blob: 8c678a30c06b2c99a58762d56258dfa12a6d879b [file] [log] [blame]
Misha Brukman13fd15c2004-01-15 00:14:41 +00001<!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 +00003<html>
4<head>
5 <title>LLVM Programmer's Manual</title>
Misha Brukman13fd15c2004-01-15 00:14:41 +00006 <link rel="stylesheet" href="llvm.css" type="text/css">
Chris Lattner261efe92003-11-25 01:02:51 +00007</head>
Misha Brukman13fd15c2004-01-15 00:14:41 +00008<body>
9
10<div class="doc_title">
11 LLVM Programmer's Manual
12</div>
13
Chris Lattner9355b472002-09-06 02:50:58 +000014<ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +000015 <li><a href="#introduction">Introduction</a></li>
Chris Lattner9355b472002-09-06 02:50:58 +000016 <li><a href="#general">General Information</a>
Chris Lattner261efe92003-11-25 01:02:51 +000017 <ul>
18 <li><a href="#stl">The C++ Standard Template Library</a><!--
Chris Lattner986e0c92002-09-22 19:38:40 +000019 <li>The <tt>-time-passes</tt> option
20 <li>How to use the LLVM Makefile system
21 <li>How to write a regression test
Chris Lattner261efe92003-11-25 01:02:51 +000022--> </li>
Chris Lattner84b7f8d2003-08-01 22:20:59 +000023 </ul>
Chris Lattner261efe92003-11-25 01:02:51 +000024 </li>
25 <li><a href="#apis">Important and useful LLVM APIs</a>
26 <ul>
27 <li><a href="#isa">The <tt>isa&lt;&gt;</tt>, <tt>cast&lt;&gt;</tt>
28and <tt>dyn_cast&lt;&gt;</tt> templates</a> </li>
29 <li><a href="#DEBUG">The <tt>DEBUG()</tt> macro &amp; <tt>-debug</tt>
30option</a>
31 <ul>
32 <li><a href="#DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt>
33and the <tt>-debug-only</tt> option</a> </li>
34 </ul>
35 </li>
36 <li><a href="#Statistic">The <tt>Statistic</tt> template &amp; <tt>-stats</tt>
37option</a><!--
Chris Lattner986e0c92002-09-22 19:38:40 +000038 <li>The <tt>InstVisitor</tt> template
39 <li>The general graph API
Chris Lattner261efe92003-11-25 01:02:51 +000040--> </li>
41 </ul>
42 </li>
Chris Lattnerae7f7592002-09-06 18:31:18 +000043 <li><a href="#common">Helpful Hints for Common Operations</a>
Chris Lattnerae7f7592002-09-06 18:31:18 +000044 <ul>
Chris Lattner261efe92003-11-25 01:02:51 +000045 <li><a href="#inspection">Basic Inspection and Traversal Routines</a>
46 <ul>
47 <li><a href="#iterate_function">Iterating over the <tt>BasicBlock</tt>s
48in a <tt>Function</tt></a> </li>
49 <li><a href="#iterate_basicblock">Iterating over the <tt>Instruction</tt>s
50in a <tt>BasicBlock</tt></a> </li>
51 <li><a href="#iterate_institer">Iterating over the <tt>Instruction</tt>s
52in a <tt>Function</tt></a> </li>
53 <li><a href="#iterate_convert">Turning an iterator into a
54class pointer</a> </li>
55 <li><a href="#iterate_complex">Finding call sites: a more
56complex example</a> </li>
57 <li><a href="#calls_and_invokes">Treating calls and invokes
58the same way</a> </li>
59 <li><a href="#iterate_chains">Iterating over def-use &amp;
60use-def chains</a> </li>
61 </ul>
62 </li>
63 <li><a href="#simplechanges">Making simple changes</a>
64 <ul>
65 <li><a href="#schanges_creating">Creating and inserting new
66 <tt>Instruction</tt>s</a> </li>
67 <li><a href="#schanges_deleting">Deleting <tt>Instruction</tt>s</a> </li>
68 <li><a href="#schanges_replacing">Replacing an <tt>Instruction</tt>
69with another <tt>Value</tt></a> </li>
70 </ul>
Chris Lattnerae7f7592002-09-06 18:31:18 +000071<!--
72 <li>Working with the Control Flow Graph
73 <ul>
74 <li>Accessing predecessors and successors of a <tt>BasicBlock</tt>
75 <li>
76 <li>
77 </ul>
Chris Lattner261efe92003-11-25 01:02:51 +000078--> </li>
79 </ul>
80 </li>
Joel Stanley9b96c442002-09-06 21:55:13 +000081 <li><a href="#coreclasses">The Core LLVM Class Hierarchy Reference</a>
Chris Lattner9355b472002-09-06 02:50:58 +000082 <ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +000083 <li><a href="#Value">The <tt>Value</tt> class</a>
84 <ul>
85 <li><a href="#User">The <tt>User</tt> class</a>
Chris Lattner9355b472002-09-06 02:50:58 +000086 <ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +000087 <li><a href="#Instruction">The <tt>Instruction</tt> class</a>
Chris Lattner261efe92003-11-25 01:02:51 +000088 <ul>
Reid Spencer096603a2004-05-26 08:41:35 +000089 <li><a href="#GetElementPtrInst">The <tt>GetElementPtrInst</tt>
90 class</a></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +000091 </ul></li>
92 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
93 <ul>
Reid Spencer096603a2004-05-26 08:41:35 +000094 <li><a href="#BasicBlock">The <tt>BasicBlock</tt>class</a></li>
95 <li><a href="#Function">The <tt>Function</tt> class</a></li>
96 <li><a href="#GlobalVariable">The <tt>GlobalVariable</tt> class
97 </a></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +000098 </ul></li>
99 <li><a href="#Module">The <tt>Module</tt> class</a></li>
Reid Spencer096603a2004-05-26 08:41:35 +0000100 <li><a href="#Constant">The <tt>Constant</tt> class</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000101 <li><a href="#Type">The <tt>Type</tt> class</a> </li>
Reid Spencer096603a2004-05-26 08:41:35 +0000102 <li><a href="#Argument">The <tt>Argument</tt> class</a></li>
103 </ul></li>
104 </ul></li>
105 <li><a href="#SymbolTable">The <tt>SymbolTable</tt> class </a></li>
106 <li>The <tt>ilist</tt> and <tt>iplist</tt> classes
107 <ul>
108 <li>Creating, inserting, moving and deleting from LLVM lists </li>
109 </ul>
110 </li>
111 <li>Important iterator invalidation semantics to be aware of.</li>
Chris Lattner261efe92003-11-25 01:02:51 +0000112 </li>
Chris Lattner9355b472002-09-06 02:50:58 +0000113</ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000114
Chris Lattner69bf8a92004-05-23 21:06:58 +0000115<div class="doc_author">
116 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
Chris Lattner94c43592004-05-26 16:52:55 +0000117 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>,
118 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a>, and
119 <a href="mailto:rspencer@x10sys.com">Reid Spencer</a></p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000120</div>
121
Chris Lattner9355b472002-09-06 02:50:58 +0000122<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000123<div class="doc_section">
124 <a name="introduction">Introduction </a>
125</div>
Chris Lattner9355b472002-09-06 02:50:58 +0000126<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000127
128<div class="doc_text">
129
130<p>This document is meant to highlight some of the important classes and
Chris Lattner261efe92003-11-25 01:02:51 +0000131interfaces available in the LLVM source-base. This manual is not
132intended to explain what LLVM is, how it works, and what LLVM code looks
133like. It assumes that you know the basics of LLVM and are interested
134in writing transformations or otherwise analyzing or manipulating the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000135code.</p>
136
137<p>This document should get you oriented so that you can find your
Chris Lattner261efe92003-11-25 01:02:51 +0000138way in the continuously growing source code that makes up the LLVM
139infrastructure. Note that this manual is not intended to serve as a
140replacement for reading the source code, so if you think there should be
141a method in one of these classes to do something, but it's not listed,
142check the source. Links to the <a href="/doxygen/">doxygen</a> sources
143are provided to make this as easy as possible.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000144
145<p>The first section of this document describes general information that is
146useful to know when working in the LLVM infrastructure, and the second describes
147the Core LLVM classes. In the future this manual will be extended with
148information describing how to use extension libraries, such as dominator
149information, CFG traversal routines, and useful utilities like the <tt><a
150href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.</p>
151
152</div>
153
Chris Lattner9355b472002-09-06 02:50:58 +0000154<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000155<div class="doc_section">
156 <a name="general">General Information</a>
157</div>
158<!-- *********************************************************************** -->
159
160<div class="doc_text">
161
162<p>This section contains general information that is useful if you are working
163in the LLVM source-base, but that isn't specific to any particular API.</p>
164
165</div>
166
167<!-- ======================================================================= -->
168<div class="doc_subsection">
169 <a name="stl">The C++ Standard Template Library</a>
170</div>
171
172<div class="doc_text">
173
174<p>LLVM makes heavy use of the C++ Standard Template Library (STL),
Chris Lattner261efe92003-11-25 01:02:51 +0000175perhaps much more than you are used to, or have seen before. Because of
176this, you might want to do a little background reading in the
177techniques used and capabilities of the library. There are many good
178pages that discuss the STL, and several books on the subject that you
Misha Brukman13fd15c2004-01-15 00:14:41 +0000179can get, so it will not be discussed in this document.</p>
180
181<p>Here are some useful links:</p>
182
183<ol>
184
185<li><a href="http://www.dinkumware.com/refxcpp.html">Dinkumware C++ Library
186reference</a> - an excellent reference for the STL and other parts of the
187standard C++ library.</li>
188
189<li><a href="http://www.tempest-sw.com/cpp/">C++ In a Nutshell</a> - This is an
190O'Reilly book in the making. It has a decent <a
191href="http://www.tempest-sw.com/cpp/ch13-libref.html">Standard Library
192Reference</a> that rivals Dinkumware's, and is actually free until the book is
193published.</li>
194
195<li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
196Questions</a></li>
197
198<li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
199Contains a useful <a
200href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
201STL</a>.</li>
202
203<li><a href="http://www.research.att.com/%7Ebs/C++.html">Bjarne Stroustrup's C++
204Page</a></li>
205
Reid Spencer096603a2004-05-26 08:41:35 +0000206<li><a href="http://www.linux.com.cn/Bruce_Eckel/TICPPv2/Contents.htm">
207Bruce Eckel's Thinking in C++, 2nd ed. Volume 2 Revision 4.0 (even better, get
208the book).</a></li>
209
Misha Brukman13fd15c2004-01-15 00:14:41 +0000210</ol>
211
212<p>You are also encouraged to take a look at the <a
213href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
214to write maintainable code more than where to put your curly braces.</p>
215
216</div>
217
218<!-- ======================================================================= -->
219<div class="doc_subsection">
220 <a name="stl">Other useful references</a>
221</div>
222
223<div class="doc_text">
224
225<p>LLVM is currently using CVS as its source versioning system. You may find
226this reference handy:</p>
227
228<ol>
229<li><a href="http://www.psc.edu/%7Esemke/cvs_branches.html">CVS
Chris Lattner261efe92003-11-25 01:02:51 +0000230Branch and Tag Primer</a></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000231</ol>
232
233</div>
234
Chris Lattner9355b472002-09-06 02:50:58 +0000235<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000236<div class="doc_section">
237 <a name="apis">Important and useful LLVM APIs</a>
238</div>
239<!-- *********************************************************************** -->
240
241<div class="doc_text">
242
243<p>Here we highlight some LLVM APIs that are generally useful and good to
244know about when writing transformations.</p>
245
246</div>
247
248<!-- ======================================================================= -->
249<div class="doc_subsection">
250 <a name="isa">The isa&lt;&gt;, cast&lt;&gt; and dyn_cast&lt;&gt; templates</a>
251</div>
252
253<div class="doc_text">
254
255<p>The LLVM source-base makes extensive use of a custom form of RTTI.
Chris Lattner261efe92003-11-25 01:02:51 +0000256These templates have many similarities to the C++ <tt>dynamic_cast&lt;&gt;</tt>
257operator, but they don't have some drawbacks (primarily stemming from
258the fact that <tt>dynamic_cast&lt;&gt;</tt> only works on classes that
259have a v-table). Because they are used so often, you must know what they
260do and how they work. All of these templates are defined in the <a
261 href="/doxygen/Casting_8h-source.html"><tt>Support/Casting.h</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000262file (note that you very rarely have to include this file directly).</p>
263
264<dl>
265 <dt><tt>isa&lt;&gt;</tt>: </dt>
266
267 <dd>The <tt>isa&lt;&gt;</tt> operator works exactly like the Java
268 "<tt>instanceof</tt>" operator. It returns true or false depending on whether
269 a reference or pointer points to an instance of the specified class. This can
270 be very useful for constraint checking of various sorts (example below).</dd>
271
272 <dt><tt>cast&lt;&gt;</tt>: </dt>
273
274 <dd>The <tt>cast&lt;&gt;</tt> operator is a "checked cast" operation. It
275 converts a pointer or reference from a base class to a derived cast, causing
276 an assertion failure if it is not really an instance of the right type. This
277 should be used in cases where you have some information that makes you believe
278 that something is of the right type. An example of the <tt>isa&lt;&gt;</tt>
279 and <tt>cast&lt;&gt;</tt> template is:
280
Chris Lattner69bf8a92004-05-23 21:06:58 +0000281 <pre>
282 static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
283 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))
284 return true;
285
286 <i>// Otherwise, it must be an instruction...</i>
287 return !L-&gt;contains(cast&lt;<a href="#Instruction">Instruction</a>&gt;(V)-&gt;getParent());
288 </pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000289
290 <p>Note that you should <b>not</b> use an <tt>isa&lt;&gt;</tt> test followed
291 by a <tt>cast&lt;&gt;</tt>, for that use the <tt>dyn_cast&lt;&gt;</tt>
292 operator.</p>
293
294 </dd>
295
296 <dt><tt>dyn_cast&lt;&gt;</tt>:</dt>
297
298 <dd>The <tt>dyn_cast&lt;&gt;</tt> operator is a "checking cast" operation. It
299 checks to see if the operand is of the specified type, and if so, returns a
300 pointer to it (this operator does not work with references). If the operand is
301 not of the correct type, a null pointer is returned. Thus, this works very
302 much like the <tt>dynamic_cast</tt> operator in C++, and should be used in the
303 same circumstances. Typically, the <tt>dyn_cast&lt;&gt;</tt> operator is used
304 in an <tt>if</tt> statement or some other flow control statement like this:
305
Chris Lattner69bf8a92004-05-23 21:06:58 +0000306 <pre>
307 if (<a href="#AllocationInst">AllocationInst</a> *AI = dyn_cast&lt;<a href="#AllocationInst">AllocationInst</a>&gt;(Val)) {
308 ...
309 }
310 </pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000311
312 <p> This form of the <tt>if</tt> statement effectively combines together a
313 call to <tt>isa&lt;&gt;</tt> and a call to <tt>cast&lt;&gt;</tt> into one
314 statement, which is very convenient.</p>
315
316 <p> Another common example is:</p>
317
Chris Lattner69bf8a92004-05-23 21:06:58 +0000318 <pre>
319 <i>// Loop over all of the phi nodes in a basic block</i>
320 BasicBlock::iterator BBI = BB-&gt;begin();
321 for (; <a href="#PhiNode">PHINode</a> *PN = dyn_cast&lt;<a href="#PHINode">PHINode</a>&gt;(BBI); ++BBI)
322 std::cerr &lt;&lt; *PN;
323 </pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000324
325 <p>Note that the <tt>dyn_cast&lt;&gt;</tt> operator, like C++'s
326 <tt>dynamic_cast</tt> or Java's <tt>instanceof</tt> operator, can be abused.
327 In particular you should not use big chained <tt>if/then/else</tt> blocks to
328 check for lots of different variants of classes. If you find yourself
329 wanting to do this, it is much cleaner and more efficient to use the
330 InstVisitor class to dispatch over the instruction type directly.</p>
331
Chris Lattner261efe92003-11-25 01:02:51 +0000332 </dd>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000333
Chris Lattner261efe92003-11-25 01:02:51 +0000334 <dt><tt>cast_or_null&lt;&gt;</tt>: </dt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000335
336 <dd>The <tt>cast_or_null&lt;&gt;</tt> operator works just like the
337 <tt>cast&lt;&gt;</tt> operator, except that it allows for a null pointer as
338 an argument (which it then propagates). This can sometimes be useful,
339 allowing you to combine several null checks into one.</dd>
340
Chris Lattner261efe92003-11-25 01:02:51 +0000341 <dt><tt>dyn_cast_or_null&lt;&gt;</tt>: </dt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000342
343 <dd>The <tt>dyn_cast_or_null&lt;&gt;</tt> operator works just like the
344 <tt>dyn_cast&lt;&gt;</tt> operator, except that it allows for a null pointer
345 as an argument (which it then propagates). This can sometimes be useful,
346 allowing you to combine several null checks into one.</dd>
347
Chris Lattner261efe92003-11-25 01:02:51 +0000348 </dl>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000349
350<p>These five templates can be used with any classes, whether they have a
351v-table or not. To add support for these templates, you simply need to add
352<tt>classof</tt> static methods to the class you are interested casting
353to. Describing this is currently outside the scope of this document, but there
354are lots of examples in the LLVM source base.</p>
355
356</div>
357
358<!-- ======================================================================= -->
359<div class="doc_subsection">
360 <a name="DEBUG">The <tt>DEBUG()</tt> macro &amp; <tt>-debug</tt> option</a>
361</div>
362
363<div class="doc_text">
364
365<p>Often when working on your pass you will put a bunch of debugging printouts
366and other code into your pass. After you get it working, you want to remove
367it... but you may need it again in the future (to work out new bugs that you run
368across).</p>
369
370<p> Naturally, because of this, you don't want to delete the debug printouts,
371but you don't want them to always be noisy. A standard compromise is to comment
372them out, allowing you to enable them if you need them in the future.</p>
373
374<p>The "<tt><a href="/doxygen/Debug_8h-source.html">Support/Debug.h</a></tt>"
375file provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to
376this problem. Basically, you can put arbitrary code into the argument of the
377<tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
378tool) is run with the '<tt>-debug</tt>' command line argument:</p>
379
Chris Lattner261efe92003-11-25 01:02:51 +0000380 <pre> ... <br> DEBUG(std::cerr &lt;&lt; "I am here!\n");<br> ...<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000381
382<p>Then you can run your pass like this:</p>
383
Chris Lattner261efe92003-11-25 01:02:51 +0000384 <pre> $ opt &lt; a.bc &gt; /dev/null -mypass<br> &lt;no output&gt;<br> $ opt &lt; a.bc &gt; /dev/null -mypass -debug<br> I am here!<br> $<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000385
386<p>Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you
387to not have to create "yet another" command line option for the debug output for
388your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
389so they do not cause a performance impact at all (for the same reason, they
390should also not contain side-effects!).</p>
391
392<p>One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
393enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
394"<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
395program hasn't been started yet, you can always just run it with
396<tt>-debug</tt>.</p>
397
398</div>
399
400<!-- _______________________________________________________________________ -->
401<div class="doc_subsubsection">
402 <a name="DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE()</tt> and
403 the <tt>-debug-only</tt> option</a>
404</div>
405
406<div class="doc_text">
407
408<p>Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
409just turns on <b>too much</b> information (such as when working on the code
410generator). If you want to enable debug information with more fine-grained
411control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
412option as follows:</p>
413
Chris Lattner261efe92003-11-25 01:02:51 +0000414 <pre> ...<br> DEBUG(std::cerr &lt;&lt; "No debug type\n");<br> #undef DEBUG_TYPE<br> #define DEBUG_TYPE "foo"<br> DEBUG(std::cerr &lt;&lt; "'foo' debug type\n");<br> #undef DEBUG_TYPE<br> #define DEBUG_TYPE "bar"<br> DEBUG(std::cerr &lt;&lt; "'bar' debug type\n");<br> #undef DEBUG_TYPE<br> #define DEBUG_TYPE ""<br> DEBUG(std::cerr &lt;&lt; "No debug type (2)\n");<br> ...<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000415
416<p>Then you can run your pass like this:</p>
417
Chris Lattner261efe92003-11-25 01:02:51 +0000418 <pre> $ opt &lt; a.bc &gt; /dev/null -mypass<br> &lt;no output&gt;<br> $ opt &lt; a.bc &gt; /dev/null -mypass -debug<br> No debug type<br> 'foo' debug type<br> 'bar' debug type<br> No debug type (2)<br> $ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=foo<br> 'foo' debug type<br> $ opt &lt; a.bc &gt; /dev/null -mypass -debug-only=bar<br> 'bar' debug type<br> $<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000419
420<p>Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of
421a file, to specify the debug type for the entire module (if you do this before
422you <tt>#include "Support/Debug.h"</tt>, you don't have to insert the ugly
423<tt>#undef</tt>'s). Also, you should use names more meaningful than "foo" and
424"bar", because there is no system in place to ensure that names do not
425conflict. If two different modules use the same string, they will all be turned
426on when the name is specified. This allows, for example, all debug information
427for instruction scheduling to be enabled with <tt>-debug-type=InstrSched</tt>,
Chris Lattner261efe92003-11-25 01:02:51 +0000428even if the source lives in multiple files.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000429
430</div>
431
432<!-- ======================================================================= -->
433<div class="doc_subsection">
434 <a name="Statistic">The <tt>Statistic</tt> template &amp; <tt>-stats</tt>
435 option</a>
436</div>
437
438<div class="doc_text">
439
440<p>The "<tt><a
441href="/doxygen/Statistic_8h-source.html">Support/Statistic.h</a></tt>" file
442provides a template named <tt>Statistic</tt> that is used as a unified way to
443keep track of what the LLVM compiler is doing and how effective various
444optimizations are. It is useful to see what optimizations are contributing to
445making a particular program run faster.</p>
446
447<p>Often you may run your pass on some big program, and you're interested to see
448how many times it makes a certain transformation. Although you can do this with
449hand inspection, or some ad-hoc method, this is a real pain and not very useful
450for big programs. Using the <tt>Statistic</tt> template makes it very easy to
451keep track of this information, and the calculated information is presented in a
452uniform manner with the rest of the passes being executed.</p>
453
454<p>There are many examples of <tt>Statistic</tt> uses, but the basics of using
455it are as follows:</p>
456
457<ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000458 <li>Define your statistic like this:
Chris Lattner261efe92003-11-25 01:02:51 +0000459 <pre>static Statistic&lt;&gt; NumXForms("mypassname", "The # of times I did stuff");<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000460
461 <p>The <tt>Statistic</tt> template can emulate just about any data-type,
462 but if you do not specify a template argument, it defaults to acting like
463 an unsigned int counter (this is usually what you want).</p></li>
464
Chris Lattner261efe92003-11-25 01:02:51 +0000465 <li>Whenever you make a transformation, bump the counter:
Chris Lattner261efe92003-11-25 01:02:51 +0000466 <pre> ++NumXForms; // I did stuff<br></pre>
Chris Lattner261efe92003-11-25 01:02:51 +0000467 </li>
468 </ol>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000469
470 <p>That's all you have to do. To get '<tt>opt</tt>' to print out the
471 statistics gathered, use the '<tt>-stats</tt>' option:</p>
472
Chris Lattner261efe92003-11-25 01:02:51 +0000473 <pre> $ opt -stats -mypassname &lt; program.bc &gt; /dev/null<br> ... statistic output ...<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000474
Chris Lattner261efe92003-11-25 01:02:51 +0000475 <p> When running <tt>gccas</tt> on a C file from the SPEC benchmark
476suite, it gives a report that looks like this:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000477
Chris Lattner261efe92003-11-25 01:02:51 +0000478 <pre> 7646 bytecodewriter - Number of normal instructions<br> 725 bytecodewriter - Number of oversized instructions<br> 129996 bytecodewriter - Number of bytecode bytes written<br> 2817 raise - Number of insts DCEd or constprop'd<br> 3213 raise - Number of cast-of-self removed<br> 5046 raise - Number of expression trees converted<br> 75 raise - Number of other getelementptr's formed<br> 138 raise - Number of load/store peepholes<br> 42 deadtypeelim - Number of unused typenames removed from symtab<br> 392 funcresolve - Number of varargs functions resolved<br> 27 globaldce - Number of global variables removed<br> 2 adce - Number of basic blocks removed<br> 134 cee - Number of branches revectored<br> 49 cee - Number of setcc instruction eliminated<br> 532 gcse - Number of loads removed<br> 2919 gcse - Number of instructions removed<br> 86 indvars - Number of canonical indvars added<br> 87 indvars - Number of aux indvars removed<br> 25 instcombine - Number of dead inst eliminate<br> 434 instcombine - Number of insts combined<br> 248 licm - Number of load insts hoisted<br> 1298 licm - Number of insts hoisted to a loop pre-header<br> 3 licm - Number of insts hoisted to multiple loop preds (bad, no loop pre-header)<br> 75 mem2reg - Number of alloca's promoted<br> 1444 cfgsimplify - Number of blocks simplified<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000479
480<p>Obviously, with so many optimizations, having a unified framework for this
481stuff is very nice. Making your pass fit well into the framework makes it more
482maintainable and useful.</p>
483
484</div>
485
486<!-- *********************************************************************** -->
487<div class="doc_section">
488 <a name="common">Helpful Hints for Common Operations</a>
489</div>
490<!-- *********************************************************************** -->
491
492<div class="doc_text">
493
494<p>This section describes how to perform some very simple transformations of
495LLVM code. This is meant to give examples of common idioms used, showing the
496practical side of LLVM transformations. <p> Because this is a "how-to" section,
497you should also read about the main classes that you will be working with. The
498<a href="#coreclasses">Core LLVM Class Hierarchy Reference</a> contains details
499and descriptions of the main classes that you should know about.</p>
500
501</div>
502
503<!-- NOTE: this section should be heavy on example code -->
504<!-- ======================================================================= -->
505<div class="doc_subsection">
506 <a name="inspection">Basic Inspection and Traversal Routines</a>
507</div>
508
509<div class="doc_text">
510
511<p>The LLVM compiler infrastructure have many different data structures that may
512be traversed. Following the example of the C++ standard template library, the
513techniques used to traverse these various data structures are all basically the
514same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
515method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
516function returns an iterator pointing to one past the last valid element of the
517sequence, and there is some <tt>XXXiterator</tt> data type that is common
518between the two operations.</p>
519
520<p>Because the pattern for iteration is common across many different aspects of
521the program representation, the standard template library algorithms may be used
522on them, and it is easier to remember how to iterate. First we show a few common
523examples of the data structures that need to be traversed. Other data
524structures are traversed in very similar ways.</p>
525
526</div>
527
528<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000529<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000530 <a name="iterate_function">Iterating over the </a><a
531 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
532 href="#Function"><tt>Function</tt></a>
533</div>
534
535<div class="doc_text">
536
537<p>It's quite common to have a <tt>Function</tt> instance that you'd like to
538transform in some way; in particular, you'd like to manipulate its
539<tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over all of
540the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>. The following is
541an example that prints the name of a <tt>BasicBlock</tt> and the number of
542<tt>Instruction</tt>s it contains:</p>
543
Chris Lattner261efe92003-11-25 01:02:51 +0000544 <pre> // func is a pointer to a Function instance<br> for (Function::iterator i = func-&gt;begin(), e = func-&gt;end(); i != e; ++i) {<br><br> // print out the name of the basic block if it has one, and then the<br> // number of instructions that it contains<br><br> cerr &lt;&lt; "Basic block (name=" &lt;&lt; i-&gt;getName() &lt;&lt; ") has " <br> &lt;&lt; i-&gt;size() &lt;&lt; " instructions.\n";<br> }<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000545
546<p>Note that i can be used as if it were a pointer for the purposes of
Joel Stanley9b96c442002-09-06 21:55:13 +0000547invoking member functions of the <tt>Instruction</tt> class. This is
548because the indirection operator is overloaded for the iterator
Chris Lattner7496ec52003-08-05 22:54:23 +0000549classes. In the above code, the expression <tt>i-&gt;size()</tt> is
Misha Brukman13fd15c2004-01-15 00:14:41 +0000550exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.</p>
551
552</div>
553
554<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000555<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000556 <a name="iterate_basicblock">Iterating over the </a><a
557 href="#Instruction"><tt>Instruction</tt></a>s in a <a
558 href="#BasicBlock"><tt>BasicBlock</tt></a>
559</div>
560
561<div class="doc_text">
562
563<p>Just like when dealing with <tt>BasicBlock</tt>s in <tt>Function</tt>s, it's
564easy to iterate over the individual instructions that make up
565<tt>BasicBlock</tt>s. Here's a code snippet that prints out each instruction in
566a <tt>BasicBlock</tt>:</p>
567
Chris Lattner261efe92003-11-25 01:02:51 +0000568 <pre> // blk is a pointer to a BasicBlock instance<br> for (BasicBlock::iterator i = blk-&gt;begin(), e = blk-&gt;end(); i != e; ++i)<br> // the next statement works since operator&lt;&lt;(ostream&amp;,...) <br> // is overloaded for Instruction&amp;<br> cerr &lt;&lt; *i &lt;&lt; "\n";<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000569
570<p>However, this isn't really the best way to print out the contents of a
571<tt>BasicBlock</tt>! Since the ostream operators are overloaded for virtually
572anything you'll care about, you could have just invoked the print routine on the
573basic block itself: <tt>cerr &lt;&lt; *blk &lt;&lt; "\n";</tt>.</p>
574
575<p>Note that currently operator&lt;&lt; is implemented for <tt>Value*</tt>, so
576it will print out the contents of the pointer, instead of the pointer value you
577might expect. This is a deprecated interface that will be removed in the
578future, so it's best not to depend on it. To print out the pointer value for
579now, you must cast to <tt>void*</tt>.</p>
580
581</div>
582
583<!-- _______________________________________________________________________ -->
Chris Lattner69bf8a92004-05-23 21:06:58 +0000584<div class="doc_subsubsection">
Misha Brukman13fd15c2004-01-15 00:14:41 +0000585 <a name="iterate_institer">Iterating over the </a><a
586 href="#Instruction"><tt>Instruction</tt></a>s in a <a
587 href="#Function"><tt>Function</tt></a>
588</div>
589
590<div class="doc_text">
591
592<p>If you're finding that you commonly iterate over a <tt>Function</tt>'s
593<tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s <tt>Instruction</tt>s,
594<tt>InstIterator</tt> should be used instead. You'll need to include <a
595href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>,
596and then instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000597small example that shows how to dump all instructions in a function to the standard error stream:<p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000598
Chris Lattner69bf8a92004-05-23 21:06:58 +0000599 <pre>#include "<a href="/doxygen/InstIterator_8h-source.html">llvm/Support/InstIterator.h</a>"<br>...<br>// Suppose F is a ptr to a function<br>for (inst_iterator i = inst_begin(F), e = inst_end(F); i != e; ++i)<br> cerr &lt;&lt; *i &lt;&lt; "\n";<br></pre>
Joel Stanleye7be6502002-09-09 15:50:33 +0000600Easy, isn't it? You can also use <tt>InstIterator</tt>s to fill a
601worklist with its initial contents. For example, if you wanted to
Chris Lattner261efe92003-11-25 01:02:51 +0000602initialize a worklist to contain all instructions in a <tt>Function</tt>
603F, all you would need to do is something like:
604 <pre>std::set&lt;Instruction*&gt; worklist;<br>worklist.insert(inst_begin(F), inst_end(F));<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000605
606<p>The STL set <tt>worklist</tt> would now contain all instructions in the
607<tt>Function</tt> pointed to by F.</p>
608
609</div>
610
611<!-- _______________________________________________________________________ -->
612<div class="doc_subsubsection">
613 <a name="iterate_convert">Turning an iterator into a class pointer (and
614 vice-versa)</a>
615</div>
616
617<div class="doc_text">
618
619<p>Sometimes, it'll be useful to grab a reference (or pointer) to a class
Joel Stanley9b96c442002-09-06 21:55:13 +0000620instance when all you've got at hand is an iterator. Well, extracting
Chris Lattner69bf8a92004-05-23 21:06:58 +0000621a reference or a pointer from an iterator is very straight-forward.
Chris Lattner261efe92003-11-25 01:02:51 +0000622Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and <tt>j</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000623is a <tt>BasicBlock::const_iterator</tt>:</p>
624
Chris Lattner261efe92003-11-25 01:02:51 +0000625 <pre> Instruction&amp; inst = *i; // grab reference to instruction reference<br> Instruction* pinst = &amp;*i; // grab pointer to instruction reference<br> const Instruction&amp; inst = *j;<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000626
627<p>However, the iterators you'll be working with in the LLVM framework are
628special: they will automatically convert to a ptr-to-instance type whenever they
629need to. Instead of dereferencing the iterator and then taking the address of
630the result, you can simply assign the iterator to the proper pointer type and
631you get the dereference and address-of operation as a result of the assignment
632(behind the scenes, this is a result of overloading casting mechanisms). Thus
633the last line of the last example,</p>
634
Chris Lattner261efe92003-11-25 01:02:51 +0000635 <pre>Instruction* pinst = &amp;*i;</pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000636
637<p>is semantically equivalent to</p>
638
Chris Lattner261efe92003-11-25 01:02:51 +0000639 <pre>Instruction* pinst = i;</pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000640
Chris Lattner69bf8a92004-05-23 21:06:58 +0000641<p>It's also possible to turn a class pointer into the corresponding iterator,
642and this is a constant time operation (very efficient). The following code
643snippet illustrates use of the conversion constructors provided by LLVM
644iterators. By using these, you can explicitly grab the iterator of something
645without actually obtaining it via iteration over some structure:</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000646
Chris Lattner261efe92003-11-25 01:02:51 +0000647 <pre>void printNextInstruction(Instruction* inst) {<br> BasicBlock::iterator it(inst);<br> ++it; // after this line, it refers to the instruction after *inst.<br> if (it != inst-&gt;getParent()-&gt;end()) cerr &lt;&lt; *it &lt;&lt; "\n";<br>}<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000648
Misha Brukman13fd15c2004-01-15 00:14:41 +0000649</div>
650
651<!--_______________________________________________________________________-->
652<div class="doc_subsubsection">
653 <a name="iterate_complex">Finding call sites: a slightly more complex
654 example</a>
655</div>
656
657<div class="doc_text">
658
659<p>Say that you're writing a FunctionPass and would like to count all the
660locations in the entire module (that is, across every <tt>Function</tt>) where a
661certain function (i.e., some <tt>Function</tt>*) is already in scope. As you'll
662learn later, you may want to use an <tt>InstVisitor</tt> to accomplish this in a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000663much more straight-forward manner, but this example will allow us to explore how
Misha Brukman13fd15c2004-01-15 00:14:41 +0000664you'd do it if you didn't have <tt>InstVisitor</tt> around. In pseudocode, this
665is what we want to do:</p>
666
Chris Lattner261efe92003-11-25 01:02:51 +0000667 <pre>initialize callCounter to zero<br>for each Function f in the Module<br> for each BasicBlock b in f<br> for each Instruction i in b<br> if (i is a CallInst and calls the given function)<br> increment callCounter<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000668
669<p>And the actual code is (remember, since we're writing a
670<tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply has to
671override the <tt>runOnFunction</tt> method...):</p>
672
Chris Lattner261efe92003-11-25 01:02:51 +0000673 <pre>Function* targetFunc = ...;<br><br>class OurFunctionPass : public FunctionPass {<br> public:<br> OurFunctionPass(): callCounter(0) { }<br><br> virtual runOnFunction(Function&amp; F) {<br> for (Function::iterator b = F.begin(), be = F.end(); b != be; ++b) {<br> for (BasicBlock::iterator i = b-&gt;begin(); ie = b-&gt;end(); i != ie; ++i) {<br> if (<a
674 href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a>&lt;<a
675 href="#CallInst">CallInst</a>&gt;(&amp;*i)) {<br> // we know we've encountered a call instruction, so we<br> // need to determine if it's a call to the<br> // function pointed to by m_func or not.<br> <br> if (callInst-&gt;getCalledFunction() == targetFunc)<br> ++callCounter;<br> }<br> }<br> }<br> <br> private:<br> unsigned callCounter;<br>};<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000676
677</div>
678
Brian Gaekef1972c62003-11-07 19:25:45 +0000679<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000680<div class="doc_subsubsection">
681 <a name="calls_and_invokes">Treating calls and invokes the same way</a>
682</div>
683
684<div class="doc_text">
685
686<p>You may have noticed that the previous example was a bit oversimplified in
687that it did not deal with call sites generated by 'invoke' instructions. In
688this, and in other situations, you may find that you want to treat
689<tt>CallInst</tt>s and <tt>InvokeInst</tt>s the same way, even though their
690most-specific common base class is <tt>Instruction</tt>, which includes lots of
691less closely-related things. For these cases, LLVM provides a handy wrapper
692class called <a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000693href="http://llvm.cs.uiuc.edu/doxygen/classCallSite.html"><tt>CallSite</tt></a>.
694It is essentially a wrapper around an <tt>Instruction</tt> pointer, with some
695methods that provide functionality common to <tt>CallInst</tt>s and
Misha Brukman13fd15c2004-01-15 00:14:41 +0000696<tt>InvokeInst</tt>s.</p>
697
Chris Lattner69bf8a92004-05-23 21:06:58 +0000698<p>This class has "value semantics": it should be passed by value, not by
699reference and it should not be dynamically allocated or deallocated using
700<tt>operator new</tt> or <tt>operator delete</tt>. It is efficiently copyable,
701assignable and constructable, with costs equivalents to that of a bare pointer.
702If you look at its definition, it has only a single pointer member.</p>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000703
704</div>
705
Chris Lattner1a3105b2002-09-09 05:49:39 +0000706<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000707<div class="doc_subsubsection">
708 <a name="iterate_chains">Iterating over def-use &amp; use-def chains</a>
709</div>
710
711<div class="doc_text">
712
713<p>Frequently, we might have an instance of the <a
714href="/doxygen/classValue.html">Value Class</a> and we want to determine which
715<tt>User</tt>s use the <tt>Value</tt>. The list of all <tt>User</tt>s of a
716particular <tt>Value</tt> is called a <i>def-use</i> chain. For example, let's
717say we have a <tt>Function*</tt> named <tt>F</tt> to a particular function
718<tt>foo</tt>. Finding all of the instructions that <i>use</i> <tt>foo</tt> is as
719simple as iterating over the <i>def-use</i> chain of <tt>F</tt>:</p>
720
Chris Lattner261efe92003-11-25 01:02:51 +0000721 <pre>Function* F = ...;<br><br>for (Value::use_iterator i = F-&gt;use_begin(), e = F-&gt;use_end(); i != e; ++i) {<br> if (Instruction *Inst = dyn_cast&lt;Instruction&gt;(*i)) {<br> cerr &lt;&lt; "F is used in instruction:\n";<br> cerr &lt;&lt; *Inst &lt;&lt; "\n";<br> }<br>}<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000722
723<p>Alternately, it's common to have an instance of the <a
724href="/doxygen/classUser.html">User Class</a> and need to know what
725<tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used by a
726<tt>User</tt> is known as a <i>use-def</i> chain. Instances of class
727<tt>Instruction</tt> are common <tt>User</tt>s, so we might want to iterate over
728all of the values that a particular instruction uses (that is, the operands of
729the particular <tt>Instruction</tt>):</p>
730
Chris Lattner261efe92003-11-25 01:02:51 +0000731 <pre>Instruction* pi = ...;<br><br>for (User::op_iterator i = pi-&gt;op_begin(), e = pi-&gt;op_end(); i != e; ++i) {<br> Value* v = *i;<br> ...<br>}<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000732
Chris Lattner1a3105b2002-09-09 05:49:39 +0000733<!--
734 def-use chains ("finding all users of"): Value::use_begin/use_end
735 use-def chains ("finding all values used"): User::op_begin/op_end [op=operand]
Misha Brukman13fd15c2004-01-15 00:14:41 +0000736-->
737
738</div>
739
740<!-- ======================================================================= -->
741<div class="doc_subsection">
742 <a name="simplechanges">Making simple changes</a>
743</div>
744
745<div class="doc_text">
746
747<p>There are some primitive transformation operations present in the LLVM
Joel Stanley753eb712002-09-11 22:32:24 +0000748infrastructure that are worth knowing about. When performing
Chris Lattner261efe92003-11-25 01:02:51 +0000749transformations, it's fairly common to manipulate the contents of basic
750blocks. This section describes some of the common methods for doing so
Misha Brukman13fd15c2004-01-15 00:14:41 +0000751and gives example code.</p>
752
753</div>
754
Chris Lattner261efe92003-11-25 01:02:51 +0000755<!--_______________________________________________________________________-->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000756<div class="doc_subsubsection">
757 <a name="schanges_creating">Creating and inserting new
758 <tt>Instruction</tt>s</a>
759</div>
760
761<div class="doc_text">
762
763<p><i>Instantiating Instructions</i></p>
764
Chris Lattner69bf8a92004-05-23 21:06:58 +0000765<p>Creation of <tt>Instruction</tt>s is straight-forward: simply call the
Misha Brukman13fd15c2004-01-15 00:14:41 +0000766constructor for the kind of instruction to instantiate and provide the necessary
767parameters. For example, an <tt>AllocaInst</tt> only <i>requires</i> a
768(const-ptr-to) <tt>Type</tt>. Thus:</p>
769
770<pre>AllocaInst* ai = new AllocaInst(Type::IntTy);</pre>
771
772<p>will create an <tt>AllocaInst</tt> instance that represents the allocation of
773one integer in the current stack frame, at runtime. Each <tt>Instruction</tt>
774subclass is likely to have varying default parameters which change the semantics
775of the instruction, so refer to the <a
776href="/doxygen/classInstruction.html">doxygen documentation for the subclass of
777Instruction</a> that you're interested in instantiating.</p>
778
779<p><i>Naming values</i></p>
780
781<p>It is very useful to name the values of instructions when you're able to, as
782this facilitates the debugging of your transformations. If you end up looking
783at generated LLVM machine code, you definitely want to have logical names
784associated with the results of instructions! By supplying a value for the
785<tt>Name</tt> (default) parameter of the <tt>Instruction</tt> constructor, you
786associate a logical name with the result of the instruction's execution at
787runtime. For example, say that I'm writing a transformation that dynamically
788allocates space for an integer on the stack, and that integer is going to be
789used as some kind of index by some other code. To accomplish this, I place an
790<tt>AllocaInst</tt> at the first point in the first <tt>BasicBlock</tt> of some
791<tt>Function</tt>, and I'm intending to use it within the same
792<tt>Function</tt>. I might do:</p>
793
794 <pre>AllocaInst* pa = new AllocaInst(Type::IntTy, 0, "indexLoc");</pre>
795
796<p>where <tt>indexLoc</tt> is now the logical name of the instruction's
797execution value, which is a pointer to an integer on the runtime stack.</p>
798
799<p><i>Inserting instructions</i></p>
800
801<p>There are essentially two ways to insert an <tt>Instruction</tt>
802into an existing sequence of instructions that form a <tt>BasicBlock</tt>:</p>
803
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000804<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000805 <li>Insertion into an explicit instruction list
806
807 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within that
808 <tt>BasicBlock</tt>, and a newly-created instruction we wish to insert
809 before <tt>*pi</tt>, we do the following: </p>
810
Alkis Evlogimenos9a5dc4f2004-05-27 00:57:51 +0000811 <pre> BasicBlock *pb = ...;<br> Instruction *pi = ...;<br> Instruction *newInst = new Instruction(...);<br> pb-&gt;getInstList().insert(pi, newInst); // inserts newInst before pi in pb<br></pre>
812
813 <p>Appending to the end of a <tt>BasicBlock</tt> is so common that
814 the <tt>Instruction</tt> class and <tt>Instruction</tt>-derived
815 classes provide constructors which take a pointer to a
816 <tt>BasicBlock</tt> to be appended to. For example code that
817 looked like: </p>
818
819 <pre> BasicBlock *pb = ...;<br> Instruction *newInst = new Instruction(...);<br> pb-&gt;getInstList().push_back(newInst); // appends newInst to pb<br></pre>
820
821 <p>becomes: </p>
822
823 <pre> BasicBlock *pb = ...;<br> Instruction *newInst = new Instruction(..., pb);<br></pre>
824
825 <p>which is much cleaner, especially if you are creating
826 long instruction streams.</p></li>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000827
828 <li>Insertion into an implicit instruction list
829
830 <p><tt>Instruction</tt> instances that are already in <tt>BasicBlock</tt>s
831 are implicitly associated with an existing instruction list: the instruction
832 list of the enclosing basic block. Thus, we could have accomplished the same
833 thing as the above code without being given a <tt>BasicBlock</tt> by doing:
834 </p>
835
836 <pre> Instruction *pi = ...;<br> Instruction *newInst = new Instruction(...);<br> pi-&gt;getParent()-&gt;getInstList().insert(pi, newInst);<br></pre>
837
838 <p>In fact, this sequence of steps occurs so frequently that the
839 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes provide
840 constructors which take (as a default parameter) a pointer to an
841 <tt>Instruction</tt> which the newly-created <tt>Instruction</tt> should
842 precede. That is, <tt>Instruction</tt> constructors are capable of
843 inserting the newly-created instance into the <tt>BasicBlock</tt> of a
844 provided instruction, immediately before that instruction. Using an
845 <tt>Instruction</tt> constructor with a <tt>insertBefore</tt> (default)
846 parameter, the above code becomes:</p>
847
848 <pre>Instruction* pi = ...;<br>Instruction* newInst = new Instruction(..., pi);<br></pre>
849
850 <p>which is much cleaner, especially if you're creating a lot of
851instructions and adding them to <tt>BasicBlock</tt>s.</p></li>
852</ul>
853
854</div>
855
856<!--_______________________________________________________________________-->
857<div class="doc_subsubsection">
858 <a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a>
859</div>
860
861<div class="doc_text">
862
863<p>Deleting an instruction from an existing sequence of instructions that form a
Chris Lattner69bf8a92004-05-23 21:06:58 +0000864<a href="#BasicBlock"><tt>BasicBlock</tt></a> is very straight-forward. First,
Misha Brukman13fd15c2004-01-15 00:14:41 +0000865you must have a pointer to the instruction that you wish to delete. Second, you
866need to obtain the pointer to that instruction's basic block. You use the
867pointer to the basic block to get its list of instructions and then use the
868erase function to remove your instruction. For example:</p>
869
Chris Lattner261efe92003-11-25 01:02:51 +0000870 <pre> <a href="#Instruction">Instruction</a> *I = .. ;<br> <a
871 href="#BasicBlock">BasicBlock</a> *BB = I-&gt;getParent();<br> BB-&gt;getInstList().erase(I);<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000872
873</div>
874
875<!--_______________________________________________________________________-->
876<div class="doc_subsubsection">
877 <a name="schanges_replacing">Replacing an <tt>Instruction</tt> with another
878 <tt>Value</tt></a>
879</div>
880
881<div class="doc_text">
882
883<p><i>Replacing individual instructions</i></p>
884
885<p>Including "<a href="/doxygen/BasicBlockUtils_8h-source.html">llvm/Transforms/Utils/BasicBlockUtils.h</a>"
Chris Lattner261efe92003-11-25 01:02:51 +0000886permits use of two very useful replace functions: <tt>ReplaceInstWithValue</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000887and <tt>ReplaceInstWithInst</tt>.</p>
888
Chris Lattner261efe92003-11-25 01:02:51 +0000889<h4><a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a></h4>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000890
Chris Lattner261efe92003-11-25 01:02:51 +0000891<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000892 <li><tt>ReplaceInstWithValue</tt>
893
894 <p>This function replaces all uses (within a basic block) of a given
895 instruction with a value, and then removes the original instruction. The
896 following example illustrates the replacement of the result of a particular
897 <tt>AllocaInst</tt> that allocates memory for a single integer with an null
898 pointer to an integer.</p>
899
900 <pre>AllocaInst* instToReplace = ...;<br>BasicBlock::iterator ii(instToReplace);<br>ReplaceInstWithValue(instToReplace-&gt;getParent()-&gt;getInstList(), ii,<br> Constant::getNullValue(PointerType::get(Type::IntTy)));<br></pre></li>
901
902 <li><tt>ReplaceInstWithInst</tt>
903
904 <p>This function replaces a particular instruction with another
905 instruction. The following example illustrates the replacement of one
906 <tt>AllocaInst</tt> with another.</p>
907
908 <pre>AllocaInst* instToReplace = ...;<br>BasicBlock::iterator ii(instToReplace);<br>ReplaceInstWithInst(instToReplace-&gt;getParent()-&gt;getInstList(), ii,<br> new AllocaInst(Type::IntTy, 0, "ptrToReplacedInt"));<br></pre></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000909</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000910
911<p><i>Replacing multiple uses of <tt>User</tt>s and <tt>Value</tt>s</i></p>
912
913<p>You can use <tt>Value::replaceAllUsesWith</tt> and
914<tt>User::replaceUsesOfWith</tt> to change more than one use at a time. See the
915doxygen documentation for the <a href="/doxygen/classValue.html">Value Class</a>
916and <a href="/doxygen/classUser.html">User Class</a>, respectively, for more
917information.</p>
918
919<!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
920include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
921ReplaceInstWithValue, ReplaceInstWithInst -->
922
923</div>
924
Chris Lattner9355b472002-09-06 02:50:58 +0000925<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +0000926<div class="doc_section">
927 <a name="coreclasses">The Core LLVM Class Hierarchy Reference </a>
928</div>
929<!-- *********************************************************************** -->
930
931<div class="doc_text">
932
933<p>The Core LLVM classes are the primary means of representing the program
Chris Lattner261efe92003-11-25 01:02:51 +0000934being inspected or transformed. The core LLVM classes are defined in
935header files in the <tt>include/llvm/</tt> directory, and implemented in
Misha Brukman13fd15c2004-01-15 00:14:41 +0000936the <tt>lib/VMCore</tt> directory.</p>
937
938</div>
939
940<!-- ======================================================================= -->
941<div class="doc_subsection">
942 <a name="Value">The <tt>Value</tt> class</a>
943</div>
944
945<div>
946
947<p><tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt>
948<br>
949doxygen info: <a href="/doxygen/classValue.html">Value Class</a></p>
950
951<p>The <tt>Value</tt> class is the most important class in the LLVM Source
952base. It represents a typed value that may be used (among other things) as an
953operand to an instruction. There are many different types of <tt>Value</tt>s,
954such as <a href="#Constant"><tt>Constant</tt></a>s,<a
955href="#Argument"><tt>Argument</tt></a>s. Even <a
956href="#Instruction"><tt>Instruction</tt></a>s and <a
957href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.</p>
958
959<p>A particular <tt>Value</tt> may be used many times in the LLVM representation
960for a program. For example, an incoming argument to a function (represented
961with an instance of the <a href="#Argument">Argument</a> class) is "used" by
962every instruction in the function that references the argument. To keep track
963of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
964href="#User"><tt>User</tt></a>s that is using it (the <a
965href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
966graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
967def-use information in the program, and is accessible through the <tt>use_</tt>*
968methods, shown below.</p>
969
970<p>Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed,
971and this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
972method. In addition, all LLVM values can be named. The "name" of the
973<tt>Value</tt> is a symbolic string printed in the LLVM code:</p>
974
Chris Lattner261efe92003-11-25 01:02:51 +0000975 <pre> %<b>foo</b> = add int 1, 2<br></pre>
Misha Brukman13fd15c2004-01-15 00:14:41 +0000976
977<p><a name="#nameWarning">The name of this instruction is "foo".</a> <b>NOTE</b>
978that the name of any value may be missing (an empty string), so names should
979<b>ONLY</b> be used for debugging (making the source code easier to read,
980debugging printouts), they should not be used to keep track of values or map
981between them. For this purpose, use a <tt>std::map</tt> of pointers to the
982<tt>Value</tt> itself instead.</p>
983
984<p>One important aspect of LLVM is that there is no distinction between an SSA
985variable and the operation that produces it. Because of this, any reference to
986the value produced by an instruction (or the value available as an incoming
Chris Lattnerd5fc4fc2004-03-18 14:58:55 +0000987argument, for example) is represented as a direct pointer to the instance of
988the class that
Misha Brukman13fd15c2004-01-15 00:14:41 +0000989represents this value. Although this may take some getting used to, it
990simplifies the representation and makes it easier to manipulate.</p>
991
992</div>
993
994<!-- _______________________________________________________________________ -->
995<div class="doc_subsubsection">
996 <a name="m_Value">Important Public Members of the <tt>Value</tt> class</a>
997</div>
998
999<div class="doc_text">
1000
Chris Lattner261efe92003-11-25 01:02:51 +00001001<ul>
1002 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the
1003use-list<br>
1004 <tt>Value::use_const_iterator</tt> - Typedef for const_iterator over
1005the use-list<br>
1006 <tt>unsigned use_size()</tt> - Returns the number of users of the
1007value.<br>
Chris Lattner9355b472002-09-06 02:50:58 +00001008 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
Chris Lattner261efe92003-11-25 01:02:51 +00001009 <tt>use_iterator use_begin()</tt> - Get an iterator to the start of
1010the use-list.<br>
1011 <tt>use_iterator use_end()</tt> - Get an iterator to the end of the
1012use-list.<br>
1013 <tt><a href="#User">User</a> *use_back()</tt> - Returns the last
1014element in the list.
1015 <p> These methods are the interface to access the def-use
1016information in LLVM. As with all other iterators in LLVM, the naming
1017conventions follow the conventions defined by the <a href="#stl">STL</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001018 </li>
1019 <li><tt><a href="#Type">Type</a> *getType() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001020 <p>This method returns the Type of the Value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001021 </li>
1022 <li><tt>bool hasName() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +00001023 <tt>std::string getName() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +00001024 <tt>void setName(const std::string &amp;Name)</tt>
1025 <p> This family of methods is used to access and assign a name to a <tt>Value</tt>,
1026be aware of the <a href="#nameWarning">precaution above</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001027 </li>
1028 <li><tt>void replaceAllUsesWith(Value *V)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001029
1030 <p>This method traverses the use list of a <tt>Value</tt> changing all <a
1031 href="#User"><tt>User</tt>s</a> of the current value to refer to
1032 "<tt>V</tt>" instead. For example, if you detect that an instruction always
1033 produces a constant value (for example through constant folding), you can
1034 replace all uses of the instruction with the constant like this:</p>
1035
Chris Lattner261efe92003-11-25 01:02:51 +00001036 <pre> Inst-&gt;replaceAllUsesWith(ConstVal);<br></pre>
Chris Lattner261efe92003-11-25 01:02:51 +00001037</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001038
1039</div>
1040
1041<!-- ======================================================================= -->
1042<div class="doc_subsection">
1043 <a name="User">The <tt>User</tt> class</a>
1044</div>
1045
1046<div class="doc_text">
1047
1048<p>
1049<tt>#include "<a href="/doxygen/User_8h-source.html">llvm/User.h</a>"</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +00001050doxygen info: <a href="/doxygen/classUser.html">User Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001051Superclass: <a href="#Value"><tt>Value</tt></a></p>
1052
1053<p>The <tt>User</tt> class is the common base class of all LLVM nodes that may
1054refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
1055that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
1056referring to. The <tt>User</tt> class itself is a subclass of
1057<tt>Value</tt>.</p>
1058
1059<p>The operands of a <tt>User</tt> point directly to the LLVM <a
1060href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
1061Single Assignment (SSA) form, there can only be one definition referred to,
1062allowing this direct connection. This connection provides the use-def
1063information in LLVM.</p>
1064
1065</div>
1066
1067<!-- _______________________________________________________________________ -->
1068<div class="doc_subsubsection">
1069 <a name="m_User">Important Public Members of the <tt>User</tt> class</a>
1070</div>
1071
1072<div class="doc_text">
1073
1074<p>The <tt>User</tt> class exposes the operand list in two ways: through
1075an index access interface and through an iterator based interface.</p>
1076
Chris Lattner261efe92003-11-25 01:02:51 +00001077<ul>
Chris Lattner261efe92003-11-25 01:02:51 +00001078 <li><tt>Value *getOperand(unsigned i)</tt><br>
1079 <tt>unsigned getNumOperands()</tt>
1080 <p> These two methods expose the operands of the <tt>User</tt> in a
Misha Brukman13fd15c2004-01-15 00:14:41 +00001081convenient form for direct access.</p></li>
1082
Chris Lattner261efe92003-11-25 01:02:51 +00001083 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand
1084list<br>
1085 <tt>User::op_const_iterator</tt> <tt>use_iterator op_begin()</tt> -
1086Get an iterator to the start of the operand list.<br>
1087 <tt>use_iterator op_end()</tt> - Get an iterator to the end of the
1088operand list.
1089 <p> Together, these methods make up the iterator based interface to
Misha Brukman13fd15c2004-01-15 00:14:41 +00001090the operands of a <tt>User</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00001091</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001092
1093</div>
1094
1095<!-- ======================================================================= -->
1096<div class="doc_subsection">
1097 <a name="Instruction">The <tt>Instruction</tt> class</a>
1098</div>
1099
1100<div class="doc_text">
1101
1102<p><tt>#include "</tt><tt><a
1103href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt><br>
1104doxygen info: <a href="/doxygen/classInstruction.html">Instruction Class</a><br>
1105Superclasses: <a href="#User"><tt>User</tt></a>, <a
1106href="#Value"><tt>Value</tt></a></p>
1107
1108<p>The <tt>Instruction</tt> class is the common base class for all LLVM
1109instructions. It provides only a few methods, but is a very commonly used
1110class. The primary data tracked by the <tt>Instruction</tt> class itself is the
1111opcode (instruction type) and the parent <a
1112href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
1113into. To represent a specific type of instruction, one of many subclasses of
1114<tt>Instruction</tt> are used.</p>
1115
1116<p> Because the <tt>Instruction</tt> class subclasses the <a
1117href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
1118way as for other <a href="#User"><tt>User</tt></a>s (with the
1119<tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
1120<tt>op_begin()</tt>/<tt>op_end()</tt> methods).</p> <p> An important file for
1121the <tt>Instruction</tt> class is the <tt>llvm/Instruction.def</tt> file. This
1122file contains some meta-data about the various different types of instructions
1123in LLVM. It describes the enum values that are used as opcodes (for example
1124<tt>Instruction::Add</tt> and <tt>Instruction::SetLE</tt>), as well as the
1125concrete sub-classes of <tt>Instruction</tt> that implement the instruction (for
1126example <tt><a href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
1127href="#SetCondInst">SetCondInst</a></tt>). Unfortunately, the use of macros in
1128this file confuses doxygen, so these enum values don't show up correctly in the
1129<a href="/doxygen/classInstruction.html">doxygen output</a>.</p>
1130
1131</div>
1132
1133<!-- _______________________________________________________________________ -->
1134<div class="doc_subsubsection">
1135 <a name="m_Instruction">Important Public Members of the <tt>Instruction</tt>
1136 class</a>
1137</div>
1138
1139<div class="doc_text">
1140
Chris Lattner261efe92003-11-25 01:02:51 +00001141<ul>
1142 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001143 <p>Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that
1144this <tt>Instruction</tt> is embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00001145 <li><tt>bool mayWriteToMemory()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001146 <p>Returns true if the instruction writes to memory, i.e. it is a
1147 <tt>call</tt>,<tt>free</tt>,<tt>invoke</tt>, or <tt>store</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00001148 <li><tt>unsigned getOpcode()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001149 <p>Returns the opcode for the <tt>Instruction</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00001150 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001151 <p>Returns another instance of the specified instruction, identical
Chris Lattner261efe92003-11-25 01:02:51 +00001152in all ways to the original except that the instruction has no parent
1153(ie it's not embedded into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>),
Misha Brukman13fd15c2004-01-15 00:14:41 +00001154and it has no name</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00001155</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001156
1157</div>
1158
1159<!-- ======================================================================= -->
1160<div class="doc_subsection">
1161 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
1162</div>
1163
1164<div class="doc_text">
1165
1166<p><tt>#include "<a href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +00001167doxygen info: <a href="/doxygen/classBasicBlock.html">BasicBlock Class</a><br>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001168Superclass: <a href="#Value"><tt>Value</tt></a></p>
1169
1170<p>This class represents a single entry multiple exit section of the code,
1171commonly known as a basic block by the compiler community. The
1172<tt>BasicBlock</tt> class maintains a list of <a
1173href="#Instruction"><tt>Instruction</tt></a>s, which form the body of the block.
1174Matching the language definition, the last element of this list of instructions
1175is always a terminator instruction (a subclass of the <a
1176href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).</p>
1177
1178<p>In addition to tracking the list of instructions that make up the block, the
1179<tt>BasicBlock</tt> class also keeps track of the <a
1180href="#Function"><tt>Function</tt></a> that it is embedded into.</p>
1181
1182<p>Note that <tt>BasicBlock</tt>s themselves are <a
1183href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
1184like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
1185<tt>label</tt>.</p>
1186
1187</div>
1188
1189<!-- _______________________________________________________________________ -->
1190<div class="doc_subsubsection">
1191 <a name="m_BasicBlock">Important Public Members of the <tt>BasicBlock</tt>
1192 class</a>
1193</div>
1194
1195<div class="doc_text">
1196
Chris Lattner261efe92003-11-25 01:02:51 +00001197<ul>
1198 <li><tt>BasicBlock(const std::string &amp;Name = "", </tt><tt><a
1199 href="#Function">Function</a> *Parent = 0)</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001200 <p>The <tt>BasicBlock</tt> constructor is used to create new basic
Chris Lattner261efe92003-11-25 01:02:51 +00001201blocks for insertion into a function. The constructor optionally takes
1202a name for the new block, and a <a href="#Function"><tt>Function</tt></a>
1203to insert it into. If the <tt>Parent</tt> parameter is specified, the
1204new <tt>BasicBlock</tt> is automatically inserted at the end of the
1205specified <a href="#Function"><tt>Function</tt></a>, if not specified,
1206the BasicBlock must be manually inserted into the <a href="#Function"><tt>Function</tt></a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001207 </li>
1208 <li><tt>BasicBlock::iterator</tt> - Typedef for instruction list
1209iterator<br>
Chris Lattner9355b472002-09-06 02:50:58 +00001210 <tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
Chris Lattner261efe92003-11-25 01:02:51 +00001211 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,<tt>size()</tt>,<tt>empty()</tt>,<tt>rbegin()</tt>,<tt>rend()
1212- </tt>STL style functions for accessing the instruction list.
1213 <p> These methods and typedefs are forwarding functions that have
1214the same semantics as the standard library methods of the same names.
1215These methods expose the underlying instruction list of a basic block in
1216a way that is easy to manipulate. To get the full complement of
1217container operations (including operations to update the list), you must
Misha Brukman13fd15c2004-01-15 00:14:41 +00001218use the <tt>getInstList()</tt> method.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00001219 <li><tt>BasicBlock::InstListType &amp;getInstList()</tt>
1220 <p> This method is used to get access to the underlying container
1221that actually holds the Instructions. This method must be used when
1222there isn't a forwarding function in the <tt>BasicBlock</tt> class for
1223the operation that you would like to perform. Because there are no
1224forwarding functions for "updating" operations, you need to use this if
Misha Brukman13fd15c2004-01-15 00:14:41 +00001225you want to update the contents of a <tt>BasicBlock</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00001226 <li><tt><a href="#Function">Function</a> *getParent()</tt>
1227 <p> Returns a pointer to <a href="#Function"><tt>Function</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001228the block is embedded into, or a null pointer if it is homeless.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00001229 <li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt>
1230 <p> Returns a pointer to the terminator instruction that appears at
1231the end of the <tt>BasicBlock</tt>. If there is no terminator
1232instruction, or if the last instruction in the block is not a
Misha Brukman13fd15c2004-01-15 00:14:41 +00001233terminator, then a null pointer is returned.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00001234</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001235
1236</div>
1237
1238<!-- ======================================================================= -->
1239<div class="doc_subsection">
1240 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
1241</div>
1242
1243<div class="doc_text">
1244
1245<p><tt>#include "<a
1246href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt><br>
1247doxygen info: <a href="/doxygen/classGlobalValue.html">GlobalValue Class</a><br>
1248Superclasses: <a href="#User"><tt>User</tt></a>, <a
1249href="#Value"><tt>Value</tt></a></p>
1250
1251<p>Global values (<a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
1252href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
1253visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
1254Because they are visible at global scope, they are also subject to linking with
1255other globals defined in different translation units. To control the linking
1256process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
1257<tt>GlobalValue</tt>s know whether they have internal or external linkage, as
1258defined by the <tt>LinkageTypes</tt> enumerator.</p>
1259
1260<p>If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
1261<tt>static</tt> in C), it is not visible to code outside the current translation
1262unit, and does not participate in linking. If it has external linkage, it is
1263visible to external code, and does participate in linking. In addition to
1264linkage information, <tt>GlobalValue</tt>s keep track of which <a
1265href="#Module"><tt>Module</tt></a> they are currently part of.</p>
1266
1267<p>Because <tt>GlobalValue</tt>s are memory objects, they are always referred to
1268by their <b>address</b>. As such, the <a href="#Type"><tt>Type</tt></a> of a
1269global is always a pointer to its contents. It is important to remember this
1270when using the <tt>GetElementPtrInst</tt> instruction because this pointer must
1271be dereferenced first. For example, if you have a <tt>GlobalVariable</tt> (a
1272subclass of <tt>GlobalValue)</tt> that is an array of 24 ints, type <tt>[24 x
1273int]</tt>, then the <tt>GlobalVariable</tt> is a pointer to that array. Although
1274the address of the first element of this array and the value of the
1275<tt>GlobalVariable</tt> are the same, they have different types. The
1276<tt>GlobalVariable</tt>'s type is <tt>[24 x int]</tt>. The first element's type
1277is <tt>int.</tt> Because of this, accessing a global value requires you to
1278dereference the pointer with <tt>GetElementPtrInst</tt> first, then its elements
1279can be accessed. This is explained in the <a href="LangRef.html#globalvars">LLVM
1280Language Reference Manual</a>.</p>
1281
1282</div>
1283
1284<!-- _______________________________________________________________________ -->
1285<div class="doc_subsubsection">
1286 <a name="m_GlobalValue">Important Public Members of the <tt>GlobalValue</tt>
1287 class</a>
1288</div>
1289
1290<div class="doc_text">
1291
Chris Lattner261efe92003-11-25 01:02:51 +00001292<ul>
1293 <li><tt>bool hasInternalLinkage() const</tt><br>
Chris Lattner9355b472002-09-06 02:50:58 +00001294 <tt>bool hasExternalLinkage() const</tt><br>
Chris Lattner261efe92003-11-25 01:02:51 +00001295 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt>
1296 <p> These methods manipulate the linkage characteristics of the <tt>GlobalValue</tt>.</p>
1297 <p> </p>
1298 </li>
1299 <li><tt><a href="#Module">Module</a> *getParent()</tt>
1300 <p> This returns the <a href="#Module"><tt>Module</tt></a> that the
Misha Brukman13fd15c2004-01-15 00:14:41 +00001301GlobalValue is currently embedded into.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00001302</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001303
1304</div>
1305
1306<!-- ======================================================================= -->
1307<div class="doc_subsection">
1308 <a name="Function">The <tt>Function</tt> class</a>
1309</div>
1310
1311<div class="doc_text">
1312
1313<p><tt>#include "<a
1314href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt><br> doxygen
1315info: <a href="/doxygen/classFunction.html">Function Class</a><br> Superclasses:
1316<a href="#GlobalValue"><tt>GlobalValue</tt></a>, <a
1317href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a></p>
1318
1319<p>The <tt>Function</tt> class represents a single procedure in LLVM. It is
1320actually one of the more complex classes in the LLVM heirarchy because it must
1321keep track of a large amount of data. The <tt>Function</tt> class keeps track
1322of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal <a
1323href="#Argument"><tt>Argument</tt></a>s, and a <a
1324href="#SymbolTable"><tt>SymbolTable</tt></a>.</p>
1325
1326<p>The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most
1327commonly used part of <tt>Function</tt> objects. The list imposes an implicit
1328ordering of the blocks in the function, which indicate how the code will be
1329layed out by the backend. Additionally, the first <a
1330href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
1331<tt>Function</tt>. It is not legal in LLVM to explicitly branch to this initial
1332block. There are no implicit exit nodes, and in fact there may be multiple exit
1333nodes from a single <tt>Function</tt>. If the <a
1334href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
1335the <tt>Function</tt> is actually a function declaration: the actual body of the
1336function hasn't been linked in yet.</p>
1337
1338<p>In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
1339<tt>Function</tt> class also keeps track of the list of formal <a
1340href="#Argument"><tt>Argument</tt></a>s that the function receives. This
1341container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
1342nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
1343the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.</p>
1344
1345<p>The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used
1346LLVM feature that is only used when you have to look up a value by name. Aside
1347from that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used
1348internally to make sure that there are not conflicts between the names of <a
1349href="#Instruction"><tt>Instruction</tt></a>s, <a
1350href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
1351href="#Argument"><tt>Argument</tt></a>s in the function body.</p>
1352
1353</div>
1354
1355<!-- _______________________________________________________________________ -->
1356<div class="doc_subsubsection">
1357 <a name="m_Function">Important Public Members of the <tt>Function</tt>
1358 class</a>
1359</div>
1360
1361<div class="doc_text">
1362
Chris Lattner261efe92003-11-25 01:02:51 +00001363<ul>
1364 <li><tt>Function(const </tt><tt><a href="#FunctionType">FunctionType</a>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001365 *Ty, bool isInternal, const std::string &amp;N = "", Module* Parent = 0)</tt>
1366
1367 <p>Constructor used when you need to create new <tt>Function</tt>s to add
1368 the the program. The constructor must specify the type of the function to
1369 create and whether or not it should start out with internal or external
1370 linkage. The&nbsp;<a href="#FunctionType"><tt>FunctionType</tt></a> argument
1371 specifies the formal arguments and return value for the function. The same
1372 <a href="#FunctionTypel"><tt>FunctionType</tt></a> value can be used to
1373 create multiple functions. The <tt>Parent</tt> argument specifies the Module
1374 in which the function is defined. If this argument is provided, the function
1375 will automatically be inserted into that module's list of
1376 functions.</p></li>
1377
Chris Lattner261efe92003-11-25 01:02:51 +00001378 <li><tt>bool isExternal()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001379
1380 <p>Return whether or not the <tt>Function</tt> has a body defined. If the
1381 function is "external", it does not have a body, and thus must be resolved
1382 by linking with a function defined in a different translation unit.</p></li>
1383
Chris Lattner261efe92003-11-25 01:02:51 +00001384 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
Chris Lattner9355b472002-09-06 02:50:58 +00001385 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001386
1387 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1388 <tt>size()</tt>, <tt>empty()</tt>, <tt>rbegin()</tt>, <tt>rend()</tt>
1389
1390 <p>These are forwarding methods that make it easy to access the contents of
1391 a <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
1392 list.</p></li>
1393
Chris Lattner261efe92003-11-25 01:02:51 +00001394 <li><tt>Function::BasicBlockListType &amp;getBasicBlockList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001395
1396 <p>Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This
1397 is necessary to use when you need to update the list or perform a complex
1398 action that doesn't have a forwarding method.</p></li>
1399
Chris Lattner261efe92003-11-25 01:02:51 +00001400 <li><tt>Function::aiterator</tt> - Typedef for the argument list
1401iterator<br>
Chris Lattner9355b472002-09-06 02:50:58 +00001402 <tt>Function::const_aiterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001403
1404 <tt>abegin()</tt>, <tt>aend()</tt>, <tt>afront()</tt>, <tt>aback()</tt>,
1405 <tt>asize()</tt>, <tt>aempty()</tt>, <tt>arbegin()</tt>, <tt>arend()</tt>
1406
1407 <p>These are forwarding methods that make it easy to access the contents of
1408 a <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a>
1409 list.</p></li>
1410
Chris Lattner261efe92003-11-25 01:02:51 +00001411 <li><tt>Function::ArgumentListType &amp;getArgumentList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001412
1413 <p>Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
1414 necessary to use when you need to update the list or perform a complex
1415 action that doesn't have a forwarding method.</p></li>
1416
Chris Lattner261efe92003-11-25 01:02:51 +00001417 <li><tt><a href="#BasicBlock">BasicBlock</a> &amp;getEntryBlock()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001418
1419 <p>Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
1420 function. Because the entry block for the function is always the first
1421 block, this returns the first block of the <tt>Function</tt>.</p></li>
1422
Chris Lattner261efe92003-11-25 01:02:51 +00001423 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
1424 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001425
1426 <p>This traverses the <a href="#Type"><tt>Type</tt></a> of the
1427 <tt>Function</tt> and returns the return type of the function, or the <a
1428 href="#FunctionType"><tt>FunctionType</tt></a> of the actual
1429 function.</p></li>
1430
Chris Lattner261efe92003-11-25 01:02:51 +00001431 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001432
Chris Lattner261efe92003-11-25 01:02:51 +00001433 <p> Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001434 for this <tt>Function</tt>.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00001435</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001436
1437</div>
1438
1439<!-- ======================================================================= -->
1440<div class="doc_subsection">
1441 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
1442</div>
1443
1444<div class="doc_text">
1445
1446<p><tt>#include "<a
1447href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt>
1448<br>
Chris Lattner261efe92003-11-25 01:02:51 +00001449doxygen info: <a href="/doxygen/classGlobalVariable.html">GlobalVariable
Misha Brukman13fd15c2004-01-15 00:14:41 +00001450Class</a><br> Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>, <a
1451href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a></p>
1452
1453<p>Global variables are represented with the (suprise suprise)
1454<tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are also
1455subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such are
1456always referenced by their address (global values must live in memory, so their
1457"name" refers to their address). See <a
1458href="#GlobalValue"><tt>GlobalValue</tt></a> for more on this. Global variables
1459may have an initial value (which must be a <a
1460href="#Constant"><tt>Constant</tt></a>), and if they have an initializer, they
1461may be marked as "constant" themselves (indicating that their contents never
1462change at runtime).</p>
1463
1464</div>
1465
1466<!-- _______________________________________________________________________ -->
1467<div class="doc_subsubsection">
1468 <a name="m_GlobalVariable">Important Public Members of the
1469 <tt>GlobalVariable</tt> class</a>
1470</div>
1471
1472<div class="doc_text">
1473
Chris Lattner261efe92003-11-25 01:02:51 +00001474<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001475 <li><tt>GlobalVariable(const </tt><tt><a href="#Type">Type</a> *Ty, bool
1476 isConstant, LinkageTypes&amp; Linkage, <a href="#Constant">Constant</a>
1477 *Initializer = 0, const std::string &amp;Name = "", Module* Parent = 0)</tt>
1478
1479 <p>Create a new global variable of the specified type. If
1480 <tt>isConstant</tt> is true then the global variable will be marked as
1481 unchanging for the program. The Linkage parameter specifies the type of
1482 linkage (internal, external, weak, linkonce, appending) for the variable. If
1483 the linkage is InternalLinkage, WeakLinkage, or LinkOnceLinkage,&nbsp; then
1484 the resultant global variable will have internal linkage. AppendingLinkage
1485 concatenates together all instances (in different translation units) of the
1486 variable into a single variable but is only applicable to arrays. &nbsp;See
1487 the <a href="LangRef.html#modulestructure">LLVM Language Reference</a> for
1488 further details on linkage types. Optionally an initializer, a name, and the
1489 module to put the variable into may be specified for the global variable as
1490 well.</p></li>
1491
Chris Lattner261efe92003-11-25 01:02:51 +00001492 <li><tt>bool isConstant() const</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001493
1494 <p>Returns true if this is a global variable that is known not to
1495 be modified at runtime.</p></li>
1496
Chris Lattner261efe92003-11-25 01:02:51 +00001497 <li><tt>bool hasInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001498
1499 <p>Returns true if this <tt>GlobalVariable</tt> has an intializer.</p></li>
1500
Chris Lattner261efe92003-11-25 01:02:51 +00001501 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001502
1503 <p>Returns the intial value for a <tt>GlobalVariable</tt>. It is not legal
1504 to call this method if there is no initializer.</p></li>
Chris Lattner261efe92003-11-25 01:02:51 +00001505</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001506
1507</div>
1508
1509<!-- ======================================================================= -->
1510<div class="doc_subsection">
1511 <a name="Module">The <tt>Module</tt> class</a>
1512</div>
1513
1514<div class="doc_text">
1515
1516<p><tt>#include "<a
1517href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt><br> doxygen info:
1518<a href="/doxygen/classModule.html">Module Class</a></p>
1519
1520<p>The <tt>Module</tt> class represents the top level structure present in LLVM
1521programs. An LLVM module is effectively either a translation unit of the
1522original program or a combination of several translation units merged by the
1523linker. The <tt>Module</tt> class keeps track of a list of <a
1524href="#Function"><tt>Function</tt></a>s, a list of <a
1525href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
1526href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
1527helpful member functions that try to make common operations easy.</p>
1528
1529</div>
1530
1531<!-- _______________________________________________________________________ -->
1532<div class="doc_subsubsection">
1533 <a name="m_Module">Important Public Members of the <tt>Module</tt> class</a>
1534</div>
1535
1536<div class="doc_text">
1537
Chris Lattner261efe92003-11-25 01:02:51 +00001538<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001539 <li><tt>Module::Module(std::string name = "")</tt></li>
Chris Lattner261efe92003-11-25 01:02:51 +00001540</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001541
1542<p>Constructing a <a href="#Module">Module</a> is easy. You can optionally
1543provide a name for it (probably based on the name of the translation unit).</p>
1544
Chris Lattner261efe92003-11-25 01:02:51 +00001545<ul>
1546 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
Chris Lattner0377de42002-09-06 14:50:55 +00001547 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001548
1549 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1550 <tt>size()</tt>, <tt>empty()</tt>, <tt>rbegin()</tt>, <tt>rend()</tt>
1551
1552 <p>These are forwarding methods that make it easy to access the contents of
1553 a <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
1554 list.</p></li>
1555
Chris Lattner261efe92003-11-25 01:02:51 +00001556 <li><tt>Module::FunctionListType &amp;getFunctionList()</tt>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001557
1558 <p> Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
1559 necessary to use when you need to update the list or perform a complex
1560 action that doesn't have a forwarding method.</p>
1561
1562 <p><!-- Global Variable --></p></li>
Chris Lattnerc75ff9a2002-10-01 23:17:09 +00001563</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001564
1565<hr>
1566
Chris Lattnerc75ff9a2002-10-01 23:17:09 +00001567<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001568 <li><tt>Module::giterator</tt> - Typedef for global variable list iterator<br>
1569
1570 <tt>Module::const_giterator</tt> - Typedef for const_iterator.<br>
1571
1572 <tt>gbegin()</tt>, <tt>gend()</tt>, <tt>gfront()</tt>, <tt>gback()</tt>,
1573 <tt>gsize()</tt>, <tt>gempty()</tt>, <tt>grbegin()</tt>, <tt>grend()</tt>
1574
1575 <p> These are forwarding methods that make it easy to access the contents of
1576 a <tt>Module</tt> object's <a
1577 href="#GlobalVariable"><tt>GlobalVariable</tt></a> list.</p></li>
1578
1579 <li><tt>Module::GlobalListType &amp;getGlobalList()</tt>
1580
1581 <p>Returns the list of <a
1582 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s. This is necessary to
1583 use when you need to update the list or perform a complex action that
1584 doesn't have a forwarding method.</p>
1585
1586 <p><!-- Symbol table stuff --> </p></li>
Chris Lattnerc75ff9a2002-10-01 23:17:09 +00001587</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001588
1589<hr>
1590
1591<ul>
1592 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
1593
1594 <p>Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
1595 for this <tt>Module</tt>.</p>
1596
1597 <p><!-- Convenience methods --></p></li>
1598</ul>
1599
1600<hr>
1601
1602<ul>
1603 <li><tt><a href="#Function">Function</a> *getFunction(const std::string
1604 &amp;Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt>
1605
1606 <p>Look up the specified function in the <tt>Module</tt> <a
1607 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
1608 <tt>null</tt>.</p></li>
1609
1610 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const
1611 std::string &amp;Name, const <a href="#FunctionType">FunctionType</a> *T)</tt>
1612
1613 <p>Look up the specified function in the <tt>Module</tt> <a
1614 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
1615 external declaration for the function and return it.</p></li>
1616
1617 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt>
1618
1619 <p>If there is at least one entry in the <a
1620 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
1621 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
1622 string.</p></li>
1623
1624 <li><tt>bool addTypeName(const std::string &amp;Name, const <a
1625 href="#Type">Type</a> *Ty)</tt>
1626
1627 <p>Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
1628 mapping <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this
1629 name, true is returned and the <a
1630 href="#SymbolTable"><tt>SymbolTable</tt></a> is not modified.</p></li>
1631</ul>
1632
1633</div>
1634
1635<!-- ======================================================================= -->
1636<div class="doc_subsection">
1637 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
1638</div>
1639
1640<div class="doc_text">
1641
1642<p>Constant represents a base class for different types of constants. It
1643is subclassed by ConstantBool, ConstantInt, ConstantSInt, ConstantUInt,
1644ConstantArray etc for representing the various types of Constants.</p>
1645
1646</div>
1647
1648<!-- _______________________________________________________________________ -->
1649<div class="doc_subsubsection">
1650 <a name="m_Value">Important Public Methods</a>
1651</div>
1652
1653<div class="doc_text">
1654
Chris Lattnerc75ff9a2002-10-01 23:17:09 +00001655<ul>
Chris Lattner261efe92003-11-25 01:02:51 +00001656 <li><tt>bool isConstantExpr()</tt>: Returns true if it is a
1657ConstantExpr
1658 <hr> Important Subclasses of Constant
1659 <p> </p>
1660 <ul>
1661 <li>ConstantSInt : This subclass of Constant represents a signed
1662integer constant.
1663 <ul>
1664 <li><tt>int64_t getValue() const</tt>: Returns the underlying value of
1665this constant. </li>
1666 </ul>
1667 </li>
1668 <li>ConstantUInt : This class represents an unsigned integer.
1669 <ul>
1670 <li><tt>uint64_t getValue() const</tt>: Returns the underlying value
1671of this constant. </li>
1672 </ul>
1673 </li>
1674 <li>ConstantFP : This class represents a floating point constant.
1675 <ul>
1676 <li><tt>double getValue() const</tt>: Returns the underlying value of
1677this constant. </li>
1678 </ul>
1679 </li>
1680 <li>ConstantBool : This represents a boolean constant.
1681 <ul>
1682 <li><tt>bool getValue() const</tt>: Returns the underlying value of
1683this constant. </li>
1684 </ul>
1685 </li>
1686 <li>ConstantArray : This represents a constant array.
1687 <ul>
1688 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>:
1689Returns a Vecotr of component constants that makeup this array. </li>
1690 </ul>
1691 </li>
1692 <li>ConstantStruct : This represents a constant struct.
1693 <ul>
1694 <li><tt>const std::vector&lt;Use&gt; &amp;getValues() const</tt>:
1695Returns a Vecotr of component constants that makeup this array. </li>
1696 </ul>
1697 </li>
1698 <li>ConstantPointerRef : This represents a constant pointer value
1699that is initialized to point to a global value, which lies at a
1700constant fixed address.
1701 <ul>
1702 <li><tt>GlobalValue *getValue()</tt>: Returns the global
1703value to which this pointer is pointing to. </li>
1704 </ul>
1705 </li>
1706 </ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001707 </li>
Chris Lattnerc75ff9a2002-10-01 23:17:09 +00001708</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001709
1710</div>
1711
1712<!-- ======================================================================= -->
1713<div class="doc_subsection">
1714 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
1715</div>
1716
1717<div class="doc_text">
1718
1719<p>Type as noted earlier is also a subclass of a Value class. Any primitive
1720type (like int, short etc) in LLVM is an instance of Type Class. All other
1721types are instances of subclasses of type like FunctionType, ArrayType
1722etc. DerivedType is the interface for all such dervied types including
1723FunctionType, ArrayType, PointerType, StructType. Types can have names. They can
1724be recursive (StructType). There exists exactly one instance of any type
1725structure at a time. This allows using pointer equality of Type *s for comparing
1726types.</p>
1727
1728</div>
1729
1730<!-- _______________________________________________________________________ -->
1731<div class="doc_subsubsection">
1732 <a name="m_Value">Important Public Methods</a>
1733</div>
1734
1735<div class="doc_text">
1736
Chris Lattnerc75ff9a2002-10-01 23:17:09 +00001737<ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001738
Misha Brukman13fd15c2004-01-15 00:14:41 +00001739 <li><tt>bool isSigned() const</tt>: Returns whether an integral numeric type
1740 is signed. This is true for SByteTy, ShortTy, IntTy, LongTy. Note that this is
1741 not true for Float and Double. </li>
1742
1743 <li><tt>bool isUnsigned() const</tt>: Returns whether a numeric type is
1744 unsigned. This is not quite the complement of isSigned... nonnumeric types
1745 return false as they do with isSigned. This returns true for UByteTy,
1746 UShortTy, UIntTy, and ULongTy. </li>
1747
1748 <li><tt>bool isInteger() const</tt>: Equilivent to isSigned() || isUnsigned(),
1749 but with only a single virtual function invocation.</li>
1750
1751 <li><tt>bool isIntegral() const</tt>: Returns true if this is an integral
1752 type, which is either Bool type or one of the Integer types.</li>
1753
1754 <li><tt>bool isFloatingPoint()</tt>: Return true if this is one of the two
1755 floating point types.</li>
1756
Misha Brukman13fd15c2004-01-15 00:14:41 +00001757 <li><tt>isLosslesslyConvertableTo (const Type *Ty) const</tt>: Return true if
1758 this type can be converted to 'Ty' without any reinterpretation of bits. For
Chris Lattner69bf8a92004-05-23 21:06:58 +00001759 example, uint to int or one pointer type to another.</li>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001760
Chris Lattner69bf8a92004-05-23 21:06:58 +00001761<br>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001762 <p>Derived Types</p>
1763
Chris Lattner261efe92003-11-25 01:02:51 +00001764 <ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001765 <li>SequentialType : This is subclassed by ArrayType and PointerType
Chris Lattner261efe92003-11-25 01:02:51 +00001766 <ul>
1767 <li><tt>const Type * getElementType() const</tt>: Returns the type of
1768each of the elements in the sequential type. </li>
1769 </ul>
1770 </li>
1771 <li>ArrayType : This is a subclass of SequentialType and defines
1772interface for array types.
1773 <ul>
1774 <li><tt>unsigned getNumElements() const</tt>: Returns the number of
1775elements in the array. </li>
1776 </ul>
1777 </li>
1778 <li>PointerType : Subclass of SequentialType for pointer types. </li>
1779 <li>StructType : subclass of DerivedTypes for struct types </li>
1780 <li>FunctionType : subclass of DerivedTypes for function types.
1781 <ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001782 <li><tt>bool isVarArg() const</tt>: Returns true if its a vararg
1783 function</li>
Chris Lattner261efe92003-11-25 01:02:51 +00001784 <li><tt> const Type * getReturnType() const</tt>: Returns the
Misha Brukman13fd15c2004-01-15 00:14:41 +00001785 return type of the function.</li>
Chris Lattner261efe92003-11-25 01:02:51 +00001786 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns
Misha Brukman13fd15c2004-01-15 00:14:41 +00001787 the type of the ith parameter.</li>
Chris Lattner261efe92003-11-25 01:02:51 +00001788 <li><tt> const unsigned getNumParams() const</tt>: Returns the
Misha Brukman13fd15c2004-01-15 00:14:41 +00001789 number of formal parameters.</li>
Chris Lattner261efe92003-11-25 01:02:51 +00001790 </ul>
1791 </li>
1792 </ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001793 </li>
Chris Lattnerc75ff9a2002-10-01 23:17:09 +00001794</ul>
Misha Brukman13fd15c2004-01-15 00:14:41 +00001795
1796</div>
1797
1798<!-- ======================================================================= -->
1799<div class="doc_subsection">
1800 <a name="Argument">The <tt>Argument</tt> class</a>
1801</div>
1802
1803<div class="doc_text">
1804
1805<p>This subclass of Value defines the interface for incoming formal
Chris Lattner261efe92003-11-25 01:02:51 +00001806arguments to a function. A Function maitanis a list of its formal
Misha Brukman13fd15c2004-01-15 00:14:41 +00001807arguments. An argument has a pointer to the parent Function.</p>
1808
1809</div>
1810
Reid Spencer096603a2004-05-26 08:41:35 +00001811<!-- ======================================================================= -->
1812<div class="doc_subsection">
1813 <a name="SymbolTable">The <tt>SymbolTable</tt> class</a>
1814</div>
1815<div class="doc_text">
1816<p>This class provides a symbol table that the
1817<a href="#Function"><tt>Function</tt></a> and <a href="#Module">
1818<tt>Module</tt></a> classes use for naming definitions. The symbol table can
1819provide a name for any <a href="#Value"><tt>Value</tt></a> or
1820<a href="#Type"><tt>Type</tt></a>. <tt>SymbolTable</tt> is an abstract data
1821type. It hides the data it contains and provides access to it through a
1822controlled interface.</p>
1823
1824<p>To use the <tt>SymbolTable</tt> well, you need to understand the
1825structure of the information it holds. The class contains two
1826<tt>std::map</tt> objects. The first, <tt>pmap</tt>, is a map of
1827<tt>Type*</tt> to maps of name (<tt>std::string</tt>) to <tt>Value*</tt>.
1828The second, <tt>tmap</tt>, is a map of names to <tt>Type*</tt>. Thus, Values
1829are stored in two-dimensions and accessed by <tt>Type</tt> and name. Types,
1830however, are stored in a single dimension and accessed only by name.</p>
1831
1832<p>The interface of this class provides three basic types of operations:
1833<ol>
1834 <li><em>Accessors</em>. Accessors provide read-only access to information
1835 such as finding a value for a name with the
1836 <a href="#SymbolTable_lookup">lookup</a> method.</li>
1837 <li><em>Mutators</em>. Mutators allow the user to add information to the
1838 <tt>SymbolTable</tt> with methods like
1839 <a href="#SymbolTable_insert"><tt>insert</tt></a>.</li>
1840 <li><em>Iterators</em>. Iterators allow the user to traverse the content
1841 of the symbol table in well defined ways, such as the method
1842 <a href="#SymbolTable_type_begin"><tt>type_begin</tt></a>.</li>
1843</ol>
1844
1845<h3>Accessors</h3>
1846<dl>
1847 <dt><tt>Value* lookup(const Type* Ty, const std::string&amp; name) const</tt>:
1848 </dt>
1849 <dd>The <tt>lookup</tt> method searches the type plane given by the
1850 <tt>Ty</tt> parameter for a <tt>Value</tt> with the provided <tt>name</tt>.
1851 If a suitable <tt>Value</tt> is not found, null is returned.</dd>
1852
1853 <dt><tt>Type* lookupType( const std::string&amp; name) const</tt>:</dt>
1854 <dd>The <tt>lookupType</tt> method searches through the types for a
1855 <tt>Type</tt> with the provided <tt>name</tt>. If a suitable <tt>Type</tt>
1856 is not found, null is returned.</dd>
1857
1858 <dt><tt>bool hasTypes() const</tt>:</dt>
1859 <dd>This function returns true if an entry has been made into the type
1860 map.</dd>
1861
1862 <dt><tt>bool isEmpty() const</tt>:</dt>
1863 <dd>This function returns true if both the value and types maps are
1864 empty</dd>
1865
1866 <dt><tt>std::string get_name(const Value*) const</tt>:</dt>
1867 <dd>This function returns the name of the Value provided or the empty
1868 string if the Value is not in the symbol table.</dd>
1869
1870 <dt><tt>std::string get_name(const Type*) const</tt>:</dt>
1871 <dd>This function returns the name of the Type provided or the empty
1872 string if the Type is not in the symbol table.</dd>
1873</dl>
1874
1875<h3>Mutators</h3>
1876<dl>
1877 <dt><tt>void insert(Value *Val)</tt>:</dt>
1878 <dd>This method adds the provided value to the symbol table. The Value must
1879 have both a name and a type which are extracted and used to place the value
1880 in the correct type plane under the value's name.</dd>
1881
1882 <dt><tt>void insert(const std::string&amp; Name, Value *Val)</tt>:</dt>
1883 <dd> Inserts a constant or type into the symbol table with the specified
1884 name. There can be a many to one mapping between names and constants
1885 or types.</dd>
1886
1887 <dt><tt>void insert(const std::string&amp; Name, Type *Typ)</tt>:</dt>
1888 <dd> Inserts a type into the symbol table with the specified name. There
1889 can be a many-to-one mapping between names and types. This method
1890 allows a type with an existing entry in the symbol table to get
1891 a new name.</dd>
1892
1893 <dt><tt>void remove(Value* Val)</tt>:</dt>
1894 <dd> This method removes a named value from the symbol table. The
1895 type and name of the Value are extracted from \p N and used to
1896 lookup the Value in the correct type plane. If the Value is
1897 not in the symbol table, this method silently ignores the
1898 request.</dd>
1899
1900 <dt><tt>void remove(Type* Typ)</tt>:</dt>
1901 <dd> This method removes a named type from the symbol table. The
1902 name of the type is extracted from \P T and used to look up
1903 the Type in the type map. If the Type is not in the symbol
1904 table, this method silently ignores the request.</dd>
1905
1906 <dt><tt>Value* remove(const std::string&amp; Name, Value *Val)</tt>:</dt>
1907 <dd> Remove a constant or type with the specified name from the
1908 symbol table.</dd>
1909
1910 <dt><tt>Type* remove(const std::string&amp; Name, Type* T)</tt>:</dt>
1911 <dd> Remove a type with the specified name from the symbol table.
1912 Returns the removed Type.</dd>
1913
1914 <dt><tt>Value *value_remove(const value_iterator&amp; It)</tt>:</dt>
1915 <dd> Removes a specific value from the symbol table.
1916 Returns the removed value.</dd>
1917
1918 <dt><tt>bool strip()</tt>:</dt>
1919 <dd> This method will strip the symbol table of its names leaving
1920 the type and values. </dd>
1921
1922 <dt><tt>void clear()</tt>:</dt>
1923 <dd>Empty the symbol table completely.</dd>
1924</dl>
1925
1926<h3>Iteration</h3>
1927<p>The following functions describe three types of iterators you can obtain
1928the beginning or end of the sequence for both const and non-const. It is
1929important to keep track of the different kinds of iterators. There are
1930three idioms worth pointing out:</p>
1931<table class="doc_table">
1932 <tr><th>Units</th><th>Iterator</th><th>Idiom</th></tr>
1933 <tr>
1934 <td>Planes Of name/Value maps</td><td>PI</td>
1935 <td><tt><pre>
1936for (SymbolTable::plane_const_iterator PI = ST.plane_begin(),
1937PE = ST.plane_end(); PI != PE; ++PI ) {
1938 PI-&gt;first // This is the Type* of the plane
1939 PI-&gt;second // This is the SymbolTable::ValueMap of name/Value pairs
1940 </pre></tt></td>
1941 </tr>
1942 <tr>
1943 <td>All name/Type Pairs</td><td>TI</td>
1944 <td><tt><pre>
1945for (SymbolTable::type_const_iterator TI = ST.type_begin(),
1946 TE = ST.type_end(); TI != TE; ++TI )
1947 TI-&gt;first // This is the name of the type
1948 TI-&gt;second // This is the Type* value associated with the name
1949 </pre></tt></td>
1950 </tr>
1951 <tr>
1952 <td>name/Value pairs in a plane</td><td>VI</td>
1953 <td><tt><pre>
1954for (SymbolTable::value_const_iterator VI = ST.value_begin(SomeType),
1955 VE = ST.value_end(SomeType); VI != VE; ++VI )
1956 VI-&gt;first // This is the name of the Value
1957 VI-&gt;second // This is the Value* value associated with the name
1958 </pre></tt></td>
1959 </tr>
1960</table>
1961<p>Using the recommended iterator names and idioms will help you avoid
1962making mistakes. Of particular note, make sure that whenever you use
1963value_begin(SomeType) that you always compare the resulting iterator
1964with value_end(SomeType) not value_end(SomeOtherType) or else you
1965will loop infinitely.</p>
1966
1967<dl>
1968
1969 <dt><tt>plane_iterator plane_begin()</tt>:</dt>
1970 <dd>Get an iterator that starts at the beginning of the type planes.
1971 The iterator will iterate over the Type/ValueMap pairs in the
1972 type planes. </dd>
1973
1974 <dt><tt>plane_const_iterator plane_begin() const</tt>:</dt>
1975 <dd>Get a const_iterator that starts at the beginning of the type
1976 planes. The iterator will iterate over the Type/ValueMap pairs
1977 in the type planes. </dd>
1978
1979 <dt><tt>plane_iterator plane_end()</tt>:</dt>
1980 <dd>Get an iterator at the end of the type planes. This serves as
1981 the marker for end of iteration over the type planes.</dd>
1982
1983 <dt><tt>plane_const_iterator plane_end() const</tt>:</dt>
1984 <dd>Get a const_iterator at the end of the type planes. This serves as
1985 the marker for end of iteration over the type planes.</dd>
1986
1987 <dt><tt>value_iterator value_begin(const Type *Typ)</tt>:</dt>
1988 <dd>Get an iterator that starts at the beginning of a type plane.
1989 The iterator will iterate over the name/value pairs in the type plane.
1990 Note: The type plane must already exist before using this.</dd>
1991
1992 <dt><tt>value_const_iterator value_begin(const Type *Typ) const</tt>:</dt>
1993 <dd>Get a const_iterator that starts at the beginning of a type plane.
1994 The iterator will iterate over the name/value pairs in the type plane.
1995 Note: The type plane must already exist before using this.</dd>
1996
1997 <dt><tt>value_iterator value_end(const Type *Typ)</tt>:</dt>
1998 <dd>Get an iterator to the end of a type plane. This serves as the marker
1999 for end of iteration of the type plane.
2000 Note: The type plane must already exist before using this.</dd>
2001
2002 <dt><tt>value_const_iterator value_end(const Type *Typ) const</tt>:</dt>
2003 <dd>Get a const_iterator to the end of a type plane. This serves as the
2004 marker for end of iteration of the type plane.
2005 Note: the type plane must already exist before using this.</dd>
2006
2007 <dt><tt>type_iterator type_begin()</tt>:</dt>
2008 <dd>Get an iterator to the start of the name/Type map.</dd>
2009
2010 <dt><tt>type_const_iterator type_begin() cons</tt>:</dt>
2011 <dd> Get a const_iterator to the start of the name/Type map.</dd>
2012
2013 <dt><tt>type_iterator type_end()</tt>:</dt>
2014 <dd>Get an iterator to the end of the name/Type map. This serves as the
2015 marker for end of iteration of the types.</dd>
2016
2017 <dt><tt>type_const_iterator type_end() const</tt>:</dt>
2018 <dd>Get a const-iterator to the end of the name/Type map. This serves
2019 as the marker for end of iteration of the types.</dd>
2020
2021 <dt><tt>plane_const_iterator find(const Type* Typ ) const</tt>:</dt>
2022 <dd>This method returns a plane_const_iterator for iteration over
2023 the type planes starting at a specific plane, given by \p Ty.</dd>
2024
2025 <dt><tt>plane_iterator find( const Type* Typ </tt>:</dt>
2026 <dd>This method returns a plane_iterator for iteration over the
2027 type planes starting at a specific plane, given by \p Ty.</dd>
2028
2029 <dt><tt>const ValueMap* findPlane( const Type* Typ ) cons</tt>:</dt>
2030 <dd>This method returns a ValueMap* for a specific type plane. This
2031 interface is deprecated and may go away in the future.</dd>
2032</dl>
2033</div>
2034
Chris Lattner9355b472002-09-06 02:50:58 +00002035<!-- *********************************************************************** -->
Misha Brukman13fd15c2004-01-15 00:14:41 +00002036<hr>
2037<address>
2038 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
2039 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
2040 <a href="http://validator.w3.org/check/referer"><img
2041 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
2042
2043 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
2044 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
2045 <a href="http://llvm.cs.uiuc.edu">The LLVM Compiler Infrastructure</a><br>
2046 Last modified: $Date$
2047</address>
2048
Chris Lattner261efe92003-11-25 01:02:51 +00002049</body>
2050</html>
Reid Spencer096603a2004-05-26 08:41:35 +00002051<!-- vim: sw=2 noai
2052-->