Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / cc377df7f5286801e5d294d1f3971095560e7bc1 / . / docs / ProgrammersManual.html
blob: 8378b5c06e28f0dc679ff9957c25deec3e4fefcc [file] [log] [blame]
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
2<html><head><title>LLVM Programmer's Manual</title></head>
3
4<body bgcolor=white>
5
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]6<table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
7<tr><td>&nbsp; <font size=+3 color="#EEEEFF" face="Georgia,Palatino,Times,Roman"><b>LLVM Programmer's Manual</b></font></td>
8</tr></table>
9
10<ol>
11 <li><a href="#introduction">Introduction</a>
12 <li><a href="#general">General Information</a>
13 <ul>
14 <li><a href="#stl">The C++ Standard Template Library</a>
Chris Lattner1d43fd42002-09-09 05:53:21 +0000[diff] [blame]15 <li><a href="#isa">The <tt>isa&lt;&gt;</tt>, <tt>cast&lt;&gt;</tt> and
16 <tt>dyn_cast&lt;&gt;</tt> templates</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]17 </ul>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]18 <li><a href="#common">Helpful Hints for Common Operations</a>
19 <ul>
20 <li><a href="#inspection">Basic Inspection and Traversal Routines</a>
21 <ul>
22 <li><a href="#iterate_function">Iterating over the <tt>BasicBlock</tt>s
23 in a <tt>Function</tt></a>
24 <li><a href="#iterate_basicblock">Iterating over the <tt>Instruction</tt>s
25 in a <tt>BasicBlock</tt></a>
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]26 <li><a href="#iterate_institer">Iterating over the <tt>Instruction</tt>s
27 in a <tt>Function</tt></a>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]28 <li><a href="#iterate_convert">Turning an iterator into a class
29 pointer</a>
Chris Lattnerf1ebdc32002-09-06 22:09:21 +0000[diff] [blame]30 <li><a href="#iterate_complex">Finding call sites: a more complex
31 example</a>
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]32 <li><a href="#iterate_chains">Iterating over def-use &amp; use-def
33 chains</a>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]34 </ul>
35 <li><a href="#simplechanges">Making simple changes</a>
36 <ul>
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]37 <li><a href="#schanges_creating">Creating and inserting new
38 <tt>Instruction</tt>s</a>
39 <li><a href="#schanges_deleting">Deleting
40 <tt>Instruction</tt>s</a>
41 <li><a href="#schanges_replacing">Replacing an
42 <tt>Instruction</tt> with another <tt>Value</tt></a>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]43 </ul>
44<!--
45 <li>Working with the Control Flow Graph
46 <ul>
47 <li>Accessing predecessors and successors of a <tt>BasicBlock</tt>
48 <li>
49 <li>
50 </ul>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]51 <li>Useful LLVM APIs
52 <ul>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]53 <li>The general graph API
54 <li>The <tt>InstVisitor</tt> template
55 <li>The DEBUG() macro
56 <li>The <tt>Statistic</tt> template
57-->
58 </ul>
59<!--
60 <li>Useful related topics
61 <ul>
62 <li>The <tt>-time-passes</tt> option
63 <li>How to use the LLVM Makefile system
64 <li>How to write a regression test
65 <li>
66 </ul>
67-->
68 </ul>
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]69 <li><a href="#coreclasses">The Core LLVM Class Hierarchy Reference</a>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]70 <ul>
71 <li><a href="#Value">The <tt>Value</tt> class</a>
72 <ul>
73 <li><a href="#User">The <tt>User</tt> class</a>
74 <ul>
75 <li><a href="#Instruction">The <tt>Instruction</tt> class</a>
76 <ul>
77 <li>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]78 </ul>
79 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
80 <ul>
81 <li><a href="#BasicBlock">The <tt>BasicBlock</tt> class</a>
82 <li><a href="#Function">The <tt>Function</tt> class</a>
83 <li><a href="#GlobalVariable">The <tt>GlobalVariable</tt> class</a>
84 </ul>
85 <li><a href="#Module">The <tt>Module</tt> class</a>
86 <li><a href="#Constant">The <tt>Constant</tt> class</a>
87 <ul>
88 <li>
89 <li>
90 </ul>
91 </ul>
92 <li><a href="#Type">The <tt>Type</tt> class</a>
93 <li><a href="#Argument">The <tt>Argument</tt> class</a>
94 </ul>
95 <li>The <tt>SymbolTable</tt> class
96 <li>The <tt>ilist</tt> and <tt>iplist</tt> classes
97 <ul>
98 <li>Creating, inserting, moving and deleting from LLVM lists
99 </ul>
100 <li>Important iterator invalidation semantics to be aware of
101 </ul>
102
Chris Lattner6b121f12002-09-10 15:20:46 +0000[diff] [blame]103 <p><b>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
104 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>, and
Chris Lattnerf1ebdc32002-09-06 22:09:21 +0000[diff] [blame]105 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a></b><p>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]106</ol>
107
108
109<!-- *********************************************************************** -->
110<table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
111<tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
112<a name="introduction">Introduction
113</b></font></td></tr></table><ul>
114<!-- *********************************************************************** -->
115
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]116This document is meant to highlight some of the important classes and interfaces
117available in the LLVM source-base. This manual is not intended to explain what
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]118LLVM is, how it works, and what LLVM code looks like. It assumes that you know
119the basics of LLVM and are interested in writing transformations or otherwise
120analyzing or manipulating the code.<p>
121
122This document should get you oriented so that you can find your way in the
123continuously growing source code that makes up the LLVM infrastructure. Note
124that this manual is not intended to serve as a replacement for reading the
125source code, so if you think there should be a method in one of these classes to
126do something, but it's not listed, check the source. Links to the <a
127href="/doxygen/">doxygen</a> sources are provided to make this as easy as
128possible.<p>
129
130The first section of this document describes general information that is useful
131to know when working in the LLVM infrastructure, and the second describes the
132Core LLVM classes. In the future this manual will be extended with information
133describing how to use extension libraries, such as dominator information, CFG
134traversal routines, and useful utilities like the <tt><a
135href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.<p>
136
137
138<!-- *********************************************************************** -->
139</ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
140<tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
141<a name="general">General Information
142</b></font></td></tr></table><ul>
143<!-- *********************************************************************** -->
144
145This section contains general information that is useful if you are working in
146the LLVM source-base, but that isn't specific to any particular API.<p>
147
148
149<!-- ======================================================================= -->
150</ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
151<tr><td>&nbsp;</td><td width="100%">&nbsp;
152<font color="#EEEEFF" face="Georgia,Palatino"><b>
153<a name="stl">The C++ Standard Template Library</a>
154</b></font></td></tr></table><ul>
155
156LLVM makes heavy use of the C++ Standard Template Library (STL), perhaps much
157more than you are used to, or have seen before. Because of this, you might want
158to do a little background reading in the techniques used and capabilities of the
159library. There are many good pages that discuss the STL, and several books on
160the subject that you can get, so it will not be discussed in this document.<p>
161
162Here are some useful links:<p>
163<ol>
164<li><a href="http://www.dinkumware.com/htm_cpl/index.html">Dinkumware C++
165Library reference</a> - an excellent reference for the STL and other parts of
166the standard C++ library.<br>
167
168<li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
169Questions</a>
170
171<li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
172Contains a useful <a
173href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
174STL</a>.
175
176<li><a href="http://www.research.att.com/~bs/C++.html">Bjarne Stroustrup's C++
177Page</a>
178
179</ol><p>
180
181You are also encouraged to take a look at the <a
182href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
183to write maintainable code more than where to put your curly braces.<p>
184
185
Chris Lattner1d43fd42002-09-09 05:53:21 +0000[diff] [blame]186<!-- ======================================================================= -->
187</ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
188<tr><td>&nbsp;</td><td width="100%">&nbsp;
189<font color="#EEEEFF" face="Georgia,Palatino"><b>
190<a name="isa">The isa&lt;&gt;, cast&lt;&gt; and dyn_cast&lt;&gt; templates</a>
191</b></font></td></tr></table><ul>
192
Chris Lattner979d9b72002-09-10 00:39:05 +0000[diff] [blame]193The LLVM source-base makes extensive use of a custom form of RTTI. These
194templates have many similarities to the C++ <tt>dynamic_cast&lt;&gt;</tt>
195operator, but they don't have some drawbacks (primarily stemming from the fact
196that <tt>dynamic_cast&lt;&gt;</tt> only works on classes that have a v-table).
197Because they are used so often, you must know what they do and how they work.
198All of these templates are defined in the <a
199href="/doxygen/Casting_8h-source.html"><tt>Support/Casting.h</tt></a> file (note
200that you very rarely have to include this file directly).<p>
Chris Lattner1d43fd42002-09-09 05:53:21 +0000[diff] [blame]201
Chris Lattner979d9b72002-09-10 00:39:05 +0000[diff] [blame]202<dl>
203
204<dt><tt>isa&lt;&gt;</tt>:
205
206<dd>The <tt>isa&lt;&gt;</tt> operator works exactly like the Java
207"<tt>instanceof</tt>" operator. It returns true or false depending on whether a
208reference or pointer points to an instance of the specified class. This can be
209very useful for constraint checking of various sorts (example below).<p>
210
211
212<dt><tt>cast&lt;&gt;</tt>:
213
214<dd>The <tt>cast&lt;&gt;</tt> operator is a "checked cast" operation. It
215converts a pointer or reference from a base class to a derived cast, causing an
216assertion failure if it is not really an instance of the right type. This
217should be used in cases where you have some information that makes you believe
218that something is of the right type. An example of the <tt>isa&lt;&gt;</tt> and
219<tt>cast&lt;&gt;</tt> template is:<p>
220
221<pre>
222static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
223 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))
224 return true;
225
226 <i>// Otherwise, it must be an instruction...</i>
227 return !L->contains(cast&lt;<a href="#Instruction">Instruction</a>&gt;(V)->getParent());
228</pre><p>
229
230Note that you should <b>not</b> use an <tt>isa&lt;&gt;</tt> test followed by a
231<tt>cast&lt;&gt;</tt>, for that use the <tt>dyn_cast&lt;&gt;</tt> operator.<p>
232
233
234<dt><tt>dyn_cast&lt;&gt;</tt>:
235
236<dd>The <tt>dyn_cast&lt;&gt;</tt> operator is a "checking cast" operation. It
237checks to see if the operand is of the specified type, and if so, returns a
238pointer to it (this operator does not work with references). If the operand is
239not of the correct type, a null pointer is returned. Thus, this works very much
240like the <tt>dynamic_cast</tt> operator in C++, and should be used in the same
Chris Lattner6b121f12002-09-10 15:20:46 +0000[diff] [blame]241circumstances. Typically, the <tt>dyn_cast&lt;&gt;</tt> operator is used in an
242<tt>if</tt> statement or some other flow control statement like this:<p>
243
244<pre>
245 if (<a href="#AllocationInst">AllocationInst</a> *AI = dyn_cast&lt;<a href="#AllocationInst">AllocationInst</a>&gt;(Val)) {
246 ...
247 }
248</pre><p>
249
250This form of the <tt>if</tt> statement effectively combines together a call to
251<tt>isa&lt;&gt;</tt> and a call to <tt>cast&lt;&gt;</tt> into one statement,
252which is very convenient.<p>
253
254Another common example is:<p>
Chris Lattner979d9b72002-09-10 00:39:05 +0000[diff] [blame]255
256<pre>
257 <i>// Loop over all of the phi nodes in a basic block</i>
258 BasicBlock::iterator BBI = BB->begin();
259 for (; <a href="#PhiNode">PHINode</a> *PN = dyn_cast&lt;<a href="#PHINode">PHINode</a>&gt;(&amp;*BBI); ++BBI)
260 cerr &lt;&lt; *PN;
261</pre><p>
262
Chris Lattner6b121f12002-09-10 15:20:46 +0000[diff] [blame]263Note that the <tt>dyn_cast&lt;&gt;</tt> operator, like C++'s
264<tt>dynamic_cast</tt> or Java's <tt>instanceof</tt> operator, can be abused. In
265particular you should not use big chained <tt>if/then/else</tt> blocks to check
266for lots of different variants of classes. If you find yourself wanting to do
267this, it is much cleaner and more efficient to use the InstVisitor class to
268dispatch over the instruction type directly.<p>
Chris Lattner979d9b72002-09-10 00:39:05 +0000[diff] [blame]269
270
Chris Lattner6b121f12002-09-10 15:20:46 +0000[diff] [blame]271<dt><tt>cast_or_null&lt;&gt;</tt>:
272
273<dd>The <tt>cast_or_null&lt;&gt;</tt> operator works just like the
274<tt>cast&lt;&gt;</tt> operator, except that it allows for a null pointer as an
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]275argument (which it then propagates). This can sometimes be useful, allowing you
Chris Lattner6b121f12002-09-10 15:20:46 +0000[diff] [blame]276to combine several null checks into one.<p>
277
278
279<dt><tt>dyn_cast_or_null&lt;&gt;</tt>:
280
281<dd>The <tt>dyn_cast_or_null&lt;&gt;</tt> operator works just like the
282<tt>dyn_cast&lt;&gt;</tt> operator, except that it allows for a null pointer as
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]283an argument (which it then propagates). This can sometimes be useful, allowing
Chris Lattner6b121f12002-09-10 15:20:46 +0000[diff] [blame]284you to combine several null checks into one.<p>
285
Chris Lattner979d9b72002-09-10 00:39:05 +0000[diff] [blame]286</dl>
Chris Lattner1d43fd42002-09-09 05:53:21 +0000[diff] [blame]287
Chris Lattner6b121f12002-09-10 15:20:46 +0000[diff] [blame]288These five templates can be used with any classes, whether they have a v-table
289or not. To add support for these templates, you simply need to add
290<tt>classof</tt> static methods to the class you are interested casting to.
291Describing this is currently outside the scope of this document, but there are
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]292lots of examples in the LLVM source base.<p>
Chris Lattner1d43fd42002-09-09 05:53:21 +0000[diff] [blame]293
294
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]295
Chris Lattnerb99344f2002-09-06 16:40:10 +0000[diff] [blame]296<!-- *********************************************************************** -->
297</ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
298<tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
299<a name="common">Helpful Hints for Common Operations
300</b></font></td></tr></table><ul>
301<!-- *********************************************************************** -->
302
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]303This section describes how to perform some very simple transformations of LLVM
304code. This is meant to give examples of common idioms used, showing the
305practical side of LLVM transformations.<p>
306
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]307Because this is a "how-to" section, you should also read about the main classes
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]308that you will be working with. The <a href="#coreclasses">Core LLVM Class
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]309Hierarchy Reference</a> contains details and descriptions of the main classes
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]310that you should know about.<p>
311
312<!-- NOTE: this section should be heavy on example code -->
313
314
315<!-- ======================================================================= -->
316</ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
317<tr><td>&nbsp;</td><td width="100%">&nbsp;
318<font color="#EEEEFF" face="Georgia,Palatino"><b>
319<a name="inspection">Basic Inspection and Traversal Routines</a>
320</b></font></td></tr></table><ul>
321
Chris Lattnercaa5d132002-09-09 19:58:18 +0000[diff] [blame]322The LLVM compiler infrastructure have many different data structures that may be
323traversed. Following the example of the C++ standard template library, the
324techniques used to traverse these various data structures are all basically the
325same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
326method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
327function returns an iterator pointing to one past the last valid element of the
328sequence, and there is some <tt>XXXiterator</tt> data type that is common
329between the two operations.<p>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]330
Chris Lattnercaa5d132002-09-09 19:58:18 +0000[diff] [blame]331Because the pattern for iteration is common across many different aspects of the
332program representation, the standard template library algorithms may be used on
333them, and it is easier to remember how to iterate. First we show a few common
334examples of the data structures that need to be traversed. Other data
335structures are traversed in very similar ways.<p>
336
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]337
338<!-- _______________________________________________________________________ -->
Chris Lattnercaa5d132002-09-09 19:58:18 +0000[diff] [blame]339</ul><h4><a name="iterate_function"><hr size=0>Iterating over the <a
340href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
341href="#Function"><tt>Function</tt></a> </h4><ul>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]342
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]343It's quite common to have a <tt>Function</tt> instance that you'd like
344to transform in some way; in particular, you'd like to manipulate its
345<tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over
346all of the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>.
347The following is an example that prints the name of a
348<tt>BasicBlock</tt> and the number of <tt>Instruction</tt>s it
349contains:
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]350
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]351<pre>
352 // func is a pointer to a Function instance
353 for(Function::iterator i = func->begin(), e = func->end(); i != e; ++i) {
354
355 // print out the name of the basic block if it has one, and then the
356 // number of instructions that it contains
357
Joel Stanley72ef35e2002-09-06 23:05:12 +0000[diff] [blame]358 cerr &lt;&lt "Basic block (name=" &lt;&lt i-&gt;getName() &lt;&lt; ") has "
359 &lt;&lt i-&gt;size() &lt;&lt " instructions.\n";
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]360 }
361</pre>
362
363Note that i can be used as if it were a pointer for the purposes of
364invoking member functions of the <tt>Instruction</tt> class. This is
365because the indirection operator is overloaded for the iterator
366classes. In the above code, the expression <tt>i->size()</tt> is
367exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]368
369<!-- _______________________________________________________________________ -->
Chris Lattnercaa5d132002-09-09 19:58:18 +0000[diff] [blame]370</ul><h4><a name="iterate_basicblock"><hr size=0>Iterating over the <a
371href="#Instruction"><tt>Instruction</tt></a>s in a <a
372href="#BasicBlock"><tt>BasicBlock</tt></a> </h4><ul>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]373
Joel Stanleyaaeb1c12002-09-06 23:42:40 +0000[diff] [blame]374Just like when dealing with <tt>BasicBlock</tt>s in
375<tt>Function</tt>s, it's easy to iterate over the individual
376instructions that make up <tt>BasicBlock</tt>s. Here's a code snippet
377that prints out each instruction in a <tt>BasicBlock</tt>:
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]378
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]379<pre>
380 // blk is a pointer to a BasicBlock instance
Chris Lattnercaa5d132002-09-09 19:58:18 +0000[diff] [blame]381 for(BasicBlock::iterator i = blk-&gt;begin(), e = blk-&gt;end(); i != e; ++i)
Chris Lattner2b763062002-09-06 22:51:10 +0000[diff] [blame]382 // the next statement works since operator&lt;&lt;(ostream&amp;,...)
383 // is overloaded for Instruction&amp;
Chris Lattnercaa5d132002-09-09 19:58:18 +0000[diff] [blame]384 cerr &lt;&lt; *i &lt;&lt; "\n";
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]385</pre>
386
387However, this isn't really the best way to print out the contents of a
388<tt>BasicBlock</tt>! Since the ostream operators are overloaded for
389virtually anything you'll care about, you could have just invoked the
Chris Lattner2b763062002-09-06 22:51:10 +0000[diff] [blame]390print routine on the basic block itself: <tt>cerr &lt;&lt; *blk &lt;&lt;
391"\n";</tt>.<p>
392
393Note that currently operator&lt;&lt; is implemented for <tt>Value*</tt>, so it
394will print out the contents of the pointer, instead of
395the pointer value you might expect. This is a deprecated interface that will
396be removed in the future, so it's best not to depend on it. To print out the
397pointer value for now, you must cast to <tt>void*</tt>.<p>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]398
Chris Lattnercaa5d132002-09-09 19:58:18 +0000[diff] [blame]399
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]400<!-- _______________________________________________________________________ -->
Chris Lattnercaa5d132002-09-09 19:58:18 +0000[diff] [blame]401</ul><h4><a name="iterate_institer"><hr size=0>Iterating over the <a
402href="#Instruction"><tt>Instruction</tt></a>s in a <a
403href="#Function"><tt>Function</tt></a></h4><ul>
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]404
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]405If you're finding that you commonly iterate over a <tt>Function</tt>'s
406<tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s
407<tt>Instruction</tt>s, <tt>InstIterator</tt> should be used instead.
Chris Lattnercaa5d132002-09-09 19:58:18 +0000[diff] [blame]408You'll need to include <a href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>, and then
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]409instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
410small example that shows how to dump all instructions in a function to
411stderr (<b>Note:</b> Dereferencing an <tt>InstIterator</tt> yields an
412<tt>Instruction*</tt>, <i>not</i> an <tt>Instruction&amp</tt>!):
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]413
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]414<pre>
Chris Lattnercaa5d132002-09-09 19:58:18 +0000[diff] [blame]415#include "<a href="/doxygen/InstIterator_8h-source.html">llvm/Support/InstIterator.h</a>"
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]416...
417// Suppose F is a ptr to a function
418for(inst_iterator i = inst_begin(F), e = inst_end(F); i != e; ++i)
419 cerr &lt;&lt **i &lt;&lt "\n";
420</pre>
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]421
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]422Easy, isn't it? You can also use <tt>InstIterator</tt>s to fill a
423worklist with its initial contents. For example, if you wanted to
424initialize a worklist to contain all instructions in a
425<tt>Function</tt> F, all you would need to do is something like:
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]426
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]427<pre>
428std::set&lt;Instruction*&gt worklist;
429worklist.insert(inst_begin(F), inst_end(F));
430</pre>
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]431
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]432The STL set <tt>worklist</tt> would now contain all instructions in
433the <tt>Function</tt> pointed to by F.
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]434
435<!-- _______________________________________________________________________ -->
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]436</ul><h4><a name="iterate_convert"><hr size=0>Turning an iterator into a class
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]437pointer (and vice-versa) </h4><ul>
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]438
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]439Sometimes, it'll be useful to grab a reference (or pointer) to a class
440instance when all you've got at hand is an iterator. Well, extracting
441a reference or a pointer from an iterator is very straightforward.
442Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and
443<tt>j</tt> is a <tt>BasicBlock::const_iterator</tt>:
444
445<pre>
Chris Lattner83b5ee02002-09-06 22:12:58 +0000[diff] [blame]446 Instruction&amp; inst = *i; // grab reference to instruction reference
447 Instruction* pinst = &amp;*i; // grab pointer to instruction reference
448 const Instruction&amp; inst = *j;
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]449</pre>
450However, the iterators you'll be working with in the LLVM framework
451are special: they will automatically convert to a ptr-to-instance type
452whenever they need to. Instead of dereferencing the iterator and then
453taking the address of the result, you can simply assign the iterator
454to the proper pointer type and you get the dereference and address-of
455operation as a result of the assignment (behind the scenes, this is a
456result of overloading casting mechanisms). Thus the last line of the
457last example,
458
Chris Lattner83b5ee02002-09-06 22:12:58 +0000[diff] [blame]459<pre>Instruction* pinst = &amp;*i;</pre>
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]460
461is semantically equivalent to
462
463<pre>Instruction* pinst = i;</pre>
464
Chris Lattner979d9b72002-09-10 00:39:05 +0000[diff] [blame]465<b>Caveat emptor</b>: The above syntax works <i>only</i> when you're <i>not</i>
466working with <tt>dyn_cast</tt>. The template definition of <tt><a
467href="#isa">dyn_cast</a></tt> isn't implemented to handle this yet, so you'll
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]468still need the following in order for things to work properly:
469
470<pre>
471BasicBlock::iterator bbi = ...;
Chris Lattnercaa5d132002-09-09 19:58:18 +0000[diff] [blame]472<a href="#BranchInst">BranchInst</a>* b = <a href="#isa">dyn_cast</a>&lt;<a href="#BranchInst">BranchInst</a>&gt;(&amp;*bbi);
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]473</pre>
474
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]475It's also possible to turn a class pointer into the corresponding
476iterator. Usually, this conversion is quite inexpensive. The
477following code snippet illustrates use of the conversion constructors
478provided by LLVM iterators. By using these, you can explicitly grab
479the iterator of something without actually obtaining it via iteration
480over some structure:
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]481
482<pre>
483void printNextInstruction(Instruction* inst) {
484 BasicBlock::iterator it(inst);
485 ++it; // after this line, it refers to the instruction after *inst.
Chris Lattnercaa5d132002-09-09 19:58:18 +0000[diff] [blame]486 if(it != inst-&gt;getParent()->end()) cerr &lt;&lt; *it &lt;&lt; "\n";
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]487}
488</pre>
Joel Stanleyaaeb1c12002-09-06 23:42:40 +0000[diff] [blame]489Of course, this example is strictly pedagogical, because it'd be much
490better to explicitly grab the next instruction directly from inst.
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]491
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]492
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]493<!--_______________________________________________________________________-->
494</ul><h4><a name="iterate_complex"><hr size=0>Finding call sites: a slightly
495more complex example </h4><ul>
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]496
497Say that you're writing a FunctionPass and would like to count all the
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]498locations in the entire module (that is, across every
Joel Stanleyd8aabb22002-09-09 16:29:58 +0000[diff] [blame]499<tt>Function</tt>) where a certain function (i.e. some
500<tt>Function</tt>*) already in scope. As you'll learn later, you may
501want to use an <tt>InstVisitor</tt> to accomplish this in a much more
502straightforward manner, but this example will allow us to explore how
503you'd do it if you didn't have <tt>InstVisitor</tt> around. In
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]504pseudocode, this is what we want to do:
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]505
506<pre>
507initialize callCounter to zero
508for each Function f in the Module
509 for each BasicBlock b in f
510 for each Instruction i in b
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]511 if(i is a CallInst and calls the given function)
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]512 increment callCounter
513</pre>
514
515And the actual code is (remember, since we're writing a
Joel Stanleyd8aabb22002-09-09 16:29:58 +0000[diff] [blame]516<tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]517has to override the <tt>runOnFunction</tt> method...):
518
519<pre>
Joel Stanleyd8aabb22002-09-09 16:29:58 +0000[diff] [blame]520Function* targetFunc = ...;
521
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]522class OurFunctionPass : public FunctionPass {
523 public:
Joel Stanleyd8aabb22002-09-09 16:29:58 +0000[diff] [blame]524 OurFunctionPass(): callCounter(0) { }
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]525
Chris Lattnercaa5d132002-09-09 19:58:18 +0000[diff] [blame]526 virtual runOnFunction(Function&amp; F) {
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]527 for(Function::iterator b = F.begin(), be = F.end(); b != be; ++b) {
528 for(BasicBlock::iterator i = b-&gt;begin(); ie = b-&gt;end(); i != ie; ++i) {
Chris Lattnera9030cb2002-09-16 22:08:07 +0000[diff] [blame]529 if (<a href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a>&lt;<a href="#CallInst">CallInst</a>&gt;(&amp;*i)) {
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]530 // we know we've encountered a call instruction, so we
531 // need to determine if it's a call to the
532 // function pointed to by m_func or not.
533
Joel Stanleyd8aabb22002-09-09 16:29:58 +0000[diff] [blame]534 if(callInst-&gt;getCalledFunction() == targetFunc)
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]535 ++callCounter;
536 }
537 }
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]538 }
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]539
540 private:
Joel Stanleyd8aabb22002-09-09 16:29:58 +0000[diff] [blame]541 unsigned callCounter;
Joel Stanleye7be6502002-09-09 15:50:33 +0000[diff] [blame]542};
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]543</pre>
544
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]545<!--_______________________________________________________________________-->
546</ul><h4><a name="iterate_chains"><hr size=0>Iterating over def-use &amp;
547use-def chains</h4><ul>
548
Joel Stanley01040b22002-09-11 20:50:04 +0000[diff] [blame]549Frequently, we might have an instance of the <a
550href="/doxygen/classValue.html">Value Class</a> and we want to
551determine which <tt>User</tt>s use the <tt>Value</tt>. The list of
552all <tt>User</tt>s of a particular <tt>Value</tt> is called a
553<i>def-use</i> chain. For example, let's say we have a
554<tt>Function*</tt> named <tt>F</tt> to a particular function
555<tt>foo</tt>. Finding all of the instructions that <i>use</i>
556<tt>foo</tt> is as simple as iterating over the <i>def-use</i> chain of
557<tt>F</tt>:
558
559<pre>
560Function* F = ...;
561
562for(Value::use_iterator i = F-&gt;use_begin(), e = F-&gt;use_end(); i != e; ++i) {
Chris Lattner24b70922002-09-17 22:43:00 +0000[diff] [blame]563 if(Instruction* Inst = dyn_cast&lt;Instruction&gt;(*i)) {
564 cerr &lt;&lt; "F is used in instruction:\n";
565 cerr &lt;&lt; *Inst &lt;&lt; "\n";
Joel Stanley01040b22002-09-11 20:50:04 +0000[diff] [blame]566 }
567}
568</pre>
569
570Alternately, it's common to have an instance of the <a
571href="/doxygen/classUser.html">User Class</a> and need to know what
572<tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used
573by a <tt>User</tt> is known as a <i>use-def</i> chain. Instances of
574class <tt>Instruction</tt> are common <tt>User</tt>s, so we might want
575to iterate over all of the values that a particular instruction uses
576(that is, the operands of the particular <tt>Instruction</tt>):
577
578<pre>
579Instruction* pi = ...;
580
581for(User::op_iterator i = pi-&gt;op_begin(), e = pi-&gt;op_end(); i != e; ++i) {
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]582 Value* v = *i;
Joel Stanley01040b22002-09-11 20:50:04 +0000[diff] [blame]583 ...
584}
585</pre>
586
587
Chris Lattner1a3105b2002-09-09 05:49:39 +0000[diff] [blame]588<!--
589 def-use chains ("finding all users of"): Value::use_begin/use_end
590 use-def chains ("finding all values used"): User::op_begin/op_end [op=operand]
591-->
592
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]593<!-- ======================================================================= -->
594</ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
595<tr><td>&nbsp;</td><td width="100%">&nbsp;
596<font color="#EEEEFF" face="Georgia,Palatino"><b>
597<a name="simplechanges">Making simple changes</a>
598</b></font></td></tr></table><ul>
599
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]600There are some primitive transformation operations present in the LLVM
601infrastructure that are worth knowing about. When performing
602transformations, it's fairly common to manipulate the contents of
603basic blocks. This section describes some of the common methods for
604doing so and gives example code.
605
606<!--_______________________________________________________________________-->
607</ul><h4><a name="schanges_creating"><hr size=0>Creating and inserting
608 new <tt>Instruction</tt>s</h4><ul>
609
610<i>Instantiating Instructions</i>
611
612<p>Creation of <tt>Instruction</tt>s is straightforward: simply call the
613constructor for the kind of instruction to instantiate and provide the
614necessary parameters. For example, an <tt>AllocaInst</tt> only
615<i>requires</i> a (const-ptr-to) <tt>Type</tt>. Thus:
616
617<pre>AllocaInst* ai = new AllocaInst(Type::IntTy);</pre>
618
619will create an <tt>AllocaInst</tt> instance that represents the
620allocation of one integer in the current stack frame, at runtime.
621Each <tt>Instruction</tt> subclass is likely to have varying default
622parameters which change the semantics of the instruction, so refer to
Chris Lattner4e1f96b2002-09-12 19:06:51 +0000[diff] [blame]623the <a href="/doxygen/classInstruction.html">doxygen documentation for
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]624the subclass of Instruction</a> that you're interested in
625instantiating.</p>
626
627<p><i>Naming values</i></p>
628
629<p>
630It is very useful to name the values of instructions when you're able
631to, as this facilitates the debugging of your transformations. If you
632end up looking at generated LLVM machine code, you definitely want to
633have logical names associated with the results of instructions! By
634supplying a value for the <tt>Name</tt> (default) parameter of the
635<tt>Instruction</tt> constructor, you associate a logical name with
636the result of the instruction's execution at runtime. For example,
637say that I'm writing a transformation that dynamically allocates space
638for an integer on the stack, and that integer is going to be used as
639some kind of index by some other code. To accomplish this, I place an
640<tt>AllocaInst</tt> at the first point in the first
641<tt>BasicBlock</tt> of some <tt>Function</tt>, and I'm intending to
642use it within the same <tt>Function</tt>. I might do:
643
644<pre>AllocaInst* pa = new AllocaInst(Type::IntTy, 0, "indexLoc");</pre>
645
646where <tt>indexLoc</tt> is now the logical name of the instruction's
647execution value, which is a pointer to an integer on the runtime
648stack.
649</p>
650
651<p><i>Inserting instructions</i></p>
652
653<p>
654There are essentially two ways to insert an <tt>Instruction</tt> into
655an existing sequence of instructions that form a <tt>BasicBlock</tt>:
656<ul>
657<li>Insertion into an explicit instruction list
658
659<p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within
660that <tt>BasicBlock</tt>, and a newly-created instruction
661we wish to insert before <tt>*pi</tt>, we do the following:
662
663<pre>
664BasicBlock* pb = ...;
665Instruction* pi = ...;
666Instruction* newInst = new Instruction(...);
667pb->getInstList().insert(pi, newInst); // inserts newInst before pi in pb
668</pre>
669</p>
670
671<li>Insertion into an implicit instruction list
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]672<p><tt>Instruction</tt> instances that are already in
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]673<tt>BasicBlock</tt>s are implicitly associated with an existing
674instruction list: the instruction list of the enclosing basic block.
675Thus, we could have accomplished the same thing as the above code
676without being given a <tt>BasicBlock</tt> by doing:
677<pre>
678Instruction* pi = ...;
679Instruction* newInst = new Instruction(...);
680pi->getParent()->getInstList().insert(pi, newInst);
681</pre>
682In fact, this sequence of steps occurs so frequently that the
683<tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes
684provide constructors which take (as a default parameter) a pointer to
685an <tt>Instruction</tt> which the newly-created <tt>Instruction</tt>
686should precede. That is, <tt>Instruction</tt> constructors are
687capable of inserting the newly-created instance into the
688<tt>BasicBlock</tt> of a provided instruction, immediately before that
689instruction. Using an <tt>Instruction</tt> constructor with a
690<tt>insertBefore</tt> (default) parameter, the above code becomes:
691<pre>
692Instruction* pi = ...;
693Instruction* newInst = new Instruction(..., pi);
694</pre>
695which is much cleaner, especially if you're creating a lot of
696instructions and adding them to <tt>BasicBlock</tt>s.
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]697 </p>
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]698</p>
Chris Lattner9ebf5162002-09-12 19:08:16 +0000[diff] [blame]699</ul>
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]700
701<!--_______________________________________________________________________-->
702</ul><h4><a name="schanges_deleting"><hr size=0>Deleting
Chris Lattner4e1f96b2002-09-12 19:06:51 +0000[diff] [blame]703<tt>Instruction</tt>s</h4><ul>
704
705Deleting an instruction from an existing sequence of instructions that form a <a
706href="#BasicBlock"><tt>BasicBlock</tt></a> is very straightforward. First, you
707must have a pointer to the instruction that you wish to delete. Second, you
708need to obtain the pointer to that instruction's basic block. You use the
709pointer to the basic block to get its list of instructions and then use the
710erase function to remove your instruction.<p>
711
712For example:<p>
713
714<pre>
715 <a href="#Instruction">Instruction</a> *I = .. ;
Chris Lattner7dbf6832002-09-18 05:14:25 +0000[diff] [blame]716 <a href="#BasicBlock">BasicBlock</a> *BB = I-&gt;getParent();
717 BB-&gt;getInstList().erase(I);
Chris Lattner4e1f96b2002-09-12 19:06:51 +0000[diff] [blame]718</pre><p>
719
Joel Stanley753eb712002-09-11 22:32:24 +0000[diff] [blame]720<!--_______________________________________________________________________-->
721</ul><h4><a name="schanges_replacing"><hr size=0>Replacing an
722 <tt>Instruction</tt> with another <tt>Value</tt></h4><ul>
723
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]724<p><i>Replacing individual instructions</i></p>
725<p>
726Including "<a
Chris Lattner7dbf6832002-09-18 05:14:25 +0000[diff] [blame]727href="/doxygen/BasicBlockUtils_8h-source.html">llvm/Transforms/Utils/BasicBlockUtils.h
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]728</a>" permits use of two very useful replace functions:
729<tt>ReplaceInstWithValue</tt> and <tt>ReplaceInstWithInst</tt>.
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]730
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]731<ul>
732
Chris Lattner7dbf6832002-09-18 05:14:25 +0000[diff] [blame]733<li><tt>ReplaceInstWithValue</tt>
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]734
735<p>This function replaces all uses (within a basic block) of a given
736instruction with a value, and then removes the original instruction.
737The following example illustrates the replacement of the result of a
738particular <tt>AllocaInst</tt> that allocates memory for a single
739integer with an null pointer to an integer.</p>
740
741<pre>
742AllocaInst* instToReplace = ...;
Chris Lattner7dbf6832002-09-18 05:14:25 +0000[diff] [blame]743ReplaceInstWithValue(*instToReplace-&gt;getParent(), instToReplace,
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]744 Constant::getNullValue(PointerType::get(Type::IntTy)));
745</pre>
746
Chris Lattner7dbf6832002-09-18 05:14:25 +0000[diff] [blame]747<li><tt>ReplaceInstWithInst</tt>
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]748
749<p>This function replaces a particular instruction with another
750instruction. The following example illustrates the replacement of one
751<tt>AllocaInst</tt> with another.<p>
752
753<pre>
754AllocaInst* instToReplace = ...;
Chris Lattner7dbf6832002-09-18 05:14:25 +0000[diff] [blame]755ReplaceInstWithInst(*instToReplace-&gt;getParent(), instToReplace,
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]756 new AllocaInst(Type::IntTy, 0, "ptrToReplacedInt");
757</pre>
758
759</ul>
760<p><i>Replacing multiple uses of <tt>User</tt>s and
761 <tt>Value</tt>s</i></p>
762
763You can use <tt>Value::replaceAllUsesWith</tt> and
764<tt>User::replaceUsesOfWith</tt> to change more than one use at a
765time. See the doxygen documentation for the <a
766href="/doxygen/classValue.html">Value Class</a> and <a
767href="/doxygen/classUser.html">User Class</a>, respectively, for more
768information.
769
770<!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
771include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
772ReplaceInstWithValue, ReplaceInstWithInst
Chris Lattnerae7f7592002-09-06 18:31:18 +0000[diff] [blame]773-->
Chris Lattnerb99344f2002-09-06 16:40:10 +0000[diff] [blame]774
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]775<!-- *********************************************************************** -->
776</ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
777<tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]778<a name="coreclasses">The Core LLVM Class Hierarchy Reference
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]779</b></font></td></tr></table><ul>
780<!-- *********************************************************************** -->
781
782The Core LLVM classes are the primary means of representing the program being
783inspected or transformed. The core LLVM classes are defined in header files in
784the <tt>include/llvm/</tt> directory, and implemented in the <tt>lib/VMCore</tt>
785directory.<p>
786
787
788<!-- ======================================================================= -->
789</ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
790<tr><td>&nbsp;</td><td width="100%">&nbsp;
791<font color="#EEEEFF" face="Georgia,Palatino"><b>
792<a name="Value">The <tt>Value</tt> class</a>
793</b></font></td></tr></table><ul>
794
795<tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt></b><br>
796doxygen info: <a href="/doxygen/classValue.html">Value Class</a><p>
797
798
799The <tt>Value</tt> class is the most important class in LLVM Source base. It
800represents a typed value that may be used (among other things) as an operand to
801an instruction. There are many different types of <tt>Value</tt>s, such as <a
802href="#Constant"><tt>Constant</tt></a>s, <a
803href="#Argument"><tt>Argument</tt></a>s, and even <a
804href="#Instruction"><tt>Instruction</tt></a>s and <a
805href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.<p>
806
807A particular <tt>Value</tt> may be used many times in the LLVM representation
808for a program. For example, an incoming argument to a function (represented
809with an instance of the <a href="#Argument">Argument</a> class) is "used" by
810every instruction in the function that references the argument. To keep track
811of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
812href="#User"><tt>User</tt></a>s that is using it (the <a
813href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
814graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
Joel Stanley9b96c442002-09-06 21:55:13 +0000[diff] [blame]815def-use information in the program, and is accessible through the <tt>use_</tt>*
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]816methods, shown below.<p>
817
818Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed, and
819this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
820method. <a name="#nameWarning">In addition, all LLVM values can be named. The
821"name" of the <tt>Value</tt> is symbolic string printed in the LLVM code:<p>
822
823<pre>
824 %<b>foo</b> = add int 1, 2
825</pre>
826
827The name of this instruction is "foo". <b>NOTE</b> that the name of any value
828may be missing (an empty string), so names should <b>ONLY</b> be used for
829debugging (making the source code easier to read, debugging printouts), they
830should not be used to keep track of values or map between them. For this
831purpose, use a <tt>std::map</tt> of pointers to the <tt>Value</tt> itself
832instead.<p>
833
834One important aspect of LLVM is that there is no distinction between an SSA
835variable and the operation that produces it. Because of this, any reference to
836the value produced by an instruction (or the value available as an incoming
837argument, for example) is represented as a direct pointer to the class that
838represents this value. Although this may take some getting used to, it
839simplifies the representation and makes it easier to manipulate.<p>
840
841
842<!-- _______________________________________________________________________ -->
843</ul><h4><a name="m_Value"><hr size=0>Important Public Members of
844the <tt>Value</tt> class</h4><ul>
845
846<li><tt>Value::use_iterator</tt> - Typedef for iterator over the use-list<br>
847 <tt>Value::use_const_iterator</tt>
848 - Typedef for const_iterator over the use-list<br>
849 <tt>unsigned use_size()</tt> - Returns the number of users of the value.<br>
850 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
851 <tt>use_iterator use_begin()</tt>
852 - Get an iterator to the start of the use-list.<br>
853 <tt>use_iterator use_end()</tt>
854 - Get an iterator to the end of the use-list.<br>
855 <tt><a href="#User">User</a> *use_back()</tt>
856 - Returns the last element in the list.<p>
857
858These methods are the interface to access the def-use information in LLVM. As with all other iterators in LLVM, the naming conventions follow the conventions defined by the <a href="#stl">STL</a>.<p>
859
860<li><tt><a href="#Type">Type</a> *getType() const</tt><p>
861This method returns the Type of the Value.
862
863<li><tt>bool hasName() const</tt><br>
864 <tt>std::string getName() const</tt><br>
865 <tt>void setName(const std::string &amp;Name)</tt><p>
866
867This family of methods is used to access and assign a name to a <tt>Value</tt>,
868be aware of the <a href="#nameWarning">precaution above</a>.<p>
869
870
871<li><tt>void replaceAllUsesWith(Value *V)</tt><p>
872
873This method traverses the use list of a <tt>Value</tt> changing all <a
Misha Brukmanc4f5bb02002-09-18 02:21:57 +0000[diff] [blame]874href="#User"><tt>User</tt>s</a> of the current value to refer to "<tt>V</tt>"
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]875instead. For example, if you detect that an instruction always produces a
876constant value (for example through constant folding), you can replace all uses
877of the instruction with the constant like this:<p>
878
879<pre>
880 Inst-&gt;replaceAllUsesWith(ConstVal);
881</pre><p>
882
883
884
885<!-- ======================================================================= -->
886</ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
887<tr><td>&nbsp;</td><td width="100%">&nbsp;
888<font color="#EEEEFF" face="Georgia,Palatino"><b>
889<a name="User">The <tt>User</tt> class</a>
890</b></font></td></tr></table><ul>
891
892<tt>#include "<a href="/doxygen/User_8h-source.html">llvm/User.h</a>"</tt></b><br>
893doxygen info: <a href="/doxygen/classUser.html">User Class</a><br>
894Superclass: <a href="#Value"><tt>Value</tt></a><p>
895
896
897The <tt>User</tt> class is the common base class of all LLVM nodes that may
898refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
899that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
900referring to. The <tt>User</tt> class itself is a subclass of
901<tt>Value</tt>.<p>
902
903The operands of a <tt>User</tt> point directly to the LLVM <a
904href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
905Single Assignment (SSA) form, there can only be one definition referred to,
906allowing this direct connection. This connection provides the use-def
907information in LLVM.<p>
908
909<!-- _______________________________________________________________________ -->
910</ul><h4><a name="m_User"><hr size=0>Important Public Members of
911the <tt>User</tt> class</h4><ul>
912
913The <tt>User</tt> class exposes the operand list in two ways: through an index
914access interface and through an iterator based interface.<p>
915
916<li><tt>Value *getOperand(unsigned i)</tt><br>
917 <tt>unsigned getNumOperands()</tt><p>
918
919These two methods expose the operands of the <tt>User</tt> in a convenient form
920for direct access.<p>
921
922<li><tt>User::op_iterator</tt> - Typedef for iterator over the operand list<br>
923 <tt>User::op_const_iterator</tt>
924 <tt>use_iterator op_begin()</tt>
925 - Get an iterator to the start of the operand list.<br>
926 <tt>use_iterator op_end()</tt>
927 - Get an iterator to the end of the operand list.<p>
928
929Together, these methods make up the iterator based interface to the operands of
930a <tt>User</tt>.<p>
931
932
933
934<!-- ======================================================================= -->
935</ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
936<tr><td>&nbsp;</td><td width="100%">&nbsp;
937<font color="#EEEEFF" face="Georgia,Palatino"><b>
938<a name="Instruction">The <tt>Instruction</tt> class</a>
939</b></font></td></tr></table><ul>
940
941<tt>#include "<a
942href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt></b><br>
943doxygen info: <a href="/doxygen/classInstruction.html">Instruction Class</a><br>
944Superclasses: <a href="#User"><tt>User</tt></a>, <a
945href="#Value"><tt>Value</tt></a><p>
946
947The <tt>Instruction</tt> class is the common base class for all LLVM
948instructions. It provides only a few methods, but is a very commonly used
949class. The primary data tracked by the <tt>Instruction</tt> class itself is the
950opcode (instruction type) and the parent <a
951href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
952into. To represent a specific type of instruction, one of many subclasses of
953<tt>Instruction</tt> are used.<p>
954
955Because the <tt>Instruction</tt> class subclasses the <a
956href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
957way as for other <a href="#User"><tt>User</tt></a>s (with the
958<tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
959<tt>op_begin()</tt>/<tt>op_end()</tt> methods).<p>
960
Chris Lattner17635252002-09-12 17:18:46 +0000[diff] [blame]961An important file for the <tt>Instruction</tt> class is the
962<tt>llvm/Instruction.def</tt> file. This file contains some meta-data about the
963various different types of instructions in LLVM. It describes the enum values
964that are used as opcodes (for example <tt>Instruction::Add</tt> and
965<tt>Instruction::SetLE</tt>), as well as the concrete sub-classes of
966<tt>Instruction</tt> that implement the instruction (for example <tt><a
967href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
968href="#SetCondInst">SetCondInst</a></tt>). Unfortunately, the use of macros in
969this file confused doxygen, so these enum values don't show up correctly in the
970<a href="/doxygen/classInstruction.html">doxygen output</a>.<p>
971
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]972
973<!-- _______________________________________________________________________ -->
974</ul><h4><a name="m_Instruction"><hr size=0>Important Public Members of
975the <tt>Instruction</tt> class</h4><ul>
976
977<li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt><p>
978
979Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that this
980<tt>Instruction</tt> is embedded into.<p>
981
982<li><tt>bool hasSideEffects()</tt><p>
983
984Returns true if the instruction has side effects, i.e. it is a <tt>call</tt>,
985<tt>free</tt>, <tt>invoke</tt>, or <tt>store</tt>.<p>
986
987<li><tt>unsigned getOpcode()</tt><p>
988
989Returns the opcode for the <tt>Instruction</tt>.<p>
990
Chris Lattner17635252002-09-12 17:18:46 +0000[diff] [blame]991<li><tt><a href="#Instruction">Instruction</a> *clone() const</tt><p>
992
993Returns another instance of the specified instruction, identical in all ways to
994the original except that the instruction has no parent (ie it's not embedded
995into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>), and it has no name.<p>
996
997
998
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]999<!--
1000
1001\subsection{Subclasses of Instruction :}
1002\begin{itemize}
1003<li>BinaryOperator : This subclass of Instruction defines a general interface to the all the instructions involvong binary operators in LLVM.
1004 \begin{itemize}
1005 <li><tt>bool swapOperands()</tt>: Exchange the two operands to this instruction. If the instruction cannot be reversed (i.e. if it's a Div), it returns true.
1006 \end{itemize}
1007<li>TerminatorInst : This subclass of Instructions defines an interface for all instructions that can terminate a BasicBlock.
1008 \begin{itemize}
1009 <li> <tt>unsigned getNumSuccessors()</tt>: Returns the number of successors for this terminator instruction.
1010 <li><tt>BasicBlock *getSuccessor(unsigned i)</tt>: As the name suggests returns the ith successor BasicBlock.
1011 <li><tt>void setSuccessor(unsigned i, BasicBlock *B)</tt>: sets BasicBlock B as the ith succesor to this terminator instruction.
1012 \end{itemize}
1013
1014<li>PHINode : This represents the PHI instructions in the SSA form.
1015 \begin{itemize}
1016 <li><tt> unsigned getNumIncomingValues()</tt>: Returns the number of incoming edges to this PHI node.
1017 <li><tt> Value *getIncomingValue(unsigned i)</tt>: Returns the ith incoming Value.
1018 <li><tt>void setIncomingValue(unsigned i, Value *V)</tt>: Sets the ith incoming Value as V
1019 <li><tt>BasicBlock *getIncomingBlock(unsigned i)</tt>: Returns the Basic Block corresponding to the ith incoming Value.
1020 <li><tt> void addIncoming(Value *D, BasicBlock *BB)</tt>:
1021 Add an incoming value to the end of the PHI list
1022 <li><tt> int getBasicBlockIndex(const BasicBlock *BB) const</tt>:
1023 Returns the first index of the specified basic block in the value list for this PHI. Returns -1 if no instance.
1024 \end{itemize}
1025<li>CastInst : In LLVM all casts have to be done through explicit cast instructions. CastInst defines the interface to the cast instructions.
1026<li>CallInst : This defines an interface to the call instruction in LLVM. ARguments to the function are nothing but operands of the instruction.
1027 \begin{itemize}
1028 <li>: <tt>Function *getCalledFunction()</tt>: Returns a handle to the function that is being called by this Function.
1029 \end{itemize}
1030<li>LoadInst, StoreInst, GetElemPtrInst : These subclasses represent load, store and getelementptr instructions in LLVM.
1031 \begin{itemize}
1032 <li><tt>Value * getPointerOperand ()</tt>: Returns the Pointer Operand which is typically the 0th operand.
1033 \end{itemize}
1034<li>BranchInst : This is a subclass of TerminatorInst and defines the interface for conditional and unconditional branches in LLVM.
1035 \begin{itemize}
1036 <li><tt>bool isConditional()</tt>: Returns true if the branch is a conditional branch else returns false
1037 <li> <tt>Value *getCondition()</tt>: Returns the condition if it is a conditional branch else returns null.
1038 <li> <tt>void setUnconditionalDest(BasicBlock *Dest)</tt>: Changes the current branch to an unconditional one targetting the specified block.
1039 \end{itemize}
1040
1041\end{itemize}
1042
1043-->
1044
1045
1046<!-- ======================================================================= -->
1047</ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1048<tr><td>&nbsp;</td><td width="100%">&nbsp;
1049<font color="#EEEEFF" face="Georgia,Palatino"><b>
1050<a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
1051</b></font></td></tr></table><ul>
1052
1053<tt>#include "<a
1054href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt></b><br>
1055doxygen info: <a href="/doxygen/classBasicBlock.html">BasicBlock Class</a><br>
1056Superclass: <a href="#Value"><tt>Value</tt></a><p>
1057
1058
1059This class represents a single entry multiple exit section of the code, commonly
1060known as a basic block by the compiler community. The <tt>BasicBlock</tt> class
1061maintains a list of <a href="#Instruction"><tt>Instruction</tt></a>s, which form
1062the body of the block. Matching the language definition, the last element of
1063this list of instructions is always a terminator instruction (a subclass of the
1064<a href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).<p>
1065
1066In addition to tracking the list of instructions that make up the block, the
1067<tt>BasicBlock</tt> class also keeps track of the <a
1068href="#Function"><tt>Function</tt></a> that it is embedded into.<p>
1069
1070Note that <tt>BasicBlock</tt>s themselves are <a
1071href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
1072like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
1073<tt>label</tt>.<p>
1074
1075
1076<!-- _______________________________________________________________________ -->
1077</ul><h4><a name="m_BasicBlock"><hr size=0>Important Public Members of
1078the <tt>BasicBlock</tt> class</h4><ul>
1079
1080<li><tt>BasicBlock(const std::string &amp;Name = "", <a
1081href="#Function">Function</a> *Parent = 0)</tt><p>
1082
1083The <tt>BasicBlock</tt> constructor is used to create new basic blocks for
1084insertion into a function. The constructor simply takes a name for the new
1085block, and optionally a <a href="#Function"><tt>Function</tt></a> to insert it
1086into. If the <tt>Parent</tt> parameter is specified, the new
1087<tt>BasicBlock</tt> is automatically inserted at the end of the specified <a
1088href="#Function"><tt>Function</tt></a>, if not specified, the BasicBlock must be
1089manually inserted into the <a href="#Function"><tt>Function</tt></a>.<p>
1090
1091<li><tt>BasicBlock::iterator</tt> - Typedef for instruction list iterator<br>
1092 <tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
1093 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1094 <tt>size()</tt>, <tt>empty()</tt>, <tt>rbegin()</tt>, <tt>rend()</tt><p>
1095
1096These methods and typedefs are forwarding functions that have the same semantics
1097as the standard library methods of the same names. These methods expose the
1098underlying instruction list of a basic block in a way that is easy to
1099manipulate. To get the full complement of container operations (including
1100operations to update the list), you must use the <tt>getInstList()</tt>
1101method.<p>
1102
1103<li><tt>BasicBlock::InstListType &amp;getInstList()</tt><p>
1104
1105This method is used to get access to the underlying container that actually
1106holds the Instructions. This method must be used when there isn't a forwarding
1107function in the <tt>BasicBlock</tt> class for the operation that you would like
1108to perform. Because there are no forwarding functions for "updating"
1109operations, you need to use this if you want to update the contents of a
1110<tt>BasicBlock</tt>.<p>
1111
1112<li><tt><A href="#Function">Function</a> *getParent()</tt><p>
1113
1114Returns a pointer to <a href="#Function"><tt>Function</tt></a> the block is
1115embedded into, or a null pointer if it is homeless.<p>
1116
1117<li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt><p>
1118
1119Returns a pointer to the terminator instruction that appears at the end of the
1120<tt>BasicBlock</tt>. If there is no terminator instruction, or if the last
1121instruction in the block is not a terminator, then a null pointer is
1122returned.<p>
1123
1124
1125<!-- ======================================================================= -->
1126</ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1127<tr><td>&nbsp;</td><td width="100%">&nbsp;
1128<font color="#EEEEFF" face="Georgia,Palatino"><b>
1129<a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
1130</b></font></td></tr></table><ul>
1131
1132<tt>#include "<a
1133href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt></b><br>
1134doxygen info: <a href="/doxygen/classGlobalValue.html">GlobalValue Class</a><br>
1135Superclasses: <a href="#User"><tt>User</tt></a>, <a
1136href="#Value"><tt>Value</tt></a><p>
1137
1138Global values (<A href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
1139href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
1140visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
1141Because they are visible at global scope, they are also subject to linking with
1142other globals defined in different translation units. To control the linking
1143process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
1144<tt>GlobalValue</tt>s know whether they have internal or external linkage.<p>
1145
1146If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
1147<tt>static</tt> in C), it is not visible to code outside the current translation
1148unit, and does not participate in linking. If it has external linkage, it is
1149visible to external code, and does participate in linking. In addition to
1150linkage information, <tt>GlobalValue</tt>s keep track of which <a
1151href="#Module"><tt>Module</tt></a> they are currently part of.<p>
1152
1153Because <tt>GlobalValue</tt>s are memory objects, they are always referred to by
1154their address. As such, the <a href="#Type"><tt>Type</tt></a> of a global is
1155always a pointer to its contents. This is explained in the LLVM Language
1156Reference Manual.<p>
1157
1158
1159<!-- _______________________________________________________________________ -->
1160</ul><h4><a name="m_GlobalValue"><hr size=0>Important Public Members of
1161the <tt>GlobalValue</tt> class</h4><ul>
1162
1163<li><tt>bool hasInternalLinkage() const</tt><br>
1164 <tt>bool hasExternalLinkage() const</tt><br>
1165 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt><p>
1166
1167These methods manipulate the linkage characteristics of the
1168<tt>GlobalValue</tt>.<p>
1169
1170<li><tt><a href="#Module">Module</a> *getParent()</tt><p>
1171
1172This returns the <a href="#Module"><tt>Module</tt></a> that the GlobalValue is
1173currently embedded into.<p>
1174
1175
1176
1177<!-- ======================================================================= -->
1178</ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1179<tr><td>&nbsp;</td><td width="100%">&nbsp;
1180<font color="#EEEEFF" face="Georgia,Palatino"><b>
1181<a name="Function">The <tt>Function</tt> class</a>
1182</b></font></td></tr></table><ul>
1183
1184<tt>#include "<a
1185href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt></b><br>
1186doxygen info: <a href="/doxygen/classFunction.html">Function Class</a><br>
1187Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>, <a
1188href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a><p>
1189
1190The <tt>Function</tt> class represents a single procedure in LLVM. It is
1191actually one of the more complex classes in the LLVM heirarchy because it must
1192keep track of a large amount of data. The <tt>Function</tt> class keeps track
1193of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal <a
1194href="#Argument"><tt>Argument</tt></a>s, and a <a
1195href="#SymbolTable"><tt>SymbolTable</tt></a>.<p>
1196
1197The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most commonly
1198used part of <tt>Function</tt> objects. The list imposes an implicit ordering
1199of the blocks in the function, which indicate how the code will be layed out by
1200the backend. Additionally, the first <a
1201href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
1202<tt>Function</tt>. It is not legal in LLVM explicitly branch to this initial
1203block. There are no implicit exit nodes, and in fact there may be multiple exit
1204nodes from a single <tt>Function</tt>. If the <a
1205href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
1206the <tt>Function</tt> is actually a function declaration: the actual body of the
1207function hasn't been linked in yet.<p>
1208
1209In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
1210<tt>Function</tt> class also keeps track of the list of formal <a
1211href="#Argument"><tt>Argument</tt></a>s that the function receives. This
1212container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
1213nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
1214the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.<p>
1215
1216The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used LLVM
1217feature that is only used when you have to look up a value by name. Aside from
1218that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used internally to
1219make sure that there are not conflicts between the names of <a
1220href="#Instruction"><tt>Instruction</tt></a>s, <a
1221href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
1222href="#Argument"><tt>Argument</tt></a>s in the function body.<p>
1223
1224
1225<!-- _______________________________________________________________________ -->
1226</ul><h4><a name="m_Function"><hr size=0>Important Public Members of
1227the <tt>Function</tt> class</h4><ul>
1228
1229<li><tt>Function(const <a href="#FunctionType">FunctionType</a> *Ty, bool isInternal, const std::string &amp;N = "")</tt><p>
1230
1231Constructor used when you need to create new <tt>Function</tt>s to add the the
1232program. The constructor must specify the type of the function to create and
1233whether or not it should start out with internal or external linkage.<p>
1234
1235<li><tt>bool isExternal()</tt><p>
1236
1237Return whether or not the <tt>Function</tt> has a body defined. If the function
1238is "external", it does not have a body, and thus must be resolved by linking
1239with a function defined in a different translation unit.<p>
1240
1241
1242<li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
1243 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
1244 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1245 <tt>size()</tt>, <tt>empty()</tt>, <tt>rbegin()</tt>, <tt>rend()</tt><p>
1246
1247These are forwarding methods that make it easy to access the contents of a
1248<tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
1249list.<p>
1250
1251<li><tt>Function::BasicBlockListType &amp;getBasicBlockList()</tt><p>
1252
1253Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This is
1254neccesary to use when you need to update the list or perform a complex action
1255that doesn't have a forwarding method.<p>
1256
1257
1258<li><tt>Function::aiterator</tt> - Typedef for the argument list iterator<br>
1259 <tt>Function::const_aiterator</tt> - Typedef for const_iterator.<br>
1260 <tt>abegin()</tt>, <tt>aend()</tt>, <tt>afront()</tt>, <tt>aback()</tt>,
1261 <tt>asize()</tt>, <tt>aempty()</tt>, <tt>arbegin()</tt>, <tt>arend()</tt><p>
1262
1263These are forwarding methods that make it easy to access the contents of a
1264<tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a> list.<p>
1265
1266<li><tt>Function::ArgumentListType &amp;getArgumentList()</tt><p>
1267
1268Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
1269neccesary to use when you need to update the list or perform a complex action
1270that doesn't have a forwarding method.<p>
1271
1272
1273
1274<li><tt><a href="#BasicBlock">BasicBlock</a> &getEntryNode()</tt><p>
1275
1276Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
1277function. Because the entry block for the function is always the first block,
1278this returns the first block of the <tt>Function</tt>.<p>
1279
1280<li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
1281 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt><p>
1282
1283This traverses the <a href="#Type"><tt>Type</tt></a> of the <tt>Function</tt>
1284and returns the return type of the function, or the <a
1285href="#FunctionType"><tt>FunctionType</tt></a> of the actual function.<p>
1286
1287
1288<li><tt>bool hasSymbolTable() const</tt><p>
1289
1290Return true if the <tt>Function</tt> has a symbol table allocated to it and if
1291there is at least one entry in it.<p>
1292
1293<li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt><p>
1294
1295Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a> for this
1296<tt>Function</tt> or a null pointer if one has not been allocated (because there
1297are no named values in the function).<p>
1298
1299<li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTableSure()</tt><p>
1300
1301Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a> for this
1302<tt>Function</tt> or allocate a new <a
1303href="#SymbolTable"><tt>SymbolTable</tt></a> if one is not already around. This
1304should only be used when adding elements to the <a
1305href="#SymbolTable"><tt>SymbolTable</tt></a>, so that empty symbol tables are
1306not left laying around.<p>
1307
1308
1309
1310<!-- ======================================================================= -->
1311</ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1312<tr><td>&nbsp;</td><td width="100%">&nbsp;
1313<font color="#EEEEFF" face="Georgia,Palatino"><b>
1314<a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
1315</b></font></td></tr></table><ul>
1316
1317<tt>#include "<a
1318href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt></b><br>
1319doxygen info: <a href="/doxygen/classGlobalVariable.html">GlobalVariable Class</a><br>
1320Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>, <a
1321href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a><p>
1322
Chris Lattner0377de42002-09-06 14:50:55 +0000[diff] [blame]1323Global variables are represented with the (suprise suprise)
1324<tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are
1325also subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such
1326are always referenced by their address (global values must live in memory, so
1327their "name" refers to their address). Global variables may have an initial
1328value (which must be a <a href="#Constant"><tt>Constant</tt></a>), and if they
1329have an initializer, they may be marked as "constant" themselves (indicating
1330that their contents never change at runtime).<p>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]1331
1332
1333<!-- _______________________________________________________________________ -->
Chris Lattner0377de42002-09-06 14:50:55 +0000[diff] [blame]1334</ul><h4><a name="m_GlobalVariable"><hr size=0>Important Public Members of the
1335<tt>GlobalVariable</tt> class</h4><ul>
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]1336
1337<li><tt>GlobalVariable(const <a href="#Type">Type</a> *Ty, bool isConstant, bool
1338isInternal, <a href="#Constant">Constant</a> *Initializer = 0, const std::string
1339&amp;Name = "")</tt><p>
1340
Chris Lattner0377de42002-09-06 14:50:55 +0000[diff] [blame]1341Create a new global variable of the specified type. If <tt>isConstant</tt> is
1342true then the global variable will be marked as unchanging for the program, and
1343if <tt>isInternal</tt> is true the resultant global variable will have internal
1344linkage. Optionally an initializer and name may be specified for the global variable as well.<p>
1345
1346
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]1347<li><tt>bool isConstant() const</tt><p>
1348
1349Returns true if this is a global variable is known not to be modified at
1350runtime.<p>
1351
Chris Lattner0377de42002-09-06 14:50:55 +0000[diff] [blame]1352
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]1353<li><tt>bool hasInitializer()</tt><p>
1354
1355Returns true if this <tt>GlobalVariable</tt> has an intializer.<p>
1356
Chris Lattner0377de42002-09-06 14:50:55 +0000[diff] [blame]1357
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]1358<li><tt><a href="#Constant">Constant</a> *getInitializer()</tt><p>
1359
Chris Lattner0377de42002-09-06 14:50:55 +0000[diff] [blame]1360Returns the intial value for a <tt>GlobalVariable</tt>. It is not legal to call
1361this method if there is no initializer.<p>
1362
1363
1364<!-- ======================================================================= -->
1365</ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1366<tr><td>&nbsp;</td><td width="100%">&nbsp;
1367<font color="#EEEEFF" face="Georgia,Palatino"><b>
1368<a name="Module">The <tt>Module</tt> class</a>
1369</b></font></td></tr></table><ul>
1370
1371<tt>#include "<a
1372href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt></b><br>
1373doxygen info: <a href="/doxygen/classModule.html">Module Class</a><p>
1374
1375The <tt>Module</tt> class represents the top level structure present in LLVM
1376programs. An LLVM module is effectively either a translation unit of the
1377original program or a combination of several translation units merged by the
1378linker. The <tt>Module</tt> class keeps track of a list of <a
1379href="#Function"><tt>Function</tt></a>s, a list of <a
1380href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
1381href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
1382helpful member functions that try to make common operations easy.<p>
1383
1384
1385<!-- _______________________________________________________________________ -->
1386</ul><h4><a name="m_Module"><hr size=0>Important Public Members of the
1387<tt>Module</tt> class</h4><ul>
1388
1389<li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
1390 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
1391 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1392 <tt>size()</tt>, <tt>empty()</tt>, <tt>rbegin()</tt>, <tt>rend()</tt><p>
1393
1394These are forwarding methods that make it easy to access the contents of a
1395<tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
1396list.<p>
1397
1398<li><tt>Module::FunctionListType &amp;getFunctionList()</tt><p>
1399
1400Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
1401neccesary to use when you need to update the list or perform a complex action
1402that doesn't have a forwarding method.<p>
1403
1404<!-- Global Variable -->
1405<hr size=0>
1406
1407<li><tt>Module::giterator</tt> - Typedef for global variable list iterator<br>
1408 <tt>Module::const_giterator</tt> - Typedef for const_iterator.<br>
1409 <tt>gbegin()</tt>, <tt>gend()</tt>, <tt>gfront()</tt>, <tt>gback()</tt>,
1410 <tt>gsize()</tt>, <tt>gempty()</tt>, <tt>grbegin()</tt>, <tt>grend()</tt><p>
1411
1412These are forwarding methods that make it easy to access the contents of a
1413<tt>Module</tt> object's <a href="#GlobalVariable"><tt>GlobalVariable</tt></a>
1414list.<p>
1415
1416<li><tt>Module::GlobalListType &amp;getGlobalList()</tt><p>
1417
1418Returns the list of <a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s.
1419This is neccesary to use when you need to update the list or perform a complex
1420action that doesn't have a forwarding method.<p>
1421
1422
1423<!-- Symbol table stuff -->
1424<hr size=0>
1425
1426<li><tt>bool hasSymbolTable() const</tt><p>
1427
1428Return true if the <tt>Module</tt> has a symbol table allocated to it and if
1429there is at least one entry in it.<p>
1430
1431<li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt><p>
1432
1433Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a> for this
1434<tt>Module</tt> or a null pointer if one has not been allocated (because there
1435are no named values in the function).<p>
1436
1437<li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTableSure()</tt><p>
1438
1439Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a> for this
1440<tt>Module</tt> or allocate a new <a
1441href="#SymbolTable"><tt>SymbolTable</tt></a> if one is not already around. This
1442should only be used when adding elements to the <a
1443href="#SymbolTable"><tt>SymbolTable</tt></a>, so that empty symbol tables are
1444not left laying around.<p>
1445
1446
1447<!-- Convenience methods -->
1448<hr size=0>
1449
1450<li><tt><a href="#Function">Function</a> *getFunction(const std::string &amp;Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt><p>
1451
1452Look up the specified function in the <tt>Module</tt> <a
1453href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
1454<tt>null</tt>.<p>
1455
1456
1457<li><tt><a href="#Function">Function</a> *getOrInsertFunction(const std::string
1458 &amp;Name, const <a href="#FunctionType">FunctionType</a> *T)</tt><p>
1459
1460Look up the specified function in the <tt>Module</tt> <a
1461href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
1462external declaration for the function and return it.<p>
1463
1464
1465<li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt><p>
1466
1467If there is at least one entry in the <a
1468href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
1469href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
1470string.<p>
1471
1472
1473<li><tt>bool addTypeName(const std::string &Name, const <a href="#Type">Type</a>
1474*Ty)</tt><p>
1475
1476Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a> mapping
1477<tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this name, true
1478is returned and the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is not
1479modified.<p>
1480
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]1481
1482<!-- ======================================================================= -->
1483</ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1484<tr><td>&nbsp;</td><td width="100%">&nbsp;
1485<font color="#EEEEFF" face="Georgia,Palatino"><b>
1486<a name="Constant">The <tt>Constant</tt> class and subclasses</a>
1487</b></font></td></tr></table><ul>
1488
1489Constant represents a base class for different types of constants. It is
1490subclassed by ConstantBool, ConstantInt, ConstantSInt, ConstantUInt,
1491ConstantArray etc for representing the various types of Constants.<p>
1492
1493
1494<!-- _______________________________________________________________________ -->
1495</ul><h4><a name="m_Value"><hr size=0>Important Public Methods</h4><ul>
1496
1497<li><tt>bool isConstantExpr()</tt>: Returns true if it is a ConstantExpr
1498
1499
1500
1501
1502\subsection{Important Subclasses of Constant}
1503\begin{itemize}
1504<li>ConstantSInt : This subclass of Constant represents a signed integer constant.
1505 \begin{itemize}
1506 <li><tt>int64_t getValue () const</tt>: Returns the underlying value of this constant.
1507 \end{itemize}
1508<li>ConstantUInt : This class represents an unsigned integer.
1509 \begin{itemize}
1510 <li><tt>uint64_t getValue () const</tt>: Returns the underlying value of this constant.
1511 \end{itemize}
1512<li>ConstantFP : This class represents a floating point constant.
1513 \begin{itemize}
1514 <li><tt>double getValue () const</tt>: Returns the underlying value of this constant.
1515 \end{itemize}
1516<li>ConstantBool : This represents a boolean constant.
1517 \begin{itemize}
1518 <li><tt>bool getValue () const</tt>: Returns the underlying value of this constant.
1519 \end{itemize}
1520<li>ConstantArray : This represents a constant array.
1521 \begin{itemize}
1522 <li><tt>const std::vector<Use> &amp;getValues() const</tt>: Returns a Vecotr of component constants that makeup this array.
1523 \end{itemize}
1524<li>ConstantStruct : This represents a constant struct.
1525 \begin{itemize}
1526 <li><tt>const std::vector<Use> &amp;getValues() const</tt>: Returns a Vecotr of component constants that makeup this array.
1527 \end{itemize}
1528<li>ConstantPointerRef : This represents a constant pointer value that is initialized to point to a global value, which lies at a constant fixed address.
1529 \begin{itemize}
1530<li><tt>GlobalValue *getValue()</tt>: Returns the global value to which this pointer is pointing to.
1531 \end{itemize}
1532\end{itemize}
1533
1534
1535<!-- ======================================================================= -->
1536</ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1537<tr><td>&nbsp;</td><td width="100%">&nbsp;
1538<font color="#EEEEFF" face="Georgia,Palatino"><b>
1539<a name="Type">The <tt>Type</tt> class and Derived Types</a>
1540</b></font></td></tr></table><ul>
1541
1542Type as noted earlier is also a subclass of a Value class. Any primitive
1543type (like int, short etc) in LLVM is an instance of Type Class. All
1544other types are instances of subclasses of type like FunctionType,
1545ArrayType etc. DerivedType is the interface for all such dervied types
1546including FunctionType, ArrayType, PointerType, StructType. Types can have
1547names. They can be recursive (StructType). There exists exactly one instance
1548of any type structure at a time. This allows using pointer equality of Type *s for comparing types.
1549
1550<!-- _______________________________________________________________________ -->
1551</ul><h4><a name="m_Value"><hr size=0>Important Public Methods</h4><ul>
1552
1553<li><tt>PrimitiveID getPrimitiveID () const</tt>: Returns the base type of the type.
1554<li><tt> bool isSigned () const</tt>: Returns whether an integral numeric type is signed. This is true for SByteTy, ShortTy, IntTy, LongTy. Note that this is not true for Float and Double.
1555<li><tt>bool isUnsigned () const</tt>: Returns whether a numeric type is unsigned. This is not quite the complement of isSigned... nonnumeric types return false as they do with isSigned. This returns true for UByteTy, UShortTy, UIntTy, and ULongTy.
1556<li><tt> bool isInteger () const</tt>: Equilivent to isSigned() || isUnsigned(), but with only a single virtual function invocation.
1557<li><tt>bool isIntegral () const</tt>: Returns true if this is an integral type, which is either Bool type or one of the Integer types.
1558
1559<li><tt>bool isFloatingPoint ()</tt>: Return true if this is one of the two floating point types.
1560<li><tt>bool isRecursive () const</tt>: Returns rue if the type graph contains a cycle.
1561<li><tt>isLosslesslyConvertableTo (const Type *Ty) const</tt>: Return true if this type can be converted to 'Ty' without any reinterpretation of bits. For example, uint to int.
1562<li><tt>bool isPrimitiveType () const</tt>: Returns true if it is a primitive type.
1563<li><tt>bool isDerivedType () const</tt>: Returns true if it is a derived type.
1564<li><tt>const Type * getContainedType (unsigned i) const</tt>:
1565This method is used to implement the type iterator. For derived types, this returns the types 'contained' in the derived type, returning 0 when 'i' becomes invalid. This allows the user to iterate over the types in a struct, for example, really easily.
1566<li><tt>unsigned getNumContainedTypes () const</tt>: Return the number of types in the derived type.
1567
1568
1569
1570\subsection{Derived Types}
1571\begin{itemize}
1572<li>SequentialType : This is subclassed by ArrayType and PointerType
1573 \begin{itemize}
1574 <li><tt>const Type * getElementType () const</tt>: Returns the type of each of the elements in the sequential type.
1575 \end{itemize}
1576<li>ArrayType : This is a subclass of SequentialType and defines interface for array types.
1577 \begin{itemize}
1578 <li><tt>unsigned getNumElements () const</tt>: Returns the number of elements in the array.
1579 \end{itemize}
1580<li>PointerType : Subclass of SequentialType for pointer types.
1581<li>StructType : subclass of DerivedTypes for struct types
1582<li>FunctionType : subclass of DerivedTypes for function types.
1583 \begin{itemize}
1584
1585 <li><tt>bool isVarArg () const</tt>: Returns true if its a vararg function
1586 <li><tt> const Type * getReturnType () const</tt>: Returns the return type of the function.
1587 <li><tt> const ParamTypes &amp;getParamTypes () const</tt>: Returns a vector of parameter types.
1588 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns the type of the ith parameter.
1589 <li><tt> const unsigned getNumParams () const</tt>: Returns the number of formal parameters.
1590 \end{itemize}
1591\end{itemize}
1592
1593
1594
1595
1596<!-- ======================================================================= -->
1597</ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1598<tr><td>&nbsp;</td><td width="100%">&nbsp;
1599<font color="#EEEEFF" face="Georgia,Palatino"><b>
1600<a name="Argument">The <tt>Argument</tt> class</a>
1601</b></font></td></tr></table><ul>
1602
1603This subclass of Value defines the interface for incoming formal arguments to a
1604function. A Function maitanis a list of its formal arguments. An argument has a
1605pointer to the parent Function.
1606
1607
1608
1609
1610<!-- *********************************************************************** -->
1611</ul>
1612<!-- *********************************************************************** -->
1613
1614<hr><font size-1>
1615<address>By: <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
1616<a href="mailto:sabre@nondot.org">Chris Lattner</a></address>
1617<!-- Created: Tue Aug 6 15:00:33 CDT 2002 -->
1618<!-- hhmts start -->
Joel Stanley9dd1ad62002-09-18 03:17:23 +0000[diff] [blame]1619Last modified: Tue Sep 17 22:16:24 CDT 2002
Chris Lattner9355b472002-09-06 02:50:58 +0000[diff] [blame]1620<!-- hhmts end -->
1621</font></body></html>