1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
5 <title>LLVM Programmer's Manual</title>
6 <link rel="stylesheet" href="llvm.css" type="text/css">
10 <div class="doc_title">
11 LLVM Programmer's Manual
15 <li><a href="#introduction">Introduction</a></li>
16 <li><a href="#general">General Information</a>
18 <li><a href="#stl">The C++ Standard Template Library</a><!--
19 <li>The <tt>-time-passes</tt> option
20 <li>How to use the LLVM Makefile system
21 <li>How to write a regression test
25 <li><a href="#apis">Important and useful LLVM APIs</a>
27 <li><a href="#isa">The <tt>isa<></tt>, <tt>cast<></tt>
28 and <tt>dyn_cast<></tt> templates</a> </li>
29 <li><a href="#DEBUG">The <tt>DEBUG()</tt> macro & <tt>-debug</tt>
32 <li><a href="#DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt>
33 and the <tt>-debug-only</tt> option</a> </li>
36 <li><a href="#Statistic">The <tt>Statistic</tt> template & <tt>-stats</tt>
38 <li>The <tt>InstVisitor</tt> template
39 <li>The general graph API
43 <li><a href="#common">Helpful Hints for Common Operations</a>
45 <li><a href="#inspection">Basic Inspection and Traversal Routines</a>
47 <li><a href="#iterate_function">Iterating over the <tt>BasicBlock</tt>s
48 in a <tt>Function</tt></a> </li>
49 <li><a href="#iterate_basicblock">Iterating over the <tt>Instruction</tt>s
50 in a <tt>BasicBlock</tt></a> </li>
51 <li><a href="#iterate_institer">Iterating over the <tt>Instruction</tt>s
52 in a <tt>Function</tt></a> </li>
53 <li><a href="#iterate_convert">Turning an iterator into a
54 class pointer</a> </li>
55 <li><a href="#iterate_complex">Finding call sites: a more
56 complex example</a> </li>
57 <li><a href="#calls_and_invokes">Treating calls and invokes
58 the same way</a> </li>
59 <li><a href="#iterate_chains">Iterating over def-use &
60 use-def chains</a> </li>
63 <li><a href="#simplechanges">Making simple changes</a>
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>
69 with another <tt>Value</tt></a> </li>
72 <li>Working with the Control Flow Graph
74 <li>Accessing predecessors and successors of a <tt>BasicBlock</tt>
81 <li><a href="#coreclasses">The Core LLVM Class Hierarchy Reference</a>
83 <li><a href="#Value">The <tt>Value</tt> class</a>
85 <li><a href="#User">The <tt>User</tt> class</a>
87 <li><a href="#Instruction">The <tt>Instruction</tt> class</a>
89 <li><a href="#GetElementPtrInst">The <tt>GetElementPtrInst</tt>
92 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
94 <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>
99 <li><a href="#Module">The <tt>Module</tt> class</a></li>
100 <li><a href="#Constant">The <tt>Constant</tt> class</a>
102 <li><a href="#Type">The <tt>Type</tt> class</a> </li>
103 <li><a href="#Argument">The <tt>Argument</tt> class</a> </li>
106 <li>The <tt>SymbolTable</tt> class </li>
107 <li>The <tt>ilist</tt> and <tt>iplist</tt> classes
109 <li>Creating, inserting, moving and deleting from LLVM lists </li>
112 <li>Important iterator invalidation semantics to be aware of.</li>
117 <div class="doc_author">
118 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
119 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>, and
120 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a></p>
123 <!-- *********************************************************************** -->
124 <div class="doc_section">
125 <a name="introduction">Introduction </a>
127 <!-- *********************************************************************** -->
129 <div class="doc_text">
131 <p>This document is meant to highlight some of the important classes and
132 interfaces available in the LLVM source-base. This manual is not
133 intended to explain what LLVM is, how it works, and what LLVM code looks
134 like. It assumes that you know the basics of LLVM and are interested
135 in writing transformations or otherwise analyzing or manipulating the
138 <p>This document should get you oriented so that you can find your
139 way in the continuously growing source code that makes up the LLVM
140 infrastructure. Note that this manual is not intended to serve as a
141 replacement for reading the source code, so if you think there should be
142 a method in one of these classes to do something, but it's not listed,
143 check the source. Links to the <a href="/doxygen/">doxygen</a> sources
144 are provided to make this as easy as possible.</p>
146 <p>The first section of this document describes general information that is
147 useful to know when working in the LLVM infrastructure, and the second describes
148 the Core LLVM classes. In the future this manual will be extended with
149 information describing how to use extension libraries, such as dominator
150 information, CFG traversal routines, and useful utilities like the <tt><a
151 href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.</p>
155 <!-- *********************************************************************** -->
156 <div class="doc_section">
157 <a name="general">General Information</a>
159 <!-- *********************************************************************** -->
161 <div class="doc_text">
163 <p>This section contains general information that is useful if you are working
164 in the LLVM source-base, but that isn't specific to any particular API.</p>
168 <!-- ======================================================================= -->
169 <div class="doc_subsection">
170 <a name="stl">The C++ Standard Template Library</a>
173 <div class="doc_text">
175 <p>LLVM makes heavy use of the C++ Standard Template Library (STL),
176 perhaps much more than you are used to, or have seen before. Because of
177 this, you might want to do a little background reading in the
178 techniques used and capabilities of the library. There are many good
179 pages that discuss the STL, and several books on the subject that you
180 can get, so it will not be discussed in this document.</p>
182 <p>Here are some useful links:</p>
186 <li><a href="http://www.dinkumware.com/refxcpp.html">Dinkumware C++ Library
187 reference</a> - an excellent reference for the STL and other parts of the
188 standard C++ library.</li>
190 <li><a href="http://www.tempest-sw.com/cpp/">C++ In a Nutshell</a> - This is an
191 O'Reilly book in the making. It has a decent <a
192 href="http://www.tempest-sw.com/cpp/ch13-libref.html">Standard Library
193 Reference</a> that rivals Dinkumware's, and is actually free until the book is
196 <li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
199 <li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
201 href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
204 <li><a href="http://www.research.att.com/%7Ebs/C++.html">Bjarne Stroustrup's C++
209 <p>You are also encouraged to take a look at the <a
210 href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
211 to write maintainable code more than where to put your curly braces.</p>
215 <!-- ======================================================================= -->
216 <div class="doc_subsection">
217 <a name="stl">Other useful references</a>
220 <div class="doc_text">
222 <p>LLVM is currently using CVS as its source versioning system. You may find
223 this reference handy:</p>
226 <li><a href="http://www.psc.edu/%7Esemke/cvs_branches.html">CVS
227 Branch and Tag Primer</a></li>
232 <!-- *********************************************************************** -->
233 <div class="doc_section">
234 <a name="apis">Important and useful LLVM APIs</a>
236 <!-- *********************************************************************** -->
238 <div class="doc_text">
240 <p>Here we highlight some LLVM APIs that are generally useful and good to
241 know about when writing transformations.</p>
245 <!-- ======================================================================= -->
246 <div class="doc_subsection">
247 <a name="isa">The isa<>, cast<> and dyn_cast<> templates</a>
250 <div class="doc_text">
252 <p>The LLVM source-base makes extensive use of a custom form of RTTI.
253 These templates have many similarities to the C++ <tt>dynamic_cast<></tt>
254 operator, but they don't have some drawbacks (primarily stemming from
255 the fact that <tt>dynamic_cast<></tt> only works on classes that
256 have a v-table). Because they are used so often, you must know what they
257 do and how they work. All of these templates are defined in the <a
258 href="/doxygen/Casting_8h-source.html"><tt>Support/Casting.h</tt></a>
259 file (note that you very rarely have to include this file directly).</p>
262 <dt><tt>isa<></tt>: </dt>
264 <dd>The <tt>isa<></tt> operator works exactly like the Java
265 "<tt>instanceof</tt>" operator. It returns true or false depending on whether
266 a reference or pointer points to an instance of the specified class. This can
267 be very useful for constraint checking of various sorts (example below).</dd>
269 <dt><tt>cast<></tt>: </dt>
271 <dd>The <tt>cast<></tt> operator is a "checked cast" operation. It
272 converts a pointer or reference from a base class to a derived cast, causing
273 an assertion failure if it is not really an instance of the right type. This
274 should be used in cases where you have some information that makes you believe
275 that something is of the right type. An example of the <tt>isa<></tt>
276 and <tt>cast<></tt> template is:
279 static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
280 if (isa<<a href="#Constant">Constant</a>>(V) || isa<<a href="#Argument">Argument</a>>(V) || isa<<a href="#GlobalValue">GlobalValue</a>>(V))
283 <i>// Otherwise, it must be an instruction...</i>
284 return !L->contains(cast<<a href="#Instruction">Instruction</a>>(V)->getParent());
287 <p>Note that you should <b>not</b> use an <tt>isa<></tt> test followed
288 by a <tt>cast<></tt>, for that use the <tt>dyn_cast<></tt>
293 <dt><tt>dyn_cast<></tt>:</dt>
295 <dd>The <tt>dyn_cast<></tt> operator is a "checking cast" operation. It
296 checks to see if the operand is of the specified type, and if so, returns a
297 pointer to it (this operator does not work with references). If the operand is
298 not of the correct type, a null pointer is returned. Thus, this works very
299 much like the <tt>dynamic_cast</tt> operator in C++, and should be used in the
300 same circumstances. Typically, the <tt>dyn_cast<></tt> operator is used
301 in an <tt>if</tt> statement or some other flow control statement like this:
304 if (<a href="#AllocationInst">AllocationInst</a> *AI = dyn_cast<<a href="#AllocationInst">AllocationInst</a>>(Val)) {
309 <p> This form of the <tt>if</tt> statement effectively combines together a
310 call to <tt>isa<></tt> and a call to <tt>cast<></tt> into one
311 statement, which is very convenient.</p>
313 <p> Another common example is:</p>
316 <i>// Loop over all of the phi nodes in a basic block</i>
317 BasicBlock::iterator BBI = BB->begin();
318 for (; <a href="#PhiNode">PHINode</a> *PN = dyn_cast<<a href="#PHINode">PHINode</a>>(BBI); ++BBI)
319 std::cerr << *PN;
322 <p>Note that the <tt>dyn_cast<></tt> operator, like C++'s
323 <tt>dynamic_cast</tt> or Java's <tt>instanceof</tt> operator, can be abused.
324 In particular you should not use big chained <tt>if/then/else</tt> blocks to
325 check for lots of different variants of classes. If you find yourself
326 wanting to do this, it is much cleaner and more efficient to use the
327 InstVisitor class to dispatch over the instruction type directly.</p>
331 <dt><tt>cast_or_null<></tt>: </dt>
333 <dd>The <tt>cast_or_null<></tt> operator works just like the
334 <tt>cast<></tt> operator, except that it allows for a null pointer as
335 an argument (which it then propagates). This can sometimes be useful,
336 allowing you to combine several null checks into one.</dd>
338 <dt><tt>dyn_cast_or_null<></tt>: </dt>
340 <dd>The <tt>dyn_cast_or_null<></tt> operator works just like the
341 <tt>dyn_cast<></tt> operator, except that it allows for a null pointer
342 as an argument (which it then propagates). This can sometimes be useful,
343 allowing you to combine several null checks into one.</dd>
347 <p>These five templates can be used with any classes, whether they have a
348 v-table or not. To add support for these templates, you simply need to add
349 <tt>classof</tt> static methods to the class you are interested casting
350 to. Describing this is currently outside the scope of this document, but there
351 are lots of examples in the LLVM source base.</p>
355 <!-- ======================================================================= -->
356 <div class="doc_subsection">
357 <a name="DEBUG">The <tt>DEBUG()</tt> macro & <tt>-debug</tt> option</a>
360 <div class="doc_text">
362 <p>Often when working on your pass you will put a bunch of debugging printouts
363 and other code into your pass. After you get it working, you want to remove
364 it... but you may need it again in the future (to work out new bugs that you run
367 <p> Naturally, because of this, you don't want to delete the debug printouts,
368 but you don't want them to always be noisy. A standard compromise is to comment
369 them out, allowing you to enable them if you need them in the future.</p>
371 <p>The "<tt><a href="/doxygen/Debug_8h-source.html">Support/Debug.h</a></tt>"
372 file provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to
373 this problem. Basically, you can put arbitrary code into the argument of the
374 <tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
375 tool) is run with the '<tt>-debug</tt>' command line argument:</p>
377 <pre> ... <br> DEBUG(std::cerr << "I am here!\n");<br> ...<br></pre>
379 <p>Then you can run your pass like this:</p>
381 <pre> $ opt < a.bc > /dev/null -mypass<br> <no output><br> $ opt < a.bc > /dev/null -mypass -debug<br> I am here!<br> $<br></pre>
383 <p>Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you
384 to not have to create "yet another" command line option for the debug output for
385 your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
386 so they do not cause a performance impact at all (for the same reason, they
387 should also not contain side-effects!).</p>
389 <p>One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
390 enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
391 "<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
392 program hasn't been started yet, you can always just run it with
397 <!-- _______________________________________________________________________ -->
398 <div class="doc_subsubsection">
399 <a name="DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE()</tt> and
400 the <tt>-debug-only</tt> option</a>
403 <div class="doc_text">
405 <p>Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
406 just turns on <b>too much</b> information (such as when working on the code
407 generator). If you want to enable debug information with more fine-grained
408 control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
409 option as follows:</p>
411 <pre> ...<br> DEBUG(std::cerr << "No debug type\n");<br> #undef DEBUG_TYPE<br> #define DEBUG_TYPE "foo"<br> DEBUG(std::cerr << "'foo' debug type\n");<br> #undef DEBUG_TYPE<br> #define DEBUG_TYPE "bar"<br> DEBUG(std::cerr << "'bar' debug type\n");<br> #undef DEBUG_TYPE<br> #define DEBUG_TYPE ""<br> DEBUG(std::cerr << "No debug type (2)\n");<br> ...<br></pre>
413 <p>Then you can run your pass like this:</p>
415 <pre> $ opt < a.bc > /dev/null -mypass<br> <no output><br> $ opt < a.bc > /dev/null -mypass -debug<br> No debug type<br> 'foo' debug type<br> 'bar' debug type<br> No debug type (2)<br> $ opt < a.bc > /dev/null -mypass -debug-only=foo<br> 'foo' debug type<br> $ opt < a.bc > /dev/null -mypass -debug-only=bar<br> 'bar' debug type<br> $<br></pre>
417 <p>Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of
418 a file, to specify the debug type for the entire module (if you do this before
419 you <tt>#include "Support/Debug.h"</tt>, you don't have to insert the ugly
420 <tt>#undef</tt>'s). Also, you should use names more meaningful than "foo" and
421 "bar", because there is no system in place to ensure that names do not
422 conflict. If two different modules use the same string, they will all be turned
423 on when the name is specified. This allows, for example, all debug information
424 for instruction scheduling to be enabled with <tt>-debug-type=InstrSched</tt>,
425 even if the source lives in multiple files.</p>
429 <!-- ======================================================================= -->
430 <div class="doc_subsection">
431 <a name="Statistic">The <tt>Statistic</tt> template & <tt>-stats</tt>
435 <div class="doc_text">
438 href="/doxygen/Statistic_8h-source.html">Support/Statistic.h</a></tt>" file
439 provides a template named <tt>Statistic</tt> that is used as a unified way to
440 keep track of what the LLVM compiler is doing and how effective various
441 optimizations are. It is useful to see what optimizations are contributing to
442 making a particular program run faster.</p>
444 <p>Often you may run your pass on some big program, and you're interested to see
445 how many times it makes a certain transformation. Although you can do this with
446 hand inspection, or some ad-hoc method, this is a real pain and not very useful
447 for big programs. Using the <tt>Statistic</tt> template makes it very easy to
448 keep track of this information, and the calculated information is presented in a
449 uniform manner with the rest of the passes being executed.</p>
451 <p>There are many examples of <tt>Statistic</tt> uses, but the basics of using
452 it are as follows:</p>
455 <li>Define your statistic like this:
456 <pre>static Statistic<> NumXForms("mypassname", "The # of times I did stuff");<br></pre>
458 <p>The <tt>Statistic</tt> template can emulate just about any data-type,
459 but if you do not specify a template argument, it defaults to acting like
460 an unsigned int counter (this is usually what you want).</p></li>
462 <li>Whenever you make a transformation, bump the counter:
463 <pre> ++NumXForms; // I did stuff<br></pre>
467 <p>That's all you have to do. To get '<tt>opt</tt>' to print out the
468 statistics gathered, use the '<tt>-stats</tt>' option:</p>
470 <pre> $ opt -stats -mypassname < program.bc > /dev/null<br> ... statistic output ...<br></pre>
472 <p> When running <tt>gccas</tt> on a C file from the SPEC benchmark
473 suite, it gives a report that looks like this:</p>
475 <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>
477 <p>Obviously, with so many optimizations, having a unified framework for this
478 stuff is very nice. Making your pass fit well into the framework makes it more
479 maintainable and useful.</p>
483 <!-- *********************************************************************** -->
484 <div class="doc_section">
485 <a name="common">Helpful Hints for Common Operations</a>
487 <!-- *********************************************************************** -->
489 <div class="doc_text">
491 <p>This section describes how to perform some very simple transformations of
492 LLVM code. This is meant to give examples of common idioms used, showing the
493 practical side of LLVM transformations. <p> Because this is a "how-to" section,
494 you should also read about the main classes that you will be working with. The
495 <a href="#coreclasses">Core LLVM Class Hierarchy Reference</a> contains details
496 and descriptions of the main classes that you should know about.</p>
500 <!-- NOTE: this section should be heavy on example code -->
501 <!-- ======================================================================= -->
502 <div class="doc_subsection">
503 <a name="inspection">Basic Inspection and Traversal Routines</a>
506 <div class="doc_text">
508 <p>The LLVM compiler infrastructure have many different data structures that may
509 be traversed. Following the example of the C++ standard template library, the
510 techniques used to traverse these various data structures are all basically the
511 same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
512 method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
513 function returns an iterator pointing to one past the last valid element of the
514 sequence, and there is some <tt>XXXiterator</tt> data type that is common
515 between the two operations.</p>
517 <p>Because the pattern for iteration is common across many different aspects of
518 the program representation, the standard template library algorithms may be used
519 on them, and it is easier to remember how to iterate. First we show a few common
520 examples of the data structures that need to be traversed. Other data
521 structures are traversed in very similar ways.</p>
525 <!-- _______________________________________________________________________ -->
526 <div class="doc_subsubsection">
527 <a name="iterate_function">Iterating over the </a><a
528 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
529 href="#Function"><tt>Function</tt></a>
532 <div class="doc_text">
534 <p>It's quite common to have a <tt>Function</tt> instance that you'd like to
535 transform in some way; in particular, you'd like to manipulate its
536 <tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over all of
537 the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>. The following is
538 an example that prints the name of a <tt>BasicBlock</tt> and the number of
539 <tt>Instruction</tt>s it contains:</p>
541 <pre> // func is a pointer to a Function instance<br> for (Function::iterator i = func->begin(), e = func->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 << "Basic block (name=" << i->getName() << ") has " <br> << i->size() << " instructions.\n";<br> }<br></pre>
543 <p>Note that i can be used as if it were a pointer for the purposes of
544 invoking member functions of the <tt>Instruction</tt> class. This is
545 because the indirection operator is overloaded for the iterator
546 classes. In the above code, the expression <tt>i->size()</tt> is
547 exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.</p>
551 <!-- _______________________________________________________________________ -->
552 <div class="doc_subsubsection">
553 <a name="iterate_basicblock">Iterating over the </a><a
554 href="#Instruction"><tt>Instruction</tt></a>s in a <a
555 href="#BasicBlock"><tt>BasicBlock</tt></a>
558 <div class="doc_text">
560 <p>Just like when dealing with <tt>BasicBlock</tt>s in <tt>Function</tt>s, it's
561 easy to iterate over the individual instructions that make up
562 <tt>BasicBlock</tt>s. Here's a code snippet that prints out each instruction in
563 a <tt>BasicBlock</tt>:</p>
565 <pre> // blk is a pointer to a BasicBlock instance<br> for (BasicBlock::iterator i = blk->begin(), e = blk->end(); i != e; ++i)<br> // the next statement works since operator<<(ostream&,...) <br> // is overloaded for Instruction&<br> cerr << *i << "\n";<br></pre>
567 <p>However, this isn't really the best way to print out the contents of a
568 <tt>BasicBlock</tt>! Since the ostream operators are overloaded for virtually
569 anything you'll care about, you could have just invoked the print routine on the
570 basic block itself: <tt>cerr << *blk << "\n";</tt>.</p>
572 <p>Note that currently operator<< is implemented for <tt>Value*</tt>, so
573 it will print out the contents of the pointer, instead of the pointer value you
574 might expect. This is a deprecated interface that will be removed in the
575 future, so it's best not to depend on it. To print out the pointer value for
576 now, you must cast to <tt>void*</tt>.</p>
580 <!-- _______________________________________________________________________ -->
581 <div class="doc_subsubsection">
582 <a name="iterate_institer">Iterating over the </a><a
583 href="#Instruction"><tt>Instruction</tt></a>s in a <a
584 href="#Function"><tt>Function</tt></a>
587 <div class="doc_text">
589 <p>If you're finding that you commonly iterate over a <tt>Function</tt>'s
590 <tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s <tt>Instruction</tt>s,
591 <tt>InstIterator</tt> should be used instead. You'll need to include <a
592 href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>,
593 and then instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
594 small example that shows how to dump all instructions in a function to the standard error stream:<p>
596 <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 << *i << "\n";<br></pre>
597 Easy, isn't it? You can also use <tt>InstIterator</tt>s to fill a
598 worklist with its initial contents. For example, if you wanted to
599 initialize a worklist to contain all instructions in a <tt>Function</tt>
600 F, all you would need to do is something like:
601 <pre>std::set<Instruction*> worklist;<br>worklist.insert(inst_begin(F), inst_end(F));<br></pre>
603 <p>The STL set <tt>worklist</tt> would now contain all instructions in the
604 <tt>Function</tt> pointed to by F.</p>
608 <!-- _______________________________________________________________________ -->
609 <div class="doc_subsubsection">
610 <a name="iterate_convert">Turning an iterator into a class pointer (and
614 <div class="doc_text">
616 <p>Sometimes, it'll be useful to grab a reference (or pointer) to a class
617 instance when all you've got at hand is an iterator. Well, extracting
618 a reference or a pointer from an iterator is very straight-forward.
619 Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and <tt>j</tt>
620 is a <tt>BasicBlock::const_iterator</tt>:</p>
622 <pre> Instruction& inst = *i; // grab reference to instruction reference<br> Instruction* pinst = &*i; // grab pointer to instruction reference<br> const Instruction& inst = *j;<br></pre>
624 <p>However, the iterators you'll be working with in the LLVM framework are
625 special: they will automatically convert to a ptr-to-instance type whenever they
626 need to. Instead of dereferencing the iterator and then taking the address of
627 the result, you can simply assign the iterator to the proper pointer type and
628 you get the dereference and address-of operation as a result of the assignment
629 (behind the scenes, this is a result of overloading casting mechanisms). Thus
630 the last line of the last example,</p>
632 <pre>Instruction* pinst = &*i;</pre>
634 <p>is semantically equivalent to</p>
636 <pre>Instruction* pinst = i;</pre>
638 <p>It's also possible to turn a class pointer into the corresponding iterator,
639 and this is a constant time operation (very efficient). The following code
640 snippet illustrates use of the conversion constructors provided by LLVM
641 iterators. By using these, you can explicitly grab the iterator of something
642 without actually obtaining it via iteration over some structure:</p>
644 <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->getParent()->end()) cerr << *it << "\n";<br>}<br></pre>
648 <!--_______________________________________________________________________-->
649 <div class="doc_subsubsection">
650 <a name="iterate_complex">Finding call sites: a slightly more complex
654 <div class="doc_text">
656 <p>Say that you're writing a FunctionPass and would like to count all the
657 locations in the entire module (that is, across every <tt>Function</tt>) where a
658 certain function (i.e., some <tt>Function</tt>*) is already in scope. As you'll
659 learn later, you may want to use an <tt>InstVisitor</tt> to accomplish this in a
660 much more straight-forward manner, but this example will allow us to explore how
661 you'd do it if you didn't have <tt>InstVisitor</tt> around. In pseudocode, this
662 is what we want to do:</p>
664 <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>
666 <p>And the actual code is (remember, since we're writing a
667 <tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply has to
668 override the <tt>runOnFunction</tt> method...):</p>
670 <pre>Function* targetFunc = ...;<br><br>class OurFunctionPass : public FunctionPass {<br> public:<br> OurFunctionPass(): callCounter(0) { }<br><br> virtual runOnFunction(Function& F) {<br> for (Function::iterator b = F.begin(), be = F.end(); b != be; ++b) {<br> for (BasicBlock::iterator i = b->begin(); ie = b->end(); i != ie; ++i) {<br> if (<a
671 href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a><<a
672 href="#CallInst">CallInst</a>>(&*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->getCalledFunction() == targetFunc)<br> ++callCounter;<br> }<br> }<br> }<br> <br> private:<br> unsigned callCounter;<br>};<br></pre>
676 <!--_______________________________________________________________________-->
677 <div class="doc_subsubsection">
678 <a name="calls_and_invokes">Treating calls and invokes the same way</a>
681 <div class="doc_text">
683 <p>You may have noticed that the previous example was a bit oversimplified in
684 that it did not deal with call sites generated by 'invoke' instructions. In
685 this, and in other situations, you may find that you want to treat
686 <tt>CallInst</tt>s and <tt>InvokeInst</tt>s the same way, even though their
687 most-specific common base class is <tt>Instruction</tt>, which includes lots of
688 less closely-related things. For these cases, LLVM provides a handy wrapper
690 href="http://llvm.cs.uiuc.edu/doxygen/classCallSite.html"><tt>CallSite</tt></a>.
691 It is essentially a wrapper around an <tt>Instruction</tt> pointer, with some
692 methods that provide functionality common to <tt>CallInst</tt>s and
693 <tt>InvokeInst</tt>s.</p>
695 <p>This class has "value semantics": it should be passed by value, not by
696 reference and it should not be dynamically allocated or deallocated using
697 <tt>operator new</tt> or <tt>operator delete</tt>. It is efficiently copyable,
698 assignable and constructable, with costs equivalents to that of a bare pointer.
699 If you look at its definition, it has only a single pointer member.</p>
703 <!--_______________________________________________________________________-->
704 <div class="doc_subsubsection">
705 <a name="iterate_chains">Iterating over def-use & use-def chains</a>
708 <div class="doc_text">
710 <p>Frequently, we might have an instance of the <a
711 href="/doxygen/classValue.html">Value Class</a> and we want to determine which
712 <tt>User</tt>s use the <tt>Value</tt>. The list of all <tt>User</tt>s of a
713 particular <tt>Value</tt> is called a <i>def-use</i> chain. For example, let's
714 say we have a <tt>Function*</tt> named <tt>F</tt> to a particular function
715 <tt>foo</tt>. Finding all of the instructions that <i>use</i> <tt>foo</tt> is as
716 simple as iterating over the <i>def-use</i> chain of <tt>F</tt>:</p>
718 <pre>Function* F = ...;<br><br>for (Value::use_iterator i = F->use_begin(), e = F->use_end(); i != e; ++i) {<br> if (Instruction *Inst = dyn_cast<Instruction>(*i)) {<br> cerr << "F is used in instruction:\n";<br> cerr << *Inst << "\n";<br> }<br>}<br></pre>
720 <p>Alternately, it's common to have an instance of the <a
721 href="/doxygen/classUser.html">User Class</a> and need to know what
722 <tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used by a
723 <tt>User</tt> is known as a <i>use-def</i> chain. Instances of class
724 <tt>Instruction</tt> are common <tt>User</tt>s, so we might want to iterate over
725 all of the values that a particular instruction uses (that is, the operands of
726 the particular <tt>Instruction</tt>):</p>
728 <pre>Instruction* pi = ...;<br><br>for (User::op_iterator i = pi->op_begin(), e = pi->op_end(); i != e; ++i) {<br> Value* v = *i;<br> ...<br>}<br></pre>
731 def-use chains ("finding all users of"): Value::use_begin/use_end
732 use-def chains ("finding all values used"): User::op_begin/op_end [op=operand]
737 <!-- ======================================================================= -->
738 <div class="doc_subsection">
739 <a name="simplechanges">Making simple changes</a>
742 <div class="doc_text">
744 <p>There are some primitive transformation operations present in the LLVM
745 infrastructure that are worth knowing about. When performing
746 transformations, it's fairly common to manipulate the contents of basic
747 blocks. This section describes some of the common methods for doing so
748 and gives example code.</p>
752 <!--_______________________________________________________________________-->
753 <div class="doc_subsubsection">
754 <a name="schanges_creating">Creating and inserting new
755 <tt>Instruction</tt>s</a>
758 <div class="doc_text">
760 <p><i>Instantiating Instructions</i></p>
762 <p>Creation of <tt>Instruction</tt>s is straight-forward: simply call the
763 constructor for the kind of instruction to instantiate and provide the necessary
764 parameters. For example, an <tt>AllocaInst</tt> only <i>requires</i> a
765 (const-ptr-to) <tt>Type</tt>. Thus:</p>
767 <pre>AllocaInst* ai = new AllocaInst(Type::IntTy);</pre>
769 <p>will create an <tt>AllocaInst</tt> instance that represents the allocation of
770 one integer in the current stack frame, at runtime. Each <tt>Instruction</tt>
771 subclass is likely to have varying default parameters which change the semantics
772 of the instruction, so refer to the <a
773 href="/doxygen/classInstruction.html">doxygen documentation for the subclass of
774 Instruction</a> that you're interested in instantiating.</p>
776 <p><i>Naming values</i></p>
778 <p>It is very useful to name the values of instructions when you're able to, as
779 this facilitates the debugging of your transformations. If you end up looking
780 at generated LLVM machine code, you definitely want to have logical names
781 associated with the results of instructions! By supplying a value for the
782 <tt>Name</tt> (default) parameter of the <tt>Instruction</tt> constructor, you
783 associate a logical name with the result of the instruction's execution at
784 runtime. For example, say that I'm writing a transformation that dynamically
785 allocates space for an integer on the stack, and that integer is going to be
786 used as some kind of index by some other code. To accomplish this, I place an
787 <tt>AllocaInst</tt> at the first point in the first <tt>BasicBlock</tt> of some
788 <tt>Function</tt>, and I'm intending to use it within the same
789 <tt>Function</tt>. I might do:</p>
791 <pre>AllocaInst* pa = new AllocaInst(Type::IntTy, 0, "indexLoc");</pre>
793 <p>where <tt>indexLoc</tt> is now the logical name of the instruction's
794 execution value, which is a pointer to an integer on the runtime stack.</p>
796 <p><i>Inserting instructions</i></p>
798 <p>There are essentially two ways to insert an <tt>Instruction</tt>
799 into an existing sequence of instructions that form a <tt>BasicBlock</tt>:</p>
802 <li>Insertion into an explicit instruction list
804 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within that
805 <tt>BasicBlock</tt>, and a newly-created instruction we wish to insert
806 before <tt>*pi</tt>, we do the following: </p>
808 <pre> BasicBlock *pb = ...;<br> Instruction *pi = ...;<br> Instruction *newInst = new Instruction(...);<br> pb->getInstList().insert(pi, newInst); // inserts newInst before pi in pb<br></pre></li>
810 <li>Insertion into an implicit instruction list
812 <p><tt>Instruction</tt> instances that are already in <tt>BasicBlock</tt>s
813 are implicitly associated with an existing instruction list: the instruction
814 list of the enclosing basic block. Thus, we could have accomplished the same
815 thing as the above code without being given a <tt>BasicBlock</tt> by doing:
818 <pre> Instruction *pi = ...;<br> Instruction *newInst = new Instruction(...);<br> pi->getParent()->getInstList().insert(pi, newInst);<br></pre>
820 <p>In fact, this sequence of steps occurs so frequently that the
821 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes provide
822 constructors which take (as a default parameter) a pointer to an
823 <tt>Instruction</tt> which the newly-created <tt>Instruction</tt> should
824 precede. That is, <tt>Instruction</tt> constructors are capable of
825 inserting the newly-created instance into the <tt>BasicBlock</tt> of a
826 provided instruction, immediately before that instruction. Using an
827 <tt>Instruction</tt> constructor with a <tt>insertBefore</tt> (default)
828 parameter, the above code becomes:</p>
830 <pre>Instruction* pi = ...;<br>Instruction* newInst = new Instruction(..., pi);<br></pre>
832 <p>which is much cleaner, especially if you're creating a lot of
833 instructions and adding them to <tt>BasicBlock</tt>s.</p></li>
838 <!--_______________________________________________________________________-->
839 <div class="doc_subsubsection">
840 <a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a>
843 <div class="doc_text">
845 <p>Deleting an instruction from an existing sequence of instructions that form a
846 <a href="#BasicBlock"><tt>BasicBlock</tt></a> is very straight-forward. First,
847 you must have a pointer to the instruction that you wish to delete. Second, you
848 need to obtain the pointer to that instruction's basic block. You use the
849 pointer to the basic block to get its list of instructions and then use the
850 erase function to remove your instruction. For example:</p>
852 <pre> <a href="#Instruction">Instruction</a> *I = .. ;<br> <a
853 href="#BasicBlock">BasicBlock</a> *BB = I->getParent();<br> BB->getInstList().erase(I);<br></pre>
857 <!--_______________________________________________________________________-->
858 <div class="doc_subsubsection">
859 <a name="schanges_replacing">Replacing an <tt>Instruction</tt> with another
863 <div class="doc_text">
865 <p><i>Replacing individual instructions</i></p>
867 <p>Including "<a href="/doxygen/BasicBlockUtils_8h-source.html">llvm/Transforms/Utils/BasicBlockUtils.h</a>"
868 permits use of two very useful replace functions: <tt>ReplaceInstWithValue</tt>
869 and <tt>ReplaceInstWithInst</tt>.</p>
871 <h4><a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a></h4>
874 <li><tt>ReplaceInstWithValue</tt>
876 <p>This function replaces all uses (within a basic block) of a given
877 instruction with a value, and then removes the original instruction. The
878 following example illustrates the replacement of the result of a particular
879 <tt>AllocaInst</tt> that allocates memory for a single integer with an null
880 pointer to an integer.</p>
882 <pre>AllocaInst* instToReplace = ...;<br>BasicBlock::iterator ii(instToReplace);<br>ReplaceInstWithValue(instToReplace->getParent()->getInstList(), ii,<br> Constant::getNullValue(PointerType::get(Type::IntTy)));<br></pre></li>
884 <li><tt>ReplaceInstWithInst</tt>
886 <p>This function replaces a particular instruction with another
887 instruction. The following example illustrates the replacement of one
888 <tt>AllocaInst</tt> with another.</p>
890 <pre>AllocaInst* instToReplace = ...;<br>BasicBlock::iterator ii(instToReplace);<br>ReplaceInstWithInst(instToReplace->getParent()->getInstList(), ii,<br> new AllocaInst(Type::IntTy, 0, "ptrToReplacedInt"));<br></pre></li>
893 <p><i>Replacing multiple uses of <tt>User</tt>s and <tt>Value</tt>s</i></p>
895 <p>You can use <tt>Value::replaceAllUsesWith</tt> and
896 <tt>User::replaceUsesOfWith</tt> to change more than one use at a time. See the
897 doxygen documentation for the <a href="/doxygen/classValue.html">Value Class</a>
898 and <a href="/doxygen/classUser.html">User Class</a>, respectively, for more
901 <!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
902 include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
903 ReplaceInstWithValue, ReplaceInstWithInst -->
907 <!-- *********************************************************************** -->
908 <div class="doc_section">
909 <a name="coreclasses">The Core LLVM Class Hierarchy Reference </a>
911 <!-- *********************************************************************** -->
913 <div class="doc_text">
915 <p>The Core LLVM classes are the primary means of representing the program
916 being inspected or transformed. The core LLVM classes are defined in
917 header files in the <tt>include/llvm/</tt> directory, and implemented in
918 the <tt>lib/VMCore</tt> directory.</p>
922 <!-- ======================================================================= -->
923 <div class="doc_subsection">
924 <a name="Value">The <tt>Value</tt> class</a>
929 <p><tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt>
931 doxygen info: <a href="/doxygen/classValue.html">Value Class</a></p>
933 <p>The <tt>Value</tt> class is the most important class in the LLVM Source
934 base. It represents a typed value that may be used (among other things) as an
935 operand to an instruction. There are many different types of <tt>Value</tt>s,
936 such as <a href="#Constant"><tt>Constant</tt></a>s,<a
937 href="#Argument"><tt>Argument</tt></a>s. Even <a
938 href="#Instruction"><tt>Instruction</tt></a>s and <a
939 href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.</p>
941 <p>A particular <tt>Value</tt> may be used many times in the LLVM representation
942 for a program. For example, an incoming argument to a function (represented
943 with an instance of the <a href="#Argument">Argument</a> class) is "used" by
944 every instruction in the function that references the argument. To keep track
945 of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
946 href="#User"><tt>User</tt></a>s that is using it (the <a
947 href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
948 graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
949 def-use information in the program, and is accessible through the <tt>use_</tt>*
950 methods, shown below.</p>
952 <p>Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed,
953 and this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
954 method. In addition, all LLVM values can be named. The "name" of the
955 <tt>Value</tt> is a symbolic string printed in the LLVM code:</p>
957 <pre> %<b>foo</b> = add int 1, 2<br></pre>
959 <p><a name="#nameWarning">The name of this instruction is "foo".</a> <b>NOTE</b>
960 that the name of any value may be missing (an empty string), so names should
961 <b>ONLY</b> be used for debugging (making the source code easier to read,
962 debugging printouts), they should not be used to keep track of values or map
963 between them. For this purpose, use a <tt>std::map</tt> of pointers to the
964 <tt>Value</tt> itself instead.</p>
966 <p>One important aspect of LLVM is that there is no distinction between an SSA
967 variable and the operation that produces it. Because of this, any reference to
968 the value produced by an instruction (or the value available as an incoming
969 argument, for example) is represented as a direct pointer to the instance of
971 represents this value. Although this may take some getting used to, it
972 simplifies the representation and makes it easier to manipulate.</p>
976 <!-- _______________________________________________________________________ -->
977 <div class="doc_subsubsection">
978 <a name="m_Value">Important Public Members of the <tt>Value</tt> class</a>
981 <div class="doc_text">
984 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the
986 <tt>Value::use_const_iterator</tt> - Typedef for const_iterator over
988 <tt>unsigned use_size()</tt> - Returns the number of users of the
990 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
991 <tt>use_iterator use_begin()</tt> - Get an iterator to the start of
993 <tt>use_iterator use_end()</tt> - Get an iterator to the end of the
995 <tt><a href="#User">User</a> *use_back()</tt> - Returns the last
997 <p> These methods are the interface to access the def-use
998 information in LLVM. As with all other iterators in LLVM, the naming
999 conventions follow the conventions defined by the <a href="#stl">STL</a>.</p>
1001 <li><tt><a href="#Type">Type</a> *getType() const</tt>
1002 <p>This method returns the Type of the Value.</p>
1004 <li><tt>bool hasName() const</tt><br>
1005 <tt>std::string getName() const</tt><br>
1006 <tt>void setName(const std::string &Name)</tt>
1007 <p> This family of methods is used to access and assign a name to a <tt>Value</tt>,
1008 be aware of the <a href="#nameWarning">precaution above</a>.</p>
1010 <li><tt>void replaceAllUsesWith(Value *V)</tt>
1012 <p>This method traverses the use list of a <tt>Value</tt> changing all <a
1013 href="#User"><tt>User</tt>s</a> of the current value to refer to
1014 "<tt>V</tt>" instead. For example, if you detect that an instruction always
1015 produces a constant value (for example through constant folding), you can
1016 replace all uses of the instruction with the constant like this:</p>
1018 <pre> Inst->replaceAllUsesWith(ConstVal);<br></pre>
1023 <!-- ======================================================================= -->
1024 <div class="doc_subsection">
1025 <a name="User">The <tt>User</tt> class</a>
1028 <div class="doc_text">
1031 <tt>#include "<a href="/doxygen/User_8h-source.html">llvm/User.h</a>"</tt><br>
1032 doxygen info: <a href="/doxygen/classUser.html">User Class</a><br>
1033 Superclass: <a href="#Value"><tt>Value</tt></a></p>
1035 <p>The <tt>User</tt> class is the common base class of all LLVM nodes that may
1036 refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
1037 that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
1038 referring to. The <tt>User</tt> class itself is a subclass of
1041 <p>The operands of a <tt>User</tt> point directly to the LLVM <a
1042 href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
1043 Single Assignment (SSA) form, there can only be one definition referred to,
1044 allowing this direct connection. This connection provides the use-def
1045 information in LLVM.</p>
1049 <!-- _______________________________________________________________________ -->
1050 <div class="doc_subsubsection">
1051 <a name="m_User">Important Public Members of the <tt>User</tt> class</a>
1054 <div class="doc_text">
1056 <p>The <tt>User</tt> class exposes the operand list in two ways: through
1057 an index access interface and through an iterator based interface.</p>
1060 <li><tt>Value *getOperand(unsigned i)</tt><br>
1061 <tt>unsigned getNumOperands()</tt>
1062 <p> These two methods expose the operands of the <tt>User</tt> in a
1063 convenient form for direct access.</p></li>
1065 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand
1067 <tt>User::op_const_iterator</tt> <tt>use_iterator op_begin()</tt> -
1068 Get an iterator to the start of the operand list.<br>
1069 <tt>use_iterator op_end()</tt> - Get an iterator to the end of the
1071 <p> Together, these methods make up the iterator based interface to
1072 the operands of a <tt>User</tt>.</p></li>
1077 <!-- ======================================================================= -->
1078 <div class="doc_subsection">
1079 <a name="Instruction">The <tt>Instruction</tt> class</a>
1082 <div class="doc_text">
1084 <p><tt>#include "</tt><tt><a
1085 href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt><br>
1086 doxygen info: <a href="/doxygen/classInstruction.html">Instruction Class</a><br>
1087 Superclasses: <a href="#User"><tt>User</tt></a>, <a
1088 href="#Value"><tt>Value</tt></a></p>
1090 <p>The <tt>Instruction</tt> class is the common base class for all LLVM
1091 instructions. It provides only a few methods, but is a very commonly used
1092 class. The primary data tracked by the <tt>Instruction</tt> class itself is the
1093 opcode (instruction type) and the parent <a
1094 href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
1095 into. To represent a specific type of instruction, one of many subclasses of
1096 <tt>Instruction</tt> are used.</p>
1098 <p> Because the <tt>Instruction</tt> class subclasses the <a
1099 href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
1100 way as for other <a href="#User"><tt>User</tt></a>s (with the
1101 <tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
1102 <tt>op_begin()</tt>/<tt>op_end()</tt> methods).</p> <p> An important file for
1103 the <tt>Instruction</tt> class is the <tt>llvm/Instruction.def</tt> file. This
1104 file contains some meta-data about the various different types of instructions
1105 in LLVM. It describes the enum values that are used as opcodes (for example
1106 <tt>Instruction::Add</tt> and <tt>Instruction::SetLE</tt>), as well as the
1107 concrete sub-classes of <tt>Instruction</tt> that implement the instruction (for
1108 example <tt><a href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
1109 href="#SetCondInst">SetCondInst</a></tt>). Unfortunately, the use of macros in
1110 this file confuses doxygen, so these enum values don't show up correctly in the
1111 <a href="/doxygen/classInstruction.html">doxygen output</a>.</p>
1115 <!-- _______________________________________________________________________ -->
1116 <div class="doc_subsubsection">
1117 <a name="m_Instruction">Important Public Members of the <tt>Instruction</tt>
1121 <div class="doc_text">
1124 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt>
1125 <p>Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that
1126 this <tt>Instruction</tt> is embedded into.</p></li>
1127 <li><tt>bool mayWriteToMemory()</tt>
1128 <p>Returns true if the instruction writes to memory, i.e. it is a
1129 <tt>call</tt>,<tt>free</tt>,<tt>invoke</tt>, or <tt>store</tt>.</p></li>
1130 <li><tt>unsigned getOpcode()</tt>
1131 <p>Returns the opcode for the <tt>Instruction</tt>.</p></li>
1132 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt>
1133 <p>Returns another instance of the specified instruction, identical
1134 in all ways to the original except that the instruction has no parent
1135 (ie it's not embedded into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>),
1136 and it has no name</p></li>
1141 <!-- ======================================================================= -->
1142 <div class="doc_subsection">
1143 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
1146 <div class="doc_text">
1148 <p><tt>#include "<a href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt><br>
1149 doxygen info: <a href="/doxygen/classBasicBlock.html">BasicBlock Class</a><br>
1150 Superclass: <a href="#Value"><tt>Value</tt></a></p>
1152 <p>This class represents a single entry multiple exit section of the code,
1153 commonly known as a basic block by the compiler community. The
1154 <tt>BasicBlock</tt> class maintains a list of <a
1155 href="#Instruction"><tt>Instruction</tt></a>s, which form the body of the block.
1156 Matching the language definition, the last element of this list of instructions
1157 is always a terminator instruction (a subclass of the <a
1158 href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).</p>
1160 <p>In addition to tracking the list of instructions that make up the block, the
1161 <tt>BasicBlock</tt> class also keeps track of the <a
1162 href="#Function"><tt>Function</tt></a> that it is embedded into.</p>
1164 <p>Note that <tt>BasicBlock</tt>s themselves are <a
1165 href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
1166 like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
1171 <!-- _______________________________________________________________________ -->
1172 <div class="doc_subsubsection">
1173 <a name="m_BasicBlock">Important Public Members of the <tt>BasicBlock</tt>
1177 <div class="doc_text">
1180 <li><tt>BasicBlock(const std::string &Name = "", </tt><tt><a
1181 href="#Function">Function</a> *Parent = 0)</tt>
1182 <p>The <tt>BasicBlock</tt> constructor is used to create new basic
1183 blocks for insertion into a function. The constructor optionally takes
1184 a name for the new block, and a <a href="#Function"><tt>Function</tt></a>
1185 to insert it into. If the <tt>Parent</tt> parameter is specified, the
1186 new <tt>BasicBlock</tt> is automatically inserted at the end of the
1187 specified <a href="#Function"><tt>Function</tt></a>, if not specified,
1188 the BasicBlock must be manually inserted into the <a href="#Function"><tt>Function</tt></a>.</p>
1190 <li><tt>BasicBlock::iterator</tt> - Typedef for instruction list
1192 <tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
1193 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,<tt>size()</tt>,<tt>empty()</tt>,<tt>rbegin()</tt>,<tt>rend()
1194 - </tt>STL style functions for accessing the instruction list.
1195 <p> These methods and typedefs are forwarding functions that have
1196 the same semantics as the standard library methods of the same names.
1197 These methods expose the underlying instruction list of a basic block in
1198 a way that is easy to manipulate. To get the full complement of
1199 container operations (including operations to update the list), you must
1200 use the <tt>getInstList()</tt> method.</p></li>
1201 <li><tt>BasicBlock::InstListType &getInstList()</tt>
1202 <p> This method is used to get access to the underlying container
1203 that actually holds the Instructions. This method must be used when
1204 there isn't a forwarding function in the <tt>BasicBlock</tt> class for
1205 the operation that you would like to perform. Because there are no
1206 forwarding functions for "updating" operations, you need to use this if
1207 you want to update the contents of a <tt>BasicBlock</tt>.</p></li>
1208 <li><tt><a href="#Function">Function</a> *getParent()</tt>
1209 <p> Returns a pointer to <a href="#Function"><tt>Function</tt></a>
1210 the block is embedded into, or a null pointer if it is homeless.</p></li>
1211 <li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt>
1212 <p> Returns a pointer to the terminator instruction that appears at
1213 the end of the <tt>BasicBlock</tt>. If there is no terminator
1214 instruction, or if the last instruction in the block is not a
1215 terminator, then a null pointer is returned.</p></li>
1220 <!-- ======================================================================= -->
1221 <div class="doc_subsection">
1222 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
1225 <div class="doc_text">
1228 href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt><br>
1229 doxygen info: <a href="/doxygen/classGlobalValue.html">GlobalValue Class</a><br>
1230 Superclasses: <a href="#User"><tt>User</tt></a>, <a
1231 href="#Value"><tt>Value</tt></a></p>
1233 <p>Global values (<a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
1234 href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
1235 visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
1236 Because they are visible at global scope, they are also subject to linking with
1237 other globals defined in different translation units. To control the linking
1238 process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
1239 <tt>GlobalValue</tt>s know whether they have internal or external linkage, as
1240 defined by the <tt>LinkageTypes</tt> enumerator.</p>
1242 <p>If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
1243 <tt>static</tt> in C), it is not visible to code outside the current translation
1244 unit, and does not participate in linking. If it has external linkage, it is
1245 visible to external code, and does participate in linking. In addition to
1246 linkage information, <tt>GlobalValue</tt>s keep track of which <a
1247 href="#Module"><tt>Module</tt></a> they are currently part of.</p>
1249 <p>Because <tt>GlobalValue</tt>s are memory objects, they are always referred to
1250 by their <b>address</b>. As such, the <a href="#Type"><tt>Type</tt></a> of a
1251 global is always a pointer to its contents. It is important to remember this
1252 when using the <tt>GetElementPtrInst</tt> instruction because this pointer must
1253 be dereferenced first. For example, if you have a <tt>GlobalVariable</tt> (a
1254 subclass of <tt>GlobalValue)</tt> that is an array of 24 ints, type <tt>[24 x
1255 int]</tt>, then the <tt>GlobalVariable</tt> is a pointer to that array. Although
1256 the address of the first element of this array and the value of the
1257 <tt>GlobalVariable</tt> are the same, they have different types. The
1258 <tt>GlobalVariable</tt>'s type is <tt>[24 x int]</tt>. The first element's type
1259 is <tt>int.</tt> Because of this, accessing a global value requires you to
1260 dereference the pointer with <tt>GetElementPtrInst</tt> first, then its elements
1261 can be accessed. This is explained in the <a href="LangRef.html#globalvars">LLVM
1262 Language Reference Manual</a>.</p>
1266 <!-- _______________________________________________________________________ -->
1267 <div class="doc_subsubsection">
1268 <a name="m_GlobalValue">Important Public Members of the <tt>GlobalValue</tt>
1272 <div class="doc_text">
1275 <li><tt>bool hasInternalLinkage() const</tt><br>
1276 <tt>bool hasExternalLinkage() const</tt><br>
1277 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt>
1278 <p> These methods manipulate the linkage characteristics of the <tt>GlobalValue</tt>.</p>
1281 <li><tt><a href="#Module">Module</a> *getParent()</tt>
1282 <p> This returns the <a href="#Module"><tt>Module</tt></a> that the
1283 GlobalValue is currently embedded into.</p></li>
1288 <!-- ======================================================================= -->
1289 <div class="doc_subsection">
1290 <a name="Function">The <tt>Function</tt> class</a>
1293 <div class="doc_text">
1296 href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt><br> doxygen
1297 info: <a href="/doxygen/classFunction.html">Function Class</a><br> Superclasses:
1298 <a href="#GlobalValue"><tt>GlobalValue</tt></a>, <a
1299 href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a></p>
1301 <p>The <tt>Function</tt> class represents a single procedure in LLVM. It is
1302 actually one of the more complex classes in the LLVM heirarchy because it must
1303 keep track of a large amount of data. The <tt>Function</tt> class keeps track
1304 of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal <a
1305 href="#Argument"><tt>Argument</tt></a>s, and a <a
1306 href="#SymbolTable"><tt>SymbolTable</tt></a>.</p>
1308 <p>The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most
1309 commonly used part of <tt>Function</tt> objects. The list imposes an implicit
1310 ordering of the blocks in the function, which indicate how the code will be
1311 layed out by the backend. Additionally, the first <a
1312 href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
1313 <tt>Function</tt>. It is not legal in LLVM to explicitly branch to this initial
1314 block. There are no implicit exit nodes, and in fact there may be multiple exit
1315 nodes from a single <tt>Function</tt>. If the <a
1316 href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
1317 the <tt>Function</tt> is actually a function declaration: the actual body of the
1318 function hasn't been linked in yet.</p>
1320 <p>In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
1321 <tt>Function</tt> class also keeps track of the list of formal <a
1322 href="#Argument"><tt>Argument</tt></a>s that the function receives. This
1323 container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
1324 nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
1325 the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.</p>
1327 <p>The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used
1328 LLVM feature that is only used when you have to look up a value by name. Aside
1329 from that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used
1330 internally to make sure that there are not conflicts between the names of <a
1331 href="#Instruction"><tt>Instruction</tt></a>s, <a
1332 href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
1333 href="#Argument"><tt>Argument</tt></a>s in the function body.</p>
1337 <!-- _______________________________________________________________________ -->
1338 <div class="doc_subsubsection">
1339 <a name="m_Function">Important Public Members of the <tt>Function</tt>
1343 <div class="doc_text">
1346 <li><tt>Function(const </tt><tt><a href="#FunctionType">FunctionType</a>
1347 *Ty, bool isInternal, const std::string &N = "", Module* Parent = 0)</tt>
1349 <p>Constructor used when you need to create new <tt>Function</tt>s to add
1350 the the program. The constructor must specify the type of the function to
1351 create and whether or not it should start out with internal or external
1352 linkage. The <a href="#FunctionType"><tt>FunctionType</tt></a> argument
1353 specifies the formal arguments and return value for the function. The same
1354 <a href="#FunctionTypel"><tt>FunctionType</tt></a> value can be used to
1355 create multiple functions. The <tt>Parent</tt> argument specifies the Module
1356 in which the function is defined. If this argument is provided, the function
1357 will automatically be inserted into that module's list of
1360 <li><tt>bool isExternal()</tt>
1362 <p>Return whether or not the <tt>Function</tt> has a body defined. If the
1363 function is "external", it does not have a body, and thus must be resolved
1364 by linking with a function defined in a different translation unit.</p></li>
1366 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
1367 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
1369 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1370 <tt>size()</tt>, <tt>empty()</tt>, <tt>rbegin()</tt>, <tt>rend()</tt>
1372 <p>These are forwarding methods that make it easy to access the contents of
1373 a <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
1376 <li><tt>Function::BasicBlockListType &getBasicBlockList()</tt>
1378 <p>Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This
1379 is necessary to use when you need to update the list or perform a complex
1380 action that doesn't have a forwarding method.</p></li>
1382 <li><tt>Function::aiterator</tt> - Typedef for the argument list
1384 <tt>Function::const_aiterator</tt> - Typedef for const_iterator.<br>
1386 <tt>abegin()</tt>, <tt>aend()</tt>, <tt>afront()</tt>, <tt>aback()</tt>,
1387 <tt>asize()</tt>, <tt>aempty()</tt>, <tt>arbegin()</tt>, <tt>arend()</tt>
1389 <p>These are forwarding methods that make it easy to access the contents of
1390 a <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a>
1393 <li><tt>Function::ArgumentListType &getArgumentList()</tt>
1395 <p>Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
1396 necessary to use when you need to update the list or perform a complex
1397 action that doesn't have a forwarding method.</p></li>
1399 <li><tt><a href="#BasicBlock">BasicBlock</a> &getEntryBlock()</tt>
1401 <p>Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
1402 function. Because the entry block for the function is always the first
1403 block, this returns the first block of the <tt>Function</tt>.</p></li>
1405 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
1406 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt>
1408 <p>This traverses the <a href="#Type"><tt>Type</tt></a> of the
1409 <tt>Function</tt> and returns the return type of the function, or the <a
1410 href="#FunctionType"><tt>FunctionType</tt></a> of the actual
1413 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
1415 <p> Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
1416 for this <tt>Function</tt>.</p></li>
1421 <!-- ======================================================================= -->
1422 <div class="doc_subsection">
1423 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
1426 <div class="doc_text">
1429 href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt>
1431 doxygen info: <a href="/doxygen/classGlobalVariable.html">GlobalVariable
1432 Class</a><br> Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>, <a
1433 href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a></p>
1435 <p>Global variables are represented with the (suprise suprise)
1436 <tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are also
1437 subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such are
1438 always referenced by their address (global values must live in memory, so their
1439 "name" refers to their address). See <a
1440 href="#GlobalValue"><tt>GlobalValue</tt></a> for more on this. Global variables
1441 may have an initial value (which must be a <a
1442 href="#Constant"><tt>Constant</tt></a>), and if they have an initializer, they
1443 may be marked as "constant" themselves (indicating that their contents never
1444 change at runtime).</p>
1448 <!-- _______________________________________________________________________ -->
1449 <div class="doc_subsubsection">
1450 <a name="m_GlobalVariable">Important Public Members of the
1451 <tt>GlobalVariable</tt> class</a>
1454 <div class="doc_text">
1457 <li><tt>GlobalVariable(const </tt><tt><a href="#Type">Type</a> *Ty, bool
1458 isConstant, LinkageTypes& Linkage, <a href="#Constant">Constant</a>
1459 *Initializer = 0, const std::string &Name = "", Module* Parent = 0)</tt>
1461 <p>Create a new global variable of the specified type. If
1462 <tt>isConstant</tt> is true then the global variable will be marked as
1463 unchanging for the program. The Linkage parameter specifies the type of
1464 linkage (internal, external, weak, linkonce, appending) for the variable. If
1465 the linkage is InternalLinkage, WeakLinkage, or LinkOnceLinkage, then
1466 the resultant global variable will have internal linkage. AppendingLinkage
1467 concatenates together all instances (in different translation units) of the
1468 variable into a single variable but is only applicable to arrays. See
1469 the <a href="LangRef.html#modulestructure">LLVM Language Reference</a> for
1470 further details on linkage types. Optionally an initializer, a name, and the
1471 module to put the variable into may be specified for the global variable as
1474 <li><tt>bool isConstant() const</tt>
1476 <p>Returns true if this is a global variable that is known not to
1477 be modified at runtime.</p></li>
1479 <li><tt>bool hasInitializer()</tt>
1481 <p>Returns true if this <tt>GlobalVariable</tt> has an intializer.</p></li>
1483 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt>
1485 <p>Returns the intial value for a <tt>GlobalVariable</tt>. It is not legal
1486 to call this method if there is no initializer.</p></li>
1491 <!-- ======================================================================= -->
1492 <div class="doc_subsection">
1493 <a name="Module">The <tt>Module</tt> class</a>
1496 <div class="doc_text">
1499 href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt><br> doxygen info:
1500 <a href="/doxygen/classModule.html">Module Class</a></p>
1502 <p>The <tt>Module</tt> class represents the top level structure present in LLVM
1503 programs. An LLVM module is effectively either a translation unit of the
1504 original program or a combination of several translation units merged by the
1505 linker. The <tt>Module</tt> class keeps track of a list of <a
1506 href="#Function"><tt>Function</tt></a>s, a list of <a
1507 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
1508 href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
1509 helpful member functions that try to make common operations easy.</p>
1513 <!-- _______________________________________________________________________ -->
1514 <div class="doc_subsubsection">
1515 <a name="m_Module">Important Public Members of the <tt>Module</tt> class</a>
1518 <div class="doc_text">
1521 <li><tt>Module::Module(std::string name = "")</tt></li>
1524 <p>Constructing a <a href="#Module">Module</a> is easy. You can optionally
1525 provide a name for it (probably based on the name of the translation unit).</p>
1528 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
1529 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
1531 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1532 <tt>size()</tt>, <tt>empty()</tt>, <tt>rbegin()</tt>, <tt>rend()</tt>
1534 <p>These are forwarding methods that make it easy to access the contents of
1535 a <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
1538 <li><tt>Module::FunctionListType &getFunctionList()</tt>
1540 <p> Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
1541 necessary to use when you need to update the list or perform a complex
1542 action that doesn't have a forwarding method.</p>
1544 <p><!-- Global Variable --></p></li>
1550 <li><tt>Module::giterator</tt> - Typedef for global variable list iterator<br>
1552 <tt>Module::const_giterator</tt> - Typedef for const_iterator.<br>
1554 <tt>gbegin()</tt>, <tt>gend()</tt>, <tt>gfront()</tt>, <tt>gback()</tt>,
1555 <tt>gsize()</tt>, <tt>gempty()</tt>, <tt>grbegin()</tt>, <tt>grend()</tt>
1557 <p> These are forwarding methods that make it easy to access the contents of
1558 a <tt>Module</tt> object's <a
1559 href="#GlobalVariable"><tt>GlobalVariable</tt></a> list.</p></li>
1561 <li><tt>Module::GlobalListType &getGlobalList()</tt>
1563 <p>Returns the list of <a
1564 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s. This is necessary to
1565 use when you need to update the list or perform a complex action that
1566 doesn't have a forwarding method.</p>
1568 <p><!-- Symbol table stuff --> </p></li>
1574 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
1576 <p>Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
1577 for this <tt>Module</tt>.</p>
1579 <p><!-- Convenience methods --></p></li>
1585 <li><tt><a href="#Function">Function</a> *getFunction(const std::string
1586 &Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt>
1588 <p>Look up the specified function in the <tt>Module</tt> <a
1589 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
1590 <tt>null</tt>.</p></li>
1592 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const
1593 std::string &Name, const <a href="#FunctionType">FunctionType</a> *T)</tt>
1595 <p>Look up the specified function in the <tt>Module</tt> <a
1596 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
1597 external declaration for the function and return it.</p></li>
1599 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt>
1601 <p>If there is at least one entry in the <a
1602 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
1603 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
1606 <li><tt>bool addTypeName(const std::string &Name, const <a
1607 href="#Type">Type</a> *Ty)</tt>
1609 <p>Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
1610 mapping <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this
1611 name, true is returned and the <a
1612 href="#SymbolTable"><tt>SymbolTable</tt></a> is not modified.</p></li>
1617 <!-- ======================================================================= -->
1618 <div class="doc_subsection">
1619 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
1622 <div class="doc_text">
1624 <p>Constant represents a base class for different types of constants. It
1625 is subclassed by ConstantBool, ConstantInt, ConstantSInt, ConstantUInt,
1626 ConstantArray etc for representing the various types of Constants.</p>
1630 <!-- _______________________________________________________________________ -->
1631 <div class="doc_subsubsection">
1632 <a name="m_Value">Important Public Methods</a>
1635 <div class="doc_text">
1638 <li><tt>bool isConstantExpr()</tt>: Returns true if it is a
1640 <hr> Important Subclasses of Constant
1643 <li>ConstantSInt : This subclass of Constant represents a signed
1646 <li><tt>int64_t getValue() const</tt>: Returns the underlying value of
1647 this constant. </li>
1650 <li>ConstantUInt : This class represents an unsigned integer.
1652 <li><tt>uint64_t getValue() const</tt>: Returns the underlying value
1653 of this constant. </li>
1656 <li>ConstantFP : This class represents a floating point constant.
1658 <li><tt>double getValue() const</tt>: Returns the underlying value of
1659 this constant. </li>
1662 <li>ConstantBool : This represents a boolean constant.
1664 <li><tt>bool getValue() const</tt>: Returns the underlying value of
1665 this constant. </li>
1668 <li>ConstantArray : This represents a constant array.
1670 <li><tt>const std::vector<Use> &getValues() const</tt>:
1671 Returns a Vecotr of component constants that makeup this array. </li>
1674 <li>ConstantStruct : This represents a constant struct.
1676 <li><tt>const std::vector<Use> &getValues() const</tt>:
1677 Returns a Vecotr of component constants that makeup this array. </li>
1680 <li>ConstantPointerRef : This represents a constant pointer value
1681 that is initialized to point to a global value, which lies at a
1682 constant fixed address.
1684 <li><tt>GlobalValue *getValue()</tt>: Returns the global
1685 value to which this pointer is pointing to. </li>
1694 <!-- ======================================================================= -->
1695 <div class="doc_subsection">
1696 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
1699 <div class="doc_text">
1701 <p>Type as noted earlier is also a subclass of a Value class. Any primitive
1702 type (like int, short etc) in LLVM is an instance of Type Class. All other
1703 types are instances of subclasses of type like FunctionType, ArrayType
1704 etc. DerivedType is the interface for all such dervied types including
1705 FunctionType, ArrayType, PointerType, StructType. Types can have names. They can
1706 be recursive (StructType). There exists exactly one instance of any type
1707 structure at a time. This allows using pointer equality of Type *s for comparing
1712 <!-- _______________________________________________________________________ -->
1713 <div class="doc_subsubsection">
1714 <a name="m_Value">Important Public Methods</a>
1717 <div class="doc_text">
1721 <li><tt>bool isSigned() const</tt>: Returns whether an integral numeric type
1722 is signed. This is true for SByteTy, ShortTy, IntTy, LongTy. Note that this is
1723 not true for Float and Double. </li>
1725 <li><tt>bool isUnsigned() const</tt>: Returns whether a numeric type is
1726 unsigned. This is not quite the complement of isSigned... nonnumeric types
1727 return false as they do with isSigned. This returns true for UByteTy,
1728 UShortTy, UIntTy, and ULongTy. </li>
1730 <li><tt>bool isInteger() const</tt>: Equilivent to isSigned() || isUnsigned(),
1731 but with only a single virtual function invocation.</li>
1733 <li><tt>bool isIntegral() const</tt>: Returns true if this is an integral
1734 type, which is either Bool type or one of the Integer types.</li>
1736 <li><tt>bool isFloatingPoint()</tt>: Return true if this is one of the two
1737 floating point types.</li>
1739 <li><tt>isLosslesslyConvertableTo (const Type *Ty) const</tt>: Return true if
1740 this type can be converted to 'Ty' without any reinterpretation of bits. For
1741 example, uint to int or one pointer type to another.</li>
1744 <p>Derived Types</p>
1747 <li>SequentialType : This is subclassed by ArrayType and PointerType
1749 <li><tt>const Type * getElementType() const</tt>: Returns the type of
1750 each of the elements in the sequential type. </li>
1753 <li>ArrayType : This is a subclass of SequentialType and defines
1754 interface for array types.
1756 <li><tt>unsigned getNumElements() const</tt>: Returns the number of
1757 elements in the array. </li>
1760 <li>PointerType : Subclass of SequentialType for pointer types. </li>
1761 <li>StructType : subclass of DerivedTypes for struct types </li>
1762 <li>FunctionType : subclass of DerivedTypes for function types.
1764 <li><tt>bool isVarArg() const</tt>: Returns true if its a vararg
1766 <li><tt> const Type * getReturnType() const</tt>: Returns the
1767 return type of the function.</li>
1768 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns
1769 the type of the ith parameter.</li>
1770 <li><tt> const unsigned getNumParams() const</tt>: Returns the
1771 number of formal parameters.</li>
1780 <!-- ======================================================================= -->
1781 <div class="doc_subsection">
1782 <a name="Argument">The <tt>Argument</tt> class</a>
1785 <div class="doc_text">
1787 <p>This subclass of Value defines the interface for incoming formal
1788 arguments to a function. A Function maitanis a list of its formal
1789 arguments. An argument has a pointer to the parent Function.</p>
1793 <!-- *********************************************************************** -->
1796 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
1797 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
1798 <a href="http://validator.w3.org/check/referer"><img
1799 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
1801 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
1802 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
1803 <a href="http://llvm.cs.uiuc.edu">The LLVM Compiler Infrastructure</a><br>
1804 Last modified: $Date$