insert an <tt>Instruction</tt>
2070into an existing sequence of instructions that form a <tt>BasicBlock</tt>:</p>
2071
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]2072<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2073 <li>Insertion into an explicit instruction list
2074
2075 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within that
2076 <tt>BasicBlock</tt>, and a newly-created instruction we wish to insert
2077 before <tt>*pi</tt>, we do the following: </p>
2078
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2079<div class="doc_code">
2080<pre>
2081BasicBlock *pb = ...;
2082Instruction *pi = ...;
2083Instruction *newInst = new Instruction(...);
2084
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2085pb-&gt;getInstList().insert(pi, newInst); // <i>Inserts newInst before pi in pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2086</pre>
2087</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]2088
2089 <p>Appending to the end of a <tt>BasicBlock</tt> is so common that
2090 the <tt>Instruction</tt> class and <tt>Instruction</tt>-derived
2091 classes provide constructors which take a pointer to a
2092 <tt>BasicBlock</tt> to be appended to. For example code that
2093 looked like: </p>
2094
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2095<div class="doc_code">
2096<pre>
2097BasicBlock *pb = ...;
2098Instruction *newInst = new Instruction(...);
2099
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2100pb-&gt;getInstList().push_back(newInst); // <i>Appends newInst to pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2101</pre>
2102</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]2103
2104 <p>becomes: </p>
2105
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2106<div class="doc_code">
2107<pre>
2108BasicBlock *pb = ...;
2109Instruction *newInst = new Instruction(..., pb);
2110</pre>
2111</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]2112
2113 <p>which is much cleaner, especially if you are creating
2114 long instruction streams.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2115
2116 <li>Insertion into an implicit instruction list
2117
2118 <p><tt>Instruction</tt> instances that are already in <tt>BasicBlock</tt>s
2119 are implicitly associated with an existing instruction list: the instruction
2120 list of the enclosing basic block. Thus, we could have accomplished the same
2121 thing as the above code without being given a <tt>BasicBlock</tt> by doing:
2122 </p>
2123
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2124<div class="doc_code">
2125<pre>
2126Instruction *pi = ...;
2127Instruction *newInst = new Instruction(...);
2128
2129pi-&gt;getParent()-&gt;getInstList().insert(pi, newInst);
2130</pre>
2131</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2132
2133 <p>In fact, this sequence of steps occurs so frequently that the
2134 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes provide
2135 constructors which take (as a default parameter) a pointer to an
2136 <tt>Instruction</tt> which the newly-created <tt>Instruction</tt> should
2137 precede. That is, <tt>Instruction</tt> constructors are capable of
2138 inserting the newly-created instance into the <tt>BasicBlock</tt> of a
2139 provided instruction, immediately before that instruction. Using an
2140 <tt>Instruction</tt> constructor with a <tt>insertBefore</tt> (default)
2141 parameter, the above code becomes:</p>
2142
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2143<div class="doc_code">
2144<pre>
2145Instruction* pi = ...;
2146Instruction* newInst = new Instruction(..., pi);
2147</pre>
2148</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2149
2150 <p>which is much cleaner, especially if you're creating a lot of
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2151 instructions and adding them to <tt>BasicBlock</tt>s.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2152</ul>
2153
2154</div>
2155
2156<!--_______________________________________________________________________-->
2157<div class="doc_subsubsection">
2158 <a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a>
2159</div>
2160
2161<div class="doc_text">
2162
2163<p>Deleting an instruction from an existing sequence of instructions that form a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]2164<a href="#BasicBlock"><tt>BasicBlock</tt></a> is very straight-forward. First,
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2165you must have a pointer to the instruction that you wish to delete. Second, you
2166need to obtain the pointer to that instruction's basic block. You use the
2167pointer to the basic block to get its list of instructions and then use the
2168erase function to remove your instruction. For example:</p>
2169
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2170<div class="doc_code">
2171<pre>
2172<a href="#Instruction">Instruction</a> *I = .. ;
Chris Lattner9f8ec252008-02-15 22:57:17 +0000[diff] [blame]2173I-&gt;eraseFromParent();
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2174</pre>
2175</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2176
2177</div>
2178
2179<!--_______________________________________________________________________-->
2180<div class="doc_subsubsection">
2181 <a name="schanges_replacing">Replacing an <tt>Instruction</tt> with another
2182 <tt>Value</tt></a>
2183</div>
2184
2185<div class="doc_text">
2186
2187<p><i>Replacing individual instructions</i></p>
2188
2189<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]2190permits use of two very useful replace functions: <tt>ReplaceInstWithValue</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2191and <tt>ReplaceInstWithInst</tt>.</p>
2192
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2193<h4><a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a></h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2194
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2195<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2196 <li><tt>ReplaceInstWithValue</tt>
2197
Nick Lewyckyb6d1f392008-09-15 06:31:52 +0000[diff] [blame]2198 <p>This function replaces all uses of a given instruction with a value,
2199 and then removes the original instruction. The following example
2200 illustrates the replacement of the result of a particular
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]2201 <tt>AllocaInst</tt> that allocates memory for a single integer with a null
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2202 pointer to an integer.</p>
2203
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2204<div class="doc_code">
2205<pre>
2206AllocaInst* instToReplace = ...;
2207BasicBlock::iterator ii(instToReplace);
2208
2209ReplaceInstWithValue(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
Daniel Dunbar58c2ac02008-10-03 22:17:25 +0000[diff] [blame]2210 Constant::getNullValue(PointerType::getUnqual(Type::Int32Ty)));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2211</pre></div></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2212
2213 <li><tt>ReplaceInstWithInst</tt>
2214
2215 <p>This function replaces a particular instruction with another
Nick Lewyckyb6d1f392008-09-15 06:31:52 +0000[diff] [blame]2216 instruction, inserting the new instruction into the basic block at the
2217 location where the old instruction was, and replacing any uses of the old
2218 instruction with the new instruction. The following example illustrates
2219 the replacement of one <tt>AllocaInst</tt> with another.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2220
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2221<div class="doc_code">
2222<pre>
2223AllocaInst* instToReplace = ...;
2224BasicBlock::iterator ii(instToReplace);
2225
2226ReplaceInstWithInst(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2227 new AllocaInst(Type::Int32Ty, 0, "ptrToReplacedInt"));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2228</pre></div></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2229</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2230
2231<p><i>Replacing multiple uses of <tt>User</tt>s and <tt>Value</tt>s</i></p>
2232
2233<p>You can use <tt>Value::replaceAllUsesWith</tt> and
2234<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]2235doxygen documentation for the <a href="/doxygen/classllvm_1_1Value.html">Value Class</a>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]2236and <a href="/doxygen/classllvm_1_1User.html">User Class</a>, respectively, for more
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2237information.</p>
2238
2239<!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
2240include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
2241ReplaceInstWithValue, ReplaceInstWithInst -->
2242
2243</div>
2244
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2245<!--_______________________________________________________________________-->
2246<div class="doc_subsubsection">
2247 <a name="schanges_deletingGV">Deleting <tt>GlobalVariable</tt>s</a>
2248</div>
2249
2250<div class="doc_text">
2251
Tanya Lattnerc5dfcdb2007-06-20 20:46:37 +0000[diff] [blame]2252<p>Deleting a global variable from a module is just as easy as deleting an
2253Instruction. First, you must have a pointer to the global variable that you wish
2254 to delete. You use this pointer to erase it from its parent, the module.
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2255 For example:</p>
2256
2257<div class="doc_code">
2258<pre>
2259<a href="#GlobalVariable">GlobalVariable</a> *GV = .. ;
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2260
Tanya Lattnerc5dfcdb2007-06-20 20:46:37 +0000[diff] [blame]2261GV-&gt;eraseFromParent();
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2262</pre>
2263</div>
2264
2265</div>
2266
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2267<!-- ======================================================================= -->
2268<div class="doc_subsection">
2269 <a name="create_types">How to Create Types</a>
2270</div>
2271
2272<div class="doc_text">
2273
2274<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]2275statically, you can use <tt>TypeBuilder&lt;...&gt;::get()</tt>, defined
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2276in <tt>llvm/Support/TypeBuilder.h</tt>, to retrieve them. <tt>TypeBuilder</tt>
2277has two forms depending on whether you're building types for cross-compilation
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2278or native library use. <tt>TypeBuilder&lt;T, true&gt;</tt> requires
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2279that <tt>T</tt> be independent of the host environment, meaning that it's built
2280out of types from
2281the <a href="/doxygen/namespacellvm_1_1types.html"><tt>llvm::types</tt></a>
2282namespace and pointers, functions, arrays, etc. built of
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2283those. <tt>TypeBuilder&lt;T, false&gt;</tt> additionally allows native C types
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2284whose size may depend on the host compiler. For example,</p>
2285
2286<div class="doc_code">
2287<pre>
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2288FunctionType *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]2289</pre>
2290</div>
2291
2292<p>is easier to read and write than the equivalent</p>
2293
2294<div class="doc_code">
2295<pre>
Owen Anderson5e8c50e2009-06-16 17:40:28 +0000[diff] [blame]2296std::vector&lt;const Type*&gt; params;
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2297params.push_back(PointerType::getUnqual(Type::Int32Ty));
2298FunctionType *ft = FunctionType::get(Type::Int8Ty, params, false);
2299</pre>
2300</div>
2301
2302<p>See the <a href="/doxygen/TypeBuilder_8h-source.html#l00001">class
2303comment</a> for more details.</p>
2304
2305</div>
2306
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]2307<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2308<div class="doc_section">
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2309 <a name="threading">Threads and LLVM</a>
2310</div>
2311<!-- *********************************************************************** -->
2312
2313<div class="doc_text">
2314<p>
2315This section describes the interaction of the LLVM APIs with multithreading,
2316both on the part of client applications, and in the JIT, in the hosted
2317application.
2318</p>
2319
2320<p>
2321Note that LLVM's support for multithreading is still relatively young. Up
2322through version 2.5, the execution of threaded hosted applications was
2323supported, but not threaded client access to the APIs. While this use case is
2324now supported, clients <em>must</em> adhere to the guidelines specified below to
2325ensure proper operation in multithreaded mode.
2326</p>
2327
2328<p>
2329Note that, on Unix-like platforms, LLVM requires the presence of GCC's atomic
2330intrinsics in order to support threaded operation. If you need a
2331multhreading-capable LLVM on a platform without a suitably modern system
2332compiler, consider compiling LLVM and LLVM-GCC in single-threaded mode, and
2333using the resultant compiler to build a copy of LLVM with multithreading
2334support.
2335</p>
2336</div>
2337
2338<!-- ======================================================================= -->
2339<div class="doc_subsection">
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2340 <a name="startmultithreaded">Entering and Exiting Multithreaded Mode</a>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2341</div>
2342
2343<div class="doc_text">
2344
2345<p>
2346In order to properly protect its internal data structures while avoiding
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2347excessive locking overhead in the single-threaded case, the LLVM must intialize
2348certain data structures necessary to provide guards around its internals. To do
2349so, the client program must invoke <tt>llvm_start_multithreaded()</tt> before
2350making any concurrent LLVM API calls. To subsequently tear down these
2351structures, use the <tt>llvm_stop_multithreaded()</tt> call. You can also use
2352the <tt>llvm_is_multithreaded()</tt> call to check the status of multithreaded
2353mode.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2354</p>
2355
2356<p>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2357Note that both of these calls must be made <em>in isolation</em>. That is to
2358say that no other LLVM API calls may be executing at any time during the
2359execution of <tt>llvm_start_multithreaded()</tt> or <tt>llvm_stop_multithreaded
2360</tt>. It's is the client's responsibility to enforce this isolation.
2361</p>
2362
2363<p>
2364The return value of <tt>llvm_start_multithreaded()</tt> indicates the success or
2365failure of the initialization. Failure typically indicates that your copy of
2366LLVM was built without multithreading support, typically because GCC atomic
2367intrinsics were not found in your system compiler. In this case, the LLVM API
2368will not be safe for concurrent calls. However, it <em>will</em> be safe for
2369hosting threaded applications in the JIT, though care must be taken to ensure
2370that side exits and the like do not accidentally result in concurrent LLVM API
2371calls.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2372</p>
2373</div>
2374
2375<!-- ======================================================================= -->
2376<div class="doc_subsection">
2377 <a name="shutdown">Ending Execution with <tt>llvm_shutdown()</tt></a>
2378</div>
2379
2380<div class="doc_text">
2381<p>
2382When you are done using the LLVM APIs, you should call <tt>llvm_shutdown()</tt>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2383to deallocate memory used for internal structures. This will also invoke
2384<tt>llvm_stop_multithreaded()</tt> if LLVM is operating in multithreaded mode.
2385As such, <tt>llvm_shutdown()</tt> requires the same isolation guarantees as
2386<tt>llvm_stop_multithreaded()</tt>.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2387</p>
2388
2389<p>
2390Note that, if you use scope-based shutdown, you can use the
2391<tt>llvm_shutdown_obj</tt> class, which calls <tt>llvm_shutdown()</tt> in its
2392destructor.
2393</div>
2394
2395<!-- ======================================================================= -->
2396<div class="doc_subsection">
2397 <a name="managedstatic">Lazy Initialization with <tt>ManagedStatic</tt></a>
2398</div>
2399
2400<div class="doc_text">
2401<p>
2402<tt>ManagedStatic</tt> is a utility class in LLVM used to implement static
2403initialization of static resources, such as the global type tables. Before the
2404invocation of <tt>llvm_shutdown()</tt>, it implements a simple lazy
2405initialization scheme. Once <tt>llvm_start_multithreaded()</tt> returns,
2406however, it uses double-checked locking to implement thread-safe lazy
2407initialization.
2408</p>
2409
2410<p>
2411Note that, because no other threads are allowed to issue LLVM API calls before
2412<tt>llvm_start_multithreaded()</tt> returns, it is possible to have
2413<tt>ManagedStatic</tt>s of <tt>llvm::sys::Mutex</tt>s.
2414</p>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2415
2416<p>
2417The <tt>llvm_acquire_global_lock()</tt> and <tt>llvm_release_global_lock</tt>
2418APIs provide access to the global lock used to implement the double-checked
2419locking for lazy initialization. These should only be used internally to LLVM,
2420and only if you know what you're doing!
2421</p>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2422</div>
2423
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]2424<!-- ======================================================================= -->
2425<div class="doc_subsection">
2426 <a name="llvmcontext">Achieving Isolation with <tt>LLVMContext</tt></a>
2427</div>
2428
2429<div class="doc_text">
2430<p>
2431<tt>LLVMContext</tt> is an opaque class in the LLVM API which clients can use
2432to operate multiple, isolated instances of LLVM concurrently within the same
2433address space. For instance, in a hypothetical compile-server, the compilation
2434of an individual translation unit is conceptually independent from all the
2435others, and it would be desirable to be able to compile incoming translation
2436units concurrently on independent server threads. Fortunately,
2437<tt>LLVMContext</tt> exists to enable just this kind of scenario!
2438</p>
2439
2440<p>
2441Conceptually, <tt>LLVMContext</tt> provides isolation. Every LLVM entity
2442(<tt>Module</tt>s, <tt>Value</tt>s, <tt>Type</tt>s, <tt>Constant</tt>s, etc.)
Chris Lattner38eee3c2009-08-20 03:10:14 +0000[diff] [blame]2443in LLVM's in-memory IR belongs to an <tt>LLVMContext</tt>. Entities in
Owen Andersone0c951a2009-08-19 17:58:52 +0000[diff] [blame]2444different contexts <em>cannot</em> interact with each other: <tt>Module</tt>s in
2445different contexts cannot be linked together, <tt>Function</tt>s cannot be added
2446to <tt>Module</tt>s in different contexts, etc. What this means is that is is
2447safe to compile on multiple threads simultaneously, as long as no two threads
2448operate on entities within the same context.
2449</p>
2450
2451<p>
2452In practice, very few places in the API require the explicit specification of a
2453<tt>LLVMContext</tt>, other than the <tt>Type</tt> creation/lookup APIs.
2454Because every <tt>Type</tt> carries a reference to its owning context, most
2455other entities can determine what context they belong to by looking at their
2456own <tt>Type</tt>. If you are adding new entities to LLVM IR, please try to
2457maintain this interface design.
2458</p>
2459
2460<p>
2461For clients that do <em>not</em> require the benefits of isolation, LLVM
2462provides a convenience API <tt>getGlobalContext()</tt>. This returns a global,
2463lazily initialized <tt>LLVMContext</tt> that may be used in situations where
2464isolation is not a concern.
2465</p>
2466</div>
2467
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2468<!-- *********************************************************************** -->
2469<div class="doc_section">
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2470 <a name="advanced">Advanced Topics</a>
2471</div>
2472<!-- *********************************************************************** -->
2473
2474<div class="doc_text">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2475<p>
2476This section describes some of the advanced or obscure API's that most clients
2477do not need to be aware of. These API's tend manage the inner workings of the
2478LLVM system, and only need to be accessed in unusual circumstances.
2479</p>
2480</div>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2481
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2482<!-- ======================================================================= -->
2483<div class="doc_subsection">
2484 <a name="TypeResolve">LLVM Type Resolution</a>
2485</div>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2486
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2487<div class="doc_text">
2488
2489<p>
2490The LLVM type system has a very simple goal: allow clients to compare types for
2491structural equality with a simple pointer comparison (aka a shallow compare).
2492This goal makes clients much simpler and faster, and is used throughout the LLVM
2493system.
2494</p>
2495
2496<p>
2497Unfortunately achieving this goal is not a simple matter. In particular,
2498recursive types and late resolution of opaque types makes the situation very
2499difficult to handle. Fortunately, for the most part, our implementation makes
2500most clients able to be completely unaware of the nasty internal details. The
2501primary case where clients are exposed to the inner workings of it are when
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]2502building a recursive type. In addition to this case, the LLVM bitcode reader,
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2503assembly parser, and linker also have to be aware of the inner workings of this
2504system.
2505</p>
2506
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]2507<p>
2508For our purposes below, we need three concepts. First, an "Opaque Type" is
2509exactly as defined in the <a href="LangRef.html#t_opaque">language
2510reference</a>. Second an "Abstract Type" is any type which includes an
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2511opaque type as part of its type graph (for example "<tt>{ opaque, i32 }</tt>").
2512Third, 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]2513float }</tt>").
2514</p>
2515
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2516</div>
2517
2518<!-- ______________________________________________________________________ -->
2519<div class="doc_subsubsection">
2520 <a name="BuildRecType">Basic Recursive Type Construction</a>
2521</div>
2522
2523<div class="doc_text">
2524
2525<p>
2526Because the most common question is "how do I build a recursive type with LLVM",
2527we answer it now and explain it as we go. Here we include enough to cause this
2528to be emitted to an output .ll file:
2529</p>
2530
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2531<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2532<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2533%mylist = type { %mylist*, i32 }
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2534</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2535</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2536
2537<p>
2538To build this, use the following LLVM APIs:
2539</p>
2540
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2541<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2542<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2543// <i>Create the initial outer struct</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2544<a href="#PATypeHolder">PATypeHolder</a> StructTy = OpaqueType::get();
2545std::vector&lt;const Type*&gt; Elts;
Daniel Dunbar58c2ac02008-10-03 22:17:25 +0000[diff] [blame]2546Elts.push_back(PointerType::getUnqual(StructTy));
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2547Elts.push_back(Type::Int32Ty);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2548StructType *NewSTy = StructType::get(Elts);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2549
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2550// <i>At this point, NewSTy = "{ opaque*, i32 }". Tell VMCore that</i>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2551// <i>the struct and the opaque type are actually the same.</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2552cast&lt;OpaqueType&gt;(StructTy.get())-&gt;<a href="#refineAbstractTypeTo">refineAbstractTypeTo</a>(NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2553
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2554// <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]2555// <i>kept up-to-date</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2556NewSTy = cast&lt;StructType&gt;(StructTy.get());
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2557
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2558// <i>Add a name for the type to the module symbol table (optional)</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2559MyModule-&gt;addTypeName("mylist", NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2560</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2561</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2562
2563<p>
2564This code shows the basic approach used to build recursive types: build a
2565non-recursive type using 'opaque', then use type unification to close the cycle.
2566The type unification step is performed by the <tt><a
Chris Lattneraff26d12007-02-03 03:06:52 +0000[diff] [blame]2567href="#refineAbstractTypeTo">refineAbstractTypeTo</a></tt> method, which is
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2568described next. After that, we describe the <a
2569href="#PATypeHolder">PATypeHolder class</a>.
2570</p>
2571
2572</div>
2573
2574<!-- ______________________________________________________________________ -->
2575<div class="doc_subsubsection">
2576 <a name="refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a>
2577</div>
2578
2579<div class="doc_text">
2580<p>
2581The <tt>refineAbstractTypeTo</tt> method starts the type unification process.
2582While this method is actually a member of the DerivedType class, it is most
2583often used on OpaqueType instances. Type unification is actually a recursive
2584process. After unification, types can become structurally isomorphic to
2585existing types, and all duplicates are deleted (to preserve pointer equality).
2586</p>
2587
2588<p>
2589In the example above, the OpaqueType object is definitely deleted.
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2590Additionally, if there is an "{ \2*, i32}" type already created in the system,
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2591the pointer and struct type created are <b>also</b> deleted. Obviously whenever
2592a type is deleted, any "Type*" pointers in the program are invalidated. As
2593such, it is safest to avoid having <i>any</i> "Type*" pointers to abstract types
2594live across a call to <tt>refineAbstractTypeTo</tt> (note that non-abstract
2595types can never move or be deleted). To deal with this, the <a
2596href="#PATypeHolder">PATypeHolder</a> class is used to maintain a stable
2597reference to a possibly refined type, and the <a
2598href="#AbstractTypeUser">AbstractTypeUser</a> class is used to update more
2599complex datastructures.
2600</p>
2601
2602</div>
2603
2604<!-- ______________________________________________________________________ -->
2605<div class="doc_subsubsection">
2606 <a name="PATypeHolder">The PATypeHolder Class</a>
2607</div>
2608
2609<div class="doc_text">
2610<p>
2611PATypeHolder is a form of a "smart pointer" for Type objects. When VMCore
2612happily goes about nuking types that become isomorphic to existing types, it
2613automatically updates all PATypeHolder objects to point to the new type. In the
2614example above, this allows the code to maintain a pointer to the resultant
2615resolved recursive type, even though the Type*'s are potentially invalidated.
2616</p>
2617
2618<p>
2619PATypeHolder is an extremely light-weight object that uses a lazy union-find
2620implementation to update pointers. For example the pointer from a Value to its
2621Type is maintained by PATypeHolder objects.
2622</p>
2623
2624</div>
2625
2626<!-- ______________________________________________________________________ -->
2627<div class="doc_subsubsection">
2628 <a name="AbstractTypeUser">The AbstractTypeUser Class</a>
2629</div>
2630
2631<div class="doc_text">
2632
2633<p>
2634Some data structures need more to perform more complex updates when types get
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2635resolved. To support this, a class can derive from the AbstractTypeUser class.
2636This class
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2637allows it to get callbacks when certain types are resolved. To register to get
2638callbacks for a particular type, the DerivedType::{add/remove}AbstractTypeUser
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]2639methods can be called on a type. Note that these methods only work for <i>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2640 abstract</i> types. Concrete types (those that do not include any opaque
2641objects) can never be refined.
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2642</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2643</div>
2644
2645
2646<!-- ======================================================================= -->
2647<div class="doc_subsection">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2648 <a name="SymbolTable">The <tt>ValueSymbolTable</tt> and
2649 <tt>TypeSymbolTable</tt> classes</a>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2650</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2651
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2652<div class="doc_text">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2653<p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1ValueSymbolTable.html">
2654ValueSymbolTable</a></tt> class provides a symbol table that the <a
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2655href="#Function"><tt>Function</tt></a> and <a href="#Module">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2656<tt>Module</tt></a> classes use for naming value definitions. The symbol table
2657can provide a name for any <a href="#Value"><tt>Value</tt></a>.
2658The <tt><a href="http://llvm.org/doxygen/classllvm_1_1TypeSymbolTable.html">
2659TypeSymbolTable</a></tt> class is used by the <tt>Module</tt> class to store
2660names for types.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2661
Reid Spencera6362242007-01-07 00:41:39 +0000[diff] [blame]2662<p>Note that the <tt>SymbolTable</tt> class should not be directly accessed
2663by most clients. It should only be used when iteration over the symbol table
2664names themselves are required, which is very special purpose. Note that not
2665all LLVM
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2666<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]2667an empty name) do not exist in the symbol table.
2668</p>
2669
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2670<p>These symbol tables support iteration over the values/types in the symbol
2671table with <tt>begin/end/iterator</tt> and supports querying to see if a
2672specific name is in the symbol table (with <tt>lookup</tt>). The
2673<tt>ValueSymbolTable</tt> class exposes no public mutator methods, instead,
2674simply call <tt>setName</tt> on a value, which will autoinsert it into the
2675appropriate symbol table. For types, use the Module::addTypeName method to
2676insert entries into the symbol table.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2677
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2678</div>
2679
2680
2681
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2682<!-- ======================================================================= -->
2683<div class="doc_subsection">
2684 <a name="UserLayout">The <tt>User</tt> and owned <tt>Use</tt> classes' memory layout</a>
2685</div>
2686
2687<div class="doc_text">
2688<p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1User.html">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2689User</a></tt> class provides a basis for expressing the ownership of <tt>User</tt>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2690towards other <tt><a href="http://llvm.org/doxygen/classllvm_1_1Value.html">
2691Value</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]2692Use</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]2693addition and removal.</p>
2694
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2695<!-- ______________________________________________________________________ -->
2696<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2697 <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]2698</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2699
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2700<div class="doc_text">
2701<p>
2702A subclass of <tt>User</tt> can choose between incorporating its <tt>Use</tt> objects
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2703or refer to them out-of-line by means of a pointer. A mixed variant
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2704(some <tt>Use</tt>s inline others hung off) is impractical and breaks the invariant
2705that the <tt>Use</tt> objects belonging to the same <tt>User</tt> form a contiguous array.
2706</p>
2707</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2708
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2709<p>
2710We have 2 different layouts in the <tt>User</tt> (sub)classes:
2711<ul>
2712<li><p>Layout a)
2713The <tt>Use</tt> object(s) are inside (resp. at fixed offset) of the <tt>User</tt>
2714object and there are a fixed number of them.</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2715
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2716<li><p>Layout b)
2717The <tt>Use</tt> object(s) are referenced by a pointer to an
2718array from the <tt>User</tt> object and there may be a variable
2719number of them.</p>
2720</ul>
2721<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2722As of v2.4 each layout still possesses a direct pointer to the
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2723start of the array of <tt>Use</tt>s. Though not mandatory for layout a),
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2724we stick to this redundancy for the sake of simplicity.
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2725The <tt>User</tt> object also stores the number of <tt>Use</tt> objects it
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2726has. (Theoretically this information can also be calculated
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2727given the scheme presented below.)</p>
2728<p>
2729Special forms of allocation operators (<tt>operator new</tt>)
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2730enforce the following memory layouts:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2731
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2732<ul>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2733<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]2734
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2735<pre>
2736...---.---.---.---.-------...
2737 | P | P | P | P | User
2738'''---'---'---'---'-------'''
2739</pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2740
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2741<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]2742<pre>
2743.-------...
2744| User
2745'-------'''
2746 |
2747 v
2748 .---.---.---.---...
2749 | P | P | P | P |
2750 '---'---'---'---'''
2751</pre>
2752</ul>
2753<i>(In the above figures '<tt>P</tt>' stands for the <tt>Use**</tt> that
2754 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]2755
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2756<!-- ______________________________________________________________________ -->
2757<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2758 <a name="Waymarking">The waymarking algorithm</a>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2759</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2760
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2761<div class="doc_text">
2762<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2763Since the <tt>Use</tt> objects are deprived of the direct (back)pointer to
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2764their <tt>User</tt> objects, there must be a fast and exact method to
2765recover it. This is accomplished by the following scheme:</p>
2766</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2767
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2768A 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]2769start of the <tt>User</tt> object:
2770<ul>
2771<li><tt>00</tt> &mdash;&gt; binary digit 0</li>
2772<li><tt>01</tt> &mdash;&gt; binary digit 1</li>
2773<li><tt>10</tt> &mdash;&gt; stop and calculate (<tt>s</tt>)</li>
2774<li><tt>11</tt> &mdash;&gt; full stop (<tt>S</tt>)</li>
2775</ul>
2776<p>
2777Given a <tt>Use*</tt>, all we have to do is to walk till we get
2778a stop and we either have a <tt>User</tt> immediately behind or
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2779we have to walk to the next stop picking up digits
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2780and calculating the offset:</p>
2781<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2782.---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.----------------
2783| 1 | s | 1 | 0 | 1 | 0 | s | 1 | 1 | 0 | s | 1 | 1 | s | 1 | S | User (or User*)
2784'---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'----------------
2785 |+15 |+10 |+6 |+3 |+1
2786 | | | | |__>
2787 | | | |__________>
2788 | | |______________________>
2789 | |______________________________________>
2790 |__________________________________________________________>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2791</pre>
2792<p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2793Only the significant number of bits need to be stored between the
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2794stops, so that the <i>worst case is 20 memory accesses</i> when there are
27951000 <tt>Use</tt> objects associated with a <tt>User</tt>.</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2796
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2797<!-- ______________________________________________________________________ -->
2798<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2799 <a name="ReferenceImpl">Reference implementation</a>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2800</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2801
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2802<div class="doc_text">
2803<p>
2804The following literate Haskell fragment demonstrates the concept:</p>
2805</div>
2806
2807<div class="doc_code">
2808<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2809> import Test.QuickCheck
2810>
2811> digits :: Int -> [Char] -> [Char]
2812> digits 0 acc = '0' : acc
2813> digits 1 acc = '1' : acc
2814> digits n acc = digits (n `div` 2) $ digits (n `mod` 2) acc
2815>
2816> dist :: Int -> [Char] -> [Char]
2817> dist 0 [] = ['S']
2818> dist 0 acc = acc
2819> dist 1 acc = let r = dist 0 acc in 's' : digits (length r) r
2820> dist n acc = dist (n - 1) $ dist 1 acc
2821>
2822> takeLast n ss = reverse $ take n $ reverse ss
2823>
2824> test = takeLast 40 $ dist 20 []
2825>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2826</pre>
2827</div>
2828<p>
2829Printing &lt;test&gt; gives: <tt>"1s100000s11010s10100s1111s1010s110s11s1S"</tt></p>
2830<p>
2831The reverse algorithm computes the length of the string just by examining
2832a certain prefix:</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> pref :: [Char] -> Int
2837> pref "S" = 1
2838> pref ('s':'1':rest) = decode 2 1 rest
2839> pref (_:rest) = 1 + pref rest
2840>
2841> decode walk acc ('0':rest) = decode (walk + 1) (acc * 2) rest
2842> decode walk acc ('1':rest) = decode (walk + 1) (acc * 2 + 1) rest
2843> decode walk acc _ = walk + acc
2844>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2845</pre>
2846</div>
2847<p>
2848Now, as expected, printing &lt;pref test&gt; gives <tt>40</tt>.</p>
2849<p>
2850We can <i>quickCheck</i> this with following property:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2851
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2852<div class="doc_code">
2853<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2854> testcase = dist 2000 []
2855> testcaseLength = length testcase
2856>
2857> identityProp n = n > 0 && n <= testcaseLength ==> length arr == pref arr
2858> where arr = takeLast n testcase
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2859>
2860</pre>
2861</div>
2862<p>
2863As expected &lt;quickCheck identityProp&gt; gives:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2864
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2865<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2866*Main> quickCheck identityProp
2867OK, passed 100 tests.
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2868</pre>
2869<p>
2870Let's be a bit more exhaustive:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2871
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2872<div class="doc_code">
2873<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2874>
2875> deepCheck p = check (defaultConfig { configMaxTest = 500 }) p
2876>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2877</pre>
2878</div>
2879<p>
2880And here is the result of &lt;deepCheck identityProp&gt;:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2881
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2882<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2883*Main> deepCheck identityProp
2884OK, passed 500 tests.
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2885</pre>
2886
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2887<!-- ______________________________________________________________________ -->
2888<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2889 <a name="Tagging">Tagging considerations</a>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2890</div>
2891
2892<p>
2893To maintain the invariant that the 2 LSBits of each <tt>Use**</tt> in <tt>Use</tt>
2894never change after being set up, setters of <tt>Use::Prev</tt> must re-tag the
2895new <tt>Use**</tt> on every modification. Accordingly getters must strip the
2896tag bits.</p>
2897<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2898For layout b) instead of the <tt>User</tt> we find a pointer (<tt>User*</tt> with LSBit set).
2899Following this pointer brings us to the <tt>User</tt>. A portable trick ensures
2900that the first bytes of <tt>User</tt> (if interpreted as a pointer) never has
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2901the LSBit set. (Portability is relying on the fact that all known compilers place the
2902<tt>vptr</tt> in the first word of the instances.)</p>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2903
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2904</div>
2905
2906 <!-- *********************************************************************** -->
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2907<div class="doc_section">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2908 <a name="coreclasses">The Core LLVM Class Hierarchy Reference </a>
2909</div>
2910<!-- *********************************************************************** -->
2911
2912<div class="doc_text">
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2913<p><tt>#include "<a href="/doxygen/Type_8h-source.html">llvm/Type.h</a>"</tt>
2914<br>doxygen info: <a href="/doxygen/classllvm_1_1Type.html">Type Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2915
2916<p>The Core LLVM classes are the primary means of representing the program
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2917being inspected or transformed. The core LLVM classes are defined in
2918header files in the <tt>include/llvm/</tt> directory, and implemented in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2919the <tt>lib/VMCore</tt> directory.</p>
2920
2921</div>
2922
2923<!-- ======================================================================= -->
2924<div class="doc_subsection">
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2925 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
2926</div>
2927
2928<div class="doc_text">
2929
2930 <p><tt>Type</tt> is a superclass of all type classes. Every <tt>Value</tt> has
2931 a <tt>Type</tt>. <tt>Type</tt> cannot be instantiated directly but only
2932 through its subclasses. Certain primitive types (<tt>VoidType</tt>,
2933 <tt>LabelType</tt>, <tt>FloatType</tt> and <tt>DoubleType</tt>) have hidden
2934 subclasses. They are hidden because they offer no useful functionality beyond
2935 what the <tt>Type</tt> class offers except to distinguish themselves from
2936 other subclasses of <tt>Type</tt>.</p>
2937 <p>All other types are subclasses of <tt>DerivedType</tt>. Types can be
2938 named, but this is not a requirement. There exists exactly
2939 one instance of a given shape at any one time. This allows type equality to
2940 be performed with address equality of the Type Instance. That is, given two
2941 <tt>Type*</tt> values, the types are identical if the pointers are identical.
2942 </p>
2943</div>
2944
2945<!-- _______________________________________________________________________ -->
2946<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2947 <a name="m_Type">Important Public Methods</a>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2948</div>
2949
2950<div class="doc_text">
2951
2952<ul>
Chris Lattner8f79df32007-01-15 01:55:32 +0000[diff] [blame]2953 <li><tt>bool isInteger() const</tt>: Returns true for any integer type.</li>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2954
2955 <li><tt>bool isFloatingPoint()</tt>: Return true if this is one of the two
2956 floating point types.</li>
2957
2958 <li><tt>bool isAbstract()</tt>: Return true if the type is abstract (contains
2959 an OpaqueType anywhere in its definition).</li>
2960
2961 <li><tt>bool isSized()</tt>: Return true if the type has known size. Things
2962 that don't have a size are abstract types, labels and void.</li>
2963
2964</ul>
2965</div>
2966
2967<!-- _______________________________________________________________________ -->
2968<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2969 <a name="derivedtypes">Important Derived Types</a>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2970</div>
2971<div class="doc_text">
2972<dl>
2973 <dt><tt>IntegerType</tt></dt>
2974 <dd>Subclass of DerivedType that represents integer types of any bit width.
2975 Any bit width between <tt>IntegerType::MIN_INT_BITS</tt> (1) and
2976 <tt>IntegerType::MAX_INT_BITS</tt> (~8 million) can be represented.
2977 <ul>
2978 <li><tt>static const IntegerType* get(unsigned NumBits)</tt>: get an integer
2979 type of a specific bit width.</li>
2980 <li><tt>unsigned getBitWidth() const</tt>: Get the bit width of an integer
2981 type.</li>
2982 </ul>
2983 </dd>
2984 <dt><tt>SequentialType</tt></dt>
2985 <dd>This is subclassed by ArrayType and PointerType
2986 <ul>
2987 <li><tt>const Type * getElementType() const</tt>: Returns the type of each
2988 of the elements in the sequential type. </li>
2989 </ul>
2990 </dd>
2991 <dt><tt>ArrayType</tt></dt>
2992 <dd>This is a subclass of SequentialType and defines the interface for array
2993 types.
2994 <ul>
2995 <li><tt>unsigned getNumElements() const</tt>: Returns the number of
2996 elements in the array. </li>
2997 </ul>
2998 </dd>
2999 <dt><tt>PointerType</tt></dt>
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]3000 <dd>Subclass of SequentialType for pointer types.</dd>
Reid Spencer9d6565a2007-02-15 02:26:10 +0000[diff] [blame]3001 <dt><tt>VectorType</tt></dt>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]3002 <dd>Subclass of SequentialType for vector types. A
3003 vector type is similar to an ArrayType but is distinguished because it is
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3004 a first class type whereas ArrayType is not. Vector types are used for
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3005 vector operations and are usually small vectors of of an integer or floating
3006 point type.</dd>
3007 <dt><tt>StructType</tt></dt>
3008 <dd>Subclass of DerivedTypes for struct types.</dd>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3009 <dt><tt><a name="FunctionType">FunctionType</a></tt></dt>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3010 <dd>Subclass of DerivedTypes for function types.
3011 <ul>
3012 <li><tt>bool isVarArg() const</tt>: Returns true if its a vararg
3013 function</li>
3014 <li><tt> const Type * getReturnType() const</tt>: Returns the
3015 return type of the function.</li>
3016 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns
3017 the type of the ith parameter.</li>
3018 <li><tt> const unsigned getNumParams() const</tt>: Returns the
3019 number of formal parameters.</li>
3020 </ul>
3021 </dd>
3022 <dt><tt>OpaqueType</tt></dt>
3023 <dd>Sublcass of DerivedType for abstract types. This class
3024 defines no content and is used as a placeholder for some other type. Note
3025 that OpaqueType is used (temporarily) during type resolution for forward
3026 references of types. Once the referenced type is resolved, the OpaqueType
3027 is replaced with the actual type. OpaqueType can also be used for data
3028 abstraction. At link time opaque types can be resolved to actual types
3029 of the same name.</dd>
3030</dl>
3031</div>
3032
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3033
3034
3035<!-- ======================================================================= -->
3036<div class="doc_subsection">
3037 <a name="Module">The <tt>Module</tt> class</a>
3038</div>
3039
3040<div class="doc_text">
3041
3042<p><tt>#include "<a
3043href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt><br> doxygen info:
3044<a href="/doxygen/classllvm_1_1Module.html">Module Class</a></p>
3045
3046<p>The <tt>Module</tt> class represents the top level structure present in LLVM
3047programs. An LLVM module is effectively either a translation unit of the
3048original program or a combination of several translation units merged by the
3049linker. The <tt>Module</tt> class keeps track of a list of <a
3050href="#Function"><tt>Function</tt></a>s, a list of <a
3051href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
3052href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
3053helpful member functions that try to make common operations easy.</p>
3054
3055</div>
3056
3057<!-- _______________________________________________________________________ -->
3058<div class="doc_subsubsection">
3059 <a name="m_Module">Important Public Members of the <tt>Module</tt> class</a>
3060</div>
3061
3062<div class="doc_text">
3063
3064<ul>
3065 <li><tt>Module::Module(std::string name = "")</tt></li>
3066</ul>
3067
3068<p>Constructing a <a href="#Module">Module</a> is easy. You can optionally
3069provide a name for it (probably based on the name of the translation unit).</p>
3070
3071<ul>
3072 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
3073 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
3074
3075 <tt>begin()</tt>, <tt>end()</tt>
3076 <tt>size()</tt>, <tt>empty()</tt>
3077
3078 <p>These are forwarding methods that make it easy to access the contents of
3079 a <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
3080 list.</p></li>
3081
3082 <li><tt>Module::FunctionListType &amp;getFunctionList()</tt>
3083
3084 <p> Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
3085 necessary to use when you need to update the list or perform a complex
3086 action that doesn't have a forwarding method.</p>
3087
3088 <p><!-- Global Variable --></p></li>
3089</ul>
3090
3091<hr>
3092
3093<ul>
3094 <li><tt>Module::global_iterator</tt> - Typedef for global variable list iterator<br>
3095
3096 <tt>Module::const_global_iterator</tt> - Typedef for const_iterator.<br>
3097
3098 <tt>global_begin()</tt>, <tt>global_end()</tt>
3099 <tt>global_size()</tt>, <tt>global_empty()</tt>
3100
3101 <p> These are forwarding methods that make it easy to access the contents of
3102 a <tt>Module</tt> object's <a
3103 href="#GlobalVariable"><tt>GlobalVariable</tt></a> list.</p></li>
3104
3105 <li><tt>Module::GlobalListType &amp;getGlobalList()</tt>
3106
3107 <p>Returns the list of <a
3108 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s. This is necessary to
3109 use when you need to update the list or perform a complex action that
3110 doesn't have a forwarding method.</p>
3111
3112 <p><!-- Symbol table stuff --> </p></li>
3113</ul>
3114
3115<hr>
3116
3117<ul>
3118 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
3119
3120 <p>Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
3121 for this <tt>Module</tt>.</p>
3122
3123 <p><!-- Convenience methods --></p></li>
3124</ul>
3125
3126<hr>
3127
3128<ul>
3129 <li><tt><a href="#Function">Function</a> *getFunction(const std::string
3130 &amp;Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt>
3131
3132 <p>Look up the specified function in the <tt>Module</tt> <a
3133 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
3134 <tt>null</tt>.</p></li>
3135
3136 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const
3137 std::string &amp;Name, const <a href="#FunctionType">FunctionType</a> *T)</tt>
3138
3139 <p>Look up the specified function in the <tt>Module</tt> <a
3140 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
3141 external declaration for the function and return it.</p></li>
3142
3143 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt>
3144
3145 <p>If there is at least one entry in the <a
3146 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
3147 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
3148 string.</p></li>
3149
3150 <li><tt>bool addTypeName(const std::string &amp;Name, const <a
3151 href="#Type">Type</a> *Ty)</tt>
3152
3153 <p>Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
3154 mapping <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this
3155 name, true is returned and the <a
3156 href="#SymbolTable"><tt>SymbolTable</tt></a> is not modified.</p></li>
3157</ul>
3158
3159</div>
3160
3161
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3162<!-- ======================================================================= -->
3163<div class="doc_subsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3164 <a name="Value">The <tt>Value</tt> class</a>
3165</div>
3166
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3167<div class="doc_text">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3168
3169<p><tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt>
3170<br>
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]3171doxygen info: <a href="/doxygen/classllvm_1_1Value.html">Value Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3172
3173<p>The <tt>Value</tt> class is the most important class in the LLVM Source
3174base. It represents a typed value that may be used (among other things) as an
3175operand to an instruction. There are many different types of <tt>Value</tt>s,
3176such as <a href="#Constant"><tt>Constant</tt></a>s,<a
3177href="#Argument"><tt>Argument</tt></a>s. Even <a
3178href="#Instruction"><tt>Instruction</tt></a>s and <a
3179href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.</p>
3180
3181<p>A particular <tt>Value</tt> may be used many times in the LLVM representation
3182for a program. For example, an incoming argument to a function (represented
3183with an instance of the <a href="#Argument">Argument</a> class) is "used" by
3184every instruction in the function that references the argument. To keep track
3185of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
3186href="#User"><tt>User</tt></a>s that is using it (the <a
3187href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
3188graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
3189def-use information in the program, and is accessible through the <tt>use_</tt>*
3190methods, shown below.</p>
3191
3192<p>Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed,
3193and this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
3194method. In addition, all LLVM values can be named. The "name" of the
3195<tt>Value</tt> is a symbolic string printed in the LLVM code:</p>
3196
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3197<div class="doc_code">
3198<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]3199%<b>foo</b> = add i32 1, 2
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3200</pre>
3201</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3202
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3203<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]3204that the name of any value may be missing (an empty string), so names should
3205<b>ONLY</b> be used for debugging (making the source code easier to read,
3206debugging printouts), they should not be used to keep track of values or map
3207between them. For this purpose, use a <tt>std::map</tt> of pointers to the
3208<tt>Value</tt> itself instead.</p>
3209
3210<p>One important aspect of LLVM is that there is no distinction between an SSA
3211variable and the operation that produces it. Because of this, any reference to
3212the value produced by an instruction (or the value available as an incoming
Chris Lattnerd5fc4fc2004-03-18 14:58:55 +0000[diff] [blame]3213argument, for example) is represented as a direct pointer to the instance of
3214the class that
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3215represents this value. Although this may take some getting used to, it
3216simplifies the representation and makes it easier to manipulate.</p>
3217
3218</div>
3219
3220<!-- _______________________________________________________________________ -->
3221<div class="doc_subsubsection">
3222 <a name="m_Value">Important Public Members of the <tt>Value</tt> class</a>
3223</div>
3224
3225<div class="doc_text">
3226
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3227<ul>
3228 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the
3229use-list<br>
3230 <tt>Value::use_const_iterator</tt> - Typedef for const_iterator over
3231the use-list<br>
3232 <tt>unsigned use_size()</tt> - Returns the number of users of the
3233value.<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3234 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3235 <tt>use_iterator use_begin()</tt> - Get an iterator to the start of
3236the use-list.<br>
3237 <tt>use_iterator use_end()</tt> - Get an iterator to the end of the
3238use-list.<br>
3239 <tt><a href="#User">User</a> *use_back()</tt> - Returns the last
3240element in the list.
3241 <p> These methods are the interface to access the def-use
3242information in LLVM. As with all other iterators in LLVM, the naming
3243conventions follow the conventions defined by the <a href="#stl">STL</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3244 </li>
3245 <li><tt><a href="#Type">Type</a> *getType() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3246 <p>This method returns the Type of the Value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3247 </li>
3248 <li><tt>bool hasName() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3249 <tt>std::string getName() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3250 <tt>void setName(const std::string &amp;Name)</tt>
3251 <p> This family of methods is used to access and assign a name to a <tt>Value</tt>,
3252be aware of the <a href="#nameWarning">precaution above</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3253 </li>
3254 <li><tt>void replaceAllUsesWith(Value *V)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3255
3256 <p>This method traverses the use list of a <tt>Value</tt> changing all <a
3257 href="#User"><tt>User</tt>s</a> of the current value to refer to
3258 "<tt>V</tt>" instead. For example, if you detect that an instruction always
3259 produces a constant value (for example through constant folding), you can
3260 replace all uses of the instruction with the constant like this:</p>
3261
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3262<div class="doc_code">
3263<pre>
3264Inst-&gt;replaceAllUsesWith(ConstVal);
3265</pre>
3266</div>
3267
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3268</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3269
3270</div>
3271
3272<!-- ======================================================================= -->
3273<div class="doc_subsection">
3274 <a name="User">The <tt>User</tt> class</a>
3275</div>
3276
3277<div class="doc_text">
3278
3279<p>
3280<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]3281doxygen info: <a href="/doxygen/classllvm_1_1User.html">User Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3282Superclass: <a href="#Value"><tt>Value</tt></a></p>
3283
3284<p>The <tt>User</tt> class is the common base class of all LLVM nodes that may
3285refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
3286that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
3287referring to. The <tt>User</tt> class itself is a subclass of
3288<tt>Value</tt>.</p>
3289
3290<p>The operands of a <tt>User</tt> point directly to the LLVM <a
3291href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
3292Single Assignment (SSA) form, there can only be one definition referred to,
3293allowing this direct connection. This connection provides the use-def
3294information in LLVM.</p>
3295
3296</div>
3297
3298<!-- _______________________________________________________________________ -->
3299<div class="doc_subsubsection">
3300 <a name="m_User">Important Public Members of the <tt>User</tt> class</a>
3301</div>
3302
3303<div class="doc_text">
3304
3305<p>The <tt>User</tt> class exposes the operand list in two ways: through
3306an index access interface and through an iterator based interface.</p>
3307
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3308<ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3309 <li><tt>Value *getOperand(unsigned i)</tt><br>
3310 <tt>unsigned getNumOperands()</tt>
3311 <p> These two methods expose the operands of the <tt>User</tt> in a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3312convenient form for direct access.</p></li>
3313
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3314 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand
3315list<br>
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]3316 <tt>op_iterator op_begin()</tt> - Get an iterator to the start of
3317the operand list.<br>
3318 <tt>op_iterator op_end()</tt> - Get an iterator to the end of the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3319operand list.
3320 <p> Together, these methods make up the iterator based interface to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3321the operands of a <tt>User</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3322</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3323
3324</div>
3325
3326<!-- ======================================================================= -->
3327<div class="doc_subsection">
3328 <a name="Instruction">The <tt>Instruction</tt> class</a>
3329</div>
3330
3331<div class="doc_text">
3332
3333<p><tt>#include "</tt><tt><a
3334href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt><br>
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3335doxygen info: <a href="/doxygen/classllvm_1_1Instruction.html">Instruction Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3336Superclasses: <a href="#User"><tt>User</tt></a>, <a
3337href="#Value"><tt>Value</tt></a></p>
3338
3339<p>The <tt>Instruction</tt> class is the common base class for all LLVM
3340instructions. It provides only a few methods, but is a very commonly used
3341class. The primary data tracked by the <tt>Instruction</tt> class itself is the
3342opcode (instruction type) and the parent <a
3343href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
3344into. To represent a specific type of instruction, one of many subclasses of
3345<tt>Instruction</tt> are used.</p>
3346
3347<p> Because the <tt>Instruction</tt> class subclasses the <a
3348href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
3349way as for other <a href="#User"><tt>User</tt></a>s (with the
3350<tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
3351<tt>op_begin()</tt>/<tt>op_end()</tt> methods).</p> <p> An important file for
3352the <tt>Instruction</tt> class is the <tt>llvm/Instruction.def</tt> file. This
3353file contains some meta-data about the various different types of instructions
3354in LLVM. It describes the enum values that are used as opcodes (for example
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3355<tt>Instruction::Add</tt> and <tt>Instruction::ICmp</tt>), as well as the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3356concrete sub-classes of <tt>Instruction</tt> that implement the instruction (for
3357example <tt><a href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3358href="#CmpInst">CmpInst</a></tt>). Unfortunately, the use of macros in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3359this file confuses doxygen, so these enum values don't show up correctly in the
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3360<a href="/doxygen/classllvm_1_1Instruction.html">doxygen output</a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3361
3362</div>
3363
3364<!-- _______________________________________________________________________ -->
3365<div class="doc_subsubsection">
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3366 <a name="s_Instruction">Important Subclasses of the <tt>Instruction</tt>
3367 class</a>
3368</div>
3369<div class="doc_text">
3370 <ul>
3371 <li><tt><a name="BinaryOperator">BinaryOperator</a></tt>
3372 <p>This subclasses represents all two operand instructions whose operands
3373 must be the same type, except for the comparison instructions.</p></li>
3374 <li><tt><a name="CastInst">CastInst</a></tt>
3375 <p>This subclass is the parent of the 12 casting instructions. It provides
3376 common operations on cast instructions.</p>
3377 <li><tt><a name="CmpInst">CmpInst</a></tt>
3378 <p>This subclass respresents the two comparison instructions,
3379 <a href="LangRef.html#i_icmp">ICmpInst</a> (integer opreands), and
3380 <a href="LangRef.html#i_fcmp">FCmpInst</a> (floating point operands).</p>
3381 <li><tt><a name="TerminatorInst">TerminatorInst</a></tt>
3382 <p>This subclass is the parent of all terminator instructions (those which
3383 can terminate a block).</p>
3384 </ul>
3385 </div>
3386
3387<!-- _______________________________________________________________________ -->
3388<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3389 <a name="m_Instruction">Important Public Members of the <tt>Instruction</tt>
3390 class</a>
3391</div>
3392
3393<div class="doc_text">
3394
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3395<ul>
3396 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3397 <p>Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that
3398this <tt>Instruction</tt> is embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3399 <li><tt>bool mayWriteToMemory()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3400 <p>Returns true if the instruction writes to memory, i.e. it is a
3401 <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]3402 <li><tt>unsigned getOpcode()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3403 <p>Returns the opcode for the <tt>Instruction</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3404 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3405 <p>Returns another instance of the specified instruction, identical
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3406in all ways to the original except that the instruction has no parent
3407(ie it's not embedded into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>),
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3408and it has no name</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3409</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3410
3411</div>
3412
3413<!-- ======================================================================= -->
3414<div class="doc_subsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3415 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3416</div>
3417
3418<div class="doc_text">
3419
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3420<p>Constant represents a base class for different types of constants. It
3421is subclassed by ConstantInt, ConstantArray, etc. for representing
3422the various types of Constants. <a href="#GlobalValue">GlobalValue</a> is also
3423a subclass, which represents the address of a global variable or function.
3424</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3425
3426</div>
3427
3428<!-- _______________________________________________________________________ -->
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3429<div class="doc_subsubsection">Important Subclasses of Constant </div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3430<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3431<ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3432 <li>ConstantInt : This subclass of Constant represents an integer constant of
3433 any width.
3434 <ul>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]3435 <li><tt>const APInt&amp; getValue() const</tt>: Returns the underlying
3436 value of this constant, an APInt value.</li>
3437 <li><tt>int64_t getSExtValue() const</tt>: Converts the underlying APInt
3438 value to an int64_t via sign extension. If the value (not the bit width)
3439 of the APInt is too large to fit in an int64_t, an assertion will result.
3440 For this reason, use of this method is discouraged.</li>
3441 <li><tt>uint64_t getZExtValue() const</tt>: Converts the underlying APInt
3442 value to a uint64_t via zero extension. IF the value (not the bit width)
3443 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]3444 For this reason, use of this method is discouraged.</li>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]3445 <li><tt>static ConstantInt* get(const APInt&amp; Val)</tt>: Returns the
3446 ConstantInt object that represents the value provided by <tt>Val</tt>.
3447 The type is implied as the IntegerType that corresponds to the bit width
3448 of <tt>Val</tt>.</li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3449 <li><tt>static ConstantInt* get(const Type *Ty, uint64_t Val)</tt>:
3450 Returns the ConstantInt object that represents the value provided by
3451 <tt>Val</tt> for integer type <tt>Ty</tt>.</li>
3452 </ul>
3453 </li>
3454 <li>ConstantFP : This class represents a floating point constant.
3455 <ul>
3456 <li><tt>double getValue() const</tt>: Returns the underlying value of
3457 this constant. </li>
3458 </ul>
3459 </li>
3460 <li>ConstantArray : This represents a constant array.
3461 <ul>
3462 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
3463 a vector of component constants that makeup this array. </li>
3464 </ul>
3465 </li>
3466 <li>ConstantStruct : This represents a constant struct.
3467 <ul>
3468 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
3469 a vector of component constants that makeup this array. </li>
3470 </ul>
3471 </li>
3472 <li>GlobalValue : This represents either a global variable or a function. In
3473 either case, the value is a constant fixed address (after linking).
3474 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3475</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3476</div>
3477
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3478
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3479<!-- ======================================================================= -->
3480<div class="doc_subsection">
3481 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
3482</div>
3483
3484<div class="doc_text">
3485
3486<p><tt>#include "<a
3487href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt><br>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]3488doxygen info: <a href="/doxygen/classllvm_1_1GlobalValue.html">GlobalValue
3489Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3490Superclasses: <a href="#Constant"><tt>Constant</tt></a>,
3491<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]3492
3493<p>Global values (<a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
3494href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
3495visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
3496Because they are visible at global scope, they are also subject to linking with
3497other globals defined in different translation units. To control the linking
3498process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
3499<tt>GlobalValue</tt>s know whether they have internal or external linkage, as
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]3500defined by the <tt>LinkageTypes</tt> enumeration.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3501
3502<p>If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
3503<tt>static</tt> in C), it is not visible to code outside the current translation
3504unit, and does not participate in linking. If it has external linkage, it is
3505visible to external code, and does participate in linking. In addition to
3506linkage information, <tt>GlobalValue</tt>s keep track of which <a
3507href="#Module"><tt>Module</tt></a> they are currently part of.</p>
3508
3509<p>Because <tt>GlobalValue</tt>s are memory objects, they are always referred to
3510by their <b>address</b>. As such, the <a href="#Type"><tt>Type</tt></a> of a
3511global is always a pointer to its contents. It is important to remember this
3512when using the <tt>GetElementPtrInst</tt> instruction because this pointer must
3513be dereferenced first. For example, if you have a <tt>GlobalVariable</tt> (a
3514subclass 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]3515i32]</tt>, then the <tt>GlobalVariable</tt> is a pointer to that array. Although
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3516the address of the first element of this array and the value of the
3517<tt>GlobalVariable</tt> are the same, they have different types. The
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]3518<tt>GlobalVariable</tt>'s type is <tt>[24 x i32]</tt>. The first element's type
3519is <tt>i32.</tt> Because of this, accessing a global value requires you to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3520dereference the pointer with <tt>GetElementPtrInst</tt> first, then its elements
3521can be accessed. This is explained in the <a href="LangRef.html#globalvars">LLVM
3522Language Reference Manual</a>.</p>
3523
3524</div>
3525
3526<!-- _______________________________________________________________________ -->
3527<div class="doc_subsubsection">
3528 <a name="m_GlobalValue">Important Public Members of the <tt>GlobalValue</tt>
3529 class</a>
3530</div>
3531
3532<div class="doc_text">
3533
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3534<ul>
3535 <li><tt>bool hasInternalLinkage() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3536 <tt>bool hasExternalLinkage() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3537 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt>
3538 <p> These methods manipulate the linkage characteristics of the <tt>GlobalValue</tt>.</p>
3539 <p> </p>
3540 </li>
3541 <li><tt><a href="#Module">Module</a> *getParent()</tt>
3542 <p> This returns the <a href="#Module"><tt>Module</tt></a> that the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3543GlobalValue is currently embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3544</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3545
3546</div>
3547
3548<!-- ======================================================================= -->
3549<div class="doc_subsection">
3550 <a name="Function">The <tt>Function</tt> class</a>
3551</div>
3552
3553<div class="doc_text">
3554
3555<p><tt>#include "<a
3556href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt><br> doxygen
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3557info: <a href="/doxygen/classllvm_1_1Function.html">Function Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3558Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
3559<a href="#Constant"><tt>Constant</tt></a>,
3560<a href="#User"><tt>User</tt></a>,
3561<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3562
3563<p>The <tt>Function</tt> class represents a single procedure in LLVM. It is
Torok Edwin87469292009-10-12 13:37:29 +0000[diff] [blame]3564actually one of the more complex classes in the LLVM hierarchy because it must
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3565keep track of a large amount of data. The <tt>Function</tt> class keeps track
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3566of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal
3567<a href="#Argument"><tt>Argument</tt></a>s, and a
3568<a href="#SymbolTable"><tt>SymbolTable</tt></a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3569
3570<p>The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most
3571commonly used part of <tt>Function</tt> objects. The list imposes an implicit
3572ordering of the blocks in the function, which indicate how the code will be
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3573laid out by the backend. Additionally, the first <a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3574href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
3575<tt>Function</tt>. It is not legal in LLVM to explicitly branch to this initial
3576block. There are no implicit exit nodes, and in fact there may be multiple exit
3577nodes from a single <tt>Function</tt>. If the <a
3578href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
3579the <tt>Function</tt> is actually a function declaration: the actual body of the
3580function hasn't been linked in yet.</p>
3581
3582<p>In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
3583<tt>Function</tt> class also keeps track of the list of formal <a
3584href="#Argument"><tt>Argument</tt></a>s that the function receives. This
3585container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
3586nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
3587the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.</p>
3588
3589<p>The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used
3590LLVM feature that is only used when you have to look up a value by name. Aside
3591from that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used
3592internally to make sure that there are not conflicts between the names of <a
3593href="#Instruction"><tt>Instruction</tt></a>s, <a
3594href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
3595href="#Argument"><tt>Argument</tt></a>s in the function body.</p>
3596
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]3597<p>Note that <tt>Function</tt> is a <a href="#GlobalValue">GlobalValue</a>
3598and therefore also a <a href="#Constant">Constant</a>. The value of the function
3599is its address (after linking) which is guaranteed to be constant.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3600</div>
3601
3602<!-- _______________________________________________________________________ -->
3603<div class="doc_subsubsection">
3604 <a name="m_Function">Important Public Members of the <tt>Function</tt>
3605 class</a>
3606</div>
3607
3608<div class="doc_text">
3609
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3610<ul>
3611 <li><tt>Function(const </tt><tt><a href="#FunctionType">FunctionType</a>
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]3612 *Ty, LinkageTypes Linkage, const std::string &amp;N = "", Module* Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3613
3614 <p>Constructor used when you need to create new <tt>Function</tt>s to add
3615 the the program. The constructor must specify the type of the function to
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]3616 create and what type of linkage the function should have. The <a
3617 href="#FunctionType"><tt>FunctionType</tt></a> argument
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3618 specifies the formal arguments and return value for the function. The same
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3619 <a href="#FunctionType"><tt>FunctionType</tt></a> value can be used to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3620 create multiple functions. The <tt>Parent</tt> argument specifies the Module
3621 in which the function is defined. If this argument is provided, the function
3622 will automatically be inserted into that module's list of
3623 functions.</p></li>
3624
Chris Lattner62810e32008-11-25 18:34:50 +0000[diff] [blame]3625 <li><tt>bool isDeclaration()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3626
3627 <p>Return whether or not the <tt>Function</tt> has a body defined. If the
3628 function is "external", it does not have a body, and thus must be resolved
3629 by linking with a function defined in a different translation unit.</p></li>
3630
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3631 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3632 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3633
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]3634 <tt>begin()</tt>, <tt>end()</tt>
3635 <tt>size()</tt>, <tt>empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3636
3637 <p>These are forwarding methods that make it easy to access the contents of
3638 a <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
3639 list.</p></li>
3640
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3641 <li><tt>Function::BasicBlockListType &amp;getBasicBlockList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3642
3643 <p>Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This
3644 is necessary to use when you need to update the list or perform a complex
3645 action that doesn't have a forwarding method.</p></li>
3646
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3647 <li><tt>Function::arg_iterator</tt> - Typedef for the argument list
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3648iterator<br>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3649 <tt>Function::const_arg_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3650
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]3651 <tt>arg_begin()</tt>, <tt>arg_end()</tt>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3652 <tt>arg_size()</tt>, <tt>arg_empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3653
3654 <p>These are forwarding methods that make it easy to access the contents of
3655 a <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a>
3656 list.</p></li>
3657
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3658 <li><tt>Function::ArgumentListType &amp;getArgumentList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3659
3660 <p>Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
3661 necessary to use when you need to update the list or perform a complex
3662 action that doesn't have a forwarding method.</p></li>
3663
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3664 <li><tt><a href="#BasicBlock">BasicBlock</a> &amp;getEntryBlock()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3665
3666 <p>Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
3667 function. Because the entry block for the function is always the first
3668 block, this returns the first block of the <tt>Function</tt>.</p></li>
3669
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3670 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
3671 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3672
3673 <p>This traverses the <a href="#Type"><tt>Type</tt></a> of the
3674 <tt>Function</tt> and returns the return type of the function, or the <a
3675 href="#FunctionType"><tt>FunctionType</tt></a> of the actual
3676 function.</p></li>
3677
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3678 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3679
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3680 <p> Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3681 for this <tt>Function</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3682</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3683
3684</div>
3685
3686<!-- ======================================================================= -->
3687<div class="doc_subsection">
3688 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
3689</div>
3690
3691<div class="doc_text">
3692
3693<p><tt>#include "<a
3694href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt>
3695<br>
Tanya Lattnera3da7772004-06-22 08:02:25 +0000[diff] [blame]3696doxygen info: <a href="/doxygen/classllvm_1_1GlobalVariable.html">GlobalVariable
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3697 Class</a><br>
3698Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
3699<a href="#Constant"><tt>Constant</tt></a>,
3700<a href="#User"><tt>User</tt></a>,
3701<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3702
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3703<p>Global variables are represented with the (surprise surprise)
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3704<tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are also
3705subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such are
3706always referenced by their address (global values must live in memory, so their
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3707"name" refers to their constant address). See
3708<a href="#GlobalValue"><tt>GlobalValue</tt></a> for more on this. Global
3709variables may have an initial value (which must be a
3710<a href="#Constant"><tt>Constant</tt></a>), and if they have an initializer,
3711they may be marked as "constant" themselves (indicating that their contents
3712never change at runtime).</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3713</div>
3714
3715<!-- _______________________________________________________________________ -->
3716<div class="doc_subsubsection">
3717 <a name="m_GlobalVariable">Important Public Members of the
3718 <tt>GlobalVariable</tt> class</a>
3719</div>
3720
3721<div class="doc_text">
3722
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3723<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3724 <li><tt>GlobalVariable(const </tt><tt><a href="#Type">Type</a> *Ty, bool
3725 isConstant, LinkageTypes&amp; Linkage, <a href="#Constant">Constant</a>
3726 *Initializer = 0, const std::string &amp;Name = "", Module* Parent = 0)</tt>
3727
3728 <p>Create a new global variable of the specified type. If
3729 <tt>isConstant</tt> is true then the global variable will be marked as
3730 unchanging for the program. The Linkage parameter specifies the type of
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]3731 linkage (internal, external, weak, linkonce, appending) for the variable.
3732 If the linkage is InternalLinkage, WeakAnyLinkage, WeakODRLinkage,
3733 LinkOnceAnyLinkage or LinkOnceODRLinkage,&nbsp; then the resultant
3734 global variable will have internal linkage. AppendingLinkage concatenates
3735 together all instances (in different translation units) of the variable
3736 into a single variable but is only applicable to arrays. &nbsp;See
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3737 the <a href="LangRef.html#modulestructure">LLVM Language Reference</a> for
3738 further details on linkage types. Optionally an initializer, a name, and the
3739 module to put the variable into may be specified for the global variable as
3740 well.</p></li>
3741
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3742 <li><tt>bool isConstant() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3743
3744 <p>Returns true if this is a global variable that is known not to
3745 be modified at runtime.</p></li>
3746
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3747 <li><tt>bool hasInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3748
3749 <p>Returns true if this <tt>GlobalVariable</tt> has an intializer.</p></li>
3750
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3751 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3752
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3753 <p>Returns the initial value for a <tt>GlobalVariable</tt>. It is not legal
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3754 to call this method if there is no initializer.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3755</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3756
3757</div>
3758
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3759
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3760<!-- ======================================================================= -->
3761<div class="doc_subsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3762 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3763</div>
3764
3765<div class="doc_text">
3766
3767<p><tt>#include "<a
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3768href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt><br>
Stefanus Du Toit24e04112009-06-17 21:12:26 +0000[diff] [blame]3769doxygen info: <a href="/doxygen/classllvm_1_1BasicBlock.html">BasicBlock
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3770Class</a><br>
3771Superclass: <a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3772
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3773<p>This class represents a single entry multiple exit section of the code,
3774commonly known as a basic block by the compiler community. The
3775<tt>BasicBlock</tt> class maintains a list of <a
3776href="#Instruction"><tt>Instruction</tt></a>s, which form the body of the block.
3777Matching the language definition, the last element of this list of instructions
3778is always a terminator instruction (a subclass of the <a
3779href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).</p>
3780
3781<p>In addition to tracking the list of instructions that make up the block, the
3782<tt>BasicBlock</tt> class also keeps track of the <a
3783href="#Function"><tt>Function</tt></a> that it is embedded into.</p>
3784
3785<p>Note that <tt>BasicBlock</tt>s themselves are <a
3786href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
3787like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
3788<tt>label</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3789
3790</div>
3791
3792<!-- _______________________________________________________________________ -->
3793<div class="doc_subsubsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3794 <a name="m_BasicBlock">Important Public Members of the <tt>BasicBlock</tt>
3795 class</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3796</div>
3797
3798<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3799<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3800
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3801<li><tt>BasicBlock(const std::string &amp;Name = "", </tt><tt><a
3802 href="#Function">Function</a> *Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3803
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3804<p>The <tt>BasicBlock</tt> constructor is used to create new basic blocks for
3805insertion into a function. The constructor optionally takes a name for the new
3806block, and a <a href="#Function"><tt>Function</tt></a> to insert it into. If
3807the <tt>Parent</tt> parameter is specified, the new <tt>BasicBlock</tt> is
3808automatically inserted at the end of the specified <a
3809href="#Function"><tt>Function</tt></a>, if not specified, the BasicBlock must be
3810manually inserted into the <a href="#Function"><tt>Function</tt></a>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3811
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3812<li><tt>BasicBlock::iterator</tt> - Typedef for instruction list iterator<br>
3813<tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
3814<tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
3815<tt>size()</tt>, <tt>empty()</tt>
3816STL-style functions for accessing the instruction list.
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3817
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3818<p>These methods and typedefs are forwarding functions that have the same
3819semantics as the standard library methods of the same names. These methods
3820expose the underlying instruction list of a basic block in a way that is easy to
3821manipulate. To get the full complement of container operations (including
3822operations to update the list), you must use the <tt>getInstList()</tt>
3823method.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3824
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3825<li><tt>BasicBlock::InstListType &amp;getInstList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3826
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3827<p>This method is used to get access to the underlying container that actually
3828holds the Instructions. This method must be used when there isn't a forwarding
3829function in the <tt>BasicBlock</tt> class for the operation that you would like
3830to perform. Because there are no forwarding functions for "updating"
3831operations, you need to use this if you want to update the contents of a
3832<tt>BasicBlock</tt>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3833
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3834<li><tt><a href="#Function">Function</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3835
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3836<p> Returns a pointer to <a href="#Function"><tt>Function</tt></a> the block is
3837embedded into, or a null pointer if it is homeless.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3838
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3839<li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3840
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3841<p> Returns a pointer to the terminator instruction that appears at the end of
3842the <tt>BasicBlock</tt>. If there is no terminator instruction, or if the last
3843instruction in the block is not a terminator, then a null pointer is
3844returned.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3845
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3846</ul>
3847
3848</div>
3849
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3850
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3851<!-- ======================================================================= -->
3852<div class="doc_subsection">
3853 <a name="Argument">The <tt>Argument</tt> class</a>
3854</div>
3855
3856<div class="doc_text">
3857
3858<p>This subclass of Value defines the interface for incoming formal
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]3859arguments to a function. A Function maintains a list of its formal
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3860arguments. An argument has a pointer to the parent Function.</p>
3861
3862</div>
3863
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3864<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3865<hr>
3866<address>
3867 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]3868 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3869 <a href="http://validator.w3.org/check/referer"><img
Gabor Greifa9c0f2b2008-06-18 14:05:31 +0000[diff] [blame]3870 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]3871
3872 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
3873 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]3874 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3875 Last modified: $Date$
3876</address>
3877
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3878</body>
3879</html>