Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 425d08c654c7893782b1215f9b2715bc1d90bd07 / . / docs / ProgrammersManual.html
blob: e920cbbc6460d54ac21bd5094e93bb824a08d58e [file] [log] [blame]
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3<html>
4<head>
Owen Anderson5e8c50e2009-06-16 17:40:28 +0000[diff] [blame]5 <meta http-equiv="Content-type" content="text/html;charset=UTF-8">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]6 <title>LLVM Programmer's Manual</title>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]7 <link rel="stylesheet" href="llvm.css" type="text/css">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]8</head>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]9<body>
10
11<div class="doc_title">
12 LLVM Programmer's Manual
13</div>
14
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]15<ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]16 <li><a href="#introduction">Introduction</a></li>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]17 <li><a href="#general">General Information</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]18 <ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]19 <li><a href="#stl">The C++ Standard Template Library</a></li>
20<!--
21 <li>The <tt>-time-passes</tt> option</li>
22 <li>How to use the LLVM Makefile system</li>
23 <li>How to write a regression test</li>
Chris Lattner61db4652004-12-08 19:05:44 +0000[diff] [blame]24
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]25-->
Chris Lattner84b7f8d2003-08-01 22:20:59 +0000[diff] [blame]26 </ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]27 </li>
28 <li><a href="#apis">Important and useful LLVM APIs</a>
29 <ul>
30 <li><a href="#isa">The <tt>isa&lt;&gt;</tt>, <tt>cast&lt;&gt;</tt>
31and <tt>dyn_cast&lt;&gt;</tt> templates</a> </li>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]32 <li><a href="#string_apis">Passing strings (the <tt>StringRef</tt>
Benjamin Kramere15192b2009-08-05 15:42:44 +0000[diff] [blame]33and <tt>Twine</tt> classes)</a>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]34 <ul>
35 <li><a href="#StringRef">The <tt>StringRef</tt> class</a> </li>
36 <li><a href="#Twine">The <tt>Twine</tt> class</a> </li>
37 </ul>
Benjamin Kramere15192b2009-08-05 15:42:44 +0000[diff] [blame]38 </li>
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]39 <li><a href="#DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]40option</a>
41 <ul>
42 <li><a href="#DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt>
43and the <tt>-debug-only</tt> option</a> </li>
44 </ul>
45 </li>
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]46 <li><a href="#Statistic">The <tt>Statistic</tt> class &amp; <tt>-stats</tt>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]47option</a></li>
48<!--
49 <li>The <tt>InstVisitor</tt> template
50 <li>The general graph API
51-->
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]52 <li><a href="#ViewGraph">Viewing graphs while debugging code</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]53 </ul>
54 </li>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]55 <li><a href="#datastructure">Picking the Right Data Structure for a Task</a>
56 <ul>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]57 <li><a href="#ds_sequential">Sequential Containers (std::vector, std::list, etc)</a>
58 <ul>
59 <li><a href="#dss_fixedarrays">Fixed Size Arrays</a></li>
60 <li><a href="#dss_heaparrays">Heap Allocated Arrays</a></li>
61 <li><a href="#dss_smallvector">"llvm/ADT/SmallVector.h"</a></li>
62 <li><a href="#dss_vector">&lt;vector&gt;</a></li>
63 <li><a href="#dss_deque">&lt;deque&gt;</a></li>
64 <li><a href="#dss_list">&lt;list&gt;</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]65 <li><a href="#dss_ilist">llvm/ADT/ilist.h</a></li>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]66 <li><a href="#dss_other">Other Sequential Container Options</a></li>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]67 </ul></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]68 <li><a href="#ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
69 <ul>
70 <li><a href="#dss_sortedvectorset">A sorted 'vector'</a></li>
71 <li><a href="#dss_smallset">"llvm/ADT/SmallSet.h"</a></li>
72 <li><a href="#dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a></li>
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]73 <li><a href="#dss_denseset">"llvm/ADT/DenseSet.h"</a></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]74 <li><a href="#dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a></li>
75 <li><a href="#dss_set">&lt;set&gt;</a></li>
76 <li><a href="#dss_setvector">"llvm/ADT/SetVector.h"</a></li>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]77 <li><a href="#dss_uniquevector">"llvm/ADT/UniqueVector.h"</a></li>
78 <li><a href="#dss_otherset">Other Set-Like ContainerOptions</a></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]79 </ul></li>
Chris Lattnerf3692522007-02-03 19:51:56 +0000[diff] [blame]80 <li><a href="#ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
81 <ul>
82 <li><a href="#dss_sortedvectormap">A sorted 'vector'</a></li>
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]83 <li><a href="#dss_stringmap">"llvm/ADT/StringMap.h"</a></li>
Chris Lattnerf3692522007-02-03 19:51:56 +0000[diff] [blame]84 <li><a href="#dss_indexedmap">"llvm/ADT/IndexedMap.h"</a></li>
85 <li><a href="#dss_densemap">"llvm/ADT/DenseMap.h"</a></li>
86 <li><a href="#dss_map">&lt;map&gt;</a></li>
87 <li><a href="#dss_othermap">Other Map-Like Container Options</a></li>
88 </ul></li>
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]89 <li><a href="#ds_string">String-like containers</a>
Benjamin Kramere15192b2009-08-05 15:42:44 +0000[diff] [blame]90 <!--<ul>
91 todo
92 </ul>--></li>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]93 <li><a href="#ds_bit">BitVector-like containers</a>
94 <ul>
95 <li><a href="#dss_bitvector">A dense bitvector</a></li>
96 <li><a href="#dss_sparsebitvector">A sparse bitvector</a></li>
97 </ul></li>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]98 </ul>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]99 </li>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]100 <li><a href="#common">Helpful Hints for Common Operations</a>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]101 <ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]102 <li><a href="#inspection">Basic Inspection and Traversal Routines</a>
103 <ul>
104 <li><a href="#iterate_function">Iterating over the <tt>BasicBlock</tt>s
105in a <tt>Function</tt></a> </li>
106 <li><a href="#iterate_basicblock">Iterating over the <tt>Instruction</tt>s
107in a <tt>BasicBlock</tt></a> </li>
108 <li><a href="#iterate_institer">Iterating over the <tt>Instruction</tt>s
109in a <tt>Function</tt></a> </li>
110 <li><a href="#iterate_convert">Turning an iterator into a
111class pointer</a> </li>
112 <li><a href="#iterate_complex">Finding call sites: a more
113complex example</a> </li>
114 <li><a href="#calls_and_invokes">Treating calls and invokes
115the same way</a> </li>
116 <li><a href="#iterate_chains">Iterating over def-use &amp;
117use-def chains</a> </li>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]118 <li><a href="#iterate_preds">Iterating over predecessors &amp;
119successors of blocks</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]120 </ul>
121 </li>
122 <li><a href="#simplechanges">Making simple changes</a>
123 <ul>
124 <li><a href="#schanges_creating">Creating and inserting new
125 <tt>Instruction</tt>s</a> </li>
126 <li><a href="#schanges_deleting">Deleting <tt>Instruction</tt>s</a> </li>
127 <li><a href="#schanges_replacing">Replacing an <tt>Instruction</tt>
128with another <tt>Value</tt></a> </li>
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]129 <li><a href="#schanges_deletingGV">Deleting <tt>GlobalVariable</tt>s</a> </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]130 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]131 </li>
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]132 <li><a href="#create_types">How to Create Types</a></li>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]133<!--
134 <li>Working with the Control Flow Graph
135 <ul>
136 <li>Accessing predecessors and successors of a <tt>BasicBlock</tt>
137 <li>
138 <li>
139 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]140-->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]141 </ul>
142 </li>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]143
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]144 <li><a href="#threading">Threads and LLVM</a>
145 <ul>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]146 <li><a href="#startmultithreaded">Entering and Exiting Multithreaded Mode
147 </a></li>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]148 <li><a href="#shutdown">Ending execution with <tt>llvm_shutdown()</tt></a></li>
149 <li><a href="#managedstatic">Lazy initialization with <tt>ManagedStatic</tt></a></li>
150 </ul>
151 </li>
152
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]153 <li><a href="#advanced">Advanced Topics</a>
154 <ul>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]155 <li><a href="#TypeResolve">LLVM Type Resolution</a>
156 <ul>
157 <li><a href="#BuildRecType">Basic Recursive Type Construction</a></li>
158 <li><a href="#refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a></li>
159 <li><a href="#PATypeHolder">The PATypeHolder Class</a></li>
160 <li><a href="#AbstractTypeUser">The AbstractTypeUser Class</a></li>
161 </ul></li>
162
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]163 <li><a href="#SymbolTable">The <tt>ValueSymbolTable</tt> and <tt>TypeSymbolTable</tt> classes</a></li>
164 <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]165 </ul></li>
166
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]167 <li><a href="#coreclasses">The Core LLVM Class Hierarchy Reference</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]168 <ul>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]169 <li><a href="#Type">The <tt>Type</tt> class</a> </li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]170 <li><a href="#Module">The <tt>Module</tt> class</a></li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]171 <li><a href="#Value">The <tt>Value</tt> class</a>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]172 <ul>
173 <li><a href="#User">The <tt>User</tt> class</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]174 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]175 <li><a href="#Instruction">The <tt>Instruction</tt> class</a></li>
176 <li><a href="#Constant">The <tt>Constant</tt> class</a>
177 <ul>
178 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]179 <ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]180 <li><a href="#Function">The <tt>Function</tt> class</a></li>
181 <li><a href="#GlobalVariable">The <tt>GlobalVariable</tt> class</a></li>
182 </ul>
183 </li>
184 </ul>
185 </li>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]186 </ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]187 </li>
188 <li><a href="#BasicBlock">The <tt>BasicBlock</tt> class</a></li>
189 <li><a href="#Argument">The <tt>Argument</tt> class</a></li>
190 </ul>
Reid Spencerfe8f4ff2004-11-01 09:02:53 +0000[diff] [blame]191 </li>
192 </ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]193 </li>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]194</ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]195
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]196<div class="doc_author">
197 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
Chris Lattner94c43592004-05-26 16:52:55 +0000[diff] [blame]198 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>,
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]199 <a href="mailto:ggreif@gmail.com">Gabor Greif</a>,
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]200 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a>,
201 <a href="mailto:rspencer@x10sys.com">Reid Spencer</a> and
202 <a href="mailto:owen@apple.com">Owen Anderson</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]203</div>
204
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]205<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]206<div class="doc_section">
207 <a name="introduction">Introduction </a>
208</div>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]209<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]210
211<div class="doc_text">
212
213<p>This document is meant to highlight some of the important classes and
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]214interfaces available in the LLVM source-base. This manual is not
215intended to explain what LLVM is, how it works, and what LLVM code looks
216like. It assumes that you know the basics of LLVM and are interested
217in writing transformations or otherwise analyzing or manipulating the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]218code.</p>
219
220<p>This document should get you oriented so that you can find your
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]221way in the continuously growing source code that makes up the LLVM
222infrastructure. Note that this manual is not intended to serve as a
223replacement for reading the source code, so if you think there should be
224a method in one of these classes to do something, but it's not listed,
225check the source. Links to the <a href="/doxygen/">doxygen</a> sources
226are provided to make this as easy as possible.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]227
228<p>The first section of this document describes general information that is
229useful to know when working in the LLVM infrastructure, and the second describes
230the Core LLVM classes. In the future this manual will be extended with
231information describing how to use extension libraries, such as dominator
232information, CFG traversal routines, and useful utilities like the <tt><a
233href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.</p>
234
235</div>
236
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]237<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]238<div class="doc_section">
239 <a name="general">General Information</a>
240</div>
241<!-- *********************************************************************** -->
242
243<div class="doc_text">
244
245<p>This section contains general information that is useful if you are working
246in the LLVM source-base, but that isn't specific to any particular API.</p>
247
248</div>
249
250<!-- ======================================================================= -->
251<div class="doc_subsection">
252 <a name="stl">The C++ Standard Template Library</a>
253</div>
254
255<div class="doc_text">
256
257<p>LLVM makes heavy use of the C++ Standard Template Library (STL),
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]258perhaps much more than you are used to, or have seen before. Because of
259this, you might want to do a little background reading in the
260techniques used and capabilities of the library. There are many good
261pages that discuss the STL, and several books on the subject that you
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]262can get, so it will not be discussed in this document.</p>
263
264<p>Here are some useful links:</p>
265
266<ol>
267
268<li><a href="http://www.dinkumware.com/refxcpp.html">Dinkumware C++ Library
269reference</a> - an excellent reference for the STL and other parts of the
270standard C++ library.</li>
271
272<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]273O'Reilly book in the making. It has a decent Standard Library
274Reference that rivals Dinkumware's, and is unfortunately no longer free since the
275book has been published.</li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]276
277<li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
278Questions</a></li>
279
280<li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
281Contains a useful <a
282href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
283STL</a>.</li>
284
285<li><a href="http://www.research.att.com/%7Ebs/C++.html">Bjarne Stroustrup's C++
286Page</a></li>
287
Tanya Lattner79445ba2004-12-08 18:34:56 +0000[diff] [blame]288<li><a href="http://64.78.49.204/">
Reid Spencer096603a2004-05-26 08:41:35 +0000[diff] [blame]289Bruce Eckel's Thinking in C++, 2nd ed. Volume 2 Revision 4.0 (even better, get
290the book).</a></li>
291
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]292</ol>
293
294<p>You are also encouraged to take a look at the <a
295href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
296to write maintainable code more than where to put your curly braces.</p>
297
298</div>
299
300<!-- ======================================================================= -->
301<div class="doc_subsection">
302 <a name="stl">Other useful references</a>
303</div>
304
305<div class="doc_text">
306
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]307<ol>
308<li><a href="http://www.psc.edu/%7Esemke/cvs_branches.html">CVS
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]309Branch and Tag Primer</a></li>
Misha Brukmana0f71e42004-06-18 18:39:00 +0000[diff] [blame]310<li><a href="http://www.fortran-2000.com/ArnaudRecipes/sharedlib.html">Using
311static and shared libraries across platforms</a></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]312</ol>
313
314</div>
315
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]316<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]317<div class="doc_section">
318 <a name="apis">Important and useful LLVM APIs</a>
319</div>
320<!-- *********************************************************************** -->
321
322<div class="doc_text">
323
324<p>Here we highlight some LLVM APIs that are generally useful and good to
325know about when writing transformations.</p>
326
327</div>
328
329<!-- ======================================================================= -->
330<div class="doc_subsection">
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]331 <a name="isa">The <tt>isa&lt;&gt;</tt>, <tt>cast&lt;&gt;</tt> and
332 <tt>dyn_cast&lt;&gt;</tt> templates</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]333</div>
334
335<div class="doc_text">
336
337<p>The LLVM source-base makes extensive use of a custom form of RTTI.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]338These templates have many similarities to the C++ <tt>dynamic_cast&lt;&gt;</tt>
339operator, but they don't have some drawbacks (primarily stemming from
340the fact that <tt>dynamic_cast&lt;&gt;</tt> only works on classes that
341have a v-table). Because they are used so often, you must know what they
342do and how they work. All of these templates are defined in the <a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]343 href="/doxygen/Casting_8h-source.html"><tt>llvm/Support/Casting.h</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]344file (note that you very rarely have to include this file directly).</p>
345
346<dl>
347 <dt><tt>isa&lt;&gt;</tt>: </dt>
348
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]349 <dd><p>The <tt>isa&lt;&gt;</tt> operator works exactly like the Java
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]350 "<tt>instanceof</tt>" operator. It returns true or false depending on whether
351 a reference or pointer points to an instance of the specified class. This can
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]352 be very useful for constraint checking of various sorts (example below).</p>
353 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]354
355 <dt><tt>cast&lt;&gt;</tt>: </dt>
356
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]357 <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]358 converts a pointer or reference from a base class to a derived class, causing
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]359 an assertion failure if it is not really an instance of the right type. This
360 should be used in cases where you have some information that makes you believe
361 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]362 and <tt>cast&lt;&gt;</tt> template is:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]363
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]364<div class="doc_code">
365<pre>
366static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
367 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))
368 return true;
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]369
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]370 // <i>Otherwise, it must be an instruction...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]371 return !L-&gt;contains(cast&lt;<a href="#Instruction">Instruction</a>&gt;(V)-&gt;getParent());
372}
373</pre>
374</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]375
376 <p>Note that you should <b>not</b> use an <tt>isa&lt;&gt;</tt> test followed
377 by a <tt>cast&lt;&gt;</tt>, for that use the <tt>dyn_cast&lt;&gt;</tt>
378 operator.</p>
379
380 </dd>
381
382 <dt><tt>dyn_cast&lt;&gt;</tt>:</dt>
383
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]384 <dd><p>The <tt>dyn_cast&lt;&gt;</tt> operator is a "checking cast" operation.
385 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]386 pointer to it (this operator does not work with references). If the operand is
387 not of the correct type, a null pointer is returned. Thus, this works very
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]388 much like the <tt>dynamic_cast&lt;&gt;</tt> operator in C++, and should be
389 used in the same circumstances. Typically, the <tt>dyn_cast&lt;&gt;</tt>
390 operator is used in an <tt>if</tt> statement or some other flow control
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]391 statement like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]392
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]393<div class="doc_code">
394<pre>
395if (<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]396 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]397}
398</pre>
399</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]400
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]401 <p>This form of the <tt>if</tt> statement effectively combines together a call
402 to <tt>isa&lt;&gt;</tt> and a call to <tt>cast&lt;&gt;</tt> into one
403 statement, which is very convenient.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]404
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]405 <p>Note that the <tt>dyn_cast&lt;&gt;</tt> operator, like C++'s
406 <tt>dynamic_cast&lt;&gt;</tt> or Java's <tt>instanceof</tt> operator, can be
407 abused. In particular, you should not use big chained <tt>if/then/else</tt>
408 blocks to check for lots of different variants of classes. If you find
409 yourself wanting to do this, it is much cleaner and more efficient to use the
410 <tt>InstVisitor</tt> class to dispatch over the instruction type directly.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]411
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]412 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]413
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]414 <dt><tt>cast_or_null&lt;&gt;</tt>: </dt>
415
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]416 <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]417 <tt>cast&lt;&gt;</tt> operator, except that it allows for a null pointer as an
418 argument (which it then propagates). This can sometimes be useful, allowing
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]419 you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]420
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]421 <dt><tt>dyn_cast_or_null&lt;&gt;</tt>: </dt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]422
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]423 <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]424 <tt>dyn_cast&lt;&gt;</tt> operator, except that it allows for a null pointer
425 as an argument (which it then propagates). This can sometimes be useful,
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]426 allowing you to combine several null checks into one.</p></dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]427
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]428</dl>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]429
430<p>These five templates can be used with any classes, whether they have a
431v-table or not. To add support for these templates, you simply need to add
432<tt>classof</tt> static methods to the class you are interested casting
433to. Describing this is currently outside the scope of this document, but there
434are lots of examples in the LLVM source base.</p>
435
436</div>
437
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]438
439<!-- ======================================================================= -->
440<div class="doc_subsection">
441 <a name="string_apis">Passing strings (the <tt>StringRef</tt>
442and <tt>Twine</tt> classes)</a>
443</div>
444
445<div class="doc_text">
446
447<p>Although LLVM generally does not do much string manipulation, we do have
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]448several important APIs which take strings. Two important examples are the
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]449Value class -- which has names for instructions, functions, etc. -- and the
450StringMap class which is used extensively in LLVM and Clang.</p>
451
452<p>These are generic classes, and they need to be able to accept strings which
453may have embedded null characters. Therefore, they cannot simply take
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]454a <tt>const char *</tt>, and taking a <tt>const std::string&amp;</tt> requires
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]455clients to perform a heap allocation which is usually unnecessary. Instead,
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]456many LLVM APIs use a <tt>const StringRef&amp;</tt> or a <tt>const
457Twine&amp;</tt> for passing strings efficiently.</p>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]458
459</div>
460
461<!-- _______________________________________________________________________ -->
462<div class="doc_subsubsection">
463 <a name="StringRef">The <tt>StringRef</tt> class</a>
464</div>
465
466<div class="doc_text">
467
468<p>The <tt>StringRef</tt> data type represents a reference to a constant string
469(a character array and a length) and supports the common operations available
470on <tt>std:string</tt>, but does not require heap allocation.</p>
471
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]472<p>It can be implicitly constructed using a C style null-terminated string,
473an <tt>std::string</tt>, or explicitly with a character pointer and length.
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]474For example, the <tt>StringRef</tt> find function is declared as:</p>
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]475
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]476<div class="doc_code">
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]477 iterator find(const StringRef &amp;Key);
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]478</div>
479
480<p>and clients can call it using any one of:</p>
481
482<div class="doc_code">
483<pre>
484 Map.find("foo"); <i>// Lookup "foo"</i>
485 Map.find(std::string("bar")); <i>// Lookup "bar"</i>
486 Map.find(StringRef("\0baz", 4)); <i>// Lookup "\0baz"</i>
487</pre>
488</div>
489
490<p>Similarly, APIs which need to return a string may return a <tt>StringRef</tt>
491instance, which can be used directly or converted to an <tt>std::string</tt>
492using the <tt>str</tt> member function. See
493"<tt><a href="/doxygen/classllvm_1_1StringRef_8h-source.html">llvm/ADT/StringRef.h</a></tt>"
494for more information.</p>
495
496<p>You should rarely use the <tt>StringRef</tt> class directly, because it contains
497pointers to external memory it is not generally safe to store an instance of the
Chris Lattner81187ae2009-07-25 07:16:59 +0000[diff] [blame]498class (unless you know that the external storage will not be freed).</p>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]499
500</div>
501
502<!-- _______________________________________________________________________ -->
503<div class="doc_subsubsection">
504 <a name="Twine">The <tt>Twine</tt> class</a>
505</div>
506
507<div class="doc_text">
508
509<p>The <tt>Twine</tt> class is an efficient way for APIs to accept concatenated
510strings. For example, a common LLVM paradigm is to name one instruction based on
511the name of another instruction with a suffix, for example:</p>
512
513<div class="doc_code">
514<pre>
515 New = CmpInst::Create(<i>...</i>, SO->getName() + ".cmp");
516</pre>
517</div>
518
519<p>The <tt>Twine</tt> class is effectively a
520lightweight <a href="http://en.wikipedia.org/wiki/Rope_(computer_science)">rope</a>
521which points to temporary (stack allocated) objects. Twines can be implicitly
522constructed as the result of the plus operator applied to strings (i.e., a C
523strings, an <tt>std::string</tt>, or a <tt>StringRef</tt>). The twine delays the
524actual concatentation of strings until it is actually required, at which point
525it can be efficiently rendered directly into a character array. This avoids
526unnecessary heap allocation involved in constructing the temporary results of
527string concatenation. See
528"<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]529for more information.</p>
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]530
531<p>As with a <tt>StringRef</tt>, <tt>Twine</tt> objects point to external memory
532and should almost never be stored or mentioned directly. They are intended
533solely for use when defining a function which should be able to efficiently
534accept concatenated strings.</p>
535
536</div>
537
538
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]539<!-- ======================================================================= -->
540<div class="doc_subsection">
Misha Brukman2c122ce2005-11-01 21:12:49 +0000[diff] [blame]541 <a name="DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt> option</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]542</div>
543
544<div class="doc_text">
545
546<p>Often when working on your pass you will put a bunch of debugging printouts
547and other code into your pass. After you get it working, you want to remove
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]548it, 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]549across).</p>
550
551<p> Naturally, because of this, you don't want to delete the debug printouts,
552but you don't want them to always be noisy. A standard compromise is to comment
553them out, allowing you to enable them if you need them in the future.</p>
554
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]555<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]556file provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to
557this problem. Basically, you can put arbitrary code into the argument of the
558<tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
559tool) is run with the '<tt>-debug</tt>' command line argument:</p>
560
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]561<div class="doc_code">
562<pre>
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]563DEBUG(errs() &lt;&lt; "I am here!\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]564</pre>
565</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]566
567<p>Then you can run your pass like this:</p>
568
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]569<div class="doc_code">
570<pre>
571$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]572<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]573$ opt &lt; a.bc &gt; /dev/null -mypass -debug
574I am here!
575</pre>
576</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]577
578<p>Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you
579to not have to create "yet another" command line option for the debug output for
580your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
581so they do not cause a performance impact at all (for the same reason, they
582should also not contain side-effects!).</p>
583
584<p>One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
585enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
586"<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
587program hasn't been started yet, you can always just run it with
588<tt>-debug</tt>.</p>
589
590</div>
591
592<!-- _______________________________________________________________________ -->
593<div class="doc_subsubsection">
Chris Lattnerc9151082005-04-26 22:57:07 +0000[diff] [blame]594 <a name="DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt> and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]595 the <tt>-debug-only</tt> option</a>
596</div>
597
598<div class="doc_text">
599
600<p>Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
601just turns on <b>too much</b> information (such as when working on the code
602generator). If you want to enable debug information with more fine-grained
603control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
604option as follows:</p>
605
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]606<div class="doc_code">
607<pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]608#undef DEBUG_TYPE
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]609DEBUG(errs() &lt;&lt; "No debug type\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]610#define DEBUG_TYPE "foo"
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]611DEBUG(errs() &lt;&lt; "'foo' debug type\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]612#undef DEBUG_TYPE
613#define DEBUG_TYPE "bar"
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]614DEBUG(errs() &lt;&lt; "'bar' debug type\n"));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]615#undef DEBUG_TYPE
616#define DEBUG_TYPE ""
Daniel Dunbar06388ae2009-07-25 01:55:32 +0000[diff] [blame]617DEBUG(errs() &lt;&lt; "No debug type (2)\n");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]618</pre>
619</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]620
621<p>Then you can run your pass like this:</p>
622
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]623<div class="doc_code">
624<pre>
625$ opt &lt; a.bc &gt; /dev/null -mypass
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]626<i>&lt;no output&gt;</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]627$ opt &lt; a.bc &gt; /dev/null -mypass -debug
628No debug type
629'foo' debug type
630'bar' debug type
631No debug type (2)
632$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=foo
633'foo' debug type
634$ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=bar
635'bar' debug type
636</pre>
637</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]638
639<p>Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of
640a file, to specify the debug type for the entire module (if you do this before
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]641you <tt>#include "llvm/Support/Debug.h"</tt>, you don't have to insert the ugly
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]642<tt>#undef</tt>'s). Also, you should use names more meaningful than "foo" and
643"bar", because there is no system in place to ensure that names do not
644conflict. If two different modules use the same string, they will all be turned
645on when the name is specified. This allows, for example, all debug information
646for instruction scheduling to be enabled with <tt>-debug-type=InstrSched</tt>,
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]647even if the source lives in multiple files.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]648
Daniel Dunbarc3c92392009-08-07 23:48:59 +0000[diff] [blame]649<p>The <tt>DEBUG_WITH_TYPE</tt> macro is also available for situations where you
650would like to set <tt>DEBUG_TYPE</tt>, but only for one specific <tt>DEBUG</tt>
651statement. It takes an additional first parameter, which is the type to use. For
652example, the preceeding example could be written as:</p>
653
654
655<div class="doc_code">
656<pre>
657DEBUG_WITH_TYPE("", errs() &lt;&lt; "No debug type\n");
658DEBUG_WITH_TYPE("foo", errs() &lt;&lt; "'foo' debug type\n");
659DEBUG_WITH_TYPE("bar", errs() &lt;&lt; "'bar' debug type\n"));
660DEBUG_WITH_TYPE("", errs() &lt;&lt; "No debug type (2)\n");
661</pre>
662</div>
663
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]664</div>
665
666<!-- ======================================================================= -->
667<div class="doc_subsection">
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]668 <a name="Statistic">The <tt>Statistic</tt> class &amp; <tt>-stats</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]669 option</a>
670</div>
671
672<div class="doc_text">
673
674<p>The "<tt><a
Chris Lattner695b78b2005-04-26 22:56:16 +0000[diff] [blame]675href="/doxygen/Statistic_8h-source.html">llvm/ADT/Statistic.h</a></tt>" file
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]676provides a class named <tt>Statistic</tt> that is used as a unified way to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]677keep track of what the LLVM compiler is doing and how effective various
678optimizations are. It is useful to see what optimizations are contributing to
679making a particular program run faster.</p>
680
681<p>Often you may run your pass on some big program, and you're interested to see
682how many times it makes a certain transformation. Although you can do this with
683hand 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]684for big programs. Using the <tt>Statistic</tt> class makes it very easy to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]685keep track of this information, and the calculated information is presented in a
686uniform manner with the rest of the passes being executed.</p>
687
688<p>There are many examples of <tt>Statistic</tt> uses, but the basics of using
689it are as follows:</p>
690
691<ol>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]692 <li><p>Define your statistic like this:</p>
693
694<div class="doc_code">
695<pre>
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]696#define <a href="#DEBUG_TYPE">DEBUG_TYPE</a> "mypassname" <i>// This goes before any #includes.</i>
697STATISTIC(NumXForms, "The # of times I did stuff");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]698</pre>
699</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]700
Chris Lattner0be6fdf2006-12-19 21:46:21 +0000[diff] [blame]701 <p>The <tt>STATISTIC</tt> macro defines a static variable, whose name is
702 specified by the first argument. The pass name is taken from the DEBUG_TYPE
703 macro, and the description is taken from the second argument. The variable
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]704 defined ("NumXForms" in this case) acts like an unsigned integer.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]705
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]706 <li><p>Whenever you make a transformation, bump the counter:</p>
707
708<div class="doc_code">
709<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]710++NumXForms; // <i>I did stuff!</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]711</pre>
712</div>
713
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]714 </li>
715 </ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]716
717 <p>That's all you have to do. To get '<tt>opt</tt>' to print out the
718 statistics gathered, use the '<tt>-stats</tt>' option:</p>
719
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]720<div class="doc_code">
721<pre>
722$ opt -stats -mypassname &lt; program.bc &gt; /dev/null
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]723<i>... statistics output ...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]724</pre>
725</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]726
Reid Spencer6b6c73e2007-02-09 16:00:28 +0000[diff] [blame]727 <p> When running <tt>opt</tt> on a C file from the SPEC benchmark
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]728suite, it gives a report that looks like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]729
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]730<div class="doc_code">
731<pre>
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]732 7646 bitcodewriter - Number of normal instructions
733 725 bitcodewriter - Number of oversized instructions
734 129996 bitcodewriter - Number of bitcode bytes written
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]735 2817 raise - Number of insts DCEd or constprop'd
736 3213 raise - Number of cast-of-self removed
737 5046 raise - Number of expression trees converted
738 75 raise - Number of other getelementptr's formed
739 138 raise - Number of load/store peepholes
740 42 deadtypeelim - Number of unused typenames removed from symtab
741 392 funcresolve - Number of varargs functions resolved
742 27 globaldce - Number of global variables removed
743 2 adce - Number of basic blocks removed
744 134 cee - Number of branches revectored
745 49 cee - Number of setcc instruction eliminated
746 532 gcse - Number of loads removed
747 2919 gcse - Number of instructions removed
748 86 indvars - Number of canonical indvars added
749 87 indvars - Number of aux indvars removed
750 25 instcombine - Number of dead inst eliminate
751 434 instcombine - Number of insts combined
752 248 licm - Number of load insts hoisted
753 1298 licm - Number of insts hoisted to a loop pre-header
754 3 licm - Number of insts hoisted to multiple loop preds (bad, no loop pre-header)
755 75 mem2reg - Number of alloca's promoted
756 1444 cfgsimplify - Number of blocks simplified
757</pre>
758</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]759
760<p>Obviously, with so many optimizations, having a unified framework for this
761stuff is very nice. Making your pass fit well into the framework makes it more
762maintainable and useful.</p>
763
764</div>
765
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]766<!-- ======================================================================= -->
767<div class="doc_subsection">
768 <a name="ViewGraph">Viewing graphs while debugging code</a>
769</div>
770
771<div class="doc_text">
772
773<p>Several of the important data structures in LLVM are graphs: for example
774CFGs made out of LLVM <a href="#BasicBlock">BasicBlock</a>s, CFGs made out of
775LLVM <a href="CodeGenerator.html#machinebasicblock">MachineBasicBlock</a>s, and
776<a href="CodeGenerator.html#selectiondag_intro">Instruction Selection
777DAGs</a>. In many cases, while debugging various parts of the compiler, it is
778nice to instantly visualize these graphs.</p>
779
780<p>LLVM provides several callbacks that are available in a debug build to do
781exactly that. If you call the <tt>Function::viewCFG()</tt> method, for example,
782the current LLVM tool will pop up a window containing the CFG for the function
783where each basic block is a node in the graph, and each node contains the
784instructions in the block. Similarly, there also exists
785<tt>Function::viewCFGOnly()</tt> (does not include the instructions), the
786<tt>MachineFunction::viewCFG()</tt> and <tt>MachineFunction::viewCFGOnly()</tt>,
787and the <tt>SelectionDAG::viewGraph()</tt> methods. Within GDB, for example,
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]788you can usually use something like <tt>call DAG.viewGraph()</tt> to pop
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]789up a window. Alternatively, you can sprinkle calls to these functions in your
790code in places you want to debug.</p>
791
792<p>Getting this to work requires a small amount of configuration. On Unix
793systems with X11, install the <a href="http://www.graphviz.org">graphviz</a>
794toolkit, and make sure 'dot' and 'gv' are in your path. If you are running on
795Mac OS/X, download and install the Mac OS/X <a
796href="http://www.pixelglow.com/graphviz/">Graphviz program</a>, and add
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]797<tt>/Applications/Graphviz.app/Contents/MacOS/</tt> (or wherever you install
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]798it) to your path. Once in your system and path are set up, rerun the LLVM
799configure script and rebuild LLVM to enable this functionality.</p>
800
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]801<p><tt>SelectionDAG</tt> has been extended to make it easier to locate
802<i>interesting</i> nodes in large complex graphs. From gdb, if you
803<tt>call DAG.setGraphColor(<i>node</i>, "<i>color</i>")</tt>, then the
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]804next <tt>call DAG.viewGraph()</tt> would highlight the node in the
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]805specified color (choices of colors can be found at <a
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]806href="http://www.graphviz.org/doc/info/colors.html">colors</a>.) More
Jim Laskey543a0ee2006-10-02 12:28:07 +0000[diff] [blame]807complex node attributes can be provided with <tt>call
808DAG.setGraphAttrs(<i>node</i>, "<i>attributes</i>")</tt> (choices can be
809found at <a href="http://www.graphviz.org/doc/info/attrs.html">Graph
810Attributes</a>.) If you want to restart and clear all the current graph
811attributes, then you can <tt>call DAG.clearGraphAttrs()</tt>. </p>
812
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]813</div>
814
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]815<!-- *********************************************************************** -->
816<div class="doc_section">
817 <a name="datastructure">Picking the Right Data Structure for a Task</a>
818</div>
819<!-- *********************************************************************** -->
820
821<div class="doc_text">
822
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]823<p>LLVM has a plethora of data structures in the <tt>llvm/ADT/</tt> directory,
824 and we commonly use STL data structures. This section describes the trade-offs
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]825 you should consider when you pick one.</p>
826
827<p>
828The first step is a choose your own adventure: do you want a sequential
829container, a set-like container, or a map-like container? The most important
830thing when choosing a container is the algorithmic properties of how you plan to
831access the container. Based on that, you should use:</p>
832
833<ul>
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]834<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]835 of an value based on another value. Map-like containers also support
836 efficient queries for containment (whether a key is in the map). Map-like
837 containers generally do not support efficient reverse mapping (values to
838 keys). If you need that, use two maps. Some map-like containers also
839 support efficient iteration through the keys in sorted order. Map-like
840 containers are the most expensive sort, only use them if you need one of
841 these capabilities.</li>
842
843<li>a <a href="#ds_set">set-like</a> container if you need to put a bunch of
844 stuff into a container that automatically eliminates duplicates. Some
845 set-like containers support efficient iteration through the elements in
846 sorted order. Set-like containers are more expensive than sequential
847 containers.
848</li>
849
850<li>a <a href="#ds_sequential">sequential</a> container provides
851 the most efficient way to add elements and keeps track of the order they are
852 added to the collection. They permit duplicates and support efficient
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]853 iteration, but do not support efficient look-up based on a key.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]854</li>
855
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]856<li>a <a href="#ds_string">string</a> container is a specialized sequential
857 container or reference structure that is used for character or byte
858 arrays.</li>
859
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]860<li>a <a href="#ds_bit">bit</a> container provides an efficient way to store and
861 perform set operations on sets of numeric id's, while automatically
862 eliminating duplicates. Bit containers require a maximum of 1 bit for each
863 identifier you want to store.
864</li>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]865</ul>
866
867<p>
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]868Once the proper category of container is determined, you can fine tune the
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]869memory use, constant factors, and cache behaviors of access by intelligently
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]870picking a member of the category. Note that constant factors and cache behavior
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]871can be a big deal. If you have a vector that usually only contains a few
872elements (but could contain many), for example, it's much better to use
873<a href="#dss_smallvector">SmallVector</a> than <a href="#dss_vector">vector</a>
874. Doing so avoids (relatively) expensive malloc/free calls, which dwarf the
875cost of adding the elements to the container. </p>
876
877</div>
878
879<!-- ======================================================================= -->
880<div class="doc_subsection">
881 <a name="ds_sequential">Sequential Containers (std::vector, std::list, etc)</a>
882</div>
883
884<div class="doc_text">
885There are a variety of sequential containers available for you, based on your
886needs. Pick the first in this section that will do what you want.
887</div>
888
889<!-- _______________________________________________________________________ -->
890<div class="doc_subsubsection">
891 <a name="dss_fixedarrays">Fixed Size Arrays</a>
892</div>
893
894<div class="doc_text">
895<p>Fixed size arrays are very simple and very fast. They are good if you know
896exactly how many elements you have, or you have a (low) upper bound on how many
897you have.</p>
898</div>
899
900<!-- _______________________________________________________________________ -->
901<div class="doc_subsubsection">
902 <a name="dss_heaparrays">Heap Allocated Arrays</a>
903</div>
904
905<div class="doc_text">
906<p>Heap allocated arrays (new[] + delete[]) are also simple. They are good if
907the number of elements is variable, if you know how many elements you will need
908before the array is allocated, and if the array is usually large (if not,
909consider a <a href="#dss_smallvector">SmallVector</a>). The cost of a heap
910allocated array is the cost of the new/delete (aka malloc/free). Also note that
911if you are allocating an array of a type with a constructor, the constructor and
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]912destructors will be run for every element in the array (re-sizable vectors only
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]913construct those elements actually used).</p>
914</div>
915
916<!-- _______________________________________________________________________ -->
917<div class="doc_subsubsection">
918 <a name="dss_smallvector">"llvm/ADT/SmallVector.h"</a>
919</div>
920
921<div class="doc_text">
922<p><tt>SmallVector&lt;Type, N&gt;</tt> is a simple class that looks and smells
923just like <tt>vector&lt;Type&gt;</tt>:
924it supports efficient iteration, lays out elements in memory order (so you can
925do pointer arithmetic between elements), supports efficient push_back/pop_back
926operations, supports efficient random access to its elements, etc.</p>
927
928<p>The advantage of SmallVector is that it allocates space for
929some number of elements (N) <b>in the object itself</b>. Because of this, if
930the SmallVector is dynamically smaller than N, no malloc is performed. This can
931be a big win in cases where the malloc/free call is far more expensive than the
932code that fiddles around with the elements.</p>
933
934<p>This is good for vectors that are "usually small" (e.g. the number of
935predecessors/successors of a block is usually less than 8). On the other hand,
936this makes the size of the SmallVector itself large, so you don't want to
937allocate lots of them (doing so will waste a lot of space). As such,
938SmallVectors are most useful when on the stack.</p>
939
940<p>SmallVector also provides a nice portable and efficient replacement for
941<tt>alloca</tt>.</p>
942
943</div>
944
945<!-- _______________________________________________________________________ -->
946<div class="doc_subsubsection">
947 <a name="dss_vector">&lt;vector&gt;</a>
948</div>
949
950<div class="doc_text">
951<p>
952std::vector is well loved and respected. It is useful when SmallVector isn't:
953when the size of the vector is often large (thus the small optimization will
954rarely be a benefit) or if you will be allocating many instances of the vector
955itself (which would waste space for elements that aren't in the container).
956vector is also useful when interfacing with code that expects vectors :).
957</p>
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]958
959<p>One worthwhile note about std::vector: avoid code like this:</p>
960
961<div class="doc_code">
962<pre>
963for ( ... ) {
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]964 std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]965 use V;
966}
967</pre>
968</div>
969
970<p>Instead, write this as:</p>
971
972<div class="doc_code">
973<pre>
Chris Lattner9bb3dbb2007-03-28 18:27:57 +0000[diff] [blame]974std::vector&lt;foo&gt; V;
Chris Lattner32d84762007-02-05 06:30:51 +0000[diff] [blame]975for ( ... ) {
976 use V;
977 V.clear();
978}
979</pre>
980</div>
981
982<p>Doing so will save (at least) one heap allocation and free per iteration of
983the loop.</p>
984
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]985</div>
986
987<!-- _______________________________________________________________________ -->
988<div class="doc_subsubsection">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]989 <a name="dss_deque">&lt;deque&gt;</a>
990</div>
991
992<div class="doc_text">
993<p>std::deque is, in some senses, a generalized version of std::vector. Like
994std::vector, it provides constant time random access and other similar
995properties, but it also provides efficient access to the front of the list. It
996does not guarantee continuity of elements within memory.</p>
997
998<p>In exchange for this extra flexibility, std::deque has significantly higher
999constant factor costs than std::vector. If possible, use std::vector or
1000something cheaper.</p>
1001</div>
1002
1003<!-- _______________________________________________________________________ -->
1004<div class="doc_subsubsection">
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1005 <a name="dss_list">&lt;list&gt;</a>
1006</div>
1007
1008<div class="doc_text">
1009<p>std::list is an extremely inefficient class that is rarely useful.
1010It performs a heap allocation for every element inserted into it, thus having an
1011extremely high constant factor, particularly for small data types. std::list
1012also only supports bidirectional iteration, not random access iteration.</p>
1013
1014<p>In exchange for this high cost, std::list supports efficient access to both
1015ends of the list (like std::deque, but unlike std::vector or SmallVector). In
1016addition, the iterator invalidation characteristics of std::list are stronger
1017than that of a vector class: inserting or removing an element into the list does
1018not invalidate iterator or pointers to other elements in the list.</p>
1019</div>
1020
1021<!-- _______________________________________________________________________ -->
1022<div class="doc_subsubsection">
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1023 <a name="dss_ilist">llvm/ADT/ilist.h</a>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1024</div>
1025
1026<div class="doc_text">
1027<p><tt>ilist&lt;T&gt;</tt> implements an 'intrusive' doubly-linked list. It is
1028intrusive, because it requires the element to store and provide access to the
1029prev/next pointers for the list.</p>
1030
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1031<p><tt>ilist</tt> has the same drawbacks as <tt>std::list</tt>, and additionally
1032requires an <tt>ilist_traits</tt> implementation for the element type, but it
1033provides some novel characteristics. In particular, it can efficiently store
1034polymorphic objects, the traits class is informed when an element is inserted or
Gabor Greif0cbcabe2009-03-12 09:47:03 +0000[diff] [blame]1035removed from the list, and <tt>ilist</tt>s are guaranteed to support a
1036constant-time splice operation.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1037
Gabor Greif0cbcabe2009-03-12 09:47:03 +0000[diff] [blame]1038<p>These properties are exactly what we want for things like
1039<tt>Instruction</tt>s and basic blocks, which is why these are implemented with
1040<tt>ilist</tt>s.</p>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1041
1042Related classes of interest are explained in the following subsections:
1043 <ul>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1044 <li><a href="#dss_ilist_traits">ilist_traits</a></li>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1045 <li><a href="#dss_iplist">iplist</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1046 <li><a href="#dss_ilist_node">llvm/ADT/ilist_node.h</a></li>
Gabor Greif6a65f422009-03-12 10:30:31 +0000[diff] [blame]1047 <li><a href="#dss_ilist_sentinel">Sentinels</a></li>
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1048 </ul>
1049</div>
1050
1051<!-- _______________________________________________________________________ -->
1052<div class="doc_subsubsection">
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1053 <a name="dss_ilist_traits">ilist_traits</a>
1054</div>
1055
1056<div class="doc_text">
1057<p><tt>ilist_traits&lt;T&gt;</tt> is <tt>ilist&lt;T&gt;</tt>'s customization
1058mechanism. <tt>iplist&lt;T&gt;</tt> (and consequently <tt>ilist&lt;T&gt;</tt>)
1059publicly derive from this traits class.</p>
1060</div>
1061
1062<!-- _______________________________________________________________________ -->
1063<div class="doc_subsubsection">
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1064 <a name="dss_iplist">iplist</a>
1065</div>
1066
1067<div class="doc_text">
1068<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]1069supports a slightly narrower interface. Notably, inserters from
1070<tt>T&amp;</tt> are absent.</p>
Gabor Greif01862502009-02-27 13:28:07 +0000[diff] [blame]1071
1072<p><tt>ilist_traits&lt;T&gt;</tt> is a public base of this class and can be
1073used for a wide variety of customizations.</p>
Gabor Greif2946d1c2009-02-27 12:02:19 +0000[diff] [blame]1074</div>
1075
1076<!-- _______________________________________________________________________ -->
1077<div class="doc_subsubsection">
Gabor Greif3899e492009-02-27 11:37:41 +0000[diff] [blame]1078 <a name="dss_ilist_node">llvm/ADT/ilist_node.h</a>
1079</div>
1080
1081<div class="doc_text">
1082<p><tt>ilist_node&lt;T&gt;</tt> implements a the forward and backward links
1083that are expected by the <tt>ilist&lt;T&gt;</tt> (and analogous containers)
1084in the default manner.</p>
1085
1086<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]1087<tt>T</tt>, usually <tt>T</tt> publicly derives from
1088<tt>ilist_node&lt;T&gt;</tt>.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1089</div>
1090
1091<!-- _______________________________________________________________________ -->
1092<div class="doc_subsubsection">
Gabor Greif6a65f422009-03-12 10:30:31 +0000[diff] [blame]1093 <a name="dss_ilist_sentinel">Sentinels</a>
1094</div>
1095
1096<div class="doc_text">
1097<p><tt>ilist</tt>s have another speciality that must be considered. To be a good
1098citizen in the C++ ecosystem, it needs to support the standard container
1099operations, such as <tt>begin</tt> and <tt>end</tt> iterators, etc. Also, the
1100<tt>operator--</tt> must work correctly on the <tt>end</tt> iterator in the
1101case of non-empty <tt>ilist</tt>s.</p>
1102
1103<p>The only sensible solution to this problem is to allocate a so-called
1104<i>sentinel</i> along with the intrusive list, which serves as the <tt>end</tt>
1105iterator, providing the back-link to the last element. However conforming to the
1106C++ convention it is illegal to <tt>operator++</tt> beyond the sentinel and it
1107also must not be dereferenced.</p>
1108
1109<p>These constraints allow for some implementation freedom to the <tt>ilist</tt>
1110how to allocate and store the sentinel. The corresponding policy is dictated
1111by <tt>ilist_traits&lt;T&gt;</tt>. By default a <tt>T</tt> gets heap-allocated
1112whenever the need for a sentinel arises.</p>
1113
1114<p>While the default policy is sufficient in most cases, it may break down when
1115<tt>T</tt> does not provide a default constructor. Also, in the case of many
1116instances of <tt>ilist</tt>s, the memory overhead of the associated sentinels
1117is wasted. To alleviate the situation with numerous and voluminous
1118<tt>T</tt>-sentinels, sometimes a trick is employed, leading to <i>ghostly
1119sentinels</i>.</p>
1120
1121<p>Ghostly sentinels are obtained by specially-crafted <tt>ilist_traits&lt;T&gt;</tt>
1122which superpose the sentinel with the <tt>ilist</tt> instance in memory. Pointer
1123arithmetic is used to obtain the sentinel, which is relative to the
1124<tt>ilist</tt>'s <tt>this</tt> pointer. The <tt>ilist</tt> is augmented by an
1125extra pointer, which serves as the back-link of the sentinel. This is the only
1126field in the ghostly sentinel which can be legally accessed.</p>
1127</div>
1128
1129<!-- _______________________________________________________________________ -->
1130<div class="doc_subsubsection">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1131 <a name="dss_other">Other Sequential Container options</a>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1132</div>
1133
1134<div class="doc_text">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1135<p>Other STL containers are available, such as std::string.</p>
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1136
1137<p>There are also various STL adapter classes such as std::queue,
1138std::priority_queue, std::stack, etc. These provide simplified access to an
1139underlying container but don't affect the cost of the container itself.</p>
1140
1141</div>
1142
1143
1144<!-- ======================================================================= -->
1145<div class="doc_subsection">
1146 <a name="ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
1147</div>
1148
1149<div class="doc_text">
1150
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1151<p>Set-like containers are useful when you need to canonicalize multiple values
1152into a single representation. There are several different choices for how to do
1153this, providing various trade-offs.</p>
1154
1155</div>
1156
1157
1158<!-- _______________________________________________________________________ -->
1159<div class="doc_subsubsection">
1160 <a name="dss_sortedvectorset">A sorted 'vector'</a>
1161</div>
1162
1163<div class="doc_text">
1164
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]1165<p>If you intend to insert a lot of elements, then do a lot of queries, a
1166great approach is to use a vector (or other sequential container) with
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1167std::sort+std::unique to remove duplicates. This approach works really well if
Chris Lattner3b23a8c2007-02-03 08:10:45 +0000[diff] [blame]1168your usage pattern has these two distinct phases (insert then query), and can be
1169coupled with a good choice of <a href="#ds_sequential">sequential container</a>.
1170</p>
1171
1172<p>
1173This combination provides the several nice properties: the result data is
1174contiguous in memory (good for cache locality), has few allocations, is easy to
1175address (iterators in the final vector are just indices or pointers), and can be
1176efficiently queried with a standard binary or radix search.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1177
1178</div>
1179
1180<!-- _______________________________________________________________________ -->
1181<div class="doc_subsubsection">
1182 <a name="dss_smallset">"llvm/ADT/SmallSet.h"</a>
1183</div>
1184
1185<div class="doc_text">
1186
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1187<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]1188are 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]1189has space for N elements in place (thus, if the set is dynamically smaller than
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1190N, no malloc traffic is required) and accesses them with a simple linear search.
1191When the set grows beyond 'N' elements, it allocates a more expensive representation that
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1192guarantees efficient access (for most types, it falls back to std::set, but for
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1193pointers it uses something far better, <a
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1194href="#dss_smallptrset">SmallPtrSet</a>).</p>
1195
1196<p>The magic of this class is that it handles small sets extremely efficiently,
1197but gracefully handles extremely large sets without loss of efficiency. The
1198drawback is that the interface is quite small: it supports insertion, queries
1199and erasing, but does not support iteration.</p>
1200
1201</div>
1202
1203<!-- _______________________________________________________________________ -->
1204<div class="doc_subsubsection">
1205 <a name="dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a>
1206</div>
1207
1208<div class="doc_text">
1209
1210<p>SmallPtrSet has all the advantages of SmallSet (and a SmallSet of pointers is
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1211transparently implemented with a SmallPtrSet), but also supports iterators. If
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1212more than 'N' insertions are performed, a single quadratically
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1213probed hash table is allocated and grows as needed, providing extremely
1214efficient access (constant time insertion/deleting/queries with low constant
1215factors) and is very stingy with malloc traffic.</p>
1216
1217<p>Note that, unlike std::set, the iterators of SmallPtrSet are invalidated
1218whenever an insertion occurs. Also, the values visited by the iterators are not
1219visited in sorted order.</p>
1220
1221</div>
1222
1223<!-- _______________________________________________________________________ -->
1224<div class="doc_subsubsection">
Chris Lattnerc28476f2007-09-30 00:58:59 +0000[diff] [blame]1225 <a name="dss_denseset">"llvm/ADT/DenseSet.h"</a>
1226</div>
1227
1228<div class="doc_text">
1229
1230<p>
1231DenseSet is a simple quadratically probed hash table. It excels at supporting
1232small values: it uses a single allocation to hold all of the pairs that
1233are currently inserted in the set. DenseSet is a great way to unique small
1234values that are not simple pointers (use <a
1235href="#dss_smallptrset">SmallPtrSet</a> for pointers). Note that DenseSet has
1236the same requirements for the value type that <a
1237href="#dss_densemap">DenseMap</a> has.
1238</p>
1239
1240</div>
1241
1242<!-- _______________________________________________________________________ -->
1243<div class="doc_subsubsection">
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1244 <a name="dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a>
1245</div>
1246
1247<div class="doc_text">
1248
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1249<p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1250FoldingSet is an aggregate class that is really good at uniquing
1251expensive-to-create or polymorphic objects. It is a combination of a chained
1252hash table with intrusive links (uniqued objects are required to inherit from
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1253FoldingSetNode) that uses <a href="#dss_smallvector">SmallVector</a> as part of
1254its ID process.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1255
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1256<p>Consider a case where you want to implement a "getOrCreateFoo" method for
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1257a complex object (for example, a node in the code generator). The client has a
1258description of *what* it wants to generate (it knows the opcode and all the
1259operands), 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]1260only to find out it already exists, at which point we would have to delete it
1261and return the node that already exists.
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1262</p>
1263
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1264<p>To support this style of client, FoldingSet perform a query with a
1265FoldingSetNodeID (which wraps SmallVector) that can be used to describe the
1266element that we want to query for. The query either returns the element
1267matching the ID or it returns an opaque ID that indicates where insertion should
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1268take place. Construction of the ID usually does not require heap traffic.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1269
1270<p>Because FoldingSet uses intrusive links, it can support polymorphic objects
1271in the set (for example, you can have SDNode instances mixed with LoadSDNodes).
1272Because the elements are individually allocated, pointers to the elements are
1273stable: inserting or removing elements does not invalidate any pointers to other
1274elements.
1275</p>
1276
1277</div>
1278
1279<!-- _______________________________________________________________________ -->
1280<div class="doc_subsubsection">
1281 <a name="dss_set">&lt;set&gt;</a>
1282</div>
1283
1284<div class="doc_text">
1285
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1286<p><tt>std::set</tt> is a reasonable all-around set class, which is decent at
1287many things but great at nothing. std::set allocates memory for each element
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1288inserted (thus it is very malloc intensive) and typically stores three pointers
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1289per element in the set (thus adding a large amount of per-element space
1290overhead). It offers guaranteed log(n) performance, which is not particularly
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1291fast from a complexity standpoint (particularly if the elements of the set are
1292expensive to compare, like strings), and has extremely high constant factors for
1293lookup, insertion and removal.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1294
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1295<p>The advantages of std::set are that its iterators are stable (deleting or
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1296inserting an element from the set does not affect iterators or pointers to other
1297elements) and that iteration over the set is guaranteed to be in sorted order.
1298If the elements in the set are large, then the relative overhead of the pointers
1299and malloc traffic is not a big deal, but if the elements of the set are small,
1300std::set is almost never a good choice.</p>
1301
1302</div>
1303
1304<!-- _______________________________________________________________________ -->
1305<div class="doc_subsubsection">
1306 <a name="dss_setvector">"llvm/ADT/SetVector.h"</a>
1307</div>
1308
1309<div class="doc_text">
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1310<p>LLVM's SetVector&lt;Type&gt; is an adapter class that combines your choice of
1311a set-like container along with a <a href="#ds_sequential">Sequential
1312Container</a>. The important property
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1313that this provides is efficient insertion with uniquing (duplicate elements are
1314ignored) with iteration support. It implements this by inserting elements into
1315both a set-like container and the sequential container, using the set-like
1316container for uniquing and the sequential container for iteration.
1317</p>
1318
1319<p>The difference between SetVector and other sets is that the order of
1320iteration is guaranteed to match the order of insertion into the SetVector.
1321This property is really important for things like sets of pointers. Because
1322pointer values are non-deterministic (e.g. vary across runs of the program on
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1323different machines), iterating over the pointers in the set will
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1324not be in a well-defined order.</p>
1325
1326<p>
1327The drawback of SetVector is that it requires twice as much space as a normal
1328set and has the sum of constant factors from the set-like container and the
1329sequential container that it uses. Use it *only* if you need to iterate over
1330the elements in a deterministic order. SetVector is also expensive to delete
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1331elements out of (linear time), unless you use it's "pop_back" method, which is
1332faster.
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1333</p>
1334
Chris Lattneredca3c52007-02-04 00:00:26 +0000[diff] [blame]1335<p>SetVector is an adapter class that defaults to using std::vector and std::set
1336for the underlying containers, so it is quite expensive. However,
1337<tt>"llvm/ADT/SetVector.h"</tt> also provides a SmallSetVector class, which
1338defaults to using a SmallVector and SmallSet of a specified size. If you use
1339this, and if your sets are dynamically smaller than N, you will save a lot of
1340heap traffic.</p>
1341
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1342</div>
1343
1344<!-- _______________________________________________________________________ -->
1345<div class="doc_subsubsection">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1346 <a name="dss_uniquevector">"llvm/ADT/UniqueVector.h"</a>
1347</div>
1348
1349<div class="doc_text">
1350
1351<p>
1352UniqueVector is similar to <a href="#dss_setvector">SetVector</a>, but it
1353retains a unique ID for each element inserted into the set. It internally
1354contains a map and a vector, and it assigns a unique ID for each value inserted
1355into the set.</p>
1356
1357<p>UniqueVector is very expensive: its cost is the sum of the cost of
1358maintaining both the map and vector, it has high complexity, high constant
1359factors, and produces a lot of malloc traffic. It should be avoided.</p>
1360
1361</div>
1362
1363
1364<!-- _______________________________________________________________________ -->
1365<div class="doc_subsubsection">
1366 <a name="dss_otherset">Other Set-Like Container Options</a>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1367</div>
1368
1369<div class="doc_text">
1370
1371<p>
1372The STL provides several other options, such as std::multiset and the various
Chris Lattnerf1a30822009-03-09 05:20:45 +0000[diff] [blame]1373"hash_set" like containers (whether from C++ TR1 or from the SGI library). We
1374never use hash_set and unordered_set because they are generally very expensive
1375(each insertion requires a malloc) and very non-portable.
1376</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1377
1378<p>std::multiset is useful if you're not interested in elimination of
Chris Lattner14868db2007-02-03 08:20:15 +0000[diff] [blame]1379duplicates, but has all the drawbacks of std::set. A sorted vector (where you
1380don't delete duplicate entries) or some other approach is almost always
1381better.</p>
Chris Lattner74c4ca12007-02-03 07:59:07 +0000[diff] [blame]1382
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1383</div>
1384
1385<!-- ======================================================================= -->
1386<div class="doc_subsection">
1387 <a name="ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
1388</div>
1389
1390<div class="doc_text">
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1391Map-like containers are useful when you want to associate data to a key. As
1392usual, there are a lot of different ways to do this. :)
1393</div>
1394
1395<!-- _______________________________________________________________________ -->
1396<div class="doc_subsubsection">
1397 <a name="dss_sortedvectormap">A sorted 'vector'</a>
1398</div>
1399
1400<div class="doc_text">
1401
1402<p>
1403If your usage pattern follows a strict insert-then-query approach, you can
1404trivially use the same approach as <a href="#dss_sortedvectorset">sorted vectors
1405for set-like containers</a>. The only difference is that your query function
1406(which uses std::lower_bound to get efficient log(n) lookup) should only compare
1407the key, not both the key and value. This yields the same advantages as sorted
1408vectors for sets.
1409</p>
1410</div>
1411
1412<!-- _______________________________________________________________________ -->
1413<div class="doc_subsubsection">
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1414 <a name="dss_stringmap">"llvm/ADT/StringMap.h"</a>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1415</div>
1416
1417<div class="doc_text">
1418
1419<p>
1420Strings are commonly used as keys in maps, and they are difficult to support
1421efficiently: they are variable length, inefficient to hash and compare when
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1422long, expensive to copy, etc. StringMap is a specialized container designed to
1423cope with these issues. It supports mapping an arbitrary range of bytes to an
1424arbitrary other object.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1425
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1426<p>The StringMap implementation uses a quadratically-probed hash table, where
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1427the buckets store a pointer to the heap allocated entries (and some other
1428stuff). The entries in the map must be heap allocated because the strings are
1429variable length. The string data (key) and the element object (value) are
1430stored in the same allocation with the string data immediately after the element
1431object. This container guarantees the "<tt>(char*)(&amp;Value+1)</tt>" points
1432to the key string for a value.</p>
1433
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1434<p>The StringMap is very fast for several reasons: quadratic probing is very
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1435cache efficient for lookups, the hash value of strings in buckets is not
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1436recomputed when lookup up an element, StringMap rarely has to touch the
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1437memory for unrelated objects when looking up a value (even when hash collisions
1438happen), hash table growth does not recompute the hash values for strings
1439already in the table, and each pair in the map is store in a single allocation
1440(the string data is stored in the same allocation as the Value of a pair).</p>
1441
Chris Lattner796f9fa2007-02-08 19:14:21 +0000[diff] [blame]1442<p>StringMap also provides query methods that take byte ranges, so it only ever
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1443copies a string if a value is inserted into the table.</p>
1444</div>
1445
1446<!-- _______________________________________________________________________ -->
1447<div class="doc_subsubsection">
1448 <a name="dss_indexedmap">"llvm/ADT/IndexedMap.h"</a>
1449</div>
1450
1451<div class="doc_text">
1452<p>
1453IndexedMap is a specialized container for mapping small dense integers (or
1454values that can be mapped to small dense integers) to some other type. It is
1455internally implemented as a vector with a mapping function that maps the keys to
1456the dense integer range.
1457</p>
1458
1459<p>
1460This is useful for cases like virtual registers in the LLVM code generator: they
1461have a dense mapping that is offset by a compile-time constant (the first
1462virtual register ID).</p>
1463
1464</div>
1465
1466<!-- _______________________________________________________________________ -->
1467<div class="doc_subsubsection">
1468 <a name="dss_densemap">"llvm/ADT/DenseMap.h"</a>
1469</div>
1470
1471<div class="doc_text">
1472
1473<p>
1474DenseMap is a simple quadratically probed hash table. It excels at supporting
1475small keys and values: it uses a single allocation to hold all of the pairs that
1476are currently inserted in the map. DenseMap is a great way to map pointers to
1477pointers, or map other small types to each other.
1478</p>
1479
1480<p>
1481There are several aspects of DenseMap that you should be aware of, however. The
1482iterators in a densemap are invalidated whenever an insertion occurs, unlike
1483map. Also, because DenseMap allocates space for a large number of key/value
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1484pairs (it starts with 64 by default), it will waste a lot of space if your keys
1485or values are large. Finally, you must implement a partial specialization of
Chris Lattner76c1b972007-09-17 18:34:04 +0000[diff] [blame]1486DenseMapInfo for the key that you want, if it isn't already supported. This
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1487is required to tell DenseMap about two special marker values (which can never be
Chris Lattnera4a264d2007-02-03 20:17:53 +0000[diff] [blame]1488inserted into the map) that it needs internally.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1489
1490</div>
1491
1492<!-- _______________________________________________________________________ -->
1493<div class="doc_subsubsection">
1494 <a name="dss_map">&lt;map&gt;</a>
1495</div>
1496
1497<div class="doc_text">
1498
1499<p>
1500std::map has similar characteristics to <a href="#dss_set">std::set</a>: it uses
1501a single allocation per pair inserted into the map, it offers log(n) lookup with
1502an extremely large constant factor, imposes a space penalty of 3 pointers per
1503pair in the map, etc.</p>
1504
1505<p>std::map is most useful when your keys or values are very large, if you need
1506to iterate over the collection in sorted order, or if you need stable iterators
1507into the map (i.e. they don't get invalidated if an insertion or deletion of
1508another element takes place).</p>
1509
1510</div>
1511
1512<!-- _______________________________________________________________________ -->
1513<div class="doc_subsubsection">
1514 <a name="dss_othermap">Other Map-Like Container Options</a>
1515</div>
1516
1517<div class="doc_text">
1518
1519<p>
1520The STL provides several other options, such as std::multimap and the various
Chris Lattnerf1a30822009-03-09 05:20:45 +0000[diff] [blame]1521"hash_map" like containers (whether from C++ TR1 or from the SGI library). We
1522never use hash_set and unordered_set because they are generally very expensive
1523(each insertion requires a malloc) and very non-portable.</p>
Chris Lattnerc5722432007-02-03 19:49:31 +0000[diff] [blame]1524
1525<p>std::multimap is useful if you want to map a key to multiple values, but has
1526all the drawbacks of std::map. A sorted vector or some other approach is almost
1527always better.</p>
1528
Chris Lattner098129a2007-02-03 03:04:03 +0000[diff] [blame]1529</div>
1530
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1531<!-- ======================================================================= -->
1532<div class="doc_subsection">
Chris Lattnerdced9fb2009-07-25 07:22:20 +0000[diff] [blame]1533 <a name="ds_string">String-like containers</a>
1534</div>
1535
1536<div class="doc_text">
1537
1538<p>
1539TODO: const char* vs stringref vs smallstring vs std::string. Describe twine,
1540xref to #string_apis.
1541</p>
1542
1543</div>
1544
1545<!-- ======================================================================= -->
1546<div class="doc_subsection">
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1547 <a name="ds_bit">Bit storage containers (BitVector, SparseBitVector)</a>
1548</div>
1549
1550<div class="doc_text">
Chris Lattner7086ce72007-09-25 22:37:50 +0000[diff] [blame]1551<p>Unlike the other containers, there are only two bit storage containers, and
1552choosing when to use each is relatively straightforward.</p>
1553
1554<p>One additional option is
1555<tt>std::vector&lt;bool&gt;</tt>: we discourage its use for two reasons 1) the
1556implementation in many common compilers (e.g. commonly available versions of
1557GCC) is extremely inefficient and 2) the C++ standards committee is likely to
1558deprecate this container and/or change it significantly somehow. In any case,
1559please don't use it.</p>
Daniel Berlin1939ace2007-09-24 17:52:25 +0000[diff] [blame]1560</div>
1561
1562<!-- _______________________________________________________________________ -->
1563<div class="doc_subsubsection">
1564 <a name="dss_bitvector">BitVector</a>
1565</div>
1566
1567<div class="doc_text">
1568<p> The BitVector container provides a fixed size set of bits for manipulation.
1569It supports individual bit setting/testing, as well as set operations. The set
1570operations take time O(size of bitvector), but operations are performed one word
1571at a time, instead of one bit at a time. This makes the BitVector very fast for
1572set operations compared to other containers. Use the BitVector when you expect
1573the number of set bits to be high (IE a dense set).
1574</p>
1575</div>
1576
1577<!-- _______________________________________________________________________ -->
1578<div class="doc_subsubsection">
1579 <a name="dss_sparsebitvector">SparseBitVector</a>
1580</div>
1581
1582<div class="doc_text">
1583<p> The SparseBitVector container is much like BitVector, with one major
1584difference: Only the bits that are set, are stored. This makes the
1585SparseBitVector much more space efficient than BitVector when the set is sparse,
1586as well as making set operations O(number of set bits) instead of O(size of
1587universe). 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
1588(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).
1589</p>
1590</div>
Chris Lattnerf623a082005-10-17 01:36:23 +0000[diff] [blame]1591
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1592<!-- *********************************************************************** -->
1593<div class="doc_section">
1594 <a name="common">Helpful Hints for Common Operations</a>
1595</div>
1596<!-- *********************************************************************** -->
1597
1598<div class="doc_text">
1599
1600<p>This section describes how to perform some very simple transformations of
1601LLVM code. This is meant to give examples of common idioms used, showing the
1602practical side of LLVM transformations. <p> Because this is a "how-to" section,
1603you should also read about the main classes that you will be working with. The
1604<a href="#coreclasses">Core LLVM Class Hierarchy Reference</a> contains details
1605and descriptions of the main classes that you should know about.</p>
1606
1607</div>
1608
1609<!-- NOTE: this section should be heavy on example code -->
1610<!-- ======================================================================= -->
1611<div class="doc_subsection">
1612 <a name="inspection">Basic Inspection and Traversal Routines</a>
1613</div>
1614
1615<div class="doc_text">
1616
1617<p>The LLVM compiler infrastructure have many different data structures that may
1618be traversed. Following the example of the C++ standard template library, the
1619techniques used to traverse these various data structures are all basically the
1620same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
1621method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
1622function returns an iterator pointing to one past the last valid element of the
1623sequence, and there is some <tt>XXXiterator</tt> data type that is common
1624between the two operations.</p>
1625
1626<p>Because the pattern for iteration is common across many different aspects of
1627the program representation, the standard template library algorithms may be used
1628on them, and it is easier to remember how to iterate. First we show a few common
1629examples of the data structures that need to be traversed. Other data
1630structures are traversed in very similar ways.</p>
1631
1632</div>
1633
1634<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1635<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1636 <a name="iterate_function">Iterating over the </a><a
1637 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
1638 href="#Function"><tt>Function</tt></a>
1639</div>
1640
1641<div class="doc_text">
1642
1643<p>It's quite common to have a <tt>Function</tt> instance that you'd like to
1644transform in some way; in particular, you'd like to manipulate its
1645<tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over all of
1646the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>. The following is
1647an example that prints the name of a <tt>BasicBlock</tt> and the number of
1648<tt>Instruction</tt>s it contains:</p>
1649
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1650<div class="doc_code">
1651<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1652// <i>func is a pointer to a Function instance</i>
1653for (Function::iterator i = func-&gt;begin(), e = func-&gt;end(); i != e; ++i)
1654 // <i>Print out the name of the basic block if it has one, and then the</i>
1655 // <i>number of instructions that it contains</i>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1656 llvm::cerr &lt;&lt; "Basic block (name=" &lt;&lt; i-&gt;getName() &lt;&lt; ") has "
1657 &lt;&lt; i-&gt;size() &lt;&lt; " instructions.\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1658</pre>
1659</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1660
1661<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]1662invoking member functions of the <tt>Instruction</tt> class. This is
1663because the indirection operator is overloaded for the iterator
Chris Lattner7496ec52003-08-05 22:54:23 +0000[diff] [blame]1664classes. In the above code, the expression <tt>i-&gt;size()</tt> is
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1665exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.</p>
1666
1667</div>
1668
1669<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1670<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1671 <a name="iterate_basicblock">Iterating over the </a><a
1672 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1673 href="#BasicBlock"><tt>BasicBlock</tt></a>
1674</div>
1675
1676<div class="doc_text">
1677
1678<p>Just like when dealing with <tt>BasicBlock</tt>s in <tt>Function</tt>s, it's
1679easy to iterate over the individual instructions that make up
1680<tt>BasicBlock</tt>s. Here's a code snippet that prints out each instruction in
1681a <tt>BasicBlock</tt>:</p>
1682
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1683<div class="doc_code">
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1684<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1685// <i>blk is a pointer to a BasicBlock instance</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1686for (BasicBlock::iterator i = blk-&gt;begin(), e = blk-&gt;end(); i != e; ++i)
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1687 // <i>The next statement works since operator&lt;&lt;(ostream&amp;,...)</i>
1688 // <i>is overloaded for Instruction&amp;</i>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1689 llvm::cerr &lt;&lt; *i &lt;&lt; "\n";
Chris Lattner55c04612005-03-06 06:00:13 +0000[diff] [blame]1690</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1691</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1692
1693<p>However, this isn't really the best way to print out the contents of a
1694<tt>BasicBlock</tt>! Since the ostream operators are overloaded for virtually
1695anything you'll care about, you could have just invoked the print routine on the
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1696basic block itself: <tt>llvm::cerr &lt;&lt; *blk &lt;&lt; "\n";</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1697
1698</div>
1699
1700<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1701<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1702 <a name="iterate_institer">Iterating over the </a><a
1703 href="#Instruction"><tt>Instruction</tt></a>s in a <a
1704 href="#Function"><tt>Function</tt></a>
1705</div>
1706
1707<div class="doc_text">
1708
1709<p>If you're finding that you commonly iterate over a <tt>Function</tt>'s
1710<tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s <tt>Instruction</tt>s,
1711<tt>InstIterator</tt> should be used instead. You'll need to include <a
1712href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>,
1713and then instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1714small 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]1715
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1716<div class="doc_code">
1717<pre>
1718#include "<a href="/doxygen/InstIterator_8h-source.html">llvm/Support/InstIterator.h</a>"
1719
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1720// <i>F is a pointer to a Function instance</i>
Chris Lattnerda021aa2008-06-04 18:20:42 +0000[diff] [blame]1721for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
1722 llvm::cerr &lt;&lt; *I &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1723</pre>
1724</div>
1725
1726<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]1727work list with its initial contents. For example, if you wanted to
1728initialize a work list to contain all instructions in a <tt>Function</tt>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1729F, all you would need to do is something like:</p>
1730
1731<div class="doc_code">
1732<pre>
1733std::set&lt;Instruction*&gt; worklist;
Chris Lattnerda021aa2008-06-04 18:20:42 +0000[diff] [blame]1734// or better yet, SmallPtrSet&lt;Instruction*, 64&gt; worklist;
1735
1736for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
1737 worklist.insert(&amp;*I);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1738</pre>
1739</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1740
1741<p>The STL set <tt>worklist</tt> would now contain all instructions in the
1742<tt>Function</tt> pointed to by F.</p>
1743
1744</div>
1745
1746<!-- _______________________________________________________________________ -->
1747<div class="doc_subsubsection">
1748 <a name="iterate_convert">Turning an iterator into a class pointer (and
1749 vice-versa)</a>
1750</div>
1751
1752<div class="doc_text">
1753
1754<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]1755instance when all you've got at hand is an iterator. Well, extracting
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1756a reference or a pointer from an iterator is very straight-forward.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1757Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and <tt>j</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1758is a <tt>BasicBlock::const_iterator</tt>:</p>
1759
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1760<div class="doc_code">
1761<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1762Instruction&amp; inst = *i; // <i>Grab reference to instruction reference</i>
1763Instruction* pinst = &amp;*i; // <i>Grab pointer to instruction reference</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1764const Instruction&amp; inst = *j;
1765</pre>
1766</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1767
1768<p>However, the iterators you'll be working with in the LLVM framework are
1769special: they will automatically convert to a ptr-to-instance type whenever they
1770need to. Instead of dereferencing the iterator and then taking the address of
1771the result, you can simply assign the iterator to the proper pointer type and
1772you get the dereference and address-of operation as a result of the assignment
1773(behind the scenes, this is a result of overloading casting mechanisms). Thus
1774the last line of the last example,</p>
1775
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1776<div class="doc_code">
1777<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1778Instruction *pinst = &amp;*i;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1779</pre>
1780</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1781
1782<p>is semantically equivalent to</p>
1783
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1784<div class="doc_code">
1785<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1786Instruction *pinst = i;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1787</pre>
1788</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1789
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1790<p>It's also possible to turn a class pointer into the corresponding iterator,
1791and this is a constant time operation (very efficient). The following code
1792snippet illustrates use of the conversion constructors provided by LLVM
1793iterators. By using these, you can explicitly grab the iterator of something
1794without actually obtaining it via iteration over some structure:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1795
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1796<div class="doc_code">
1797<pre>
1798void printNextInstruction(Instruction* inst) {
1799 BasicBlock::iterator it(inst);
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1800 ++it; // <i>After this line, it refers to the instruction after *inst</i>
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1801 if (it != inst-&gt;getParent()-&gt;end()) llvm::cerr &lt;&lt; *it &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1802}
1803</pre>
1804</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1805
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1806</div>
1807
1808<!--_______________________________________________________________________-->
1809<div class="doc_subsubsection">
1810 <a name="iterate_complex">Finding call sites: a slightly more complex
1811 example</a>
1812</div>
1813
1814<div class="doc_text">
1815
1816<p>Say that you're writing a FunctionPass and would like to count all the
1817locations in the entire module (that is, across every <tt>Function</tt>) where a
1818certain function (i.e., some <tt>Function</tt>*) is already in scope. As you'll
1819learn 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]1820much more straight-forward manner, but this example will allow us to explore how
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]1821you'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]1822is what we want to do:</p>
1823
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1824<div class="doc_code">
1825<pre>
1826initialize callCounter to zero
1827for each Function f in the Module
1828 for each BasicBlock b in f
1829 for each Instruction i in b
1830 if (i is a CallInst and calls the given function)
1831 increment callCounter
1832</pre>
1833</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1834
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1835<p>And the actual code is (remember, because we're writing a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1836<tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply has to
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1837override the <tt>runOnFunction</tt> method):</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1838
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1839<div class="doc_code">
1840<pre>
1841Function* targetFunc = ...;
1842
1843class OurFunctionPass : public FunctionPass {
1844 public:
1845 OurFunctionPass(): callCounter(0) { }
1846
1847 virtual runOnFunction(Function&amp; F) {
1848 for (Function::iterator b = F.begin(), be = F.end(); b != be; ++b) {
Eric Christopher203e71d2008-11-08 08:20:49 +0000[diff] [blame]1849 for (BasicBlock::iterator i = b-&gt;begin(), ie = b-&gt;end(); i != ie; ++i) {
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1850 if (<a href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a>&lt;<a
1851 href="#CallInst">CallInst</a>&gt;(&amp;*i)) {
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1852 // <i>We know we've encountered a call instruction, so we</i>
1853 // <i>need to determine if it's a call to the</i>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1854 // <i>function pointed to by m_func or not.</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1855 if (callInst-&gt;getCalledFunction() == targetFunc)
1856 ++callCounter;
1857 }
1858 }
1859 }
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1860 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1861
1862 private:
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1863 unsigned callCounter;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1864};
1865</pre>
1866</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1867
1868</div>
1869
Brian Gaekef1972c62003-11-07 19:25:45 +0000[diff] [blame]1870<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1871<div class="doc_subsubsection">
1872 <a name="calls_and_invokes">Treating calls and invokes the same way</a>
1873</div>
1874
1875<div class="doc_text">
1876
1877<p>You may have noticed that the previous example was a bit oversimplified in
1878that it did not deal with call sites generated by 'invoke' instructions. In
1879this, and in other situations, you may find that you want to treat
1880<tt>CallInst</tt>s and <tt>InvokeInst</tt>s the same way, even though their
1881most-specific common base class is <tt>Instruction</tt>, which includes lots of
1882less closely-related things. For these cases, LLVM provides a handy wrapper
1883class called <a
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]1884href="http://llvm.org/doxygen/classllvm_1_1CallSite.html"><tt>CallSite</tt></a>.
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1885It is essentially a wrapper around an <tt>Instruction</tt> pointer, with some
1886methods that provide functionality common to <tt>CallInst</tt>s and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1887<tt>InvokeInst</tt>s.</p>
1888
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]1889<p>This class has "value semantics": it should be passed by value, not by
1890reference and it should not be dynamically allocated or deallocated using
1891<tt>operator new</tt> or <tt>operator delete</tt>. It is efficiently copyable,
1892assignable and constructable, with costs equivalents to that of a bare pointer.
1893If you look at its definition, it has only a single pointer member.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1894
1895</div>
1896
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]1897<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1898<div class="doc_subsubsection">
1899 <a name="iterate_chains">Iterating over def-use &amp; use-def chains</a>
1900</div>
1901
1902<div class="doc_text">
1903
1904<p>Frequently, we might have an instance of the <a
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]1905href="/doxygen/classllvm_1_1Value.html">Value Class</a> and we want to
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]1906determine which <tt>User</tt>s use the <tt>Value</tt>. The list of all
1907<tt>User</tt>s of a particular <tt>Value</tt> is called a <i>def-use</i> chain.
1908For example, let's say we have a <tt>Function*</tt> named <tt>F</tt> to a
1909particular function <tt>foo</tt>. Finding all of the instructions that
1910<i>use</i> <tt>foo</tt> is as simple as iterating over the <i>def-use</i> chain
1911of <tt>F</tt>:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1912
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1913<div class="doc_code">
1914<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1915Function *F = ...;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1916
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1917for (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]1918 if (Instruction *Inst = dyn_cast&lt;Instruction&gt;(*i)) {
Bill Wendling832171c2006-12-07 20:04:42 +0000[diff] [blame]1919 llvm::cerr &lt;&lt; "F is used in instruction:\n";
1920 llvm::cerr &lt;&lt; *Inst &lt;&lt; "\n";
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1921 }
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1922</pre>
1923</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1924
1925<p>Alternately, it's common to have an instance of the <a
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]1926href="/doxygen/classllvm_1_1User.html">User Class</a> and need to know what
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1927<tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used by a
1928<tt>User</tt> is known as a <i>use-def</i> chain. Instances of class
1929<tt>Instruction</tt> are common <tt>User</tt>s, so we might want to iterate over
1930all of the values that a particular instruction uses (that is, the operands of
1931the particular <tt>Instruction</tt>):</p>
1932
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1933<div class="doc_code">
1934<pre>
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1935Instruction *pi = ...;
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1936
1937for (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]1938 Value *v = *i;
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]1939 // <i>...</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]1940}
1941</pre>
1942</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1943
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]1944<!--
1945 def-use chains ("finding all users of"): Value::use_begin/use_end
1946 use-def chains ("finding all values used"): User::op_begin/op_end [op=operand]
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1947-->
1948
1949</div>
1950
Chris Lattner2e438ca2008-01-03 16:56:04 +0000[diff] [blame]1951<!--_______________________________________________________________________-->
1952<div class="doc_subsubsection">
1953 <a name="iterate_preds">Iterating over predecessors &amp;
1954successors of blocks</a>
1955</div>
1956
1957<div class="doc_text">
1958
1959<p>Iterating over the predecessors and successors of a block is quite easy
1960with the routines defined in <tt>"llvm/Support/CFG.h"</tt>. Just use code like
1961this to iterate over all predecessors of BB:</p>
1962
1963<div class="doc_code">
1964<pre>
1965#include "llvm/Support/CFG.h"
1966BasicBlock *BB = ...;
1967
1968for (pred_iterator PI = pred_begin(BB), E = pred_end(BB); PI != E; ++PI) {
1969 BasicBlock *Pred = *PI;
1970 // <i>...</i>
1971}
1972</pre>
1973</div>
1974
1975<p>Similarly, to iterate over successors use
1976succ_iterator/succ_begin/succ_end.</p>
1977
1978</div>
1979
1980
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1981<!-- ======================================================================= -->
1982<div class="doc_subsection">
1983 <a name="simplechanges">Making simple changes</a>
1984</div>
1985
1986<div class="doc_text">
1987
1988<p>There are some primitive transformation operations present in the LLVM
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]1989infrastructure that are worth knowing about. When performing
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1990transformations, it's fairly common to manipulate the contents of basic
1991blocks. This section describes some of the common methods for doing so
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1992and gives example code.</p>
1993
1994</div>
1995
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1996<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]1997<div class="doc_subsubsection">
1998 <a name="schanges_creating">Creating and inserting new
1999 <tt>Instruction</tt>s</a>
2000</div>
2001
2002<div class="doc_text">
2003
2004<p><i>Instantiating Instructions</i></p>
2005
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]2006<p>Creation of <tt>Instruction</tt>s is straight-forward: simply call the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2007constructor for the kind of instruction to instantiate and provide the necessary
2008parameters. For example, an <tt>AllocaInst</tt> only <i>requires</i> a
2009(const-ptr-to) <tt>Type</tt>. Thus:</p>
2010
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2011<div class="doc_code">
2012<pre>
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2013AllocaInst* ai = new AllocaInst(Type::Int32Ty);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2014</pre>
2015</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2016
2017<p>will create an <tt>AllocaInst</tt> instance that represents the allocation of
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]2018one integer in the current stack frame, at run time. Each <tt>Instruction</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2019subclass is likely to have varying default parameters which change the semantics
2020of the instruction, so refer to the <a
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]2021href="/doxygen/classllvm_1_1Instruction.html">doxygen documentation for the subclass of
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2022Instruction</a> that you're interested in instantiating.</p>
2023
2024<p><i>Naming values</i></p>
2025
2026<p>It is very useful to name the values of instructions when you're able to, as
2027this facilitates the debugging of your transformations. If you end up looking
2028at generated LLVM machine code, you definitely want to have logical names
2029associated with the results of instructions! By supplying a value for the
2030<tt>Name</tt> (default) parameter of the <tt>Instruction</tt> constructor, you
2031associate a logical name with the result of the instruction's execution at
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]2032run time. For example, say that I'm writing a transformation that dynamically
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2033allocates space for an integer on the stack, and that integer is going to be
2034used as some kind of index by some other code. To accomplish this, I place an
2035<tt>AllocaInst</tt> at the first point in the first <tt>BasicBlock</tt> of some
2036<tt>Function</tt>, and I'm intending to use it within the same
2037<tt>Function</tt>. I might do:</p>
2038
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2039<div class="doc_code">
2040<pre>
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2041AllocaInst* pa = new AllocaInst(Type::Int32Ty, 0, "indexLoc");
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2042</pre>
2043</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2044
2045<p>where <tt>indexLoc</tt> is now the logical name of the instruction's
Reid Spencer128a7a72007-02-03 21:06:43 +0000[diff] [blame]2046execution value, which is a pointer to an integer on the run time stack.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2047
2048<p><i>Inserting instructions</i></p>
2049
2050<p>There are essentially two ways to insert an <tt>Instruction</tt>
2051into an existing sequence of instructions that form a <tt>BasicBlock</tt>:</p>
2052
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]2053<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2054 <li>Insertion into an explicit instruction list
2055
2056 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within that
2057 <tt>BasicBlock</tt>, and a newly-created instruction we wish to insert
2058 before <tt>*pi</tt>, we do the following: </p>
2059
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2060<div class="doc_code">
2061<pre>
2062BasicBlock *pb = ...;
2063Instruction *pi = ...;
2064Instruction *newInst = new Instruction(...);
2065
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2066pb-&gt;getInstList().insert(pi, newInst); // <i>Inserts newInst before pi in pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2067</pre>
2068</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]2069
2070 <p>Appending to the end of a <tt>BasicBlock</tt> is so common that
2071 the <tt>Instruction</tt> class and <tt>Instruction</tt>-derived
2072 classes provide constructors which take a pointer to a
2073 <tt>BasicBlock</tt> to be appended to. For example code that
2074 looked like: </p>
2075
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2076<div class="doc_code">
2077<pre>
2078BasicBlock *pb = ...;
2079Instruction *newInst = new Instruction(...);
2080
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2081pb-&gt;getInstList().push_back(newInst); // <i>Appends newInst to pb</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2082</pre>
2083</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]2084
2085 <p>becomes: </p>
2086
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2087<div class="doc_code">
2088<pre>
2089BasicBlock *pb = ...;
2090Instruction *newInst = new Instruction(..., pb);
2091</pre>
2092</div>
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000[diff] [blame]2093
2094 <p>which is much cleaner, especially if you are creating
2095 long instruction streams.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2096
2097 <li>Insertion into an implicit instruction list
2098
2099 <p><tt>Instruction</tt> instances that are already in <tt>BasicBlock</tt>s
2100 are implicitly associated with an existing instruction list: the instruction
2101 list of the enclosing basic block. Thus, we could have accomplished the same
2102 thing as the above code without being given a <tt>BasicBlock</tt> by doing:
2103 </p>
2104
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2105<div class="doc_code">
2106<pre>
2107Instruction *pi = ...;
2108Instruction *newInst = new Instruction(...);
2109
2110pi-&gt;getParent()-&gt;getInstList().insert(pi, newInst);
2111</pre>
2112</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2113
2114 <p>In fact, this sequence of steps occurs so frequently that the
2115 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes provide
2116 constructors which take (as a default parameter) a pointer to an
2117 <tt>Instruction</tt> which the newly-created <tt>Instruction</tt> should
2118 precede. That is, <tt>Instruction</tt> constructors are capable of
2119 inserting the newly-created instance into the <tt>BasicBlock</tt> of a
2120 provided instruction, immediately before that instruction. Using an
2121 <tt>Instruction</tt> constructor with a <tt>insertBefore</tt> (default)
2122 parameter, the above code becomes:</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(..., pi);
2128</pre>
2129</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2130
2131 <p>which is much cleaner, especially if you're creating a lot of
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2132 instructions and adding them to <tt>BasicBlock</tt>s.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2133</ul>
2134
2135</div>
2136
2137<!--_______________________________________________________________________-->
2138<div class="doc_subsubsection">
2139 <a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a>
2140</div>
2141
2142<div class="doc_text">
2143
2144<p>Deleting an instruction from an existing sequence of instructions that form a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000[diff] [blame]2145<a href="#BasicBlock"><tt>BasicBlock</tt></a> is very straight-forward. First,
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2146you must have a pointer to the instruction that you wish to delete. Second, you
2147need to obtain the pointer to that instruction's basic block. You use the
2148pointer to the basic block to get its list of instructions and then use the
2149erase function to remove your instruction. For example:</p>
2150
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2151<div class="doc_code">
2152<pre>
2153<a href="#Instruction">Instruction</a> *I = .. ;
Chris Lattner9f8ec252008-02-15 22:57:17 +0000[diff] [blame]2154I-&gt;eraseFromParent();
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2155</pre>
2156</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2157
2158</div>
2159
2160<!--_______________________________________________________________________-->
2161<div class="doc_subsubsection">
2162 <a name="schanges_replacing">Replacing an <tt>Instruction</tt> with another
2163 <tt>Value</tt></a>
2164</div>
2165
2166<div class="doc_text">
2167
2168<p><i>Replacing individual instructions</i></p>
2169
2170<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]2171permits use of two very useful replace functions: <tt>ReplaceInstWithValue</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2172and <tt>ReplaceInstWithInst</tt>.</p>
2173
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2174<h4><a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a></h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2175
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2176<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2177 <li><tt>ReplaceInstWithValue</tt>
2178
Nick Lewyckyb6d1f392008-09-15 06:31:52 +0000[diff] [blame]2179 <p>This function replaces all uses of a given instruction with a value,
2180 and then removes the original instruction. The following example
2181 illustrates the replacement of the result of a particular
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]2182 <tt>AllocaInst</tt> that allocates memory for a single integer with a null
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2183 pointer to an integer.</p>
2184
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2185<div class="doc_code">
2186<pre>
2187AllocaInst* instToReplace = ...;
2188BasicBlock::iterator ii(instToReplace);
2189
2190ReplaceInstWithValue(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
Daniel Dunbar58c2ac02008-10-03 22:17:25 +0000[diff] [blame]2191 Constant::getNullValue(PointerType::getUnqual(Type::Int32Ty)));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2192</pre></div></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2193
2194 <li><tt>ReplaceInstWithInst</tt>
2195
2196 <p>This function replaces a particular instruction with another
Nick Lewyckyb6d1f392008-09-15 06:31:52 +0000[diff] [blame]2197 instruction, inserting the new instruction into the basic block at the
2198 location where the old instruction was, and replacing any uses of the old
2199 instruction with the new instruction. The following example illustrates
2200 the replacement of one <tt>AllocaInst</tt> with another.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2201
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2202<div class="doc_code">
2203<pre>
2204AllocaInst* instToReplace = ...;
2205BasicBlock::iterator ii(instToReplace);
2206
2207ReplaceInstWithInst(instToReplace-&gt;getParent()-&gt;getInstList(), ii,
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2208 new AllocaInst(Type::Int32Ty, 0, "ptrToReplacedInt"));
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2209</pre></div></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2210</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2211
2212<p><i>Replacing multiple uses of <tt>User</tt>s and <tt>Value</tt>s</i></p>
2213
2214<p>You can use <tt>Value::replaceAllUsesWith</tt> and
2215<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]2216doxygen documentation for the <a href="/doxygen/classllvm_1_1Value.html">Value Class</a>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]2217and <a href="/doxygen/classllvm_1_1User.html">User Class</a>, respectively, for more
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2218information.</p>
2219
2220<!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
2221include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
2222ReplaceInstWithValue, ReplaceInstWithInst -->
2223
2224</div>
2225
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2226<!--_______________________________________________________________________-->
2227<div class="doc_subsubsection">
2228 <a name="schanges_deletingGV">Deleting <tt>GlobalVariable</tt>s</a>
2229</div>
2230
2231<div class="doc_text">
2232
Tanya Lattnerc5dfcdb2007-06-20 20:46:37 +0000[diff] [blame]2233<p>Deleting a global variable from a module is just as easy as deleting an
2234Instruction. First, you must have a pointer to the global variable that you wish
2235 to delete. You use this pointer to erase it from its parent, the module.
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2236 For example:</p>
2237
2238<div class="doc_code">
2239<pre>
2240<a href="#GlobalVariable">GlobalVariable</a> *GV = .. ;
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2241
Tanya Lattnerc5dfcdb2007-06-20 20:46:37 +0000[diff] [blame]2242GV-&gt;eraseFromParent();
Tanya Lattnerb011c662007-06-20 18:33:15 +0000[diff] [blame]2243</pre>
2244</div>
2245
2246</div>
2247
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2248<!-- ======================================================================= -->
2249<div class="doc_subsection">
2250 <a name="create_types">How to Create Types</a>
2251</div>
2252
2253<div class="doc_text">
2254
2255<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]2256statically, you can use <tt>TypeBuilder&lt;...&gt;::get()</tt>, defined
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2257in <tt>llvm/Support/TypeBuilder.h</tt>, to retrieve them. <tt>TypeBuilder</tt>
2258has two forms depending on whether you're building types for cross-compilation
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2259or native library use. <tt>TypeBuilder&lt;T, true&gt;</tt> requires
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2260that <tt>T</tt> be independent of the host environment, meaning that it's built
2261out of types from
2262the <a href="/doxygen/namespacellvm_1_1types.html"><tt>llvm::types</tt></a>
2263namespace and pointers, functions, arrays, etc. built of
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2264those. <tt>TypeBuilder&lt;T, false&gt;</tt> additionally allows native C types
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2265whose size may depend on the host compiler. For example,</p>
2266
2267<div class="doc_code">
2268<pre>
Misha Brukman1af789f2009-05-01 20:40:51 +0000[diff] [blame]2269FunctionType *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]2270</pre>
2271</div>
2272
2273<p>is easier to read and write than the equivalent</p>
2274
2275<div class="doc_code">
2276<pre>
Owen Anderson5e8c50e2009-06-16 17:40:28 +0000[diff] [blame]2277std::vector&lt;const Type*&gt; params;
Jeffrey Yasskin714257f2009-04-30 22:33:41 +0000[diff] [blame]2278params.push_back(PointerType::getUnqual(Type::Int32Ty));
2279FunctionType *ft = FunctionType::get(Type::Int8Ty, params, false);
2280</pre>
2281</div>
2282
2283<p>See the <a href="/doxygen/TypeBuilder_8h-source.html#l00001">class
2284comment</a> for more details.</p>
2285
2286</div>
2287
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]2288<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2289<div class="doc_section">
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2290 <a name="threading">Threads and LLVM</a>
2291</div>
2292<!-- *********************************************************************** -->
2293
2294<div class="doc_text">
2295<p>
2296This section describes the interaction of the LLVM APIs with multithreading,
2297both on the part of client applications, and in the JIT, in the hosted
2298application.
2299</p>
2300
2301<p>
2302Note that LLVM's support for multithreading is still relatively young. Up
2303through version 2.5, the execution of threaded hosted applications was
2304supported, but not threaded client access to the APIs. While this use case is
2305now supported, clients <em>must</em> adhere to the guidelines specified below to
2306ensure proper operation in multithreaded mode.
2307</p>
2308
2309<p>
2310Note that, on Unix-like platforms, LLVM requires the presence of GCC's atomic
2311intrinsics in order to support threaded operation. If you need a
2312multhreading-capable LLVM on a platform without a suitably modern system
2313compiler, consider compiling LLVM and LLVM-GCC in single-threaded mode, and
2314using the resultant compiler to build a copy of LLVM with multithreading
2315support.
2316</p>
2317</div>
2318
2319<!-- ======================================================================= -->
2320<div class="doc_subsection">
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2321 <a name="startmultithreaded">Entering and Exiting Multithreaded Mode</a>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2322</div>
2323
2324<div class="doc_text">
2325
2326<p>
2327In order to properly protect its internal data structures while avoiding
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2328excessive locking overhead in the single-threaded case, the LLVM must intialize
2329certain data structures necessary to provide guards around its internals. To do
2330so, the client program must invoke <tt>llvm_start_multithreaded()</tt> before
2331making any concurrent LLVM API calls. To subsequently tear down these
2332structures, use the <tt>llvm_stop_multithreaded()</tt> call. You can also use
2333the <tt>llvm_is_multithreaded()</tt> call to check the status of multithreaded
2334mode.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2335</p>
2336
2337<p>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2338Note that both of these calls must be made <em>in isolation</em>. That is to
2339say that no other LLVM API calls may be executing at any time during the
2340execution of <tt>llvm_start_multithreaded()</tt> or <tt>llvm_stop_multithreaded
2341</tt>. It's is the client's responsibility to enforce this isolation.
2342</p>
2343
2344<p>
2345The return value of <tt>llvm_start_multithreaded()</tt> indicates the success or
2346failure of the initialization. Failure typically indicates that your copy of
2347LLVM was built without multithreading support, typically because GCC atomic
2348intrinsics were not found in your system compiler. In this case, the LLVM API
2349will not be safe for concurrent calls. However, it <em>will</em> be safe for
2350hosting threaded applications in the JIT, though care must be taken to ensure
2351that side exits and the like do not accidentally result in concurrent LLVM API
2352calls.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2353</p>
2354</div>
2355
2356<!-- ======================================================================= -->
2357<div class="doc_subsection">
2358 <a name="shutdown">Ending Execution with <tt>llvm_shutdown()</tt></a>
2359</div>
2360
2361<div class="doc_text">
2362<p>
2363When you are done using the LLVM APIs, you should call <tt>llvm_shutdown()</tt>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2364to deallocate memory used for internal structures. This will also invoke
2365<tt>llvm_stop_multithreaded()</tt> if LLVM is operating in multithreaded mode.
2366As such, <tt>llvm_shutdown()</tt> requires the same isolation guarantees as
2367<tt>llvm_stop_multithreaded()</tt>.
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2368</p>
2369
2370<p>
2371Note that, if you use scope-based shutdown, you can use the
2372<tt>llvm_shutdown_obj</tt> class, which calls <tt>llvm_shutdown()</tt> in its
2373destructor.
2374</div>
2375
2376<!-- ======================================================================= -->
2377<div class="doc_subsection">
2378 <a name="managedstatic">Lazy Initialization with <tt>ManagedStatic</tt></a>
2379</div>
2380
2381<div class="doc_text">
2382<p>
2383<tt>ManagedStatic</tt> is a utility class in LLVM used to implement static
2384initialization of static resources, such as the global type tables. Before the
2385invocation of <tt>llvm_shutdown()</tt>, it implements a simple lazy
2386initialization scheme. Once <tt>llvm_start_multithreaded()</tt> returns,
2387however, it uses double-checked locking to implement thread-safe lazy
2388initialization.
2389</p>
2390
2391<p>
2392Note that, because no other threads are allowed to issue LLVM API calls before
2393<tt>llvm_start_multithreaded()</tt> returns, it is possible to have
2394<tt>ManagedStatic</tt>s of <tt>llvm::sys::Mutex</tt>s.
2395</p>
Owen Anderson1ad70e32009-06-16 18:04:19 +0000[diff] [blame]2396
2397<p>
2398The <tt>llvm_acquire_global_lock()</tt> and <tt>llvm_release_global_lock</tt>
2399APIs provide access to the global lock used to implement the double-checked
2400locking for lazy initialization. These should only be used internally to LLVM,
2401and only if you know what you're doing!
2402</p>
Owen Anderson8bc1b3b2009-06-16 01:17:16 +0000[diff] [blame]2403</div>
2404
2405<!-- *********************************************************************** -->
2406<div class="doc_section">
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2407 <a name="advanced">Advanced Topics</a>
2408</div>
2409<!-- *********************************************************************** -->
2410
2411<div class="doc_text">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2412<p>
2413This section describes some of the advanced or obscure API's that most clients
2414do not need to be aware of. These API's tend manage the inner workings of the
2415LLVM system, and only need to be accessed in unusual circumstances.
2416</p>
2417</div>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2418
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2419<!-- ======================================================================= -->
2420<div class="doc_subsection">
2421 <a name="TypeResolve">LLVM Type Resolution</a>
2422</div>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2423
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2424<div class="doc_text">
2425
2426<p>
2427The LLVM type system has a very simple goal: allow clients to compare types for
2428structural equality with a simple pointer comparison (aka a shallow compare).
2429This goal makes clients much simpler and faster, and is used throughout the LLVM
2430system.
2431</p>
2432
2433<p>
2434Unfortunately achieving this goal is not a simple matter. In particular,
2435recursive types and late resolution of opaque types makes the situation very
2436difficult to handle. Fortunately, for the most part, our implementation makes
2437most clients able to be completely unaware of the nasty internal details. The
2438primary case where clients are exposed to the inner workings of it are when
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]2439building a recursive type. In addition to this case, the LLVM bitcode reader,
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2440assembly parser, and linker also have to be aware of the inner workings of this
2441system.
2442</p>
2443
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]2444<p>
2445For our purposes below, we need three concepts. First, an "Opaque Type" is
2446exactly as defined in the <a href="LangRef.html#t_opaque">language
2447reference</a>. Second an "Abstract Type" is any type which includes an
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2448opaque type as part of its type graph (for example "<tt>{ opaque, i32 }</tt>").
2449Third, 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]2450float }</tt>").
2451</p>
2452
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2453</div>
2454
2455<!-- ______________________________________________________________________ -->
2456<div class="doc_subsubsection">
2457 <a name="BuildRecType">Basic Recursive Type Construction</a>
2458</div>
2459
2460<div class="doc_text">
2461
2462<p>
2463Because the most common question is "how do I build a recursive type with LLVM",
2464we answer it now and explain it as we go. Here we include enough to cause this
2465to be emitted to an output .ll file:
2466</p>
2467
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2468<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2469<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2470%mylist = type { %mylist*, i32 }
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2471</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2472</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2473
2474<p>
2475To build this, use the following LLVM APIs:
2476</p>
2477
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2478<div class="doc_code">
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2479<pre>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2480// <i>Create the initial outer struct</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2481<a href="#PATypeHolder">PATypeHolder</a> StructTy = OpaqueType::get();
2482std::vector&lt;const Type*&gt; Elts;
Daniel Dunbar58c2ac02008-10-03 22:17:25 +0000[diff] [blame]2483Elts.push_back(PointerType::getUnqual(StructTy));
Nick Lewycky10d64b92007-12-03 01:52:52 +0000[diff] [blame]2484Elts.push_back(Type::Int32Ty);
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2485StructType *NewSTy = StructType::get(Elts);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2486
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2487// <i>At this point, NewSTy = "{ opaque*, i32 }". Tell VMCore that</i>
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2488// <i>the struct and the opaque type are actually the same.</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2489cast&lt;OpaqueType&gt;(StructTy.get())-&gt;<a href="#refineAbstractTypeTo">refineAbstractTypeTo</a>(NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2490
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2491// <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]2492// <i>kept up-to-date</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2493NewSTy = cast&lt;StructType&gt;(StructTy.get());
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2494
Bill Wendling82e2eea2006-10-11 18:00:22 +0000[diff] [blame]2495// <i>Add a name for the type to the module symbol table (optional)</i>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2496MyModule-&gt;addTypeName("mylist", NewSTy);
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2497</pre>
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]2498</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2499
2500<p>
2501This code shows the basic approach used to build recursive types: build a
2502non-recursive type using 'opaque', then use type unification to close the cycle.
2503The type unification step is performed by the <tt><a
Chris Lattneraff26d12007-02-03 03:06:52 +0000[diff] [blame]2504href="#refineAbstractTypeTo">refineAbstractTypeTo</a></tt> method, which is
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2505described next. After that, we describe the <a
2506href="#PATypeHolder">PATypeHolder class</a>.
2507</p>
2508
2509</div>
2510
2511<!-- ______________________________________________________________________ -->
2512<div class="doc_subsubsection">
2513 <a name="refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a>
2514</div>
2515
2516<div class="doc_text">
2517<p>
2518The <tt>refineAbstractTypeTo</tt> method starts the type unification process.
2519While this method is actually a member of the DerivedType class, it is most
2520often used on OpaqueType instances. Type unification is actually a recursive
2521process. After unification, types can become structurally isomorphic to
2522existing types, and all duplicates are deleted (to preserve pointer equality).
2523</p>
2524
2525<p>
2526In the example above, the OpaqueType object is definitely deleted.
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2527Additionally, if there is an "{ \2*, i32}" type already created in the system,
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2528the pointer and struct type created are <b>also</b> deleted. Obviously whenever
2529a type is deleted, any "Type*" pointers in the program are invalidated. As
2530such, it is safest to avoid having <i>any</i> "Type*" pointers to abstract types
2531live across a call to <tt>refineAbstractTypeTo</tt> (note that non-abstract
2532types can never move or be deleted). To deal with this, the <a
2533href="#PATypeHolder">PATypeHolder</a> class is used to maintain a stable
2534reference to a possibly refined type, and the <a
2535href="#AbstractTypeUser">AbstractTypeUser</a> class is used to update more
2536complex datastructures.
2537</p>
2538
2539</div>
2540
2541<!-- ______________________________________________________________________ -->
2542<div class="doc_subsubsection">
2543 <a name="PATypeHolder">The PATypeHolder Class</a>
2544</div>
2545
2546<div class="doc_text">
2547<p>
2548PATypeHolder is a form of a "smart pointer" for Type objects. When VMCore
2549happily goes about nuking types that become isomorphic to existing types, it
2550automatically updates all PATypeHolder objects to point to the new type. In the
2551example above, this allows the code to maintain a pointer to the resultant
2552resolved recursive type, even though the Type*'s are potentially invalidated.
2553</p>
2554
2555<p>
2556PATypeHolder is an extremely light-weight object that uses a lazy union-find
2557implementation to update pointers. For example the pointer from a Value to its
2558Type is maintained by PATypeHolder objects.
2559</p>
2560
2561</div>
2562
2563<!-- ______________________________________________________________________ -->
2564<div class="doc_subsubsection">
2565 <a name="AbstractTypeUser">The AbstractTypeUser Class</a>
2566</div>
2567
2568<div class="doc_text">
2569
2570<p>
2571Some data structures need more to perform more complex updates when types get
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2572resolved. To support this, a class can derive from the AbstractTypeUser class.
2573This class
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2574allows it to get callbacks when certain types are resolved. To register to get
2575callbacks for a particular type, the DerivedType::{add/remove}AbstractTypeUser
Chris Lattner0f876db2005-04-25 15:47:57 +0000[diff] [blame]2576methods can be called on a type. Note that these methods only work for <i>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]2577 abstract</i> types. Concrete types (those that do not include any opaque
2578objects) can never be refined.
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2579</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2580</div>
2581
2582
2583<!-- ======================================================================= -->
2584<div class="doc_subsection">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2585 <a name="SymbolTable">The <tt>ValueSymbolTable</tt> and
2586 <tt>TypeSymbolTable</tt> classes</a>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2587</div>
Chris Lattnerf1b200b2005-04-23 17:27:36 +0000[diff] [blame]2588
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2589<div class="doc_text">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2590<p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1ValueSymbolTable.html">
2591ValueSymbolTable</a></tt> class provides a symbol table that the <a
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2592href="#Function"><tt>Function</tt></a> and <a href="#Module">
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2593<tt>Module</tt></a> classes use for naming value definitions. The symbol table
2594can provide a name for any <a href="#Value"><tt>Value</tt></a>.
2595The <tt><a href="http://llvm.org/doxygen/classllvm_1_1TypeSymbolTable.html">
2596TypeSymbolTable</a></tt> class is used by the <tt>Module</tt> class to store
2597names for types.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2598
Reid Spencera6362242007-01-07 00:41:39 +0000[diff] [blame]2599<p>Note that the <tt>SymbolTable</tt> class should not be directly accessed
2600by most clients. It should only be used when iteration over the symbol table
2601names themselves are required, which is very special purpose. Note that not
2602all LLVM
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2603<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]2604an empty name) do not exist in the symbol table.
2605</p>
2606
Chris Lattner263a98e2007-02-16 04:37:31 +0000[diff] [blame]2607<p>These symbol tables support iteration over the values/types in the symbol
2608table with <tt>begin/end/iterator</tt> and supports querying to see if a
2609specific name is in the symbol table (with <tt>lookup</tt>). The
2610<tt>ValueSymbolTable</tt> class exposes no public mutator methods, instead,
2611simply call <tt>setName</tt> on a value, which will autoinsert it into the
2612appropriate symbol table. For types, use the Module::addTypeName method to
2613insert entries into the symbol table.</p>
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2614
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2615</div>
2616
2617
2618
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2619<!-- ======================================================================= -->
2620<div class="doc_subsection">
2621 <a name="UserLayout">The <tt>User</tt> and owned <tt>Use</tt> classes' memory layout</a>
2622</div>
2623
2624<div class="doc_text">
2625<p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1User.html">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2626User</a></tt> class provides a basis for expressing the ownership of <tt>User</tt>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2627towards other <tt><a href="http://llvm.org/doxygen/classllvm_1_1Value.html">
2628Value</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]2629Use</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]2630addition and removal.</p>
2631
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2632<!-- ______________________________________________________________________ -->
2633<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2634 <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]2635</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2636
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2637<div class="doc_text">
2638<p>
2639A subclass of <tt>User</tt> can choose between incorporating its <tt>Use</tt> objects
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2640or refer to them out-of-line by means of a pointer. A mixed variant
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2641(some <tt>Use</tt>s inline others hung off) is impractical and breaks the invariant
2642that the <tt>Use</tt> objects belonging to the same <tt>User</tt> form a contiguous array.
2643</p>
2644</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2645
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2646<p>
2647We have 2 different layouts in the <tt>User</tt> (sub)classes:
2648<ul>
2649<li><p>Layout a)
2650The <tt>Use</tt> object(s) are inside (resp. at fixed offset) of the <tt>User</tt>
2651object and there are a fixed number of them.</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2652
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2653<li><p>Layout b)
2654The <tt>Use</tt> object(s) are referenced by a pointer to an
2655array from the <tt>User</tt> object and there may be a variable
2656number of them.</p>
2657</ul>
2658<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2659As of v2.4 each layout still possesses a direct pointer to the
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2660start of the array of <tt>Use</tt>s. Though not mandatory for layout a),
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2661we stick to this redundancy for the sake of simplicity.
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2662The <tt>User</tt> object also stores the number of <tt>Use</tt> objects it
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2663has. (Theoretically this information can also be calculated
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2664given the scheme presented below.)</p>
2665<p>
2666Special forms of allocation operators (<tt>operator new</tt>)
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2667enforce the following memory layouts:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2668
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2669<ul>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2670<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]2671
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2672<pre>
2673...---.---.---.---.-------...
2674 | P | P | P | P | User
2675'''---'---'---'---'-------'''
2676</pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2677
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2678<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]2679<pre>
2680.-------...
2681| User
2682'-------'''
2683 |
2684 v
2685 .---.---.---.---...
2686 | P | P | P | P |
2687 '---'---'---'---'''
2688</pre>
2689</ul>
2690<i>(In the above figures '<tt>P</tt>' stands for the <tt>Use**</tt> that
2691 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]2692
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2693<!-- ______________________________________________________________________ -->
2694<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2695 <a name="Waymarking">The waymarking algorithm</a>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2696</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2697
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2698<div class="doc_text">
2699<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2700Since the <tt>Use</tt> objects are deprived of the direct (back)pointer to
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2701their <tt>User</tt> objects, there must be a fast and exact method to
2702recover it. This is accomplished by the following scheme:</p>
2703</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2704
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2705A 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]2706start of the <tt>User</tt> object:
2707<ul>
2708<li><tt>00</tt> &mdash;&gt; binary digit 0</li>
2709<li><tt>01</tt> &mdash;&gt; binary digit 1</li>
2710<li><tt>10</tt> &mdash;&gt; stop and calculate (<tt>s</tt>)</li>
2711<li><tt>11</tt> &mdash;&gt; full stop (<tt>S</tt>)</li>
2712</ul>
2713<p>
2714Given a <tt>Use*</tt>, all we have to do is to walk till we get
2715a stop and we either have a <tt>User</tt> immediately behind or
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2716we have to walk to the next stop picking up digits
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2717and calculating the offset:</p>
2718<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2719.---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.----------------
2720| 1 | s | 1 | 0 | 1 | 0 | s | 1 | 1 | 0 | s | 1 | 1 | s | 1 | S | User (or User*)
2721'---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'----------------
2722 |+15 |+10 |+6 |+3 |+1
2723 | | | | |__>
2724 | | | |__________>
2725 | | |______________________>
2726 | |______________________________________>
2727 |__________________________________________________________>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2728</pre>
2729<p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2730Only the significant number of bits need to be stored between the
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2731stops, so that the <i>worst case is 20 memory accesses</i> when there are
27321000 <tt>Use</tt> objects associated with a <tt>User</tt>.</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2733
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2734<!-- ______________________________________________________________________ -->
2735<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2736 <a name="ReferenceImpl">Reference implementation</a>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2737</div>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2738
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2739<div class="doc_text">
2740<p>
2741The following literate Haskell fragment demonstrates the concept:</p>
2742</div>
2743
2744<div class="doc_code">
2745<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2746> import Test.QuickCheck
2747>
2748> digits :: Int -> [Char] -> [Char]
2749> digits 0 acc = '0' : acc
2750> digits 1 acc = '1' : acc
2751> digits n acc = digits (n `div` 2) $ digits (n `mod` 2) acc
2752>
2753> dist :: Int -> [Char] -> [Char]
2754> dist 0 [] = ['S']
2755> dist 0 acc = acc
2756> dist 1 acc = let r = dist 0 acc in 's' : digits (length r) r
2757> dist n acc = dist (n - 1) $ dist 1 acc
2758>
2759> takeLast n ss = reverse $ take n $ reverse ss
2760>
2761> test = takeLast 40 $ dist 20 []
2762>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2763</pre>
2764</div>
2765<p>
2766Printing &lt;test&gt; gives: <tt>"1s100000s11010s10100s1111s1010s110s11s1S"</tt></p>
2767<p>
2768The reverse algorithm computes the length of the string just by examining
2769a certain prefix:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2770
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2771<div class="doc_code">
2772<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2773> pref :: [Char] -> Int
2774> pref "S" = 1
2775> pref ('s':'1':rest) = decode 2 1 rest
2776> pref (_:rest) = 1 + pref rest
2777>
2778> decode walk acc ('0':rest) = decode (walk + 1) (acc * 2) rest
2779> decode walk acc ('1':rest) = decode (walk + 1) (acc * 2 + 1) rest
2780> decode walk acc _ = walk + acc
2781>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2782</pre>
2783</div>
2784<p>
2785Now, as expected, printing &lt;pref test&gt; gives <tt>40</tt>.</p>
2786<p>
2787We can <i>quickCheck</i> this with following property:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2788
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2789<div class="doc_code">
2790<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2791> testcase = dist 2000 []
2792> testcaseLength = length testcase
2793>
2794> identityProp n = n > 0 && n <= testcaseLength ==> length arr == pref arr
2795> where arr = takeLast n testcase
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2796>
2797</pre>
2798</div>
2799<p>
2800As expected &lt;quickCheck identityProp&gt; gives:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2801
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2802<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2803*Main> quickCheck identityProp
2804OK, passed 100 tests.
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2805</pre>
2806<p>
2807Let's be a bit more exhaustive:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2808
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2809<div class="doc_code">
2810<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2811>
2812> deepCheck p = check (defaultConfig { configMaxTest = 500 }) p
2813>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2814</pre>
2815</div>
2816<p>
2817And here is the result of &lt;deepCheck identityProp&gt;:</p>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2818
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2819<pre>
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2820*Main> deepCheck identityProp
2821OK, passed 500 tests.
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2822</pre>
2823
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2824<!-- ______________________________________________________________________ -->
2825<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2826 <a name="Tagging">Tagging considerations</a>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2827</div>
2828
2829<p>
2830To maintain the invariant that the 2 LSBits of each <tt>Use**</tt> in <tt>Use</tt>
2831never change after being set up, setters of <tt>Use::Prev</tt> must re-tag the
2832new <tt>Use**</tt> on every modification. Accordingly getters must strip the
2833tag bits.</p>
2834<p>
Gabor Greifd41720a2008-06-25 00:10:22 +0000[diff] [blame]2835For layout b) instead of the <tt>User</tt> we find a pointer (<tt>User*</tt> with LSBit set).
2836Following this pointer brings us to the <tt>User</tt>. A portable trick ensures
2837that the first bytes of <tt>User</tt> (if interpreted as a pointer) never has
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2838the LSBit set. (Portability is relying on the fact that all known compilers place the
2839<tt>vptr</tt> in the first word of the instances.)</p>
Gabor Greifdfed1182008-06-18 13:44:57 +0000[diff] [blame]2840
Gabor Greife98fc272008-06-16 21:06:12 +0000[diff] [blame]2841</div>
2842
2843 <!-- *********************************************************************** -->
Chris Lattnerd9d6e102005-04-23 16:10:52 +0000[diff] [blame]2844<div class="doc_section">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2845 <a name="coreclasses">The Core LLVM Class Hierarchy Reference </a>
2846</div>
2847<!-- *********************************************************************** -->
2848
2849<div class="doc_text">
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2850<p><tt>#include "<a href="/doxygen/Type_8h-source.html">llvm/Type.h</a>"</tt>
2851<br>doxygen info: <a href="/doxygen/classllvm_1_1Type.html">Type Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2852
2853<p>The Core LLVM classes are the primary means of representing the program
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2854being inspected or transformed. The core LLVM classes are defined in
2855header files in the <tt>include/llvm/</tt> directory, and implemented in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]2856the <tt>lib/VMCore</tt> directory.</p>
2857
2858</div>
2859
2860<!-- ======================================================================= -->
2861<div class="doc_subsection">
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2862 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
2863</div>
2864
2865<div class="doc_text">
2866
2867 <p><tt>Type</tt> is a superclass of all type classes. Every <tt>Value</tt> has
2868 a <tt>Type</tt>. <tt>Type</tt> cannot be instantiated directly but only
2869 through its subclasses. Certain primitive types (<tt>VoidType</tt>,
2870 <tt>LabelType</tt>, <tt>FloatType</tt> and <tt>DoubleType</tt>) have hidden
2871 subclasses. They are hidden because they offer no useful functionality beyond
2872 what the <tt>Type</tt> class offers except to distinguish themselves from
2873 other subclasses of <tt>Type</tt>.</p>
2874 <p>All other types are subclasses of <tt>DerivedType</tt>. Types can be
2875 named, but this is not a requirement. There exists exactly
2876 one instance of a given shape at any one time. This allows type equality to
2877 be performed with address equality of the Type Instance. That is, given two
2878 <tt>Type*</tt> values, the types are identical if the pointers are identical.
2879 </p>
2880</div>
2881
2882<!-- _______________________________________________________________________ -->
2883<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2884 <a name="m_Type">Important Public Methods</a>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2885</div>
2886
2887<div class="doc_text">
2888
2889<ul>
Chris Lattner8f79df32007-01-15 01:55:32 +0000[diff] [blame]2890 <li><tt>bool isInteger() const</tt>: Returns true for any integer type.</li>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2891
2892 <li><tt>bool isFloatingPoint()</tt>: Return true if this is one of the two
2893 floating point types.</li>
2894
2895 <li><tt>bool isAbstract()</tt>: Return true if the type is abstract (contains
2896 an OpaqueType anywhere in its definition).</li>
2897
2898 <li><tt>bool isSized()</tt>: Return true if the type has known size. Things
2899 that don't have a size are abstract types, labels and void.</li>
2900
2901</ul>
2902</div>
2903
2904<!-- _______________________________________________________________________ -->
2905<div class="doc_subsubsection">
Gabor Greiffd095b62009-01-05 16:05:32 +0000[diff] [blame]2906 <a name="derivedtypes">Important Derived Types</a>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2907</div>
2908<div class="doc_text">
2909<dl>
2910 <dt><tt>IntegerType</tt></dt>
2911 <dd>Subclass of DerivedType that represents integer types of any bit width.
2912 Any bit width between <tt>IntegerType::MIN_INT_BITS</tt> (1) and
2913 <tt>IntegerType::MAX_INT_BITS</tt> (~8 million) can be represented.
2914 <ul>
2915 <li><tt>static const IntegerType* get(unsigned NumBits)</tt>: get an integer
2916 type of a specific bit width.</li>
2917 <li><tt>unsigned getBitWidth() const</tt>: Get the bit width of an integer
2918 type.</li>
2919 </ul>
2920 </dd>
2921 <dt><tt>SequentialType</tt></dt>
2922 <dd>This is subclassed by ArrayType and PointerType
2923 <ul>
2924 <li><tt>const Type * getElementType() const</tt>: Returns the type of each
2925 of the elements in the sequential type. </li>
2926 </ul>
2927 </dd>
2928 <dt><tt>ArrayType</tt></dt>
2929 <dd>This is a subclass of SequentialType and defines the interface for array
2930 types.
2931 <ul>
2932 <li><tt>unsigned getNumElements() const</tt>: Returns the number of
2933 elements in the array. </li>
2934 </ul>
2935 </dd>
2936 <dt><tt>PointerType</tt></dt>
Chris Lattner302da1e2007-02-03 03:05:57 +0000[diff] [blame]2937 <dd>Subclass of SequentialType for pointer types.</dd>
Reid Spencer9d6565a2007-02-15 02:26:10 +0000[diff] [blame]2938 <dt><tt>VectorType</tt></dt>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2939 <dd>Subclass of SequentialType for vector types. A
2940 vector type is similar to an ArrayType but is distinguished because it is
2941 a first class type wherease ArrayType is not. Vector types are used for
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2942 vector operations and are usually small vectors of of an integer or floating
2943 point type.</dd>
2944 <dt><tt>StructType</tt></dt>
2945 <dd>Subclass of DerivedTypes for struct types.</dd>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]2946 <dt><tt><a name="FunctionType">FunctionType</a></tt></dt>
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]2947 <dd>Subclass of DerivedTypes for function types.
2948 <ul>
2949 <li><tt>bool isVarArg() const</tt>: Returns true if its a vararg
2950 function</li>
2951 <li><tt> const Type * getReturnType() const</tt>: Returns the
2952 return type of the function.</li>
2953 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns
2954 the type of the ith parameter.</li>
2955 <li><tt> const unsigned getNumParams() const</tt>: Returns the
2956 number of formal parameters.</li>
2957 </ul>
2958 </dd>
2959 <dt><tt>OpaqueType</tt></dt>
2960 <dd>Sublcass of DerivedType for abstract types. This class
2961 defines no content and is used as a placeholder for some other type. Note
2962 that OpaqueType is used (temporarily) during type resolution for forward
2963 references of types. Once the referenced type is resolved, the OpaqueType
2964 is replaced with the actual type. OpaqueType can also be used for data
2965 abstraction. At link time opaque types can be resolved to actual types
2966 of the same name.</dd>
2967</dl>
2968</div>
2969
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]2970
2971
2972<!-- ======================================================================= -->
2973<div class="doc_subsection">
2974 <a name="Module">The <tt>Module</tt> class</a>
2975</div>
2976
2977<div class="doc_text">
2978
2979<p><tt>#include "<a
2980href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt><br> doxygen info:
2981<a href="/doxygen/classllvm_1_1Module.html">Module Class</a></p>
2982
2983<p>The <tt>Module</tt> class represents the top level structure present in LLVM
2984programs. An LLVM module is effectively either a translation unit of the
2985original program or a combination of several translation units merged by the
2986linker. The <tt>Module</tt> class keeps track of a list of <a
2987href="#Function"><tt>Function</tt></a>s, a list of <a
2988href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
2989href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
2990helpful member functions that try to make common operations easy.</p>
2991
2992</div>
2993
2994<!-- _______________________________________________________________________ -->
2995<div class="doc_subsubsection">
2996 <a name="m_Module">Important Public Members of the <tt>Module</tt> class</a>
2997</div>
2998
2999<div class="doc_text">
3000
3001<ul>
3002 <li><tt>Module::Module(std::string name = "")</tt></li>
3003</ul>
3004
3005<p>Constructing a <a href="#Module">Module</a> is easy. You can optionally
3006provide a name for it (probably based on the name of the translation unit).</p>
3007
3008<ul>
3009 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
3010 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
3011
3012 <tt>begin()</tt>, <tt>end()</tt>
3013 <tt>size()</tt>, <tt>empty()</tt>
3014
3015 <p>These are forwarding methods that make it easy to access the contents of
3016 a <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
3017 list.</p></li>
3018
3019 <li><tt>Module::FunctionListType &amp;getFunctionList()</tt>
3020
3021 <p> Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
3022 necessary to use when you need to update the list or perform a complex
3023 action that doesn't have a forwarding method.</p>
3024
3025 <p><!-- Global Variable --></p></li>
3026</ul>
3027
3028<hr>
3029
3030<ul>
3031 <li><tt>Module::global_iterator</tt> - Typedef for global variable list iterator<br>
3032
3033 <tt>Module::const_global_iterator</tt> - Typedef for const_iterator.<br>
3034
3035 <tt>global_begin()</tt>, <tt>global_end()</tt>
3036 <tt>global_size()</tt>, <tt>global_empty()</tt>
3037
3038 <p> These are forwarding methods that make it easy to access the contents of
3039 a <tt>Module</tt> object's <a
3040 href="#GlobalVariable"><tt>GlobalVariable</tt></a> list.</p></li>
3041
3042 <li><tt>Module::GlobalListType &amp;getGlobalList()</tt>
3043
3044 <p>Returns the list of <a
3045 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s. This is necessary to
3046 use when you need to update the list or perform a complex action that
3047 doesn't have a forwarding method.</p>
3048
3049 <p><!-- Symbol table stuff --> </p></li>
3050</ul>
3051
3052<hr>
3053
3054<ul>
3055 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
3056
3057 <p>Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
3058 for this <tt>Module</tt>.</p>
3059
3060 <p><!-- Convenience methods --></p></li>
3061</ul>
3062
3063<hr>
3064
3065<ul>
3066 <li><tt><a href="#Function">Function</a> *getFunction(const std::string
3067 &amp;Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt>
3068
3069 <p>Look up the specified function in the <tt>Module</tt> <a
3070 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
3071 <tt>null</tt>.</p></li>
3072
3073 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const
3074 std::string &amp;Name, const <a href="#FunctionType">FunctionType</a> *T)</tt>
3075
3076 <p>Look up the specified function in the <tt>Module</tt> <a
3077 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
3078 external declaration for the function and return it.</p></li>
3079
3080 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt>
3081
3082 <p>If there is at least one entry in the <a
3083 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
3084 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
3085 string.</p></li>
3086
3087 <li><tt>bool addTypeName(const std::string &amp;Name, const <a
3088 href="#Type">Type</a> *Ty)</tt>
3089
3090 <p>Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
3091 mapping <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this
3092 name, true is returned and the <a
3093 href="#SymbolTable"><tt>SymbolTable</tt></a> is not modified.</p></li>
3094</ul>
3095
3096</div>
3097
3098
Reid Spencer303c4b42007-01-12 17:26:25 +0000[diff] [blame]3099<!-- ======================================================================= -->
3100<div class="doc_subsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3101 <a name="Value">The <tt>Value</tt> class</a>
3102</div>
3103
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3104<div class="doc_text">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3105
3106<p><tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt>
3107<br>
Chris Lattner00815172007-01-04 22:01:45 +0000[diff] [blame]3108doxygen info: <a href="/doxygen/classllvm_1_1Value.html">Value Class</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3109
3110<p>The <tt>Value</tt> class is the most important class in the LLVM Source
3111base. It represents a typed value that may be used (among other things) as an
3112operand to an instruction. There are many different types of <tt>Value</tt>s,
3113such as <a href="#Constant"><tt>Constant</tt></a>s,<a
3114href="#Argument"><tt>Argument</tt></a>s. Even <a
3115href="#Instruction"><tt>Instruction</tt></a>s and <a
3116href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.</p>
3117
3118<p>A particular <tt>Value</tt> may be used many times in the LLVM representation
3119for a program. For example, an incoming argument to a function (represented
3120with an instance of the <a href="#Argument">Argument</a> class) is "used" by
3121every instruction in the function that references the argument. To keep track
3122of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
3123href="#User"><tt>User</tt></a>s that is using it (the <a
3124href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
3125graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
3126def-use information in the program, and is accessible through the <tt>use_</tt>*
3127methods, shown below.</p>
3128
3129<p>Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed,
3130and this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
3131method. In addition, all LLVM values can be named. The "name" of the
3132<tt>Value</tt> is a symbolic string printed in the LLVM code:</p>
3133
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3134<div class="doc_code">
3135<pre>
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]3136%<b>foo</b> = add i32 1, 2
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3137</pre>
3138</div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3139
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3140<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]3141that the name of any value may be missing (an empty string), so names should
3142<b>ONLY</b> be used for debugging (making the source code easier to read,
3143debugging printouts), they should not be used to keep track of values or map
3144between them. For this purpose, use a <tt>std::map</tt> of pointers to the
3145<tt>Value</tt> itself instead.</p>
3146
3147<p>One important aspect of LLVM is that there is no distinction between an SSA
3148variable and the operation that produces it. Because of this, any reference to
3149the value produced by an instruction (or the value available as an incoming
Chris Lattnerd5fc4fc2004-03-18 14:58:55 +0000[diff] [blame]3150argument, for example) is represented as a direct pointer to the instance of
3151the class that
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3152represents this value. Although this may take some getting used to, it
3153simplifies the representation and makes it easier to manipulate.</p>
3154
3155</div>
3156
3157<!-- _______________________________________________________________________ -->
3158<div class="doc_subsubsection">
3159 <a name="m_Value">Important Public Members of the <tt>Value</tt> class</a>
3160</div>
3161
3162<div class="doc_text">
3163
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3164<ul>
3165 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the
3166use-list<br>
3167 <tt>Value::use_const_iterator</tt> - Typedef for const_iterator over
3168the use-list<br>
3169 <tt>unsigned use_size()</tt> - Returns the number of users of the
3170value.<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3171 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3172 <tt>use_iterator use_begin()</tt> - Get an iterator to the start of
3173the use-list.<br>
3174 <tt>use_iterator use_end()</tt> - Get an iterator to the end of the
3175use-list.<br>
3176 <tt><a href="#User">User</a> *use_back()</tt> - Returns the last
3177element in the list.
3178 <p> These methods are the interface to access the def-use
3179information in LLVM. As with all other iterators in LLVM, the naming
3180conventions follow the conventions defined by the <a href="#stl">STL</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3181 </li>
3182 <li><tt><a href="#Type">Type</a> *getType() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3183 <p>This method returns the Type of the Value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3184 </li>
3185 <li><tt>bool hasName() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3186 <tt>std::string getName() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3187 <tt>void setName(const std::string &amp;Name)</tt>
3188 <p> This family of methods is used to access and assign a name to a <tt>Value</tt>,
3189be aware of the <a href="#nameWarning">precaution above</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3190 </li>
3191 <li><tt>void replaceAllUsesWith(Value *V)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3192
3193 <p>This method traverses the use list of a <tt>Value</tt> changing all <a
3194 href="#User"><tt>User</tt>s</a> of the current value to refer to
3195 "<tt>V</tt>" instead. For example, if you detect that an instruction always
3196 produces a constant value (for example through constant folding), you can
3197 replace all uses of the instruction with the constant like this:</p>
3198
Bill Wendling3cd5ca62006-10-11 06:30:10 +0000[diff] [blame]3199<div class="doc_code">
3200<pre>
3201Inst-&gt;replaceAllUsesWith(ConstVal);
3202</pre>
3203</div>
3204
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3205</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3206
3207</div>
3208
3209<!-- ======================================================================= -->
3210<div class="doc_subsection">
3211 <a name="User">The <tt>User</tt> class</a>
3212</div>
3213
3214<div class="doc_text">
3215
3216<p>
3217<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]3218doxygen info: <a href="/doxygen/classllvm_1_1User.html">User Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3219Superclass: <a href="#Value"><tt>Value</tt></a></p>
3220
3221<p>The <tt>User</tt> class is the common base class of all LLVM nodes that may
3222refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
3223that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
3224referring to. The <tt>User</tt> class itself is a subclass of
3225<tt>Value</tt>.</p>
3226
3227<p>The operands of a <tt>User</tt> point directly to the LLVM <a
3228href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
3229Single Assignment (SSA) form, there can only be one definition referred to,
3230allowing this direct connection. This connection provides the use-def
3231information in LLVM.</p>
3232
3233</div>
3234
3235<!-- _______________________________________________________________________ -->
3236<div class="doc_subsubsection">
3237 <a name="m_User">Important Public Members of the <tt>User</tt> class</a>
3238</div>
3239
3240<div class="doc_text">
3241
3242<p>The <tt>User</tt> class exposes the operand list in two ways: through
3243an index access interface and through an iterator based interface.</p>
3244
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3245<ul>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3246 <li><tt>Value *getOperand(unsigned i)</tt><br>
3247 <tt>unsigned getNumOperands()</tt>
3248 <p> These two methods expose the operands of the <tt>User</tt> in a
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3249convenient form for direct access.</p></li>
3250
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3251 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand
3252list<br>
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]3253 <tt>op_iterator op_begin()</tt> - Get an iterator to the start of
3254the operand list.<br>
3255 <tt>op_iterator op_end()</tt> - Get an iterator to the end of the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3256operand list.
3257 <p> Together, these methods make up the iterator based interface to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3258the operands of a <tt>User</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3259</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3260
3261</div>
3262
3263<!-- ======================================================================= -->
3264<div class="doc_subsection">
3265 <a name="Instruction">The <tt>Instruction</tt> class</a>
3266</div>
3267
3268<div class="doc_text">
3269
3270<p><tt>#include "</tt><tt><a
3271href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt><br>
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3272doxygen info: <a href="/doxygen/classllvm_1_1Instruction.html">Instruction Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3273Superclasses: <a href="#User"><tt>User</tt></a>, <a
3274href="#Value"><tt>Value</tt></a></p>
3275
3276<p>The <tt>Instruction</tt> class is the common base class for all LLVM
3277instructions. It provides only a few methods, but is a very commonly used
3278class. The primary data tracked by the <tt>Instruction</tt> class itself is the
3279opcode (instruction type) and the parent <a
3280href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
3281into. To represent a specific type of instruction, one of many subclasses of
3282<tt>Instruction</tt> are used.</p>
3283
3284<p> Because the <tt>Instruction</tt> class subclasses the <a
3285href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
3286way as for other <a href="#User"><tt>User</tt></a>s (with the
3287<tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
3288<tt>op_begin()</tt>/<tt>op_end()</tt> methods).</p> <p> An important file for
3289the <tt>Instruction</tt> class is the <tt>llvm/Instruction.def</tt> file. This
3290file contains some meta-data about the various different types of instructions
3291in LLVM. It describes the enum values that are used as opcodes (for example
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3292<tt>Instruction::Add</tt> and <tt>Instruction::ICmp</tt>), as well as the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3293concrete sub-classes of <tt>Instruction</tt> that implement the instruction (for
3294example <tt><a href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3295href="#CmpInst">CmpInst</a></tt>). Unfortunately, the use of macros in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3296this file confuses doxygen, so these enum values don't show up correctly in the
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3297<a href="/doxygen/classllvm_1_1Instruction.html">doxygen output</a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3298
3299</div>
3300
3301<!-- _______________________________________________________________________ -->
3302<div class="doc_subsubsection">
Reid Spencerc92d25d2006-12-19 19:47:19 +0000[diff] [blame]3303 <a name="s_Instruction">Important Subclasses of the <tt>Instruction</tt>
3304 class</a>
3305</div>
3306<div class="doc_text">
3307 <ul>
3308 <li><tt><a name="BinaryOperator">BinaryOperator</a></tt>
3309 <p>This subclasses represents all two operand instructions whose operands
3310 must be the same type, except for the comparison instructions.</p></li>
3311 <li><tt><a name="CastInst">CastInst</a></tt>
3312 <p>This subclass is the parent of the 12 casting instructions. It provides
3313 common operations on cast instructions.</p>
3314 <li><tt><a name="CmpInst">CmpInst</a></tt>
3315 <p>This subclass respresents the two comparison instructions,
3316 <a href="LangRef.html#i_icmp">ICmpInst</a> (integer opreands), and
3317 <a href="LangRef.html#i_fcmp">FCmpInst</a> (floating point operands).</p>
3318 <li><tt><a name="TerminatorInst">TerminatorInst</a></tt>
3319 <p>This subclass is the parent of all terminator instructions (those which
3320 can terminate a block).</p>
3321 </ul>
3322 </div>
3323
3324<!-- _______________________________________________________________________ -->
3325<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3326 <a name="m_Instruction">Important Public Members of the <tt>Instruction</tt>
3327 class</a>
3328</div>
3329
3330<div class="doc_text">
3331
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3332<ul>
3333 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3334 <p>Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that
3335this <tt>Instruction</tt> is embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3336 <li><tt>bool mayWriteToMemory()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3337 <p>Returns true if the instruction writes to memory, i.e. it is a
3338 <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]3339 <li><tt>unsigned getOpcode()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3340 <p>Returns the opcode for the <tt>Instruction</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3341 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3342 <p>Returns another instance of the specified instruction, identical
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3343in all ways to the original except that the instruction has no parent
3344(ie it's not embedded into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>),
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3345and it has no name</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3346</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3347
3348</div>
3349
3350<!-- ======================================================================= -->
3351<div class="doc_subsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3352 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3353</div>
3354
3355<div class="doc_text">
3356
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3357<p>Constant represents a base class for different types of constants. It
3358is subclassed by ConstantInt, ConstantArray, etc. for representing
3359the various types of Constants. <a href="#GlobalValue">GlobalValue</a> is also
3360a subclass, which represents the address of a global variable or function.
3361</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3362
3363</div>
3364
3365<!-- _______________________________________________________________________ -->
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3366<div class="doc_subsubsection">Important Subclasses of Constant </div>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3367<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3368<ul>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3369 <li>ConstantInt : This subclass of Constant represents an integer constant of
3370 any width.
3371 <ul>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]3372 <li><tt>const APInt&amp; getValue() const</tt>: Returns the underlying
3373 value of this constant, an APInt value.</li>
3374 <li><tt>int64_t getSExtValue() const</tt>: Converts the underlying APInt
3375 value to an int64_t via sign extension. If the value (not the bit width)
3376 of the APInt is too large to fit in an int64_t, an assertion will result.
3377 For this reason, use of this method is discouraged.</li>
3378 <li><tt>uint64_t getZExtValue() const</tt>: Converts the underlying APInt
3379 value to a uint64_t via zero extension. IF the value (not the bit width)
3380 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]3381 For this reason, use of this method is discouraged.</li>
Reid Spencer97b4ee32007-03-01 21:05:33 +0000[diff] [blame]3382 <li><tt>static ConstantInt* get(const APInt&amp; Val)</tt>: Returns the
3383 ConstantInt object that represents the value provided by <tt>Val</tt>.
3384 The type is implied as the IntegerType that corresponds to the bit width
3385 of <tt>Val</tt>.</li>
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3386 <li><tt>static ConstantInt* get(const Type *Ty, uint64_t Val)</tt>:
3387 Returns the ConstantInt object that represents the value provided by
3388 <tt>Val</tt> for integer type <tt>Ty</tt>.</li>
3389 </ul>
3390 </li>
3391 <li>ConstantFP : This class represents a floating point constant.
3392 <ul>
3393 <li><tt>double getValue() const</tt>: Returns the underlying value of
3394 this constant. </li>
3395 </ul>
3396 </li>
3397 <li>ConstantArray : This represents a constant array.
3398 <ul>
3399 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
3400 a vector of component constants that makeup this array. </li>
3401 </ul>
3402 </li>
3403 <li>ConstantStruct : This represents a constant struct.
3404 <ul>
3405 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>: Returns
3406 a vector of component constants that makeup this array. </li>
3407 </ul>
3408 </li>
3409 <li>GlobalValue : This represents either a global variable or a function. In
3410 either case, the value is a constant fixed address (after linking).
3411 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3412</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3413</div>
3414
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3415
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3416<!-- ======================================================================= -->
3417<div class="doc_subsection">
3418 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
3419</div>
3420
3421<div class="doc_text">
3422
3423<p><tt>#include "<a
3424href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt><br>
Misha Brukman384047f2004-06-03 23:29:12 +0000[diff] [blame]3425doxygen info: <a href="/doxygen/classllvm_1_1GlobalValue.html">GlobalValue
3426Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3427Superclasses: <a href="#Constant"><tt>Constant</tt></a>,
3428<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]3429
3430<p>Global values (<a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
3431href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
3432visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
3433Because they are visible at global scope, they are also subject to linking with
3434other globals defined in different translation units. To control the linking
3435process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
3436<tt>GlobalValue</tt>s know whether they have internal or external linkage, as
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]3437defined by the <tt>LinkageTypes</tt> enumeration.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3438
3439<p>If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
3440<tt>static</tt> in C), it is not visible to code outside the current translation
3441unit, and does not participate in linking. If it has external linkage, it is
3442visible to external code, and does participate in linking. In addition to
3443linkage information, <tt>GlobalValue</tt>s keep track of which <a
3444href="#Module"><tt>Module</tt></a> they are currently part of.</p>
3445
3446<p>Because <tt>GlobalValue</tt>s are memory objects, they are always referred to
3447by their <b>address</b>. As such, the <a href="#Type"><tt>Type</tt></a> of a
3448global is always a pointer to its contents. It is important to remember this
3449when using the <tt>GetElementPtrInst</tt> instruction because this pointer must
3450be dereferenced first. For example, if you have a <tt>GlobalVariable</tt> (a
3451subclass 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]3452i32]</tt>, then the <tt>GlobalVariable</tt> is a pointer to that array. Although
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3453the address of the first element of this array and the value of the
3454<tt>GlobalVariable</tt> are the same, they have different types. The
Reid Spencer06565dc2007-01-12 17:11:23 +0000[diff] [blame]3455<tt>GlobalVariable</tt>'s type is <tt>[24 x i32]</tt>. The first element's type
3456is <tt>i32.</tt> Because of this, accessing a global value requires you to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3457dereference the pointer with <tt>GetElementPtrInst</tt> first, then its elements
3458can be accessed. This is explained in the <a href="LangRef.html#globalvars">LLVM
3459Language Reference Manual</a>.</p>
3460
3461</div>
3462
3463<!-- _______________________________________________________________________ -->
3464<div class="doc_subsubsection">
3465 <a name="m_GlobalValue">Important Public Members of the <tt>GlobalValue</tt>
3466 class</a>
3467</div>
3468
3469<div class="doc_text">
3470
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3471<ul>
3472 <li><tt>bool hasInternalLinkage() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3473 <tt>bool hasExternalLinkage() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3474 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt>
3475 <p> These methods manipulate the linkage characteristics of the <tt>GlobalValue</tt>.</p>
3476 <p> </p>
3477 </li>
3478 <li><tt><a href="#Module">Module</a> *getParent()</tt>
3479 <p> This returns the <a href="#Module"><tt>Module</tt></a> that the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3480GlobalValue is currently embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3481</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3482
3483</div>
3484
3485<!-- ======================================================================= -->
3486<div class="doc_subsection">
3487 <a name="Function">The <tt>Function</tt> class</a>
3488</div>
3489
3490<div class="doc_text">
3491
3492<p><tt>#include "<a
3493href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt><br> doxygen
Misha Brukman31ca1de2004-06-03 23:35:54 +0000[diff] [blame]3494info: <a href="/doxygen/classllvm_1_1Function.html">Function Class</a><br>
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3495Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
3496<a href="#Constant"><tt>Constant</tt></a>,
3497<a href="#User"><tt>User</tt></a>,
3498<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3499
3500<p>The <tt>Function</tt> class represents a single procedure in LLVM. It is
3501actually one of the more complex classes in the LLVM heirarchy because it must
3502keep track of a large amount of data. The <tt>Function</tt> class keeps track
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3503of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal
3504<a href="#Argument"><tt>Argument</tt></a>s, and a
3505<a href="#SymbolTable"><tt>SymbolTable</tt></a>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3506
3507<p>The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most
3508commonly used part of <tt>Function</tt> objects. The list imposes an implicit
3509ordering of the blocks in the function, which indicate how the code will be
3510layed out by the backend. Additionally, the first <a
3511href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
3512<tt>Function</tt>. It is not legal in LLVM to explicitly branch to this initial
3513block. There are no implicit exit nodes, and in fact there may be multiple exit
3514nodes from a single <tt>Function</tt>. If the <a
3515href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
3516the <tt>Function</tt> is actually a function declaration: the actual body of the
3517function hasn't been linked in yet.</p>
3518
3519<p>In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
3520<tt>Function</tt> class also keeps track of the list of formal <a
3521href="#Argument"><tt>Argument</tt></a>s that the function receives. This
3522container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
3523nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
3524the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.</p>
3525
3526<p>The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used
3527LLVM feature that is only used when you have to look up a value by name. Aside
3528from that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used
3529internally to make sure that there are not conflicts between the names of <a
3530href="#Instruction"><tt>Instruction</tt></a>s, <a
3531href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
3532href="#Argument"><tt>Argument</tt></a>s in the function body.</p>
3533
Reid Spencer8b2da7a2004-07-18 13:10:31 +0000[diff] [blame]3534<p>Note that <tt>Function</tt> is a <a href="#GlobalValue">GlobalValue</a>
3535and therefore also a <a href="#Constant">Constant</a>. The value of the function
3536is its address (after linking) which is guaranteed to be constant.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3537</div>
3538
3539<!-- _______________________________________________________________________ -->
3540<div class="doc_subsubsection">
3541 <a name="m_Function">Important Public Members of the <tt>Function</tt>
3542 class</a>
3543</div>
3544
3545<div class="doc_text">
3546
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3547<ul>
3548 <li><tt>Function(const </tt><tt><a href="#FunctionType">FunctionType</a>
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]3549 *Ty, LinkageTypes Linkage, const std::string &amp;N = "", Module* Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3550
3551 <p>Constructor used when you need to create new <tt>Function</tt>s to add
3552 the the program. The constructor must specify the type of the function to
Chris Lattnerac479e52004-08-04 05:10:48 +0000[diff] [blame]3553 create and what type of linkage the function should have. The <a
3554 href="#FunctionType"><tt>FunctionType</tt></a> argument
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3555 specifies the formal arguments and return value for the function. The same
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3556 <a href="#FunctionType"><tt>FunctionType</tt></a> value can be used to
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3557 create multiple functions. The <tt>Parent</tt> argument specifies the Module
3558 in which the function is defined. If this argument is provided, the function
3559 will automatically be inserted into that module's list of
3560 functions.</p></li>
3561
Chris Lattner62810e32008-11-25 18:34:50 +0000[diff] [blame]3562 <li><tt>bool isDeclaration()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3563
3564 <p>Return whether or not the <tt>Function</tt> has a body defined. If the
3565 function is "external", it does not have a body, and thus must be resolved
3566 by linking with a function defined in a different translation unit.</p></li>
3567
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3568 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3569 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3570
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]3571 <tt>begin()</tt>, <tt>end()</tt>
3572 <tt>size()</tt>, <tt>empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3573
3574 <p>These are forwarding methods that make it easy to access the contents of
3575 a <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
3576 list.</p></li>
3577
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3578 <li><tt>Function::BasicBlockListType &amp;getBasicBlockList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3579
3580 <p>Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This
3581 is necessary to use when you need to update the list or perform a complex
3582 action that doesn't have a forwarding method.</p></li>
3583
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3584 <li><tt>Function::arg_iterator</tt> - Typedef for the argument list
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3585iterator<br>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3586 <tt>Function::const_arg_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3587
Chris Lattner77d69242005-03-15 05:19:20 +0000[diff] [blame]3588 <tt>arg_begin()</tt>, <tt>arg_end()</tt>
Chris Lattner89cc2652005-03-15 04:48:32 +0000[diff] [blame]3589 <tt>arg_size()</tt>, <tt>arg_empty()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3590
3591 <p>These are forwarding methods that make it easy to access the contents of
3592 a <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a>
3593 list.</p></li>
3594
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3595 <li><tt>Function::ArgumentListType &amp;getArgumentList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3596
3597 <p>Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
3598 necessary to use when you need to update the list or perform a complex
3599 action that doesn't have a forwarding method.</p></li>
3600
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3601 <li><tt><a href="#BasicBlock">BasicBlock</a> &amp;getEntryBlock()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3602
3603 <p>Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
3604 function. Because the entry block for the function is always the first
3605 block, this returns the first block of the <tt>Function</tt>.</p></li>
3606
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3607 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
3608 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3609
3610 <p>This traverses the <a href="#Type"><tt>Type</tt></a> of the
3611 <tt>Function</tt> and returns the return type of the function, or the <a
3612 href="#FunctionType"><tt>FunctionType</tt></a> of the actual
3613 function.</p></li>
3614
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3615 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3616
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3617 <p> Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3618 for this <tt>Function</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3619</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3620
3621</div>
3622
3623<!-- ======================================================================= -->
3624<div class="doc_subsection">
3625 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
3626</div>
3627
3628<div class="doc_text">
3629
3630<p><tt>#include "<a
3631href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt>
3632<br>
Tanya Lattnera3da7772004-06-22 08:02:25 +0000[diff] [blame]3633doxygen info: <a href="/doxygen/classllvm_1_1GlobalVariable.html">GlobalVariable
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3634 Class</a><br>
3635Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
3636<a href="#Constant"><tt>Constant</tt></a>,
3637<a href="#User"><tt>User</tt></a>,
3638<a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3639
3640<p>Global variables are represented with the (suprise suprise)
3641<tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are also
3642subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such are
3643always referenced by their address (global values must live in memory, so their
Reid Spencerbe5e85e2006-04-14 14:11:48 +0000[diff] [blame]3644"name" refers to their constant address). See
3645<a href="#GlobalValue"><tt>GlobalValue</tt></a> for more on this. Global
3646variables may have an initial value (which must be a
3647<a href="#Constant"><tt>Constant</tt></a>), and if they have an initializer,
3648they may be marked as "constant" themselves (indicating that their contents
3649never change at runtime).</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3650</div>
3651
3652<!-- _______________________________________________________________________ -->
3653<div class="doc_subsubsection">
3654 <a name="m_GlobalVariable">Important Public Members of the
3655 <tt>GlobalVariable</tt> class</a>
3656</div>
3657
3658<div class="doc_text">
3659
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3660<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3661 <li><tt>GlobalVariable(const </tt><tt><a href="#Type">Type</a> *Ty, bool
3662 isConstant, LinkageTypes&amp; Linkage, <a href="#Constant">Constant</a>
3663 *Initializer = 0, const std::string &amp;Name = "", Module* Parent = 0)</tt>
3664
3665 <p>Create a new global variable of the specified type. If
3666 <tt>isConstant</tt> is true then the global variable will be marked as
3667 unchanging for the program. The Linkage parameter specifies the type of
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]3668 linkage (internal, external, weak, linkonce, appending) for the variable.
3669 If the linkage is InternalLinkage, WeakAnyLinkage, WeakODRLinkage,
3670 LinkOnceAnyLinkage or LinkOnceODRLinkage,&nbsp; then the resultant
3671 global variable will have internal linkage. AppendingLinkage concatenates
3672 together all instances (in different translation units) of the variable
3673 into a single variable but is only applicable to arrays. &nbsp;See
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3674 the <a href="LangRef.html#modulestructure">LLVM Language Reference</a> for
3675 further details on linkage types. Optionally an initializer, a name, and the
3676 module to put the variable into may be specified for the global variable as
3677 well.</p></li>
3678
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3679 <li><tt>bool isConstant() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3680
3681 <p>Returns true if this is a global variable that is known not to
3682 be modified at runtime.</p></li>
3683
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3684 <li><tt>bool hasInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3685
3686 <p>Returns true if this <tt>GlobalVariable</tt> has an intializer.</p></li>
3687
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3688 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3689
3690 <p>Returns the intial value for a <tt>GlobalVariable</tt>. It is not legal
3691 to call this method if there is no initializer.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3692</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3693
3694</div>
3695
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3696
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3697<!-- ======================================================================= -->
3698<div class="doc_subsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3699 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3700</div>
3701
3702<div class="doc_text">
3703
3704<p><tt>#include "<a
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3705href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt><br>
Stefanus Du Toit24e04112009-06-17 21:12:26 +0000[diff] [blame]3706doxygen info: <a href="/doxygen/classllvm_1_1BasicBlock.html">BasicBlock
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3707Class</a><br>
3708Superclass: <a href="#Value"><tt>Value</tt></a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3709
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3710<p>This class represents a single entry multiple exit section of the code,
3711commonly known as a basic block by the compiler community. The
3712<tt>BasicBlock</tt> class maintains a list of <a
3713href="#Instruction"><tt>Instruction</tt></a>s, which form the body of the block.
3714Matching the language definition, the last element of this list of instructions
3715is always a terminator instruction (a subclass of the <a
3716href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).</p>
3717
3718<p>In addition to tracking the list of instructions that make up the block, the
3719<tt>BasicBlock</tt> class also keeps track of the <a
3720href="#Function"><tt>Function</tt></a> that it is embedded into.</p>
3721
3722<p>Note that <tt>BasicBlock</tt>s themselves are <a
3723href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
3724like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
3725<tt>label</tt>.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3726
3727</div>
3728
3729<!-- _______________________________________________________________________ -->
3730<div class="doc_subsubsection">
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3731 <a name="m_BasicBlock">Important Public Members of the <tt>BasicBlock</tt>
3732 class</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3733</div>
3734
3735<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3736<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3737
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3738<li><tt>BasicBlock(const std::string &amp;Name = "", </tt><tt><a
3739 href="#Function">Function</a> *Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3740
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3741<p>The <tt>BasicBlock</tt> constructor is used to create new basic blocks for
3742insertion into a function. The constructor optionally takes a name for the new
3743block, and a <a href="#Function"><tt>Function</tt></a> to insert it into. If
3744the <tt>Parent</tt> parameter is specified, the new <tt>BasicBlock</tt> is
3745automatically inserted at the end of the specified <a
3746href="#Function"><tt>Function</tt></a>, if not specified, the BasicBlock must be
3747manually inserted into the <a href="#Function"><tt>Function</tt></a>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3748
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3749<li><tt>BasicBlock::iterator</tt> - Typedef for instruction list iterator<br>
3750<tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
3751<tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
3752<tt>size()</tt>, <tt>empty()</tt>
3753STL-style functions for accessing the instruction list.
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3754
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3755<p>These methods and typedefs are forwarding functions that have the same
3756semantics as the standard library methods of the same names. These methods
3757expose the underlying instruction list of a basic block in a way that is easy to
3758manipulate. To get the full complement of container operations (including
3759operations to update the list), you must use the <tt>getInstList()</tt>
3760method.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3761
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3762<li><tt>BasicBlock::InstListType &amp;getInstList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3763
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3764<p>This method is used to get access to the underlying container that actually
3765holds the Instructions. This method must be used when there isn't a forwarding
3766function in the <tt>BasicBlock</tt> class for the operation that you would like
3767to perform. Because there are no forwarding functions for "updating"
3768operations, you need to use this if you want to update the contents of a
3769<tt>BasicBlock</tt>.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3770
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3771<li><tt><a href="#Function">Function</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3772
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3773<p> Returns a pointer to <a href="#Function"><tt>Function</tt></a> the block is
3774embedded into, or a null pointer if it is homeless.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3775
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3776<li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3777
Chris Lattner2b78d962007-02-03 20:02:25 +0000[diff] [blame]3778<p> Returns a pointer to the terminator instruction that appears at the end of
3779the <tt>BasicBlock</tt>. If there is no terminator instruction, or if the last
3780instruction in the block is not a terminator, then a null pointer is
3781returned.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3782
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3783</ul>
3784
3785</div>
3786
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3787
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3788<!-- ======================================================================= -->
3789<div class="doc_subsection">
3790 <a name="Argument">The <tt>Argument</tt> class</a>
3791</div>
3792
3793<div class="doc_text">
3794
3795<p>This subclass of Value defines the interface for incoming formal
Chris Lattner58360822005-01-17 00:12:04 +0000[diff] [blame]3796arguments to a function. A Function maintains a list of its formal
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3797arguments. An argument has a pointer to the parent Function.</p>
3798
3799</div>
3800
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]3801<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3802<hr>
3803<address>
3804 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]3805 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3806 <a href="http://validator.w3.org/check/referer"><img
Gabor Greifa9c0f2b2008-06-18 14:05:31 +0000[diff] [blame]3807 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]3808
3809 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
3810 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]3811 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000[diff] [blame]3812 Last modified: $Date$
3813</address>
3814
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3815</body>
3816</html>